Ijef v5n3 All

Editorial Board
Editor-in-Chief
Olajide. S. Oladipo, The City University of New York, USA
Associate Editors
Alain Devalle, University of Turin, Italy
Ali Yaftian, Deakin University, Australia
Augustion Chuck Arize, Texas A&M University, United States
Joe Lee, University of London, UK
Kim Man Lui, The Hong Kong Polytechnic University, Hong Kong
Lydia Gan, University of North Carolina, USA
Mattia Iotti, University of Parma, Italy
Vasileios Bill Kallinterakis, University of Liverpool, UK
Editorial Assistant
Cindy Xu, Canadian Center of Science and Education, Canada
Editorial Board Members

Adam Steen, Australia Esen Onur, USA Putu Mahardika A. Saputra, Indonesia
Adam Szyszka, Poland Evelyn Devadason, Malaysia Richard Gregory, USA
Alina Badulescu, Romania Francesco Zirilli, Italy Rita Yi Man Li, Hong Kong
América I. Zamora, México Gabriela Baicu, Romania Saktinil Roy, Canada
Andreea Stoian, Romania Gabriela Leite, Portugal Samuel Imarhiagbe, USA
Anthony Rezitis, Greece Giuseppe Scandurra, Italy Shahram Fattahi, Iran
Antonio Ruiz-Porras, Mexico Hussain A. Bekhet, Malaysia Sooriyakumar Krishnapillai, Nigeria
Antony Gautier, France Ian McFarlane, UK Sorinel Căpuşneanu, Romania
Augustin Mbemba, USA Jeng-Yan Tsai, Taiwan Suleyman Degirmen, Turkey
Biqing Huang, USA Juan Carlos Molero, Spain Tajul Ariffin Masron, Malaysia
Carl B. McGowan, USA Junaina Muhammad, Malaysia Taoufik Bouraoui, France
Carlo Alberto Magni, Italy Lonnie K. Stevans, USA Thanos Verousis, UK
Carlos Machado-Santos, Portugal Maria do Céu Alves, Portugal Tinh Doan, New Zealand
Carmen Gabriela Sirbu, Romania Marisa Faggini, Italy Tobias Basse, Germany
Cheng-Te Lee, Taiwan Martin Gregor, Czech Republic Tristan Nguyen, Germany
Chien-Ping Chen, USA Michael J. Seiler, USA Turan Subasat, UK
Choudhry Moorad, UK Mohamed Sherif, UK Valdir de Jesus Lameira, Portugal
Christopher S. Wright, Australia Mohammed Ziaul Haider, Malaysia Veysel Ulusoy, Turkey
Cosimo Magazzino, Italy Patrice Charlier, France Wael Moustafa, Egypt
Cristian Marian Barbu, Romania Periklis Gogas, Greece Wei Wan, USA
Daniel Badulescu, Romania Pierre Nguimkeu, USA Wen Qu, Australia
Dror Parnes, USA Piotr Misztal, Poland Yinghong Zhang, USA
Ehsan Khansalar, UK Prasad Vemala, USA Younes MOUATASSIM, Morocco
www.ccsenet.org/ijef International Journal of Economics and Finance Vol. 5, No. 3; 2013
Contents
Using Morningstar’s Analytical Measures to Evaluate Investment Recommendations Made by Consumer 1
Reports
C. Edward Chang & Thomas M. Krueger
Smiley or Frowney: The Effect of Emotions and Empathy Framing in a Downstream Water Pollution 9
Game
Natalia V. Czap, Hans J. Czap, Marianna Khachaturyan, Mark E. Burbach & Gary D. Lynne
The Stock Market/Unemployment Relationship in USA, China and Japan 24
Farzad Farsio & Shokoofeh Fazel
Saudi Financial Structure and Economic Growth: A Macroeconometric Approach 30
Mohammed Moosa Ageli & Shatha Mousa Zaidan
Managerial Incentives and the Risk-Taking Behavior of Hedge Fund Managers 36
Serge Patrick Amvella Motaze
Behaviour of Bank Share Prices and Their Impact on National Stock Market Indices: Comparing 49
Countries at Different Levels of Economic Development during Recessionary and Non-Recessionary
Periods
Lilian W. Komo & Isaac K. Ngugi
The Determinants of Informal Capital in the Financing of Small Firms at Start-Up: An Ethnic Comparison 62
of Small Firms in Sweden
Darush Yazdanfar & Saeid Abbasian
Examining the Value of Money in England over the Long Term (1259-2009) 73
Adam Abdullah
Consequential Effects of Budget Deficit on Economic Growth: Empirical Evidence from Ghana 90
Samuel Antwi, Xicang Zhao & Ebenezer Fiifi Emire Atta Mills
The Optimal Licensing Strategy of an Outside Patentee in Vertically-Related Markets 102
Ming-Chung Chang, Jin-Li Hu & Chin-Hung Lin
Corporate Sustainability Reporting and Analysis of Sustainability Reports in Turkey 113
Rabia AKTAŞ, Koray KAYALIDERE & Mahmut KARĞIN
Top Companies Ranking Based on Financial Ratio with AHP-TOPSIS Combined Approach and Indices 126
of Tehran Stock Exchange - A Comparative Study
Seyed-Hasan Hosseini, Mohammad Esmaeil Ezazi, Mohamad Rasoul Heshmati & Seyed-Mohamad Reza
Hosseini Moghadam
Migrant Workers' Remittances and External Trade Balance in Sub-Sahara African Countries 134
Henry Okodua & KWumi Olayiwola
The Profitability of Banking Sector in Republic of Macedonia 143
Meri Boshkoska
Can Financial Ratios Reliably Measure the Performance of Banks in Bahrain? 152
Naser J. Najjar
Contents
Testing the CAPM for the Brazilian Stock Market: A Study of Dynamic Beta Using Multivariate GARCH 164
Lucas Lúcio Godeiro
The Regional Integration Agreements: A New Face of Protectionism 183
Montej Abida
Monetary Policy and Its Implications for Balance of Payments Stability in Nigeria: 1980-2010 196
Anthony Ilegbinosa Imoisi, Lekan Moses Olatunji & Bosco Itoro Ekpeyong
Does Fiscal Policy Matters for Growth? Empirical Evidence from Pakistan 205
Rabia Nazir, Mumtaz Anwar, Mamoona Irshad & Ayza Shoukat
The Role of Corporate Governance in Reducing the Negative Effect of Earnings Management 213
Nopphon Tangjitprom
An Investment Strategy Based on Stochastic Unit Root Models 221
Mamadou A. Konté
International Journal of Economics and Finance; Vol. 5, No. 3; 2013
ISSN 1916-971X E-ISSN 1916-9728
Published by Canadian Center of Science and Education
Using Morningstar’s Analytical Measures to Evaluate Investment

Recommendations Made by Consumer Reports
C. Edward Chang1 & Thomas M. Krueger2
1
College of Business, Missouri State University, Springfield, Missouri, U.S.A.
2
College of Business Administration, Texas A&M-Kingsville, Kingsville, Texas, U.S.A.
Correspondence: Thomas M. Krueger, College of Business Administration, Texas A&M-Kingsville, Kingsville,
TX, 78363, USA. Tel: 1-361-397-3787. E-mail: [email protected]
Received: November 17, 2012 Accepted: January 2, 2013 Online Published: February 22, 2013
doi:10.5539/ijef.v5n3p1 URL: http://dx.doi.org/10.5539/ijef.v5n3p1
Abstract
Instead of its typical analysis of automobiles and household gadgets, Consumer Reports ventured into the murky
world of predicting investments performance in 2005 and 2007. Enough time has passed since these
recommendations to conduct a rigorous investigation of Consumer Reports’ success. This study analyzes
recommendations through examination of return, risk, and risk-adjusted return using the popular Morningstar
ratings. Consumer Reports’ choices will be judged in relationship to mutual funds in the same category and
index funds. It appears as though Consumer Reports provided a valuable service to its investing readership both
times.
Keywords: mutual funds, investment recommendations, investment performance, Morningstar, Consumer
Reports
1. Introduction
1.1 Consumer Reports
Making wise investment decisions is one of the most critical and challenging decisions faced by individuals. One
way to maintain and build wealth is through astute selection of mutual funds. In 2012, 53.8 million households,
which is 44.4 percent of all households, owned mutual fund (Investment Company Institute, 2012). The
objective of this research is to evaluate the mutual fund recommendations made by Consumer Reports in terms
of return, risk, and risk-adjusted return using the popular Morningstar rankings.
Consumer Reports is a monthly publication of Consumers Union, a New York-based association of engineers,
professors, labor leaders, and journalists. Its claim to fame lies in the quality of its articles, based on independent
analysis of a wide variety of consumer goods. It does not print advertisements or accept free samples, allowing
Consumers Union to claim that it is an unbiased, independent rating agency. Since its origin in 1933, Consumer
Reports has only made two sets of recommendations regarding investment choices, one in 2005 and another in
2007.
This investigation analyzes Consumer Reports’ ability to adapt its highly-regarded analytical skills to investment
choices. Morningstar’s proprietary return, risk, and rating (also known as “star”) rankings are used to judge the
value of using Consumer Reports as an investment advisor. Significant findings indicate that the information
used by Consumer Reports is not fully reflected in stock prices. Instead of attempting to determine and then
adopt the investment criteria employed by Consumer Reports, investors may be able to simply use Consumer
Reports’ recommendations as the basis of their investment selection.
1.2 Literature Review
1.2.1 Performance of Consumer Reports' Recommended Mutual Funds
Two prior articles have evaluated the return performance of Consumer Reports’ recommended mutual funds.
Interestingly, each article only studied one mutual fund recommendation. Index Investor (undated) studies the
period prior to the first recommendation, while Chen (2011) examines the period surrounding the second
recommendation. Studying the 1995 to 2004 period, Index Investor finds that the 52 funds in the 2005 Consumer
Reports article had an average alpha of only 0.17%, and many were negative. This led Index Investor to the
1
conclusion that Consumer Reports’ 2005 mutual fund recommendation was “a big mistake by an organization
that is generally extremely reliable.” By comparison our study examines subsequent investment returns, which
would be available to investors.
Chen (2011) examines 60 mutual funds presented in the Consumer Reports’ February 2007 issue. The November
30, 2006 and September 30, 2008 time periods were analyzed in her study, in order to exclude the major declines
that occurred in the stock market during the last quarter of 2008. In addition to reporting on a wide variety of
fund characteristics, Chen studies return performance to determine whether Consumer Reports is a reliable
information source for investors seeking to purchase mutual funds. Based on two limited sample periods,
including time both before and after fund recommendation, Chen concludes that Consumer Reports is not a good
investment advisor. However, her post-publication sample period is too short to effectively test the utility of
using Consumer Reports’ 2007 recommendation.
Extensive study has been done of the relationship between mutual fund advertising and rankings as well as
investor reaction to a mutual fund being included in the rankings. Reuter and Zitzewitz (2006) report a positive
relationship between advertising expenditures and fund recommendations by personal finance publications. In
order to adjust for the potential that past fund advertising has created a demand for articles concerning
advertisers, as exposed by George and Waldfogel (2003), Reuter and Zitzewitz conduct comparisons across
personal finance publications. Between 1998 and 2002, Reuter and Zitzewitz find that 83.8 percent of the mutual
fund families that spent over $1 million in Money magazine advertising increased the probability of receiving a
positive mention by twenty percent. However, Reuter and Zietzewitz’ findings do not reveal a correlation
between advertising and subsequent returns. Nonetheless, investors respond to fund mentions in Consumer
Reports with an economically significant increase in fund investment. These findings are consistent with Sirri
and Tufano’s (1998) and Jain and Wu’s (2002) findings that mutual funds receiving more media attention
receive correspondingly higher inflows, because investors have lower search costs.
Our research is a better measure of the value of Consumer Reports’ recommendations for six reasons. One, we
use a longer sample period as the investment time horizon. Two, we do not exclude any expansionary or
recessionary stock market period. Three, we include both the 2005 and 2007 Consumer Reports’
recommendations in our analysis. Four, use of Morningstar star statistics brings with it some additional
advantages, which will be outlined in the next section. Five, unlike Chen’s study of a fund’s net assets, manager
tenures, and the like, we focus purely on returns and risk because these are the factors that impact investor
wealth. Six, we do not deviate from a focus on return performance to consider such issues as flows to and from
the fund or perception of fund performance based on the provision of supplemental information as done by
Kozup, Howlett, and Pagano (2008). After all, the core issue is the economic benefit of using a given source of
investment information. Therefore, we believe that this analysis presents a clearer, more comprehensive
investigation of Consumer Reports’ recommendations.
1.2.2 Morningstar and the Benefits of Using Its “Star” Rating System
Some information is provided about Morningstar, Inc. because it is the supplier of the data used in this analysis.
Like Consumer Reports, Morningstar is a widely-respected, independent investment research company. The
Chicago-based firm provides analysis of over 350,000 investments, and is most well known for its analysis of
stocks and mutual funds (Morningstar, 2012). The most popular information produced by Morningstar is its star
ranking system. There are actually three different rankings assigned by Morningstar based on a fund’s monthly
performance including the effects of sales charges, loads, and redemption fees. Unfortunately, Morningstar’s
assessment process is proprietary and not disseminated. Morningstar reportedly places more emphasis on
downward variation and rewards consistent performance, which are both prized by mutual fund investors.
Information shared with the general public is that the Morningstar ranking system is based on performance
during the past three, five, and ten years (where available), with more import given to recent results. Hence, it
focuses on the period since Consumer Reports made its recommendations. It also states that the top 10% of funds
in each category receive five stars, the next 22.5% receive four stars, the middle 35% receive three stars, the next
22.5% receive two stars, and the bottom 10% receive one star, providing a crude bell-shaped distribution.
Despite these limitations, many mutual funds report their performance using Morningstar’s star rating system
(for example Aston Asset Management (2012) and T. Rowe Price (2012)). In the Wall Street Journal’s Investing
in Funds & ETFs quarterly analysis in early 2013, Morningstar’s star rankings were viewed as being both
well-known and a “straight-forward and intuitive approach for gauging fund performance” (Espinoza, 2013).
The article goes on to warn readers that other rating systems, such as Morningstar’s gold, silver, bronze, neutral
and negative award system, are based on human judgment.
2
Del Guercio and Tkac (2008) provide evidence that money flowing to mutual funds responds to changes in
Morningstar ratings. However, Morningstar rankings are not infallible. In fact, while money is pouring into the
funds with four and five stars, Morningstar’s own research shows that its stars are worth little as investment
guides. According to Quinn (2010), this is especially true of market turns. During tranquil years, there is a
limited amount of forecasting ability arising from Morningstar’s stars. For instance, four-star performance tops
that of the three stars by 0.5 percentage points or less. However, any individual fund, in either group, could do
much better or much worse. During years of market changes, the star pattern is in disarray. For instance, among
bond funds and balanced stock-and-bond funds, lowly one-star and two-star funds did far better in the following
two years than the five-star funds. Russel Kinnel (2009), Morningstar's director of mutual fund research observes
lower returns to funds with more stars in the high-yield bond, technology, high price/earnings ratio, and large
growth sectors. In short, Morningstar is a respected, valid measure of past performance, which is the focus of
this study, without any guarantee regarding future performance.
2. Method
2.1 Consumer Reports’ Recommendations
Seventy mutual funds were recommended in Consumer Reports’ March 2005 issue. In the February 2007 issue,
sixty funds were recommended. Forty-six of the funds proposed in 2007 were new funds. Four of the seventy
funds recommended in 2005 (CR 2005) and two of the new funds recommended in 2007 (CR 2007) have closed,
leaving sixty-six and forty-four funds, respectively. We focus on the new funds in the 2007 recommendation
because they represent new Consumer Reports’ selections and why the publication felt an updated
recommendation was necessary. Consumer Reports does not explicitly tell readers to replace their 2005 choices
with the 2007 funds, and there is no overlap of mutual funds in our analysis.
Fourteen Morningstar categories are represented by the sixty-six remaining CR 2005 funds. Eighteen
Morningstar categories are represented by the forty-four CR 2007 funds. The biggest change in new funds
chosen is that the CR 2007 set includes selections from five foreign mutual funds categories (i.e., large blend,
large growth, large value, small/mid value, and world allocation), and a few domestic sectors (i.e., financial,
utilities).
Out of the 9,213 funds in existence in 2005, Consumer Reports picked 70 funds, or 0.76 percent of the possible
total. In 2007, out of the 11,488 funds, Consumer Reports picked 44 new funds, or 0.38 percent, in 2007.
Obviously, the chosen mutual funds are a very select group.
2.2 Mutual Fund Data and Methodology
All data for individual mutual funds and category averages of mutual funds were collected from Morningstar’s
Principia database as of March 31, 2010. The return performance of the Consumer Reports’ selections is
compared to the performance of the category average and a Vanguard index fund in each Morningstar category.
The comparative numbers are only based on those categories wherein Consumer Reports found good fund
investment choices. For instance, CR 2005’s performance benchmarks include performance of the real estate
sector, but not the foreign large blend sector.
Using Morningstar has the additional benefit of benefiting from its categorizing of mutual fund type on the basis
of the securities in each portfolio. A fund’s prospectus may not be a completely accurate description of how the
fund actually invests. Given its excellent reputation and widespread use by investors, we rely on Morningstar for
proper categorization of funds.
A paired t-test is the primary statistical testing method used in this report. There is a matching of the average
return of the mutual funds chosen by Consumer Reports and the average of the mutual funds placed in the same
category by Morningstar. Measurements are aggregated across fund categories. The null hypothesis is always
equality of means, with the alternative hypothesis being inequality of means. The lower the t-test statistics the
greater the apparent skills of the Consumer Reports investment research team.
3. Results
3.1 Morningstar Return Rankings
Morningstar helps investors gain an understanding of how mutual funds fall within a normal distribution by
presenting its own performance metric. With only a five-step scale, and clustering of funds around the mean, the
Morningstar ranking provides an impression of how a given fund performed relative to a crude bell-shaped curve
and thereby provides additional insight to the investor. An average Morningstar return rankings is 3, as
3
illustrated by this value found in the “Category Averages” row for all Consumer Reports recommendations in
Table 1.
Across the Consumer Reports fund selections, the 5.00 ranking for the CR 2007 recommendation of foreign
large blend funds indicates that this choice was in the top 10 percent of all funds in this category. Among the CR
2005 selections, the best recommendations were the two small value funds, which had an average Morningstar
return ranking of 3.50 during the April 2007 to March 2010 period, and 4.00 average during the April 2005 to
March 2010 period. The 4.00 value indicates that this selection is within the best 10.01 to 32.50 percent of funds
in the small value category.
Table 1. Morningstar return rankings

Panel A. Results within Morningstar categories and average of category results
Investment Time Horizon
2005 CR Recommendations 2007 CR Recommendations
Morningstar Categories 3 Years 5 Years 3 Years
Conservative Allocation 4.50
Financial 3.00
Foreign Large Blend 5.00
Foreign Large Growth 2.00
Foreign Large Value 5.00
Foreign Small/Mid Value 5.00
Health 3.00 3.00
Industrials 3.00 3.00
Large Blend 3.13 3.38 2.50
Large Growth 3.50 3.50
Large Value 3.63 3.44 3.50
Mid-Cap Blend 3.88 3.63 2.50
Mid-Cap Growth 3.00 2.00
Mid-Cap Value 3.00 3.00 3.50
Moderate Allocation 3.36 3.82 3.20
Natural Resources 3.00 2.00 2.00
Real Estate 2.67 3.00
Small Blend 3.60 3.20 3.80
Small Growth 3.00
Small Value 3.50 4.00 3.00
Utilities 3.00
World Allocation 3.50
World Stock 3.00 2.50 2.60
Average 3.23 3.11 3.37
Panel B. Benchmark Portfolio Morningstar Return Ratings and statistical significance of Consumer Reports’ recommendations
Category Averages 3.00 3.00 3.00
T-test statistic 0.012** 0.264 0.061*
Vanguard index funds 3.36 3.50 3.61
T-test statistic 0.230 0.047** 0.130
Notes: ***, **, *: Significant at the 0.01, 0.05, and 0.10 level
A good example of the value of Morningstar rankings is provided by the small value funds. Their average annual
return was a loss of 0.61 percent, which seems unremarkable. However, with a Morningstar ranking of 4.0, on a
relative basis the performance of CR 2005’s recommendations in the small value category were well above
normal.
Another good example of the relevant, additional informative aspect of Morningstar’s return rankings can be
witnessed in the mid-cap growth category. The performance of this fund was one of the top two performers
among Consumer Reports’ choices over three years and had mid-level performance over five years. However, in
Table 1, one can see that the Morningstar return ranking is only an average 3.00 stars during the shorter period
and a below-average, 2.00 stars during the longer period.
4
Average Morningstar return rankings for Consumer Reports’ recommendations fall between the category
averages and Vanguard index fund benchmark in every examined instance. The CR 2005 and CR 2007 instances
were significantly better at the category average at the 0.01 and 0.10 levels, respectively. However, if one would
have invested in the CR 2005 funds over the April 2005 to March 2010 period, their performance would have
been worse than the Vanguard index fund benchmark at the 0.05 level.
3.2 Morningstar Risk Rankings
Findings based on Morningstar’s risk rankings are an alternative to the common analysis of standard deviations
and beta. One advantage of using Morningstar’s risk rankings, displayed in Table 2, is that they are a
standardized measure of risk, as implied by the 3.00 values found in the “Category Averages” row. Furthermore,
as described above, negative mutual fund return variance is up-weighted in comparison to upside risk.
Undesirable higher levels of risk result in higher Morningstar risk rankings. With nine risk rankings below 2.75
versus none above 3.25, it is not surprising that that CR 2005 funds have an average of 2.40. This is significantly
lower than the 3.00 category average and 3.64 Vanguard index fund benchmark at the 0.01 level.
Table 2. Morningstar risk rankings

Panel A. Results within Morningstar categories and average of category results
Financial 3.00
Health 2.00 2.00
Large Blend 3.13 3.13 3.00
Large Value 2.81 2.75 3.00
Mid-Cap Blend 2.50 2.50 3.50
Mid-Cap Value 3.00 2.80 3.00
Small Blend 1.80 1.60 2.40
Small Growth 1.00
Small Value 2.00 2.00 2.33
Utilities 2.00
World Stock 2.50 2.50 2.80
Average 2.40 2.35 2.45
Panel B. Benchmark Portfolio Morningstar Risk Ratings and statistical significance of Consumer Reports’ recommendations
T-test statistic 0.000*** 0.000*** 0.007***
T-test statistic 0.000*** 0.000*** 0.001***
Notes: ***, **, *: Significant at the 0.01, 0.05, and 0.10 level
Expanding the sample period to five years slightly reduces the CR 2005 average Morningstar risk rankings. Only
the moderate allocation category experiences a risk ranking rise. Although the Vanguard index fund
benchmark’s Morningstar risk rankings is also lower over the longer period, the CR 2005 recommendation is
still superior, in terms of having a lower level of risk, at the 0.01 level.
5
Similar comments can be made for the CR 2007 funds. At 1.00, the foreign large value fund TBGVX (Tweedy
Browne Global Value), foreign small/mid value fund SGOVX (First Eagle Overseas), and two small growth
funds, BGRFX (Baron Growth Retail) and SAGWX (Sentinel Small Capital), have the lowest possible
Morningstar risk ranking within their respective categories. Only the CR 2007 funds in the conservative
allocation and mid-cap blend categories exceed the 3.0 average star value. However, choices in these categories
have an average Morningstar risk ranking of only 3.50 stars. The resulting low Morningstar risk ranking is
significantly less than the category average or Vanguard index fund benchmark at the 0.01 level.
3.3 Morningstar Rating Rankings
The widely-distributed Morningstar rating rankings are a function of both the Morningstar return and
Morningstar risk rankings. As before, Morningstar specifies a value of 3.00 as the typical value, which is
exemplified by the value of 3.00 found in the “Category Averages” row. Higher star frequency values in Table 3
are preferred.
Over the April 2007 to March 2010 period, CR 2005 recommendation provide a Morningstar rating ranking
exceeding 3.00 nine times. Only the mid-cap value fund has a Morningstar rating ranking less than 3.00. Though
not excessively large at 3.32, the CR 2005 Morningstar ratings over the three-year period are significantly higher
than the category average at the 0.01 level and the Vanguard index fund benchmark at the 0.10 level.
Extending the investment period to April 2005 through March 2010 enhances the Morningstar rating ranking to
3.47. Morningstar rating rankings of 4.00 points are earned by CR 2005 funds in the health, industrials, large
growth, and small value categories. Only the mid-cap value fund, at 2.60 registers a value less than 3.00.
Consequently, CR 2005 funds exceed their category average at the 0.01 level and are no longer underperforming
the Vanguard index fund benchmark, as observed in Table 1.
The best average Morningstar rating rankings are earned by the CR 2007 funds, with an average value of 3.57.
Leading the way were the foreign large blend, foreign large value, and foreign small/mid value categories, with
Morningstar rating rankings of 5.00, the highest possible score. With only two individual category Morningstar
rating rankings less than 3.00, the average CR 2007 fund average is statistically better than the category average
at the 0.01 level. Although the CR 2007 Morningstar return ranking was 0.24 lower than the Vanguard index
fund benchmark, as shown at the bottom right of Table 1, the CR 2007 Morningstar rating ranking is higher by
0.13. Not significant, but not a bad choice!
4. Conclusion
Our research generally supports the utilization of Consumer Reports in mutual fund selection. Morningstar
rankings were used to provide an impression of the relative performance of CR funds on a return, risk, and
risk-adjusted return basis. These findings support Consumer Reports’ claim regarding an ability to select funds
with superior performance when compared to Morningstar category averages. Whereas, Consumer Reports’
2005 recommendation earned a lower Morningstar return ranking than Vanguard index funds in one instance,
Consumer Reports’ selections had significantly lower risk in all comparisons. In combination, there was a
limited amount of dominance by the Consumer Reports’ selections on a risk-adjusted return, Morningstar rating
basis.
As of early-2013, Consumer Reports has not published a follow-up listing to its 2007 recommendations. This
research implies that additional recommendations are likely to be of value to the investors and Consumer
Reports’ readers. New Consumer Report investment recommendations, as well as the predictive power of the
2005 and 2007 recommendations over longer periods of time, would be a fertile area for future research.
6
Table 3. Morningstar rating rankings

Panel A: Results within Morningstar categories and average of category results
Financial 3.00
Health 3.00 4.00
Large Blend 3.13 3.25 3.00
Large Value 3.56 3.56 3.50
Mid-Cap Blend 3.88 3.88 2.50
Mid-Cap Value 2.80 2.60 3.50
Small Blend 3.80 3.80 4.00
Small Growth 4.00
Small Value 3.50 4.00 3.00
Utilities 3.00
World Stock 3.50 3.00 2.80
Average 3.32 3.47 3.57

Panel B: Benchmark Portfolio Morningstar Ratings and statistical significance of Consumer Reports’ recommendations
T-test statistic 0.002*** 0.002*** 0.004***
T-test statistic 0.080* 0.437 0.232
Notes: ***, **, *: Significant at the 0.01, 0.05, and 0.10 level.
References
Aston Asset Management. (2012, March 31). Morningstar Ratings & Rankings. Retrieved from
http://astonfunds.com/performance-assets/morningstar
Chen, H. (2011). Analysis of Consumer Reports' Recommended Mutual Funds Compared to Actual Performance.
Journal of Financial Services Marketing, 16(1), 42-49. http://dx.doi.org/10.1057/fsm.2011.5
Consumer Reports. (2005, March). Stock Mutual Funds: 70 All-weather Winners. Consumer Reports, 28-33.
Consumer Reports. (2007, February). 60 Funds you can count on. Consumer Reports, 18-22.
Consumer Reports. (2011, August 12). About US. Retrieved from
http://www.consumerreports.org/cro/aboutus/mission/overview/index.htm
Del Guercio, D., & Tkac, P. (2008). Star Power: The Effect of Morningstar Ratings on Mutual Fund Flow.
Journal of Financial and Quantitative Analysis, 43(4), 907-936.
http://dx.doi.org/10.1017/S0022109000014393
Espinoza, J. (2013, January 7). Ratings Game: What Do Those Stars and Fund Awards Mean? Wall Street
Journal, p. R1.
George, L., & Waldfogel, J. (2003). Who Affects Whom in Daily Newspaper Markets? Journal of Political
Economy, 111(4), 765-784. http://dx.doi.org/10.1086/375380
7
Investment Company Institute. (2012). Characteristics of Mutual Fund Investors. ICI Research Perspective,
18(7), 1-16.
Jain, P. C., & Wu, J. S. (2002). Truth of Mutual Fund Advertising: Evidence on Future Performance and Fund
Flows. Journal of Finance, 55(2), 937-958. http://dx.doi.org/10.1111/0022-1082.00232
Kinnel, R. (2009, November 30). Star Rating at its Best and Worst. Morningstar. Retrieved from
http://corporate.morningstar.com/us/documents/MethodologyDocuments/ResearchStudies/Star_Rating_Fun
ds_Reprint.pdf
Kozup, J., Howlett, E., & Pagano, M. (2008). The Effects of Summary Information on Consumer Perception of
Mutual Fund Characteristics. Journal of Consumer Affairs, 42(1), 37-59.
http://dx.doi.org/10.1111/j.1745-6606.2007.00093.x
Morningstar. (2012, May 1). Independent. Insightful. Trusted. Retrieved from
http://corporate.morningstar.com/us/asp/subject.aspx?xmlfile=178.xml
Quinn, J. (2010, May 6). Morningstar Rankings Fail Mutual Fund Investors. CBS News. Retrieved from
http://www.cbsnews.com/8301-505123_162-41240156/morningstar-star-rankings-fail-mutual-fund-investors
Reuter, J., & Zitzewitz, E. (2006). Do Ads Influence Editors? Advertising and Bias in the Financial Media.
Quarterly Journal of Economics, 121(1), 197-227. http://dx.doi.org/10.1093/qje/121.1.197
Sirri, E., & Tufano, P. (1998). Costly Search and Mutual Fund Flows. Journal of Finance, 53(5), 1589-1622.
http://dx.dot.org/10.111/0022-1082.00066
T. Rowe Price. (2012, March 31). Mutual Funds Rated Four Stars or Five Stars by Morningstar. Retrieved from
http://individual.troweprice.com/public/Retail/Mutual-Funds/Our-Mutual-Fund-Family/Funds-Rated-4-and-
5-Stars-by-Morningstar?adcode=7209&PlacementGUID=C9F68E39-B059-417C-A38F-1E920BCB0C5D
The Index Investor. (undated). Consumer Reports: Mutual Fund Review–Critique. Retrieved from
http://www.indexinvestor.com/Free/consumerReports.php
8
ISSN 1916-971X E-ISSN 1916-9728
Smiley or Frowney: The Effect of Emotions and Empathy Framing in

a Downstream Water Pollution Game
Natalia V. Czap1, Hans J. Czap2, Marianna Khachaturyan3, Mark E. Burbach4 & Gary D. Lynne5
1
Department of Social Sciences (Economics), University of Michigan-Dearborn, U.S.A.
2
Department of Management Studies, University of Michigan-Dearborn, U.S.A.
3
Department of Agricultural Economics, University of Nebraska-Lincoln, U.S.A.
4
School of Natural Resources, University of Nebraska-Lincoln, U.S.A.
5
Department of Agricultural Economics and School of Natural Resources, University of Nebraska-Lincoln,
U.S.A.
Correspondence: Natalia V. Czap, Department of Social Sciences (Economics), University of Michigan-Dearborn,
4901 Evergreen Rd., Dearborn, MI 48103, U.S.A. Tel: 1-313-583-6354. E-mail: [email protected]
Abstract
Common-pool resources and other shared resources frequently suffer from overextraction/overuse and associated
negative externalities. In this paper we design a framed laboratory experiment on downstream water pollution to
investigate (a) the importance of empathy vs. self-interest framing in determining the behavior of upstreamers
regarding the negative externalities, and (b) the potential of downstreamers to influence the choices of
upstreamers using non-monetary sanctions and rewards, alleviating the need for intervention by the local
governments and regulatory institutions. Our results show that empathy framing has a much more significant
impact on individual behavior than self-interest framing. Overall subjects behaved more profit-oriented in the
self-interest framing and more egalitarian in the empathy framing. Lastly, negative emotional feedback is a
powerful tool for changing behavior of subjects towards more environmentally friendly actions. Interestingly,
positive emotional feedback is counterproductive for that. In general our results indicate that explicit emotional
feedback, even though not expressed by everyone, works to the same degree as the implicit appeal to emotions
through framing.
Keywords: empathy framing, self-interest framing, emotional feedback, water pollution, environmental
experiment, reward and punishment
1. Introduction
The problem of shared resources, especially common pool resources (CPRs) (e.g., fisheries and forests), has been
studied for decades. In case of CPRs, if the users are driven solely by self-interest and do not cooperate/coordinate
their actions, overextraction occurs. However, over the years experimental researchers (see, among others, Ostrom
& Walker, 1991; Sally, 1995; Balliet 2010; Cardenas et al., 2000; Ostrom, 2010) showed in the laboratory and in
the field that users of CPRs are not always driven by strict self-interest and often manage to prevent the overuse
(Note 1) of resources through (self-imposed) regulations and sanctions.
In the context of other shared resources unchecked self-interest and lack of cooperation/coordination may lead to
negative externalities through pollution in addition to/instead of overextraction. For many flowing water resources
(such as rivers and creeks) usage creates an upstream-downstream problem. If the upstreamers are motivated
exclusively by self-interest and profit maximization, they are likely to impose negative externalities (in the case of
farmers, for example, water pollution through chemical runoffs and soil erosion) on individuals living downstream.
In contrast, if the upstreamers are motivated by empathetic considerations towards nature and/or the
downstreamers, they may reduce negative externalities by undertaking costly actions.
This situation raises two interesting empirical questions. The first question asks whether there is a difference in
behavior of the upstreamers, if their decision is framed as a decision mainly regarding profit maximization, taking
into account the externalities of pollution, versus if their decision is framed as a decision mostly on preserving the
9
cleanliness of water resources, taking into account profit maximization. The former case brings to the forefront
self-interest considerations conditioned by empathy, whereas the latter emphasizes empathetic considerations
tempered by self-interest.
The second question deals with the potential of downstreamers to influence the behavior of upstreamers without
involvement of local governments and regulatory institutions via non-monetary sanctions and rewards (such as
positive/negative word-of-mouth, campaigns to increase awareness about the problem, display of social
(dis)approval, etc.). If such sanctions work and are powerful enough to significantly change the behavior of the
upstreamers, then public policies and environmental organizations should focus on providing venues that facilitate
the flow of information from downstreamers to upstreamers and communication between them.
In this paper we design a downstream water pollution game to investigate these two questions. The rest of the
paper is structured as follows: Section 2 discusses the theoretical underpinning of our study, Section 3 describes
the experimental design and hypotheses, Section 4 analyses the results, Section 5 concludes.
2. Framing and (Dis)Approval Effects
2.1 Framing Effects
Experimental economists are rightfully concerned about the potential framing effect of instructions. The majority
of economic experiments is framed neutrally and presents a context-free environment to minimize social cues of
how one should behave and to minimize the influence of homegrown values on individual decisions in the
laboratory. Neutral wording allows researchers to retain control over the factors that influence subjects’ choices.
However, in certain contexts it is impossible to avoid framing without making the task irrelevant. A number of
studies examine the effect of framing in the instructions on individual behavior in the context of corruption
(Abbink & Hennig-Schmidt, 2006; Barr & Serra, 2009), public goods games (Andreoni, 1995; Cookson, 2000;
Park, 2000; Fujimoto & Park, 2010; Cubitt, Drouvelis & Gächter, 2011; Cubitt, et al. 2011), sequential bargaining
games (Brosig et al., 2003), and altruistic giving in dictator games (Duffy & Kornienko 2010). The existing
experiments manipulating the framing of instructions can be separated roughly into two categories:
procedural-oriented framing (which compares the different ways of representing the problem, e.g. neutral context
vs. contextualized, or the type of payoff presentation) and priming-oriented framing (which suggests a particular
value judgment, e.g. selfish vs. altruistic, positive vs. negative, take vs. give).
The effect of procedural-oriented framing is ambiguous. In bribing games, Abbink & Hennig-Schmidt (2006)
showed that there is no difference between individual behavior exposed to neutral and framed/contextualized
instructions, while Barr & Serra (2009) reported the opposite finding. In public goods games, Cookson (2000)
found that changing the presentational variables (e.g. the formulation of the payoff function or the type of
comprehension task) has a strong effect on experimental results. In sequential bargaining games, Brosig et al.
(2003) reported the existence of a hot (ordinary sequential manner, i.e. the subjects respond to the choices made by
their partner) versus cold effect (strategy method, i.e. the subjects submit a complete set of choices before play
commences) (Note 2), while earlier research by Brandts & Charness (2000) did not find it.
The effect of priming-oriented framing is more robust. Andreoni (1995) and follow-up studies by Park (2000) and
Fujimoto & Park (2010) demonstrated that subjects contributed significantly more to a public good under positive
framing (investing in a public good) compared to negative framing (investing in a private good). Along the same
lines Cubitt et al. (2011) showed that there is a framing effect when expressing moral judgment about contributions,
with free-riding in a Give treatment being condemned more strongly than in a Take treatment. However, when
one controls for contributions, monetary punishment and self-reported emotions are not sensitive to Give versus
Take manipulation. In the dictator game Duffy & Kornienko (2010) showed that the subjects give more in an
altruistic treatment (dictators are ranked in descending order according to the amount they give) as compared to a
control and to a selfish treatment (dictators are ranked in descending order according to the amount they keep for
themselves).
This paper contributes to the ongoing discussion in the experimental literature about framing effects by examining
the priming-oriented framing effects in the context of environmental protection.
2.2 Social (Dis)Approval through Reward/Punishment and Emotions
Rewards and punishments are used to express social (dis)approval and increase fairness and cooperation in social
dilemmas and situations involving distribution of income and assets. Experimental papers report that punishment
is more effective than rewards in dictator games (Andreoni et al., 2003), public goods games (Dickinson, 2001;
Sefton et al., 2008; Sutter et al., 2010, but not in Walker & Halloran, 2004), and common pool resource games (van
Soest & Vyrastekova, 2006). Furthermore, Andreoni et al. (2003) show that both methods are complements rather
10
than substitutes in enforcing the ideals/norms and reaching the specific objectives. This effect is particularly strong
when the punishments and rewards are decided and implemented by the group members rather than by an
exogenous institution (Sutter et al., 2010).
Actions of individuals, both helping and hurting, are strongly linked to positive and negative emotions of the
affected party (Offerman, 2002; Bosman & van Widen, 2002; Xiao & Houser, 2005). Rewards and punishments
represent positive and negative reciprocity and as such they are behavioral means to express positive and negative
emotions. According to Fehr & Gächter (2002, p.139) negative emotions are the proximate sources of punishment:
“Free riding may cause strong negative emotions among the cooperators and these emotions, in turn, may trigger
their willingness to punish the free riders.” Similarly, “costly punishment might itself be used to express negative
emotions” (Xiao & Houser, 2005, p.7398); anger (regarding norm violations) and guilt are sufficient to induce
punishment (Nelissen & Zeelenberg, 2009), and a higher intensity of reported positive/negative emotions
increases the likelihood of reward/punishment (Offerman, 2002). As positive and negative emotions are expressed
in terms of reward and punishment, they, in turn, may also trigger emotions in a receiving party and cause a second
round of reciprocation.
To the best of our knowledge there is no systematic investigation of the impact of expressed negative and (explicit)
positive emotions (not just an absence of disapproval) by the affected party back on the affecting party. This link
does not seem to be straightforward. Hopfensitz & Reuben (2009) showed that the effectiveness of negative
emotions, expressed by the punishing party, depends on the reaction of those who are punished. Specifically, if the
expressed negative emotions lead to anger this may cause retaliation, whereas if shame and guilt are the result, it
may restrain the desire of the punished party to fight back (Hopfensitz & Reuben, 2009). It is very plausible that
the effect of positive emotions will be similarly contradictory. Positive emotions, expressed by the rewarding party,
can trigger positive reciprocity in the rewarded party and further increase cooperation and altruistic actions. At the
same time positive emotions signal that the rewarding party is satisfied with the outcome, which may encourage
the rewarded party to be a little more selfish next time around and test a lower threshold of acceptability.
In this paper we conduct a 2-round game in which after the first round the affected party can express their emotions
regarding the outcome/behavior to the affecting party. This allows testing how much this emotional response will
influence the behavior of the affecting party in the second round.
2.3 Monetary versus Non-Monetary Rewards and Punishments
Several experimental studies demonstrate that non-monetary sanctions and rewards can be as efficient as monetary
incentives in inducing cooperation and fair outcomes. Masclet et al. (2003) showed that contribution levels to
public goods from both monetary and non-monetary sanctions (expressing a degree of disapproval) are similar,
with non-monetary sanctions working better under partner, as compared to stranger matching. Bochet et al. (2006)
found that both face-to-face and verbal communication through a chat room (with anonymity and with no facial
expressions) have stronger effects on increasing contributions to public goods, than monetary punishment. Xiao &
Houser (2009) showed in an ultimatum bargaining game that monetary punishment is more effective than informal
sanctions (written messages) in promoting fairness. However, their results, taken together with Xiao & Houser
(2005), indicate that as a sanctioning mechanism expressing emotions is overall more beneficial for social welfare
than costly monetary punishment, as it eliminates (pecuniary) welfare losses. López-Pérez & Vorsatz (2010)
demonstrated that approval/disapproval (Note 3) feedback fosters cooperation in the prisoner’s dilemma games as
it increases awareness about the other’s feelings and because individuals are disapproval-averse. However, the
effect of positive and negative feedback is not the same. According to Dugar (2010) costless ratings of disapproval
assigned by group members lead to the most efficient equilibrium in the coordination game, whereas ratings of
approval cause just the opposite. Finally, Noussair & Tucker (2005) found that monetary and non-monetary
punishments are complements rather than substitutes as they appeal to different populations and at different time
horizons.
In our experiment we test the effectiveness of non-monetary reward and punishment through a discrete and
standardized expression of emotions in inducing a more equal distribution.
3. Experimental Design and Procedures
3.1 Design of the Downstream Water Pollution Game
We model the environmental protection context by using a framed laboratory experiment on downstream water
pollution. This is a three-player game. The first player is a farmer performing agricultural operations upstream
(henceforth called Upstream Farmer or UF) who decides on the usage of conservation technology. This
technology (called conservation tillage or CT, see Note 4) lessens the negative impact of farming on the water
11
quality in downstream rivers and lakes, but is more costly than the alternative intensive tillage (IT, see Note 5) for
the UF. The descriptions of agricultural technologies presented to the subjects in the instructions were significantly
modified and simplified to avoid any misunderstanding and ambiguity. The UF’s payoff ( ) in tokens depends
negatively on her usage of conservation technology, specifically the amount of land (out of her endowment of 500
acres) she places under conservation tillage ( ∈ 0, 500 ):
500 2 ∗ (500 ) (1)
The second player is an individual who is drawing her drinking water from the lakes and rivers downstream
(henceforth called Downstream Water User or DWU) and incurs the cost of cleaning the polluted water. The
DWU’s payoff ( ) in tokens positively depends on the cleanliness of the downstream lake:
500 10 ∗ (% ) (2)
The third player has a dual role: he performs farming operations upstream (thus deciding on the usage of
conservation technology) and, at the same time, lives on farms tied to rural water supply systems that draw water
from the very stream passing through his own farmland, and is, hence, using drinking water from downstream
(those individuals are called Upstream Farmer/ Downstream Water User or UF/DWU). The UF/DWU’s payoff
( / ) in tokens negatively depends on his usage of conservation technology / ∈ 0, 500 , and
positively on the cleanliness of the lake:
/ 2 ∗ 500 / 10 ∗ (% ) (3)
The cleanliness of the lake is determined as the proportion of land placed under conservation tillage by the two
farmers (UF and UF/DWU, see Note 6):
/
% ∗ 100% (4)
The game represents a zero-sum game, where the total group payoff is independent of the payoff distribution: the
three players were sharing 3000 tokens. The structure of the payoffs was chosen so that each of the three players
gets an equal payoff of 1000 tokens if the farmers choose / 250 acres, which leads to
% 50% (for selected combinations of strategies and payoffs see the top half of
Table 1).
The Nash equilibrium (conditional on the standard assumption of profit-maximizing behavior) for both UF and
UF/DWU is to choose a zero level of conservation technology since for the UF/DWU the return on conservation
technology is twice the return on the lake cleanliness. In this sense the decision of UF and UF/DWU to use
non-zero levels of conservation technology is similar to altruistic giving in a dictator game.
Table 1. Selected strategies and payoffs of the players

Upstream Farmer/ Downstream Water User
EMPATHY FRAME CTUF/DWU =500 CTUF/DWU =250 CTUF/DWU =0
CTUF =500 500, 1000, 1500 500, 1250, 1250 500, 1500, 1000
Upstream Farmer CTUF =250 1000, 750, 1250 1000, 1000, 1000 1000, 1250, 750
CTUF =0 1500, 500, 1000 1500, 750, 750 1500, 1000, 500
SELF-INTEREST
ITUF/DWU =500 ITUF/DWU =250 ITUF/DWU =0
FRAME
ITUF =500 1500, 1000, 500 1500, 750, 750 1500, 500, 1000
Upstream Farmer ITUF =250 1000, 1250, 750 1000, 1000, 1000 1000, 750, 1250
ITUF =0 500, 1500, 1000 500, 1250, 1250 500, 1000, 1500
Description: The table provides selected decisions of the UF and UF/DWU and their outcomes in EMPATHY FRAME and SELF-INTEREST FRAME
treatments. The numbers in each cell represent the payoffs of the three players given the choices by UF and UF/DWU. The payoffs are written
in the following order: UF, UF/DWU, DWU.
3.2 Experimental Treatments

We implemented three manipulations: two framing (empathy and self-interest) treatments using loaded language
and one no-framing treatment using neutral language. In the framing treatments the players were referred to as UF,
DWU, and UF/DWU. They were presented with the description of the upstream-downstream situation and asked
to make decisions regarding the choice of tillage on their land. In the NEUTRAL FRAME treatment the players were
12
www.ccsennet.org/ijef Inteernational Journaal of Economicss and Finance Vol. 5, No. 3; 2013
referred too as Player 1, 2,

2 and 3, and w
were presented w
with a situatioon described inn a neutral langguage in which
h they
had to alloocate chips betwween Options A and B.
The two frraming treatmeents differed inn how the situattion was presennted to the subj bjects. In the EM
MPATHY FRAM ME the
context waas described frrom the point oof view empathhetic to the DW WU and the paarticipants had to choose the level
of conservvation tillage thhat will affect thhe cleanliness of the lake as ddiscussed in thhe previous subbsection. Accorrding
to Singer ((2009, p.252) “when we emppathize, we shhare another peerson’s feelinggs”. In line witth that our emp pathy
(Note 7) frraming stressess the effect of tthe farmer’s chhoice on the w water user’s outtcomes and thrrough that incre eases
the undersstanding and seensitivity of th he former to thee feelings of thhe latter.
In the SELLF-INTEREST FRAME
R the context was descriibed from a self-interest/proffit-maximizingg point of view w and
the particippants had to chhoose the levell of intensive ttillage (IT, lesss environmentaally friendly teechnology) that will
affect the pollution of thet lake. The description unnderlining selff-interest refleects the traditiional public policy
framing inn which monetaary profit-maxximization by ffarmers is a deefault assumptiion of conservvation policies. This
thus posess a natural commparison point for the empathhy frame.
The payofff of the three players
p in the SELF-INTERESST FRAME was mirroring the payoffs in thee EMPATHY FRAME R
and was caalculated basedd on IT and % pollution of thhe lake. The deecisions of the farmers in twoo treatments caan be
converted as: 5000 and % 100% % . Sim
milar
to the EMPPATHY FRAME, in SELF-INTEEREST FRAME tthe players gett an equal payyoff of 1000 tookens if the farrmers
choose / 250 accres which leeads to % 500% (for sele ected
combinatioons of strategiies and payofffs see the botttom half of Taable 1). To sim mplify the anaalysis, all data
a was
converted and the resultss below will bee discussed in terms of CT aand the cleanlinness of the lakke.
3.3 Roundds of the Game and Emotionaal Feedback
The game was played foor two rounds. In each rounnd the UF and UF/DWU (orr Player 1 and 2 in the NEUT TRAL
FRAME maanipulation) chose
c the acreeage under CT T (EMPATHY FRAME) or IT (SELF-INTEREEST FRAME) or the
number off chips to place into Option B (NEUTRAL FR RAME). Next, thhe DWU receivved informatioon about the lev
vel of
cleanlinesss or pollution of
o the lake, buut not the actuaal CT/IT choicees of UF and U
UF/DWU (Notte 8).
At the endd of the first roound the DWU U was given thee opportunity tto express her positive or neegative emotion ns by
sending em moticons: a sm miley/happy facce  or a frow wney/unhappy face  to bothh farmers (Notte 9). Sending each
face was ccostly for the DWU
D (50 tokenns). If the emooticons were seent, they appeaared on the screeen of both farrmers
(Figure 1aa, b). In order too minimally poollute the data,, the payoffs off subjects weree not shown to them until afte
er the
game. Thee UF and the DWUD could havve potentially ccalculated theiir earnings, butt not the UF/D
DWU as she wa as not
aware of tthe cleanlinesss of the lake. We specifically did not waant to show thhe cleanliness of the lake to o the
UF/DWU,, as from the cleanliness
c theey could have inferred the C CT of the UF,, and this know wledge could have
affected thheir decisions ini the second rround. After thhat the game prroceeded to thee second (and last) round.
Figure 1. a,
a b. Smiley annd frowney em
moticons receivved by the farm
mers from the D
DWU
Description: These are the z-Tree
z screen shoots that were shoown to the UF annd UF/DWU if the DWU, after observing the level of
cleanliness/ppollution of the lakke, sent them (a) – a smiling or (b) – a frowning face.
Although oone-shot gamees are preferablle for this typee of analysis beecause they elimminate strategiic consideratio
ons of
players in the repeated game
g context, we believe that our two-rouund game achiieves the samee objective. Ne either
within onee round nor acrross the two roounds were theere any strategiic consideratioons: the decisioon of the UF afffects
the payofff of the UF/DW
WU, but not vicce versa; the deecision of the U UF affects thee payoff of the DWU, but nott vice
13
versa; finally, the decision of the UF/DWU affects the DWU’s payoff, but not vice versa (in the instructions the
players were not told that the DWU will be able to send an emotional feedback and, thus, could not have reacted
strategically to avoid/solicit it).
3.4 Hypotheses
The hypotheses deal with the framing effect, the difference between the decisions of UF and UF/DWU, and the
effect of the discrete emotions expressed by the DWU towards the decisions of the UF and the UF/DWU.
The framing effect, if present, will affect two decisions: the usage of conservation technology by the farmers and
the likelihood to send a smiley or frowney emoticon by the water user. The null hypothesis is that there is no
difference neither between the EMPATHY and SELF-INTEREST FRAME nor between the loaded manipulations and
the NEUTRAL FRAME. For the farmers’ decisions, the alternative hypothesis is that the framing affects farmers’
choice of the CT level and the following pattern will be observed:
∙
. The rejection of the null
hypothesis would be in line with findings reported in similar situations by Andreoni (1995), Park (2000), Fujimoto
& Park (2010), Cubitt et al. (2011), and Duffy & Kornienko (2010). As for the likelihood to send an emoticon, the
alternative hypothesis is that the loaded framing will result in a higher propensity to send a face than the NEUTRAL
FRAME, and there will be relatively more smileys in the EMPATHY FRAME and relatively more frowneys in the
SELF-INTEREST FRAME. However, in the case of emotional response, it is reasonable to expect that the null
hypothesis will not be rejected as the DWU is more likely to be affected by the lake cleanliness and react on that,
rather than on framing. This expectation is supported by the findings of Cubitt, Drouvelis & Gächter (2011), who
showed that the propensity to punish in the public goods game is driven by the contributions rather than by the
framing effect.
From the self-interest perspective and even accounting for empathy towards the DWU (since both farmers equally
affect the lake cleanliness), the decisions of UF and UF/DWU about conservation technology are conceptually
equivalent. The null hypothesis is that there will be no difference between the CT chosen by the two farmers within
one treatment. The alternative hypothesis is that within one treatment / . The UF/DWU, by the
nature of the situation, is better able to empathize, to “walk-in-the-shoes” of the DWU, while the UF can only
imagine how it would feel like. In that sense, the position of the UF/DWU is similar to a situation in which people
play non-symmetric games with switching roles [following the discussion by Brosig et al. (2003) about the
comparison of simple bargaining game and a game with switching roles]. The UF/DWU is more likely to engage in
self-reflection and empathetic considerations and will have more realistic beliefs on how it feels like being a
DWU. The beliefs of the UF, on the other hand, may be less realistic and, thus, their decisions are expected to be
less empathetic. In this case we would expect to reject the null hypothesis in favor of the alternative. A rejection
would be corroborated by the difference between bargaining and dictator games in which the players are switching
roles and the games in which they do not [see the discussion in Brosig et al. (2003, p.85) comparing their results to
Brandts & Charness (2000)].
Finally, consider the effect of positive or negative emotions expressed by the DWU on the decisions of the UF and
UF/DWU in the second round. The null hypothesis is that the emotional response will have no effect on the
farmer’s choices. The alternative hypothesis is that both smiley and frowney faces will encourage the farmers to
increase their CT usage, with frowney being more effective in achieving that goal. As positive and negative
emotions are shown to be linked to monetary rewards and punishments (Offerman, 2002; Bosman & van Widen,
2002) and those monetary rewards and punishments affect behavior, it is reasonable to expect that we will reject
the null hypothesis. Along the same lines, as punishments were shown to be more effective than rewards (Andreoni
et al., 2003; Dickinson, 2001; Sefton et al., 2008; Sutter et al., 2010; van Soest & Vyrastekova, 2006), we would
expect that expressing negative emotions will increase CT usage more than expressing positive emotions.
3.5 Procedures and Subjects
Each subject was assigned a 5-digit random number to assure anonymity. The experimental instructions were read
to the participants aloud and also presented on their computer screen. In addition, each subject received two
handouts: a 11X11 payoffs table with some of the possible combinations of strategies and a list of formulas to
calculate the payoffs. Following the instructions the experimenter answered questions and the participants were
presented with a quiz checking their understanding of the instructions and the calculation of the payoffs. After all
subjects successfully completed the quiz they proceeded to the experiment.
In total, 216 subjects participated in the experiment: 84 in each of the EMPATHY and SELF-INTEREST FRAME and 48
in the NEUTRAL FRAME treatments, resulting in 28, 28 and 16 independent observations per treatment respectively.
All subjects were recruited at the University of Nebraska-Lincoln (the majority were students, 45% females,
14
average agge was 27 yeears). The expperiment was conducted in the Experimeental and Behavioral Econo omics
Laboratoryy. All sessions were computeerized and adm ministered usingg the software z-Tree (Fischbbacher, 2007). Each
session waas 70-100 minuutes long. The tokens that thee participants eearned during tthe experimentt (sum of payoffs in
both roundds) were conveerted into dollarrs ($1=70 tokeens) and paid too the participannts privately inn cash, with ave
erage
earnings oof $28.9 (Note 10).
4. Experim
mental Resultts and Discusssion
4.1 Decisions of the UF and UF/DWU
U in the First R
Round
4.1.1 Lakee Cleanliness and
a Conservatiion Technologgy
The decisiion of the farmmers about the conservation technology ledd to quite highh levels of lakee cleanliness in the
first roundd: 46.3%, 34.7% % and 29.4% in the EMPATH HY, NEUTRAL, and SELF-INT TEREST FRAME E, respectively. This
result is inn line with whhat one would expect given the findings reeported in the previous literrature, as discu
ussed
above. Acccording to the non-parametric Kruskal-W Wallis test (χ2(2)=7.74, p-vaalue=0.02) theere is a signifficant
difference between the treatments.
t Thhe analysis of each pair of ttreatments usinng the Mann-W Whitney-Wilcoxon
rank-sum test indicated that the onlyy significant (aat 5% or less)) difference iss between thee cleanliness in n the
EMPATHY and the defaullt position reprresented by thee SELF-INTEREEST FRAME (p= =0.00899).
Result 1. E
Empathy framiing leads to higgher usage of cconservation teechnology as ccompared to seelf-interest fram
ming.
Based on tthe Kruskal-W
Wallis test the trreatment effectt on the levels oof conservationn tillage chosenn by the UF an
nd the
fferent (χ2(2)=112.26, p-value=
UF/DWU (Figure 2) waas statistically ssignificant diff =0.0023). Simmilarly to abovee, the
Mann-Whhitney-Wilcoxoon rank-sum ttest showed a significant diifference onlyy between the CT by UF in n the
EMPATHY and SELF-INTEEREST FRAME (p=0.0005). A As for the CT leevels that weree chosen by thee UF/DWU, ne either
the Kruskaal-Wallis nor the
t Mann-Whiitney-Wilcoxoon tests indicatted statisticallyy significant ddifferences betw ween
the treatmeents.
Figgure 2. Conserrvation Tillagee (CT) chosen bby the farmerss in round 1

Description: The figure reflectts the decisions off the UF and UF/D
DWU in Round 1 ddepending on the treatment and thee type of the farme
er. The
farmers coulld place a maximuum of 500 acres under Conservatiion Tillage. If botth farmers place 2250 acres for CT
T, all three playerss (UF,
UF/DWU, annd DWU) will gett an equal payoff oof 1000 tokens.
Overall thhis leads to thee rejection of tthe null hypotthesis regardinng the absencee of a framingg effect on the lake
cleanlinesss, with the em
mphasis that thee difference inn the lake cleaanliness is drivven by the UF’s behavior. In
n this
game the ratio of UF to t UF/DWU iis 1:1, and onne would expeect more pronnounced differrences between n the
manipulatiions the more the ratio in thee community iss skewed towaards UF.
4.1.2 Diffeerence betweenn UF and UF/D
DWU Choicess
The hypotthesis regardinng no differencce between thee decisions of UF and UF/DDWU is rejecteed in one treattment
and not rej
ejected in two others. As is eevident from F
Figure 2 there is hardly any difference bettween the leveels of
conservation tillage chossen by UF andd UF/DWU in the EMPATHY Y FRAME. The 2 238 and 225 aacres of land placed
under the conservation tillage
t by UF and UF/DWU U resulted in appproximately eequal payoffs for UF, UF/D DWU,
and DWU: 1024, 1013, and a 963 tokenns respectively.. In the NEUTR RAL FRAME thee difference is noticeable (20
03 for
UF/DWU and 144 for UF), howeverr, not statisticcally significannt according tto the Mann-W Whitney-Wilcoxon
15
rank-sum test. SELF-IN NTEREST FRAM

ME is the onnly case wheere the differrence is statisstically signifficant
(Mann-Whhitney-Wilcoxxon rank-sum ttest p-value=0.0369) with U
UF/DWU placiing 196 and UUF placing 98 acres
under consservation tillagge.
Result 2. T walking-in-the-sshoes-of-otherrs” tempers sellf-interested coonsiderations.
The forced “w
This is in line with the conjecture
c by Singer & Fehhr (2005, p.3433) that “the veery ability to empathize may y also
underminee purely self-iinterested choiices”. This result also provvides further ssupport for thee dual-interestt and
metaeconoomic framewoork (Hayes & Lynne, 20004; Lynne, 20006). It suggests that peopple are imagining
themselvess in the state of others in tthe communitty, and, as a rresult, are connditioning theiir own interna alized
pursuit of self-interest with
w the shared other-interest. This result, taaken together with the aforementioned framing
effect, sugggests that undder the self-intterest frame raaising the stakees of the upstrreamers in the downstream water
w
quality (suuch as the casee with UF/DWU more equal payyoffs distributiion with a maggnitude comparable
U) inspires a m
to empathyy framing. In that
t sense salieent incentives aare as powerfuul as empathy cconsiderationss. In contrast, under
u
the empathhy frame this strategy does not lead to a significant chaange in behavvior, because thhe subjects alrready
have reachhed a near equuitable distribuution and “wallking-in-the-shhoes-of-others”” does not neeed to “temper” ” any
longer.
4.2 Expresssion of Emotioons by the DW
WU
Downstreaamers demonsstrated substanntial willingneess to (costly)) express theiir emotions: ssmiley or frow wney
emoticons were sent inn more than 660% of the ccases in the fframing treatm ments (consisteent with impo osing
monetary rrewards and puunishments in the Carrot-Stiick treatment iin Andreoni et al. (2003)) annd 56% of the cases
c
in the neuttral scenario. In
I the SELF-INTTEREST FRAMEE treatment 9 oout of 10 emotticons sent weere frowneys, while
w
in the EMPPATHY FRAME only 6 out of 10 were frownneys (Figure 3)).
Figure 3. Perrcentage of DW
WUs sending aan emotional feeedback
Description: The figure reflectts the proportion of the DWUs whoo sent to the farm
mers a smiley or a frowney in Rounnd 1 in each of the
e three
treatments. S
Sending a face cost the DWU 50 tokkens.
Positive emmotions were expressed whhen the cleanlliness of the laake was 50-700%, depending on the treattment
(Figure 4),, suggesting thhat the distribbution of payooffs has to be biased in favvor of the DWWU to trigger their
positive emmotions througgh . There is also a spreadd of about 20 percentage pooints of lake ccleanliness betw
ween
the treatm
ments in the caases when emootions were noot expressed. H However, therre is a consenssus regarding what
must be puunished by neggative emotionns – a cleanlineess of the lake of, on averagee, between 19.7% and 26.8%
% will
result in  being sent.
we constructedd multinomial logit
To test forr the statisticall significance oof the differennces between thhe treatments w
regressionns in which wee controlled foor the cleanlineess of the lakee (Table 2). T The dummies ffor treatment in n the
linear form
m (Model 2) as a well as an innteraction withh cleanliness ((Model 3) are not significannt. Given this result
r
we cannott reject the nulll hypothesis oon the absencee of a framing effect on sending emotionaal feedback. In n line
with the fiindings reporteed by Cubitt, D Drouvelis & G Gächter (2011),, in our experim
ment emotionaal punishments and
rewards arre based on the more immeddiate payoff-reelevant inform mation (such ass lake cleanlineess), rather tha
an on
framing.
16
Figure 4. Average
A cleanliiness of the lakke that triggerss emotion exprression by treaatment
Description: The figure reflectts the level of cleannliness of the lakee in Round 2 in thee three treatments tabulated by the fface sent by the DW
WU to
the farmers.
Multinomial loggit model withh the dependennt variable “Possitive or Negattive Emotions”” ( or )
Table 2. M
Model 1 Mod del 2 Model 3
INTERCEPT
 -3.19*** -2.800** -3.29***
 2.30*** 1.588* 2.55***
CLEANLINESS
 4.77** 5.466** 6.08**
 -6.41*** -7.177*** -9.98***
EMPATHY FRA AME
 -0.779
 1.333
SELF-INTERESTT FRAME
 -1.553
 1.222
CLEANLINESS x EMPATHY FRAM ME
 -0.95
 3.30
CLEANLINESS x SELF-INTEREST FRAME
 -2.37
 3.73
Log-Likelihoodd -56.7 -533.6 -54.6
Nagelkerke R2 0.45 0.550 0.49
Description: Significance levells: *** - 1%, ** - 5%, * - 10%. Thee table shows the rresult of the regressions estimating the influence of th
he
treatment andd the level of lake cleanliness on thee expression of em
motions (smiley orr frowney) by the DWU.
Cleanlinesss of the lake isi the only varriable that affeects the expresssion of emotions (Table 2).. Positive emootions
(in the forrm of ) are expressed
e signnificantly (at 5%-level) less often than no emotions, as iis evident from
m the
negative sign in front off INTERCEPT ffor  in all thhree models. H However, the ddownstreamerss are quite eag ger to
express thheir negative em for ). The cle
motions (posittive and signifficant coefficieent in front off INTERCEPT fo eaner
the lake is, the higher is the probabbility that the DWU expressses positive eemotions (in Model 1, for lake
cleanlinesss the odds ratiio of sending  over not exxpressing emottions is 1.049, see Note 11). Consequently y, the
cleaner thee lake is, the loower is the proobability that the DWU exppresses negativve emotions (inn Model 1, forr lake
cleanlinesss the odds ratioo of sending  over not exprressing emotioons is 0.938).
Result 3. T
The expressionn of emotions is driven by thhe outcomes annd it is indepenndent of framinng.
These finddings suggest that
t subjects aare eager to exxpress directlyy their emotionns even thoughh this expression is
costly for them. Positivee emotions aree expressed whhen the outcom mes are in theirr favor (higherr than equal sh
hare).
Negative eemotions are expressed
e wheen the resultingg payoff falls below 30% off the equal payyoff. Expressio on of
emotion iss motivated by the payoff, raather than by frraming.
17
4.3 Effectiiveness of Emootions: The UF

F and UF/DWU
U Choices in thhe Second Rouund
The comparison of the effectiveness
e oof positive andd negative emootions in prom
moting environm mentally consc cious
behavior leed to (perhaps) an unsurprisiing result: the expression of negative emottions worked w well and betterr than
the expresssion of positiive emotions. This is in linne with findinngs that punisshments are m more effective than
rewards (A a 2003; Dicckinson, 20011; Sefton et aal., 2008; Suutter et al., 20010; van Soest &
Andreoni et al.,
Vyrastekova, 2006). Thee cleanliness oof the lake in the second rouund increased,, on average, bby 14.7 percen ntage
points in rresponse to .
 Expressing ppositive emotiions, on the otther hand, bacckfired: the faarmers realized d that
they would be able to squeeze
s out a little more proofit and as a rresult the cleaanliness of the lake decreaseed on
average byy 6.0 percentagge points in reesponse to . The cleanliness stayed apprroximately the same (on ave erage,
0.6 percenntage points redduction) if the emotions werre not expresseed. Intriguinglyy, the reaction of the two farrmers
on emotionns was quite different
d (Figurre 5): while the frowney emooticons led to a similar increease in the usage of
conservation technologyy, smiley emotticons negativeely affected thhe UF/DWUs, but not UFs. T This rather striking
result sugggests that posiitive emotionss signaled thatt the rewardingg party was satisfied with tthe outcome which
w
encouraged the rewardedd party (UF/DW WU) to yield tto the temptation to be more selfish in the second round.
Figuure 5. Absolutee change in thee usage of consservation technnology in respoonse to emotioonal feedback
Description: The figure reflectts the absolute chaange in the level oof Conservation T
Tillage (CT, in acrres) chosen by thee farmers in Roun
nd 2 as
compared to Round 1.
The regression analysiss (Table 3) prrovides a morre detailed piccture regardinng the influencce of emotion ns on
outcomes in the second round. The deependent variaable in Modelss 4 & 5 is the absolute diffeerence between the
cleanlinesss of the lake inn the second annd the first rouunds. Only negative emotionss can explain thhis difference – the
coefficientt in front of thee dummy for  is significannt at 1% (Modeel 4), whereas tthe one for  iis not. The ana alysis
of the interraction betweeen the treatmennts and emotionns (Model 5) rreveals that neggative emotionns play a signifficant
role only inn the SELF-INTTEREST FRAMEE (as the rest off the coefficiennts are not signnificant). In thiss treatment sen
nding
a frowningg face led, , on average, to an increase in thee lake cleanliness by 21.8 perrcentage points. Coupled witth the
fact that thhe first round average
a cleanlliness was 29.44%, this suggeests that the exxpression of neegative emotio ons is
extremely effective and can, on its ownn, guarantee thhat the DWU ggets their ‘fair//equal’ share oof 1000 tokens. The
seeming inneffectiveness of the emotionnal sanctions aand rewards inn the EMPATHY Y FRAME was, most likely, driven
by the disttributions of the average payooffs in the firstt round which w
were already aalmost equal (1024, 1013, and d 963
tokens resppectively). In the
t second rouund emotional feedback was therefore unabble to sway thee farmers to giv ve up
more of thheir payoffs inn favor of the D DWU (in the second round the average payoffs in EMPPATHY FRAME were
1018, 10166, and 966 for the UF, UF/D DWU, and DW WU, respectivelyy).
Models 6--9 are estimatted with the ddifference betw ween the usagge of conservaation technology by the UF F and
UF/DWU in the second anda the first roound as the deppendent variablle. As shown aabove, for bothh farmers a neg
gative
emotional feedback led to
t a significantt increase in thhe CT usage onnly in the SELFF-INTEREST FRAAME. However, the
effect of  on UF/DWU U was more prronounced: cetteris paribus thhe UF/DWU rresponded by pplacing 128.1 acres
more undeer CT, while the UF increasedd CT “only” byy 92.3 acres. GGiven that each of the farmerss had a maximuum of
500 acres, those numberrs are quite higgh.
18
Table 3. T
Tobit regressionn with dependeent variables ““absolute changge in round 2 aas compared too round 1”
Modeel 4† Model 5† Mod
del 6‡ Mod
del 7‡ Moodel 8‡ Mo
odel 9‡
Dependent V
Variable Abbs. ch. Cleanlinesss Abs.ch. CT by UF
F A
Abs.ch. CT by UF/D
DWU
INTERCEPT -0.612 -0.6612 299.43 299.38 -334.62 -3
34.61
 (dummy) -5.350 -1.8888 -311.35 -16.88 -225.88 -2
2.885
 (dummy) 15.288*** -0.4408 41.30′ -19.58 1111.7*** 14.62
1
 x EMPATH
HY FRAME -10.36 -41.07 -6
67.76
 x SELF-INTTEREST FRAME 13..75 500.00 87.50
8
 x EMPATH
HY FRAME 13..34 400.20 93.18
9
 x SELF-INTTEREST FRAME 21.811*** 922.26* 12
28.1*
Log-Likelihoood -2455.8 -241.3 -3772.4 -3770.1 -3382.4 -3

379.4
Nagelkerke R2 0.18 0.228 0..06 00.12 00.17 0.23
Description: Significance levells: *** - 1%, ** - 5%, * - 10%, ′ - 116%. † - the depenndent variable is ccensored between -100 and 100; ‡ - the
dependent vaariable is censoredd between -500 annd 500. The table sshows the results oof the regression eestimating the inflluence of the treatment
and emotionss on the change in the cleanliness off the lake and the level of conservatiion tillage chosen by the farmers in R
Round 2 as compa
ared to
Round 1.
Result 4. E
Expressing neggative emotionns is effective in moving peoople towards m
more equal disttributions.
From the aactions of UFs and UF/DWU Us in the seconnd round we geet another conffirmation of the power of acttually
“walking-iin-the-shoes” ofo other peoplee: the UF/DWU Us are more reesponsive to soocial cues than the UFs in term
ms of
initial deciisions and reacction on the D
DWUs’ disapprroval. All but one coefficiennt in Models 6-9 are insignifficant
and thus definite conclussions cannot bee made. Neverrtheless, we obbserve that posiitive coefficiennts in Models 8 & 9
are larger than the resppective coefficients in Modeels 6 & 7, andd all but one nnegative coeff fficient are smaller,
suggestingg higher sensitiivity of UF/DW with a larger sample
WUs as compaared to UFs. It iis reasonable too expect that w
size our coonfidence in thhis conjecture w
would increasee.
Figuure 6. Averagee lake cleanlineess by round

Description: The figure compaares the average lakke cleanliness in th
the two rounds undder the three treatm
ments. A lake cleaanliness of 50% pro
ovides
the DWU witth a payoff of 10000 tokens.
On a largeer scale the eff

ffectiveness off expressing emmotions is commparable to thee empathetic fframing (Figurre 6).
Neither thhe non-parameetric Kruskal-W Wallis test (χχ2(2)=2.40, p-vvalue=0.30) noor the Mann-W Whitney-Wilcoxon
rank-sum test showed any a statisticallyy significant ddifference betw
ween the treatmments in the ssecond round. This
suggests thhat explicit em
motional feedbaack, even though not expresssed by everyonne, works simillarly to the imp
plicit
appeal to eemotions throuugh framing.
5. Conclusions and Imp
plications
Three impportant implicaations can be derived from ouur study. First,, our experimeent supports thaat priming-orie
ented
framing iss highly effecttive in changinng the behavioor of subjects.. This supports Singer & Feehr’s (2005, p.340)
argument that “empathhy renders ourr emotions otther-regardingg, which provvides the motiivational basis for
other-regaarding behavioor”. Empathy framing leadds to an almoost equitable ddistribution off income, whereas
self-interest framing leaads to more pprofit-maximizzing behavior. Both of thesse are differennt from the ne eutral
19
scenario. In line with the comment of a referee on the Alpizar et al. (2008, p.312) paper that “the framing effects do
not only cause problems; sometimes they may be seen as an asset for the researcher” we argue that rather than
avoiding framing in all experiments we should study its impact as interactions/communications in the field do not
happen in neutral language and without context. In particular our results support that environmental policy should
appeal to the empathy of polluters.
Second, perhaps not surprisingly, the subjects with a higher stake in the quality of the lake water contributed much
more to the cleanliness of the lake than those with no stake. We observe this effect despite the identical Nash
equilibrium for both types. This finding is in line with both, the neuroeconomics findings that the degree to which
individuals have empathetic feelings depends on the situational factors (Singer, 2009) and the dual-interest and
metaeconomics framework (Hayes & Lynne, 2004; Lynne, 2006) in that the UF/DWUs were placed into the
situation of “walking-in-the-shoes-of” downstream water users and, thus, displayed more empathetic behavior
than the UFs who could only imagine how it feels. In terms of policy this calls for increased exposure of
upstreamers to the negative externalities of their behavior. In contrast to a traditional approach of fully
internalizing the externality, our results support the contention that even a minor internalization, maybe even
symbolic in nature, may suffice to significantly change behavior. In addition, typically UFs will be downstream to
other negative environmental externalities, such as global warming, air pollution, and wind erosion. Pointing
out/educating about the similarities of the consequence of their own actions to other people and the impact of these
other negative externalities on themselves (asking them to project ownself into that dowsnstream situation) may
then help to change the upstreamer’s behavior.
Third, negative emotional feedback (non-monetary punishment), although costly, is an effective tool in positively
influencing environmental choices. As Xiao & Houser (2005, p.7401) noted “the desire to express emotions, and
constraints on that demand, are a ubiquitous feature of human social interaction” so it is worth utilizing this desire
in addition to the standard policies of monetary sanctions and rewards and putting more emphasis on the
dual-interest underlying such decisions. Our findings provide additional support for the effectiveness of policies
that promote social punishment and public shaming as strategies to achieve lawful and/or cooperative behavior.
Acknowledgements
The data collection was funded through the USDA-CSREES (program name changed to NIFA) National
Integrated Water Quality program. We greatly appreciate the financial support. The experiment was conducted in
the Experimental and Behavioral Economics Laboratory at the University of Nebraska-Lincoln. We wish to thank
Darin Dolberg (University of Nebraska-Lincoln) for computer support. We also thank Jonathan Bertin (University
of Michigan-Dearborn) for his assistance in administering the experiment.
References
Abbink, K., & Hennig-Schmidt, H. (2006). Neutral versus loaded instructions in a bribery experiment.
Experimental Economics, 9(1), 103-121. http://dx.doi.org/10.1007/s10683-006-5385-z
Alipizar, F., Carlsson, F., & Johansson-Stenman, O. (2008). Does context matter more for hypothetical than for
actual contributions? Evidence from a natural field experiment. Experimental Economics 11(3), 299-314.
http://dx.doi.org/10.1007/s10683-007-9194-9
Andreoni, J. (1995). Warm-glow versus cold-prickle: the effects of positive and negative framing on cooperation
in experiments. Quarterly Journal of Economics, 110(1), 1-21. http://dx.doi.org/10.2307/2118508
Andreoni, J., Harbaugh, W., & Vesterlund, L. (2003). The carrot or the stick: rewards, punishments, and
cooperation. American Economic Review, 93(3), 983-902. http://dx.doi.org/10.1257/000282803322157142
Balliet, D. (2010). Communication and cooperation in social dilemmas: a Meta-analytic review. Journal of
Conflict Resolution, 54(1), 39-57. http://dx.doi.org/10.1177/0022002709352443
Barr, A., & Serra, D. (2009). The effects of externalities and framing on bribery in a petty corruption experiment.
Experimental Economics, 12(4), 488-503. http://dx.doi.org/10.1007/s10683-009-9225-9
Ben-Shakhar, G., Bornstein, G., Hopfensitz, A., & Van Winden, F. (2007). Reciprocity and emotions in bargaining
using physiological and self-report measures. Journal of Economic Psychology, 28(3), 314-323.
http://dx.doi.org/10.1016/j.joep.2007.02.005
Benz, M., & Meier, S. (2008). Do people behave in experiments as in the field?—Evidence from donations.
Experimental Economics, 11(3), 268-281. http://dx.doi.org/10.1007/s10683-007-9192-y
20
Bochet, O., Page, T., & Putterman, L. (2006). Communication and punishment in voluntary contribution
experiments. Journal of Economic Behavior & Organization, 60(1), 11-26.
http://dx.doi.org/10.1016/j.jebo.2003.06.006
Bosman, R., & Van Winden, F. (2002). Emotional hazard in a power-to-take experiment. Economic Journal,
112(476), 147-169. http://dx.doi.org/10.1111/1468-0297.0j677
Brandts, J., & Charness, G. (2000). Hot vs. cold: Sequential responses and preference stability in experimental
games. Experimental Economics, 2(3), 227-238. http://dx.doi.org/10.1007/BF01669197
Brosig, J., Weimann, J., & Yang, C. (2003). The hot versus cold effect in a simple bargaining experiment.
Experimental Economics, 6(1), 75-90. http://dx.doi.org/10.1023/A:1024204826499
Cardenas, J. C., & Stranlund, J., & Willis, C. (2000). Local environmental control and institutional crowding-out.
World Development, 28(10), 1719-1733. http://dx.doi.org/10.1016/S0305-750X(00)00055-3
Carpenter, J., & Matthews, P. H. (2009). What norms trigger punishment? Experimental Economics, 12(3),
272-288. http://dx.doi.org/10.1007/s10683-009-9214-z
Cookson, R. (2000). Framing effects in public goods experiments. Experimental Economics, 3(1), 55-79.
http://dx.doi.org/10.1007/BF01669207
Cubitt, R. P., Drouvelis, M., & Gächter, S. (2011). Framing and free riding: emotional responses and punishment
in social dilemma games. Experimental Economics, 14(2), 254-272.
Cubitt, R. P., Drouvelis, M., Gächter, S., & Kabalin, R. (2011). Moral judgments in social dilemmas: How bad is
free riding? Journal of Public Economics, 95(3-4), 253-264.
http://dx.doi.org/10.1016/j.jpubeco.2010.10.011
Dickinson, D. L. (2001). The carrot vs. the stick in work team motivation. Experimental Economics, 4(1), 107-124.
http://dx.doi.org/10.1007/BF01669275
Duffy, J., & Kornienko, T. (2010). Does competition affect giving? An experimental study. Journal of Economic
Behavior & Organization, 74(1-2), 82-103. http://dx.doi.org/10.1016/j.jebo.2010.02.001
Dugar, S. (2010). Nonmonetary sanctions and rewards in an experimental coordination game. Journal of
Economic Behavior & Organization, 73(3), 377-386. http://dx.doi.org/10.1016/j.jebo.2009.11.003
Fehr, E., & Gächter, S. (2002). Altruistic punishment in humans. Nature, 415(6868), 137-140.
http://dx.doi.org/10.1038/415137a
Fischbacher, U. (2007). z-Tree: Zurich toolbox for ready-made economic experiments. Experimental Economics,
10(2), 171-178. http://dx.doi.org/10.1007/s10683-006-9159-4
Fujimoto, H., & Park, E. (2010). Framing effects and gender differences in voluntary public goods provision
experiments. Journal of Socio-Economics, 39(4), 455-457. http://dx.doi.org/10.1016/j.socec.2010.03.002
Gächter, S., & Fehr, E. (2000). Cooperation and punishment in public goods experiments. American Economic
Review, 90(4), 980–994. http://dx.doi.org/10.1257/aer.90.4.980
Hayes, W. H., & Lynne, G. D. (2004). Towards a centerpiece for ecological economics. Ecological Economics,
49(3), 287-301. http://dx.doi.org/10.1016/j.ecolecon.2004.01.014
Henrich, J., McElreath, R., Barr, A., Ensminger, J., Barrett, C., Bolyanatz, A., . . . Ziker, J. (2003). Costly
punishment across human societies. Science, 312(5781), 1767-1770.
http://dx.doi.org/10.1126/science.1127333
Hopfensitz, A., & Reuben, E. (2009). The importance of emotions for the effectiveness of social punishment.
Economic Journal, 119(540), 1534-1559. http://dx.doi.org/10.1111/j.1468-0297.2009.02288.x
Kühberger, A., Schulte-Mecklenbeck, M., & Perner, J. (2002). Framing decisions: hypothetical and real.
Organizational Behavior and Human Decision Processes, 89(2), 1162-1175.
http://dx.doi.org/10.1016/S0749-5978(02)00021-3
López-Pérez, R., & Vorsatz, M. (2010). On approval and disapproval: theory and experiments. Journal of
Economic Psychology, 31(4), 527-541. http://dx.doi.org/10.1016/j.joep.2010.03.016
Lynne, G. D. (2006). Toward a dual motive metaeconomic theory. Journal of Socio-Economics, 35, 634-651.
http://dx.doi.org/10.1016/j.socec.2005.12.019
21
Masclet, D., Noussair, C., Tucker, S., & Villeval, M. (2003). Monetary and nonmonetary punishment in the
voluntary contributions mechanism. American Economic Review, 93(1), 366-380.
http://dx.doi.org/10.1257/000282803321455359
Nelissen, R. M. A., & Zeelenberg, M. (2009). Moral emotions as determinants of third-party punishment: anger,
guilt, and the functions of altruistic sanctions. Judgment and Decision Making, 4(7), 543-553.
Nikiforakis, N. (2010). Feedback, punishment and cooperation in public good experiments. Games and Economic
Behavior, 68(2), 689-702. http://dx.doi.org/10.1016/j.geb.2009.09.004
Noussair, C., & Tucker, S. J. (2005). Combining monetary and social sanctions to promote cooperation. Economic
Inquiry, 43(3), 649-660. http://dx.doi.org/10.1093/ei/cbi045
Offerman, T. (2002). Hurting hurts more than helping helps. European Economic Review, 46(8), 1423-1437.
Ostrom, E. (2010). Beyond markets and states: polycentric governance of complex economic systems. American
Economic Review, 100(3), 641-672. http://dx.doi.org/10.1257/aer.100.3.641
Ostrom, E., & Walker, J. (1991). Communication in a commons: cooperation without external enforcement. In
Palfrey, T. R. (Ed.), Laboratory research in political economy (pp. 287-322). Ann Arbor: University of
Michigan Press.
Park, E. (2000). Warm-glow versus cold-prickle: a further experimental study of framing effects on free-riding.
Journal of Economic Behavior & Organization 43(4), 405-421.
Sally, D. (1995). Conservation and cooperation in social dilemmas: a Metaanalysis of experiments from 1958 to
1992. Rationality and Society, 7(1), 58-92. http://dx.doi.org/10.1177/1043463195007001004
Sefton, M., Shupp, R., & Walker, J. M. (2008). The effect of rewards and sanctions in provision of public goods.
Economic Inquiry, 45(4), 671-690. http://dx.doi.org/10.1111/j.1465-7295.2007.00051.x
Singer, T. (2009). Understanding others: brain mechanisms of theory of mind and empathy. In Glimcher, P. W.,
Camerer, C. F., Fehr, E., Poldrack, R. A. (Eds.), Neuroeconomics: Decision Making and the Brain (pp.
251-268). New York: Academic Press.
Singer, T., & Fehr, E. (2005). The neuroeconomics of mind reading and empathy. American Economic Review,
95(2), 340-345. http://dx.doi.org/10.1257/000282805774670103
Sutter, M., Haigner, S., & Kocher, M. G. (2010). Choosing the carrot or the stick? Endogenous institutional
choice in social dilemma situations. Review of Economic Studies, 77(4), 1540-1566.
http://dx.doi.org/10.1111/j.1467-937X.2010.00608.x
Van Soest, D., & Vyrastekova, J. (2006). Peer enforcement in CPR experiments: The relative effectiveness of
sanctions and rewards, and the role of behavioral types. In List, J. A. (Ed.), Using experimental methods in
environmental and resource economics (pp. 113-136). Cheltenham: Edward Elgar.
Walker, J. M., & Halloran, W. A. (2004). Rewards and sanctions and the provision of public goods in one-shot
settings. Experimental Economics, 7(3), 235-247. http://dx.doi.org/10.1023/B:EXEC.0000040559.08652.51
Xiao, E., & Houser, D. (2005). Emotion expression in human punishment behavior. Proceedings of the National
Academy of Sciences of the United States of America, 102(20), 7398-7401.
http://dx.doi.org/10.1073/pnas.0502399102
Notes
Note 1. For example, a number of studies of irrigation systems in Nepal as well as in Japan, India and Sri Lanka
and studies on forests all over the world reported about their effective farmer–designed systems (Ostrom, 2010).
Note 2. In the hot version of the game visceral factors (including strongly felt emotions) lead to a strong impulse to
punish unfair behavior and dominate the decision making process (Brosig et al., 2003).
Note 3. A subject could select to send one of the following three messages: “Your choice was (1) good, (2) neither
good nor bad, and (3) bad” (López-Pérez & Vorsatz 2010, p.532).
Note 4. The CT (minimal tilling of the land) was presented as a relatively lower profit practice. The CT was also
presented as the tillage with relatively mild environmental impact: under CT the land is disturbed minimally
leading to less soil erosion, lower chemical runoff, and overall higher drinking water quality of the downstream
rivers and lakes.
22
Note 5. The IT was presented as a relatively higher profit tillage practice for the agricultural operator. The IT was
also presented as the tillage with relatively higher negative environmental impact: IT leads to soil erosion and
greater chemical runoff, and, thus, significantly reduces the drinking water quality of the downstream rivers and
lakes.
Note 6. If each of the farmers will choose (independently) to place all their 500 acres of land under CT (
/ 500), then the lake cleanliness will be 100%.
Note 7. In Merriam-Webster dictionary empathy is defined as “the action of understanding, being aware of, being
sensitive to, and vicariously experiencing the feelings, thoughts, and experience of another of either the past or
present without having the feelings, thoughts, and experience fully communicated in an objectively explicit
manner” (http://www.merriam-webster.com accessed on 09/04/2011).
Note 8. Nikiforakis (2010) shows that the type of information that is presented to the subjects (e.g. contribution
versus earning of the fellow group members in the public goods game) matters for their play. Along similar lines,
as noted by the referee of Cubitt, Drouvelis & Gächter’s (2011) paper (footnote 6) it may be problematic to assume
that a player is able to differentiate his emotions towards the other two players individually. We showed to the
DWU only the outcome (lake cleanliness/pollution) of both choices (but not the actual CT/IT choices), so that this
outcome can be judged by the DWU against some “ideal” and/or expected outcome rather than the choices of the
two farmers being judged in comparison to each other. This also reflects the real life situation in which the
downstream water users can observe only the outcomes of the actions of the upstream farmers, but not their actual
actions/choices.
Note 9. Faces were identified as smiley/unhappy and frowny/unhappy based on the common usage of emoticons in
electronic communications. Google search for a respective image out of the first 24 faces, 21 (smiley) and 22
(happy) clearly resemble ; 19 (frowney) and 18 (unhappy) clearly resemble . Results of the search are available
from the authors upon request.
Note 10. This roughly corresponds to the incentive payments in recent experiments (e.g. Cubitt, Drouvelis &
Gächter, 2011; Duffy & Kornienko, 2010). Opportunity costs represented by the reported average hourly wage of
subjects were $8.95, which is calculated without an outlier – one subject recorded an hourly wage of $1,320.
Note 11. An odds ratio below one means that the event is less likely to occur than the reference event.
23
ISSN 1916-971X E-ISSN 1916-9728
The Stock Market/Unemployment Relationship in

USA, China and Japan
Farzad Farsio1,2 & Shokoofeh Fazel1,2
1
Zayed University, Dubai, UAE
2
Montana State University-Billings, Montana, United States
Correspondence: Farzad Farsio, Montana State University-Billings, United States. Tel: 1-406-657-2033. E-mail:
[email protected]
Received: December 25, 2012 Accepted: January 29, 2013 Online Published: February 22, 2013
Abstract
This study investigates the relationship between unemployment rate and stock prices in USA, China and Japan;
the top three world economies. Recently, there have been some articles by financial analysts asserting that
unemployment rate is a strong predictor of stock prices. They refer to certain short-term periods and posit a
negative causal relation from unemployment rate to stock prices. They argue that declining (rising)
unemployment would display an upturn (a downturn) in the economy, an increase (a decrease) in demand for
goods and services, and would therefore lead to higher (lower) profits and stock prices. In this paper, using
logical analysis, we argue that these views are misleading to potential investors. We hypothesize that there is no
stable long-term causal relationship from unemployment rate to stock prices. Furthermore, using quarterly data
in US, China and Japan over the 1970-2011 period, we provide empirical support for our hypothesis. The
empirical analysis of this paper is based on cointegration and Granger Causality tests. Our findings have one
important implication: it would be a mistake to rely on unemployment rate data to make investment decisions in
the stock market.
Keywords: unemployment, stock prices, causal relation
1. Introduction
This study focuses on the relationship between unemployment and stock prices in US, China and Japan. The
objective of this study is to present logical and empirical evidence against views presented by some financial
analysts (see for example, Miller, 2010 and Wojdylo, 2009) of a stable negative causal relation from
unemployment rate to stock prices. These views assert that rising (falling) unemployment rates are followed by
falling (rising) stock prices, and that unemployment rate can be used to predict stock prices. Referring to
short-term periods, it is rationalized that rising (falling) unemployment leads to a(an) decrease (increase) in
demand for goods and services, and as a result, firms’ revenues, profits, and stock prices will decline (increase).
In this study, we present logical and empirical evidence against these views and hypothesize that unemployment
rate and stock prices do not hold any stable long-term relationship and that there is no causal effect from
unemployment to stock prices. Our findings have one important implication: it would be a mistake to rely on
data for unemployment rate forecasts and trends to make investment decisions in the stock market. The empirical
analysis of this study is focused on US, China and Japan, covers the 1970-2011 period, and is based on
cointegration, and Granger causality tests.
Many academic scholars have examined the relationship between unemployment and stock prices. Blanchard
(1981) showed that in equilibrium, the same news about unemployment can sometimes be good and sometimes
bad for financial assets, depending on the state of the economy. Orphanides (1992) gave empirical support for
this view by showing that stock price responses to macroeconomic news may depend on the state of the
economy. In particular he showed that the stock price response to unemployment news depends on the average
unemployment rate during the previous year. McQueen and Roley (1993) found a strong relationship between
stock prices and macroeconomic news, such as news about inflation, industrial production, and the
unemployment rate based on their own definition of business conditions. However their purpose was to
demonstrate the dependence of stock price responses to all macroeconomic news. Krueger (1996) studied the
market rationality of bond price responses to labor market news. His focus was on the market reaction to the
24
availabilityy of more reliable informatiion, as the uneemployment ddata were revissed. His studyy found that market
prices werre strongly afffected by thee unemploymeent announcem ments. Veroneesi (1999), baased on theore etical
argumentss, showed that bad news in ggood times andd good news inn bad times woould generally be associated with
increased uuncertainty annd hence an inccrease in the eequity risk premmium investorrs require for iinvesting in stocks.
Jagannathaan and Wang (1993) foundd that monthlyy stock returnss are negativelly correlated w with the per capita
c
labor incom me growth rate. Jagannathann, Kubota and Takehara (19998) report simiilar findings using Japanese data.
Since mosst of the variattion in per cappita labor incoome arises fromm variation in hours workedd and not the wagew
rate, thesee findings are consistent witth the positivee correlation bbetween the grrowth rate in uunemploymentt and
stock returrns.
Most of thhe above studiies attempt to display a cauusal relation frrom unemployyment rate to sstock prices. While
W
either a neegative or a poositive correlaation may exisst during certaain periods, it would be a m
mistake to assume a
stable longg-term causal relation
r from uunemploymentt rate to stock prices. Such aassumption maay mislead pote ential
investors tto believe that movements inn unemploymeent rate can prredict movemeents in stock pprices. The factt that
both unemmployment andd stock prices aare endogenouus variables annd that their m movements depend on a varie ety of
exogenouss variables, makes
m it essenttial to conduct deeper analyysis in identiffying the preddominant causes of
movementts in these variiables before aassuming any ccausal relationnships. This stuudy presents a logical analyssis of
the relatioonship and shoows that no sstable long-term m causal relaation from uneemployment raate to stock prices
p
should exist. We will thent provide empirical resuults supporting this view. Our results have one impo ortant
implicationn: it would be a mistake to rrely on movem ments in unempployment rate to make invesstment decision ns on
stocks. In this paper, foollowing logiccal analysis off the relationshhip in 2, sectiion 3 presentss methodologyy and
empirical rresults which are
a followed bby concluding rremarks in secction 4.
2. Method
d
Unemployyment rate andd stock prices are both endoogenous variabbles whose levvels and movem ments depend on a
variety of exogenous faactors. Before positing any ccorrelation bettween these enndogenous varriables, one sh hould
first analyzze and identifyy exogenous faactors that mayy affect these vvariables. Uneemployment raate is defined as
a the
ratio of inndividuals whoo do not have a job to total number of inddividuals in thhe in the labor force. Labor force
does not innclude those who
w are not loooking for a jobb or are unablee to work. Anyy change in suppply or deman nd for
labor can cchange unempployment rate. Changes in w willingness to w work or abilityy to work willl change the su
upply
of labor annd thus unempployment rate. It is importannt to note that changes in wiillingness and ability to worrk for
already unnemployed inndividuals do not change tthe actual num mber of workking people bbut will chang ge in
unemploym ment rate. Forr example, if ssome unemplooyed individuaals quit lookinng for a job orr become disa abled,
there will be a reductioon of equal aamount in botth numerator aand denominaator of the unnemployment ratio, r
resulting inn lower unempployment rate. On the other hand, changess in technologyy, aggregate deemand, and intterest
rate can chhange the demaand for labor aand thus changge unemploym ment rate.
In order too identify facttors that can cause changes in stock pricees, we may anaalyze the Disccounted Cash Flow
model (DC CF). Accordinng to this com mmonly used model, a com mpany’s intrinssic value (exppected share price)
p
should be equal to the prresent value off its future cashh flows as show
wn in equationn (1).
(1)
Where:
Po = Intrinnsic Value (Exppected Share P
Price)
CF = Cashh Flow Per Shaare
r = Discouunt Rate.
The DCF model is one of the most ccommonly usedd models for stock valuatioon. CF represents free cash flow; f
cash flow available afterr all operating and investing needs of the ffirm are met. r is also called tthe required ra
ate of
return on tthe stock which is most oftenn measured byy the firm’s weeighted averagge cost of capittal (WACC). Based
B
on this moodel, future caash flows and the discount rrate that is useed to calculate the present value of future cash
flows are the two mainn factor determ minants of stoock prices. Higher future prrofits (cash fllows) and/or lower
discount raates (interest rates)
r increasee stock prices. These two maain factors of the stock valuuation model are a in
turn determmined by manny other factors. Profit is tootal revenue m minus total cosst. Total revennue is equal to o the
number off units of the good the firm m sells times the price of tthe good. Thee number of uunits sold (qua antity
demandedd) in turn depennds on the pricce of the good, consumers’ iincome as welll as income elaasticity of dem mand,
25
price of related products as well as their cross elasticity of demand, consumers’ taste, and number of buyers of
the product.
Total cost is another element affecting profits and thus the stock price. Wages and interest expenses are part of
total cost. Interest rate affects stock prices through two different channels. First, it affects the discount rate that is
used to estimate the present value of future cash flows. It is important to note that the discount rate (required rate
of return) also depends on stock’s beta coefficient and stock market’s expected rate of return (CAPM). The
second channel through which interest rate affects stock prices is its direct negative impact on cost of borrowing,
and thus total cost, and cash flows.
In short, the list of factors that may influence the stock price of a particular company should include the
following: the price of the good, consumers’ income, the product’s income elasticity of demand, price of related
products, the product’s cross elasticity of demand, consumers’ taste, number of buyers of the product, interest
rate, wages, other components of total cost, beta coefficient, and stock market’s expected rate of return. Some of
these factors can be classified as internal factors while others are grouped as external factors. For example, the
product’s income elasticity of demand and cross elasticity of demand are internal factors that can be controlled
by the management through efforts to enhance the quality of the product and advertisement. Non-interest costs
and beta coefficient are also controllable by firm’s management and are therefore internal factors. On the other
hand, in a competitive market, wages, price of the product, and prices of related products are determined
externally by forces of demand and supply. Also, consumers’ income, consumers’ tastes and preferences,
number of buyers (population), level of interest rates, and stock market’s expected rate of return are all out of the
control of managers and therefore considered external factors.
Based on the above lists of factor determinants of unemployment rate and stock prices, one could clearly deduct
that there might be many situations in which a change in a factor that causes unemployment rate to change may
not affect stock prices. As discussed earlier, changes in willingness and ability to work for already unemployed
individuals would change the unemployment rate without changing the actual number of working people and
will therefore have no impact on profits and stock prices. Even changes in technology, aggregate demand, and
interest rates may not always lead to a causal relation from unemployment to stock prices. For example, one
scenario in which a relationship (and not necessarily a causal relationship) may be created is when interest rates
change. According to the theory of investment, all else equal, a reduction in interest rates causes an increase in
firm’s investment and production, increases demand for labor, and thus leads to a decline in unemployment.
Lower interest rates means lowers interest expenses, lower total costs, higher profits and thus higher stock prices.
In this scenario, rising stock prices would be associated with (and not caused by) falling unemployment rate.
However, there are uncertainties associated with this scenario. There are ample of evidence suggesting that lower
interest rates do not always lead to lower unemployment. For one thing, if firms’ expectations about future
profitability and the overall state of economy are not positive, no matter how low interest rates are, they may not
increase their investment and as a result unemployment rate many not change at all. Even if firms’ expectations
are positive and they expand their investment and production, they may simply increase the working hours of
current employees instead of hiring more workers. In either case, unemployment rate remains unchanged while
stock prices increase. In addition to these uncertainties, there is another powerful uncertainty associated with the
“all else equal” assumption. It is rare if not impossible for everything else in the economy to remain unchanged.
Given the above analysis, one could conclude that the assumption of a stable causal relation from unemployment
to stock prices is illogical. Even if a short-term relationship existed in a certain period, it would not be a causal
relation. Instead, it would be the result of both endogenous variables reacting to one or more exogenous factors.
In short, we can posit that there is no causal relation from unemployment to stock prices. In the next section, we
present the results of our empirical analysis that support this view.
3. Results
For US and Japan, quarterly data for unemployment rates, S&P 500 index and the Nikkei Index from 1970-2011
were used. For China we used quarterly data for unemployment rate and the Shanghai index from 2002-2011.
Data prior to 2002 were not available for China. Data for all six variables were obtained from Trading economics
data base. Our empirical analysis includes three tests; the Augmented Dickey-Fuller test, the Engle-Granger
cointegration test, and the Granger causality test.
3.1 Augmented Dickey-Fuller Test
The first step in performing the cointegration test is to test for the presence of a unit root in the individual series.
To do so we employ the Augmented Dickey-Fuller test (1979). Tables 1-3 display the results of the Augmented
Dickey-Fuller tests. In each case, a log polynomial in first difference of the variable was taken out six periods to
26
render the residuals approximate white noise. Ljung-Box “Q” statistic was used to test the hypothesis that all of
the autocorrelations are zero. For all six variables, absolute values of calculated t-statistics were lower than the
MacKinnon critical values. Consequently, the null hypothesis of difference-stationary could not be rejected at
any standard significance level. Although this does not prove there are unit roots in each of the variables, the
consequences of specifying spurious deterministic trends convinced us that defining the variables as the first
difference in the logs was the prudent way to proceed.
3.2 Engle-Granger Cointegration Test
The concept of correlation in a growing economy is that of common stochastic trend or cointegration. Many
economic time series are not stationary. If, however, the first difference of a series is stationary, the original
series is said to be integrated of order one. As described in Engle and Granger (1987), two or more variables are
said to be cointegrated if individually each is non-stationary (has one or more unit roots) but there exists a linear
combination of the variables that is stationary. Table 4 reports the results of the Engle-Granger test for
cointegration between stock prices and unemployment rate in the three countries. The absolute value of
calculated t-statistic in below the conventional MacKinnon critical values for all three countries. That is, there is
no evidence of cointegration or common stochastic trends among stock prices and unemployment rate. This
contradicts the view that unemployment rate and stock prices hold a stable long-run relationship. The
cointegration results further confirm that unemployment rate does not have any long-run explanatory power in
predicting movements in stock prices, and provide support for the hypothesis presented in this paper.
3.3 Granger Causality Test
In performing the Granger causality tests, the hypothesized dependent variable is regressed on its lagged values.
The lag length in the regression equation must be selected in such a way that the resulting residuals are white
noise, and therefore any first order serial correlations are eliminated. Next, the lagged values of the hypothesized
independent variable are added to the right side of the regression equation and the new regression is executed.
Using an F test, the resulting sums of squared residuals from the two regression equations are compared. A
relatively large difference between the two sums of squared residuals (a large F) would provide evidence that the
hypothesized independent variable Granger causes the dependent variable. The Granger Causality test results for
the three countries are shown in Table 5. The small F statistics of the Granger Causality test (1.22,1.09, and 1.23)
which are significantly lower than the critical F value at the 5% confidence level (4.66 for US and Japan and
5.23 for China)) support the view that there is no causal relation from unemployment rate to stock prices in any
of the three countries under the study. Our test results conform well to our logical explanations presented earlier
that unemployment rate does not have any explanatory power in predicting changes in stock prices.
Table 1. Augmented Dickey-Fuller (ADF) unit root tests in USA 1970.1—2011.3

Variable ADF Coefficient t-stat
S&P 500 Index -.005 -.95
Unemployment -.01 -1.86
Table 2. Augmented Dickey-Fuller (ADF) unit root tests in Japan 1970.1—2011.3

Nikkei Index -.08 -.1.13
MacKinnon Critical Values (167 observations) 1% -4.12
5% -3.48
10% -3.19
Table 3. Augmented Dickey-Fuller (ADF) unit root tests in China 2002.1—2011.3

Shanghai Index -.004 -.94
MacKinnon Critical Values (39 observations) 1% -4.19
5% -3.68
10% -3.51
27
Table 4. Cointegration test between stock prices and unemployment

Variables Coefficient t-stat
S&P, Unemployment -.05 -1.27
Nikkei, Unemployment -.03 -1.12
Shanghai, Unemployment -.14 -1.26
MacKinnon Critical Values 1% -6.32
5% -4.43
10% -3.92
Table 5. Granger causality tests

Null Hypothesis: F-Stat Probability
Unemployment does not cause S&P causeS&P S&PTOCK PRICES S&P 1.22 0.31
Unemployment does not cause Nikkei causeS&P S&PTOCK PRICES S&P 1.09 .40
Unemployment does not cause Shanghai Shanghai Shanghai Shanghai Nikkei 1.23 0.30
Lag:3
4. Conclusion
The objective of the study was to investigate the assertion by some financial analysts that a negative casual
relation exists from unemployment to stock prices, and that unemployment rate can be used to predict future
stock prices. In this paper, we analyzed factor determinants of unemployment rate and stock prices, and
hypothesized that there would be no stable long-term causal relationship from unemployment rate to stock
prices. Furthermore, using quarterly data covering the 1970-2011 period, we provided empirical support for our
hypothesis in the three largest world economies. The empirical analysis of this paper was based on cointegration
and Granger Causality tests. Our findings have one important implication for investors: it would be a mistake to
rely on data for unemployment rate forecasts and trends to make investment decisions in the stock market.
References
Blanchard, O. J. (1981). Output, the Stock Market, and Interest Rates. The American Economic Review, 71(1),
132-143.
Boyd, J. H., Hu, J., & Jagannathan, R. (2005). The Stock Market’s Reaction to Unemployment News: Why Bad
News Is Usually Good for Stocks. Journal of Finance, 60(2), 649-672.
Chen, N. F., Roll, R., & Ross, S. A. (1986). Economic forces and the stock market. Journal of Business, 59,
383-403. http://dx.doi.org/10.1086/296344
Dicky, D. A., & Fuller, W. A. (1979). Distribution of the Estimators for Autoregressive Time Series with a Unit
Root. Journal of the American Statistical Association, 74(336), 427-431.
Engle, R. F., & Granger, C. W. J. (1987). Cointegration and Error Correction: Representation, Estimation, and
Testing. Econometrica, 55, 251-276. http://dx.doi.org/10.2307/1913236
Farsio, F. F., & Fazel, S. (2010). Does Unemployment Rate Predict Stock Prices? European Journal of
Management, 10(3), 97-102.
Gertler, M., & Grinols, E. L. (1982). Unemployment, Inflation, and Common Stock Returns. Journal of Money,
Credit and Banking, 14, 216-233. http://dx.doi.org/10.2307/1991640
Jagannathan, R., & Wang, Z. (1993). The CAPM is Alive and Well. Staff Report #165, Federal Reserve Bank of
Minneapolis.
Jagannathan, R., Kubota, K., & Takehara, H. (1998). Relationship between Labor-Income Risk and Average
Return: Empirical Evidence from the Japanese Stock Market. Journal of Business, 71(3), 319-347.
Krueger, A. B. (1996). Do Markets Respond More to More Reliable Labor Market Data? A Test of Market
Rationality. NBER Working Paper 5769.
Little, K. (Feb.10, 2010). Rising Unemployment Hurts Stocks. About.com:stocks.
McQueen, G., & Roley, V. V. (1993). Stock Prices, News, and Business Condition. Review of Financial Studies,
28
6, 683-707. http://dx.doi.org/10.1093/rfs/6.3.683
Orphanides, A. (1992). When Good News Is Bad News: Macroeconomic News and the Stock Market. Working
paper, Board of Governors of the Federal Reserve System.
Veronesi, P. (1999). Stock Market Overreaction to Bad New in Good Times: A Rational Expectations
Equilibrium Model. Review of Financial Studies, 12(5), 975-1007. http://dx.doi.org/10.1093/rfs/12.5.975
Wojdylo, J. (March 17, 2009). Best Prediction of the Stock Market and Economy-Unemployment rate. Subprime
Blogger.
29
ISSN 1916-971X E-ISSN 1916-9728
Saudi Financial Structure and Economic Growth:

A Macroeconometric Approach
Mohammed Moosa Ageli1 & Shatha Mousa Zaidan2
1
Assistant Professor of Economics, King Saud University, Riyadh, Saudi Arabia
2
Lecturer of Business Administration and Finance, Princess Nora University, Riyadh, Saudi Arabia
Correspondence: Mohammed Moosa Ageli, Assistant Professor of Economics, King Saud University, Riyadh,
Saudi Arabia. E-mail: [email protected]
Abstract
This paper investigates the nexus between financial sector development and economic growth in the Saudi
economy over the period 1970-2012 by using four alternative proxies for financial development and several
techniques including unit root tests, the co-integration test, the Granger Causality Test, and the Vector Error
Correction Model (VECM). We used time series econometrics techniques to examine the causal relationship
between financial sector development and economic growth in the Saudi economy. The results obtained from the
analyses show that there is a positive relationship between financial sector development and economic growth in
Saudi Arabia. The development of the financial system will thus have a positive impact on the growth of the Saudi
economy.
Keywords: financial sector development, unit root test, co-integration test, Vector Error Correction Model
(VECM), Augmented Dickey Fuller (ADF), economic growth
1. Introduction
The financial sector in Saudi Arabia is one of the most important sectors and acts as a device to encourage the
development process. The growth of the Saudi economy in the period of 1970-2012 was not homogeneous.
According to the main economic theory of developing countries, Saudi Arabia fits into this classification.
Some research has found that financial sector development is a major sector in developing countries, as
less-developed financial systems can affect the economic growth in those countries. On the other hand, other
research has found that a negative relationship exists between finance and growth in developed countries (Aghion
et al., 1999).
The role of financial expenditure in promoting economic growth remains a debatable subject in both developing
and industrial countries. The size and role of the public and financial sectors in the Saudi Arabian economy has
changed over the last thirty years since the 1980s when the public sector was the most prominent sector. This study
has the following objectives:
1) To examine the effect of the stock market index and the total exports on economic growth, taking into account
the positive effect of money supply and bank credit development on economic growth.
2) To apply the unit root test whether the variables are non-stationary.
3) To apply the johansen’s co-integration analysis to exmamines of a long-run equilibrium relationship.
The rest of this paper is structured as follows: section two will present some historical studies; section three will
derive the formulae and data; then, we will describe the methodology and results and perform the unit root tests
and Johansen’s co-integration test to confirm the uniqueness of the co-integration vectors among the variables
under study. Using annual data for Saudi Arabia for the period 1970-2012, we will investigate the relationship
between the development of the financial sector and the economic growth using appropriate estimation methods.
Finally, we will use the Error Correction Model (ECM) to test the short-run relationship between financial sector
development and economic growth (GDP) during the period 1970-2012. The final section will present the results
and the conclusions.
30
2. Literature Review
Previous economic studies have performed some case studies on this topic, thus this paper will review some
important related studies on the issue of financial sector development and economic growth in economic history.
Kiran et al (2009) used Fully Modified OLS (FMOLS) methods for ten emerging countries and found that there is
a positive relationship between financial development and economic growth using time series data over the period
1968–2007.
In the same line, Apergis et al. (2007) argued that there is a positive relationship between financial sector
development and economic growth and showed this using data for “15 countries of the Organization for Economic
Co-operation and Development (OECD) and 50 non-OECD countries”. In another study, Guryay et al (2007) used
the Granger causality test to present their results and found a positive effect of financial development on the
economic growth of Cyprus.
In another study, Neusser and Kugler (1998) analyzed the relationship between financial development and
economic growth for 13 OECD countries using time series for the period 1970-1991. Time series analysis showed
a positive correlation between financial development and growth.
A study by Rousseau and Wachtel (1998) in five industrialized countries found a positive and significant
relationship between financial development and economic growth. Levine et al. (2000) supported the view that
there is a positive correlation between financial system and economic growth using time series data for the period
1960-1995.
Sanusi and Salleh (2007) analyzed the casual relationship between financial development and economic growth in
Malaysia, using time series data for the period 1960-2002. The results indicated a positive relationship between
financial development and economic growth in the long run.
Fatima (2004) conducted a study in Morocco using the Granger causality test to investigate the relationship
between financial development and economic growth from time series data for the period 1970-2000. The study
found a positive relationship between financial development and economic growth. Ndebbio (2004) found a
positive impact of growth rate in per capita real money balances on real per capita GDP growth in African
countries.
Similarly, Khan et al (2005) found a positive impact on economic growth in the long run but found that the
relationship was insignificant in the short run by testing the link between financial development and economic
growth in Pakistan using time series data for the period 1971-2004.
Pradhan (2009) examined the causal nexus between financial development and economic growth in India using
monthly data for the period 1993-2008 in the VAR model. He found no relationship between financial
development and economic growth in India. He used the index of industrial production as a proxy to economic
development.
3. The Model and Data
To test the nexus relationship between financial sector development and economic growth, we use the following
model (equation 1):
GDPt  0  1STMI t   2 M 3t  3BCPSt   4 EX t  5BDt  6CCPSt  7 FD2002t  8 FD2007t  9 FD2010t  t
We have used Gross Domestic Product (GDP) as a proxy for economic growth, and four indicators and measures
of financial sector development, namely STMI, which measures the stock market index; M3, which is money
supply; BCPS, which is a measure of bank credits to the private sector from government specialized and banks;
EX, which is a measure of total exports; FD, which is a financial dummy variables; BD, which is a measured
bank deposits; and CCPS, which measures bank credits given to the private sector from commercial banks. The
data used in this study consist of the following variables (table 1):
31
Table 1. Variables definitions

Variable Definition
The symbol (LN) Natural Logarithm
GDP Gross Domestic Product
STMI Stock Market Index
M3 Money Supply
BCPS Bank credits to the private sector from government specialized and bank
EX Total Export
FD Financial dummy variables (2002) and (2007)
BD Bank Deposit
CCPS Bank credits to the private sector from commercial bank
The study will cover Saudi Arabia for the period from 1970 to 2012. The data sources are the Saudi Arabian
Monetary Agency (SAMA), the IMF and the International Financial Statistical Yearbook, (IFS). The time period
has started and included 1970 because it was the year when the capital market and the financial sector in Saudi
Arabia actually started and it was when the powerful influence of monetary policy in the Saudi economy
emerged. However, time series data in this paper is reported annually, and so we used the natural logarithm with
the variables. The FD dummy variable is used for 2002, to cover the collapse of the Saudi stock market, for 2007,
to cover the global financial crisis, and for 2012 to cover the Arab Spring effect on Saudi stock market.
4. Methodology and Results
4.1 Unit Root Tests
If the variables under investigation are stationary, it means that they do not have unit roots and that the series is
said to be 1(0). If the variables under investigation are non-stationary in their level form but stationary in their
first-difference form, meaning that they do have unit roots, they are said to be 1(1). To test the nexus between
financial sector development and economic growth for Saudi economy; we used the Augmented Dickey Fuller
(ADF) (1979, 1981) method to test the unit root.
k
yt     yt1   yti  t
ti
Table (2) presents unit root test for Augmented Dickey-Fuller test (ADF) table. It shows that all the variables
tested (LNGDP, LNISTM, LNM3, LNBCPS, LNEX, LNBD and LNCCPS) have a unit root significant level of
5% for ADF Unit Root in their First Difference. These results are consistent with the standard theory, which
assumes that most macroeconomic variables are not static, but become stationary in the first difference (Enders,
1995).
lnGDPt  0  1 ln ISTM t   2 ln M 3t  3 ln BCPSt   4 ln EX t  5 ln BDt  6 lnCCPSt  7 FD2002t 8 FD2007t  9 FD2010t  t
Table 2. Augmented Dickey-Fuller for stationary unit root tests

Variables ADF (Level) ADF (Differenced)
LNGDP -0.155641 -2.631043
LNSTMI -0.917964 -2.104423
LNM3 -0.480012 -2.416632
LNBCPS -0.760121 -3.262110
LNEX -0.762910 -2.120712
LNBD -0.798551 -3.338131
LNCCPS -1.720314 -4.411701
1% C-Value -3.6752 -2.6486
5% C-Value -2.9665 -1.9535
10% C-Value -2.6220 -1.6221
32
Table (2) presents the stationary tests results showing that the variables are non-stationary in levels, but become
stationary with the first difference; in other words, they are integrated in order one, when their first differences 1(1)
are stationary.
Table 3. Regression analysis

Variable Coefficient T-Statistic Prob.
C 3.401901 4.186696 0.0004
LNSTMI 0.015350 1.114512 0.0966
LNM3 0.005619 1.436135 0.0115
LNBCPS 0.155593 1.167616 0.0749
LNEX 0.387234 10.05235 0.0000
FD 0.037867 1.906983 0.0038
LNBD 0.008719 1.050167 0.1604
LNCCPS 0.096527 2.153945 0.0420
R-squared 0.992192
Table (3) shows that the financial sector development has a significant and positive effect on economic growth.
4.2 Johansen Co-Integration Analysis
Co-integration tests are used to test the relationship between economic growth and financial sector development.
Granger (1981) was the first to propose a connection between non-stationary series and long-run equilibrium. The
purpose of conducting co-integration is to explore whether the data exhibit a long-run relationship. Engle and
Granger (1987) developed and introduced the theory of co-integration.
r s
xt   0    xti xti    yti yti  t
i1 i1
r s
yt   0    yti yti    xti xti  t
i1 i1
Table 4. Johansen co-integration test

Hypothesized Trace 5% 1%
Eigen-value
No. of CE(s) Statistic Critical Value Critical Value
None ** 0.968178 347.9647 156.00 168.36
At most 1 ** 0.952413 247.9847 124.24 133.57
At most 2 ** 0.852362 159.6743 94.15 103.18
At most 3 ** 0.798038 104.1975 68.52 76.07
At most 4 ** 0.648014 57.80693 47.21 54.46
At most 5 0.397198 37.52614 29.68 35.65
At most 6 0.323098 21.84733 15.41 20.04
At most 7 0.051413 7.530685 3.76 6.65
Table (4) shows that there is a long-run equilibrium relationship between the variables at 5% levels. We can
reject the null hypothesis of the co-integration because the Trace Statistic values are greater than the critical
values at 5% levels, which indicate there is a co-integrating equation(s) at 5%. In other words, there is a long-run
relationship among the variables, meaning there is a co-integration between the real GDP and the four proxies
for financial sector development. The results indicate that there is a nexus between financial sector development
and economic growth for the Saudi economy.
4.3 Vector Error Correction Model (VECM)
The aim of Vector Error-Correction Models is to determine whether co–integration exists between two variables;
for this to be true there must be Granger causality in at least one direction, but the most valuable aspect is that
co-integration does not reflect the direction of the causality between the variables.
Engle and Granger (1987) provide such a procedure, which is known as the “Error-Correction Models”. It explains
the long-run equilibrium state in each time period.
33
This model forces the long-run behavior of the variables to meet their co-integrated relationships. The Vector Error
Correction Models (VECM) are given in equations:
r s
xt  1  1ECTt1    xti yti    yti yti  t
i1 i1
r s
yt  1  1ECTt1    yti xti    yti yti   t
i1 i1
Where: (ECTt-1), the error correction term lagged by one period, is equivalent to ( et  yt  a   yt ), this represents
the disequilibrium residual of the co-integration equation.
Table 5. Vector Error Correction Model (VECM)

Independent Variable Estimated coefficients T-test
LNSTMI(-1) -15.24646 -12.2956
LNM3(-1) -12.86076 -11.4142
LNBCPS(-1) -5.081837 -9.83308
LNEX(-1) -1.902907 -12.0622
LNBD(-1) -16.59397 -12.2698
LNCCPS(-1) -0.113395 -0.71630
Table (5) indicates that there is a long-run causality between the variables. This conclusion can be drawn
because the variables are statistically significant at a level of 5%. Thus, the model of financial sector
development and economic growth is found to hold in the case of Saudi Arabia.
5. Conclusion
Our major aim in this paper was to investigate the relationship between financial sector development and
economic growth.
We used several methods to examine the relationship between financial sector development and economic
growth, including unit root tests; we have found that the outcome for each variable indicated that the series are
non-stationary in levels, but stationary after the first difference (i.e., order one or 1(1)). We therefore conclude
that the variables have a unit root and thus are integrated in process 1(1). Johansen’s co-integration analysis was
applied to examine whether the variables are co-integrated in the same order, and evidence found that this was
true.
The Error Correction Model (ECM) was also used on the Saudi economic data from 1970-2012. The concept of
error correction is related to co-integration because the co-integration relationship describes the long-run
equilibrium. If sets of variables are co-integrated, then an error correction model exists to describe the short-run
adjustments to equilibrium. Thus, the relationship between financial sector development and economic growth is
found to hold in the case of Saudi Arabia.
In conclusion, there is a positive relationship between financial sector development and economic growth over
the long run, and the development of the financial system will have a positive impact on the growth of the Saudi
economy. The development plan must take into account how the function of each group complements the
functions of the others.
References
Aghion, P., Banerjee, A., & Piketty, T. (1999). Dualism and macroeconomic volatility. Quarterly Journal of
Economics, 114(4), 1359-1397. http://dx.doi.org/10.1162/003355399556296
Apergis, N., Filippidis, I., & Economidou, C. (2007). Financial deepening and economic growth linkages: A panel
data analysis. Review of World Economics, 143, 179-198. http://dx.doi.org/10.1007/s10290-007-0102-3
Dickey, D. A., & Fuller, W. A. (1979). Distribution of the Estimators for Autoregressive Time Series with a Unit
Root. Journal of the American Statistical Association, 74, 427-431.
Dickey, D. A., & Fuller, W. A. (1981). Likelihood Ratio Statistics for Autoregressive Time Series with a Unit
Root. Econometrica, 49, 1057-1072. http://dx.doi.org/10.2307/1912517
Engle, R. F., & Granger, C. W. J. (1987). Co-Integration and Error Correction: Representation, Estimation, and
Testing. Econometrica, 55, 251-276. http://dx.doi.org/10.2307/1913236
34
Fatima, A. M. (2004). Does financial development cause economic growth? An empirical investigation drawing
on the Moroccan experience. Working Papers 000295, Lancaster University Management School, Economics
Department.
Granger, C. W. J. (1980). Testing for Causality. Journal of Economic Dynamics and Control, 2, 329-352.
http://dx.doi.org/10.1016/0165-1889(80)90069-X
Gurley, J. G., & Shaw, E. S. (1955). Financial aspects of economic development. American Economic Review, 45,
515-538.
International Monetary Fund (IMF). (2001-2010). Government Financial Statistics Yearbook. Washington:
International Monetary Fund.
Johansen, S. (1988). Statistical Analysis of Co-integration Vectors. Journal of Economic Dynamic and Control, 12,
231-254. http://dx.doi.org/10.1016/0165-1889(88)90041-3
Khan, M. A., Qayyum, A., & Saeed, A. S. (2005). Financial development and economic growth: the case of
Pakistan. The Pakistan Development Review, 44(2), 819-837.
Kiran, B., Yavus, N. C., & Guris, B. (2009). Financial development and economic growth: A panel data analysis of
emerging countries. International Research Journal of Finance and Economics, 30, 1450-2887.
Levine, R., Loayza, N., & Beck, T. (2000). Financial intermediation and growth: Causality and causes. Journal of
Monetary Economics, 46, 31-77. http://dx.doi.org/10.1016/S0304-3932(00)00017-9
Neusser, K., & Kugler, M. (1998). Manufacturing growth and financial development: evidence from OECD
countries. Review of Economics and Statistics, 80, 638-646. http://dx.doi.org/10.1162/003465398557726
Pradhan, R. (2009). The Nexus between Financial Development and Economic Growth in India: Evidence from
multivariate VAR Model. International Journal of Research and Reviews in Applied Sciences, 1, 141-151.
Samudram, M., Nair, M., & Vaithilingam, S. (2009). Keynes and Wagner on Government Expenditures and
Economic Development: The Case of a Developing Economy. Empirical Economics, 36, 697-712.
Sanusi, N. A., & Sallah, N. H. M. (2007). Financial development and economic growth in Malaysia: An
application of ARDL approach. Retrieved from
http://www.ibacnet.org/bai2007/proceedings/Papers/2007bai7443.doc
The Saudi Arabia Monetary Agency (SAMA). (2001 to 2009). Annual Report. SAMA.
35
ISSN 1916-971X E-ISSN 1916-9728
Managerial Incentives and the Risk-Taking Behavior of

Hedge Fund Managers
Serge Patrick Amvella Motaze1
1
Finance Department, HEC Montreal, Montreal, QC, Canada
Correspondence: Serge P. Amvella Motaze, Finance Department, HEC Montreal, 3000 Cote Sainte-Catherine,
Montreal, QC, H3T 2A7, Canada. Tel: 1-514-340-6609. E-mail: [email protected]
Abstract
We use a binomial model to derive the optimal trading strategy for a hedge fund manager facing different
constraints such as the possibility of the fund liquidation and a minimum net-of-fees return to deliver in order to
meet investors expectations. Our model enables us to link the optimal trading strategy and the optimal volatility
of the fund to the management and incentive fee rates, the minimum net-of-fees return required by investors, the
size of the fund and the moneyness of the option-like contract held by the manager. We find that even if the
optimal volatility of the fund increases when the option is out of the money, there is a certain point at which it
starts to decrease; and this point can vary from one manager to another given that it is related to factors such as
his ownership in the fund and his tolerance for risk. These results give further insights on why some empirical
studies such as Brown, Goetzmann and Park (2001) and Clare and Motson (2009) fail to find a general trend in
hedge fund managers’ risk-taking behavior when their option is out of the money.
Keywords: hedge funds, option-type compensation, high water mark, volatility, binomial model
1. Introduction
Several studies have addressed the issue of the risk-taking behavior of managers who have asymmetric
compensation packages (incentive fees) and the conflicting results show how difficult it is to come up with a
satisfactory answer. The incentive contract resembles a call option given that the manager receives incentive fees
when the asset value exceeds the high water mark. Using a TASS database, Brown, Goetzmann and Park (2001)
find that hedge fund managers tend to increase (decrease) risk in response to performance relative to their peers,
but not in response to their absolute performance. They argue that the positive relationship between volatility and
termination provides a disincentive for fund managers to gamble excessively when their option is out of the
money (OTM). Carpenter (2000) examines the behavior of a mutual fund manager who has an option-like
contract on the asset under management. She finds that the manager will increase the risk of the fund if the
fund's return is below the hurdle rate and will decrease the risk if the return is above the hurdle rate, which also
suggests a "locking in" behavior when the option-like contract is ITM. However, her analysis is over one period
and ignores the possibility of the fund being liquidated in response to poor performance unless its value becomes
zero. A liquidation boundary has been specified by other authors such as Goetzmann, Ingersoll and Ross (2003),
Hodder and Jackwerth (2007) or Panageas and Westerfield (2009). Their models show that the manager has the
incentive to increase risk given that the value of the option increases with the volatility, but as the fund's value
approaches the liquidation boundary, the manager will decrease risk.
In this study, we use a binomial model to assess the risk-taking behavior of a hedge fund manager. As in Hodder
and Jackwerth (2007), we present a theoretical and discrete time-model. Our approach enables us first to assess
the impact that the level of volatility has on the expected manager's fees and the expected investors' wealth. Our
results suggest the manager may not benefit from excessive risk-taking because a high level of volatility reduces
the expected investors' wealth, which may lead to outflows when their objectives are not accomplished and
therefore, may reduce the future compensation of the manager.
Second, we solve for the optimal trading strategy from the manager's prospect on a one-period basis. The
one-period setting is justified because even though the manager's horizon is longer, he implements his optimal
trading strategy period by period to maximize his compensation. He has an incentive to increase risk because his
expected fees increase with volatility, while investors would like to inhibit excessive risk-taking because it
36
impacts negatively on their expected wealth and increases the magnitude of potential losses. To account for this
conflict of interest or agency problem, we derive the optimal trading strategy by setting first a liquidation
boundary for the fund, and unlike previous studies, we add a constraint related to a minimum net-of-fees return
to be delivered by the manager in order to renew his contract or avoid outflows due to poor performance. This
constraint can be considered as a benchmark for the manager and shows that the fund should bear a minimum
risk in order to meet investors’ expectations in terms of returns. We also add a third constraint related to the
manager' tolerance for risk. Our results show that the optimal allocation in the risky asset is negatively related to
management and incentive fee rates, and to the size of the fund. However, it is positively related to the minimum
net-of-fees return required by investors and to the moneyness of the option contract, the latter defined herein as
the distance between the high water mark and the fund value. This means that the more the option is OTM, the
higher the incentive to take risk will be. However, because of the liquidation boundary, there is a level of asset
under management where the risk exposure of the fund starts to decrease.
The contribution of our paper to the hedge fund literature is to give further insights on the reason why the
relationship between the volatility strategy of the manager and the moneyness of the option is complex
especially when the latter is OTM. Indeed, Brown et al. (2001) and Clare and Motson (2009) fail to empirically
find a clear behavior of hedge fund managers in response to absolute performance. Clare and Motson (2009) find
that sometimes managers increase risk when their option is OTM, but they do not "put it all on black" in other to
"win" back earlier losses. Our model shows that the manager increases the volatility up to a certain point when
the fund is below the high water mark. Above this point, he decreases the volatility of the fund due to the fear of
liquidation. This point depends on factors such as his tolerance for risk or his ownership in the fund and it is the
reason why the risk-taking behavior will differ from one manager to another, and it will remain difficult to find a
clear increase or decrease of risk when the option-like contract is OTM.
2. The Model
Compensation contracts for hedge fund managers include incentive fees which can be thought of as a call option
on the portfolio that they manage. At the end of each period, the portfolio can take on a continuum of values.
However, whatever the end value, there are only two possible outcomes: either the option is exercised or it is not.
The goal of this study is not to implement a sophisticated model that describes the precise portfolio value at the
end of each period. Our study aims to model the two possible outcomes for the option through time. To achieve
this, we use a binomial model. The binomial approach offers an intuitive and tractable setting for evaluating
options, and provides results that are easy to interpret. It has the advantage of providing a closed-form solution
for the optimal trading strategy and the optimal volatility of the fund.
The binomial model assumes that the asset value can either go up or down by a known amount and we use this
approach for the intuition it offers in modeling the risk-taking behavior of the manager. Indeed, one can consider
a binomial world setting where the manager bets on the market's direction. He can either win or lose. The initial
size of the fund is S0. We consider the problem of a hedge fund manager's optimal allocation of portfolio value
into a risky and a riskless investment opportunity. The riskless asset grows at the risk free rate rf and has a
volatility of zero, while the risky asset has a mean μe and a volatility σe. It follows a binomial random walk with
the up and down parameters u and d defined as:
u  1  e T (1)
d  1 e T (2)
The probability of a rise pe is defined as follows:
1 e T
pe   (3)
2 2 e
and the probability of a fall is 1- pe. The choice of these constants is far from unique and we choose this
specification for estimation purposes. (Note 1)
The manager selects a dynamic strategy by allocating a fraction xe of the fund in the risky asset and the
remaining part (1- xe), is invested in the riskless asset. We allow the manager to control for xe which can be
superior to 1 when he uses leverage. In that case he borrows at the risk free rate rf. The asset under management
S follows a binomial random walk with a drift μ = xe μe + (1- xe)rf and a volatility σ = xe σe, and at the end of
the evaluation period, its value ST can have two possible outcomes, either
ST ,u  xe S 0 (1   e T )  (1  xe S 0 )(1  r f T ) (4)
37
or
S T ,d  xe S 0 (1   e T )  (1  xe S 0 )(1  r f T ) (5)
The manager owns a fraction a of the fund with 0 ≤ a ≤ 1, and investors hold the remaining portion (1 - a) of the
assets. Management fees equal a proportion m ≥ 0 of fund value (1- a)ST at the end of the evaluation period and
incentive fees are a percentage k ≥ 0 of the fund's performance in excess of the high water mark HT, that is
(1-a)kmax{ST - HT,0}. For some incentive contracts, the high water mark grows at a certain rate of interest. As
Hodder and Jackwerth (2007), we assume that it grows at the risk free rate during the evaluation period such that
HT = H0(1+ rfT). The total wealth of the manager WT is the sum of his ownership in the fund and his fees. At the
end of the evaluation period, it is expressed as follows:
WT  aS T  (1  a)mST  k maxS T  H T ,0 (6)
Investors wealth is expressed as follows:

I T  (1  a)(1  m)ST  k maxST  HT ,0 (7)
From the above variables, the expected value of the manager's wealth at the end of the period is:
EWT   aEST   (1  a)mEST   kEmaxST  H T ,0
From the binomial model, it follows that:
 
E WT   aS 0 (1  T )  (1  a) mS 0 (1  T )  kpe max xe S 0 ( e T  r f T )  L0 (1  r f T ),0  (8)
with
L0 = H0 – S0, the moneyness of the option. (Note 2)
We can note that L0 ≥ 0 because the option is always either at the money (H0 = S0) or out of the money (H0 > S0)
at the beginning of the year.
The expected investors' wealth is:
  
EI T   (1  a) (1  m) S 0 (1  T )  kpe max xe S 0 ( e T  r f T )  L0 (1  r f T ),0 (9)
The moneyness (or more precisely the "out of the moneyness") L0 is an important variable of the model, because
it determines how far the fund value is from the high water mark. In other words, it is a measure of how far the
option is out of the money. This measure is interesting because it increases with previous losses and it shows the
path-dependent nature of the manager's payoffs. The larger the value of L0, the smaller the expected
compensation of the manager.
3. Impact of the Fund's Volatility Level on Expected Values
Replacing the probability pe by its expression in equation (3), we obtain, depending on whether the option is
expected to be exercised or not, the following expressions for the derivatives of E[WT] and E[IT] with respect to
the volatility of the risky asset:
EWT  1 
 (1  a)1k  xe S 0 T 

 e T xe S 0 r f T  L0 (1  r f T ) 
 (10)
 e 2   e2 
and
38
E I T  1 
  (1  a)1k  xe S 0 T 

 e T xe S 0 r f T  L0 (1  r f T )  (11)

 e 2   e2 
where 1k is an indicator variable such that:

1k = k if the option is exercised
1k = 0 otherwise.
This shows that EWT   e  0 and EI T   e  0 . The sensitivity of the expected manager's wealth and that
of the expected investors' wealth with respect to volatility are equal in amplitude, but of opposite signs; the first
being positive and the latter being negative. Then, for a given expected return, a higher level of risk has a
positive impact on the manager's wealth. In fact, a high level volatility does not change the expected value of the
manager's ownership in the fund, nor that of the management fees at time T. For a given expected return, these
values remain constant whatever the level of volatility and we have EaST   e  0 and EMfessT   e  0 .
A higher level of volatility only increases the expected value of incentive fees because the distribution of the
possible values of the option-like payoff is truncated at zero and we have EWT   e  EIfessT   e . As a
consequence, this reduces the expected investors' wealth of the fund by the same amount. The above equations
show that the manager may have an incentive to increase risk, but if he has to choose between two assets having
the same expected return, a rational manager will invest in the one with the lower volatility; because it has a
better risk-adjusted return and also because even if investing in the asset with the higher volatility increases the
expected value of his incentive fees, it reduces the expected value of investors wealth. If the manager wants to
increase his fees, a solution is to increase the proportion of the fund invested in the asset with the best
risk-adjusted return. This will increase the volatility of the whole fund as well as its expected return. As result,
the manager will increase both the expected value of his wealth and the expected investors' wealth. The
question is what is the optimal proportion to be invested in the risky asset, because increasing the risk of the fund
also increases the magnitude of losses in the case where his bet does not come true.
4. Other Factors That Can Impact the Manager's Risk-Taking Behavior
The above results suggest that even if, a priori, we may have reason to believe that hedge fund managers have a
strong incentive to take a high level of risk, especially when their incentive contract is OTM, they may not
necessarily do so given the potential impact that it has on investors wealth. Furthermore, their volatility strategy
may not only be driven by the willingness to exercise their option or to increase the value of their option contract,
but also by other factors. Theory suggests that the value of an option increases with the volatility of the
underlying asset. But this theory does not address the issue of overall wealth when the holder of the option is
also the manager of the underlying asset, and this is the case for hedge fund managers. As Goetzmann et al.
(2003) mention, hedge fund managers cannot do any form of delta hedging. They hold option-like contracts on
assets they manage. If you hold an option contract on an underlying asset traded in the market, you will benefit
from an increase in the volatility of the asset because it implies that your contract is worth more. However, if you
have some interest in the underlying asset, you will be concerned by its risk level, and this is the case for hedge
fund managers because a high volatility of the fund is also associated with a high probability of termination.
Many other factors may affect the risk-taking behavior of the manager such as his relative performance, his
ownership in the fund, the size of the fund and the average risk of his peers. The various categories of hedge
funds do not have the same exposure to risk factors and therefore, do not have the same level of risk. For
instance, the average monthly volatility for Fixed Income Arbitrage funds in the TASS database from January
1994 to June 2011 was 2.60%, while it was 5.93% for Emerging Market hedge funds. (Note 3)
As for the size of the fund, incentive fees do not have the same importance in the manager's compensation.
Incentive fees might be more important for a small fund because management fees (paid from the asset under
management) might sometimes be insufficient to cover operating expenses. On the other hand, incentive fees
might not be as important for a large fund which can cover its operating costs with management fees. Hence, a
manager of a small fund is constrained to assume a higher level of risk in order to survive because for him, the
39
relative importance of incentive fees is higher than it is for a manager of a large fund. As the fund grows, the
relative importance of incentive fees decreases and thereby, the need to take riskier positions decreases.
5. Optimal Trading Strategy
The previous equations show that for a given expected return, a higher level of risk has a positive impact on
manager incentive fees and a negative impact on investors wealth. On one hand, the manager has an incentive to
manage a fund with a high level of risk because his option value at each period increases with volatility. On the
other hand however, a higher volatility increases the magnitude of losses and this could be harmful to the
manager's reputation as well as to his ownership in the fund and his compensation because of outflows due to
bad performance. Outflows represent a powerful tool that investors can use in order to prevent excessive
risk-taking by managers. Outflows lower managerial compensation because fewer assets under management
imply smaller management and incentive fees. Outflows can also be costly for the manager because he could be
forced to liquidate certain positions at a suboptimal price in order to meet redemptions. Outflows can even lead
to the liquidation of the fund. The 2008 financial crisis is an example where many hedge funds have been forced
to liquidate due to redemptions. In order to avoid withdrawals, and even attract new investments, the manager
should maintain a good track record. For each period, he should provide the investor with a return after fees
superior to that the latter can achieve on his own, otherwise he is better off investing elsewhere. Indeed, the
reason why investors accept liquidity constraints imposed by hedge funds and the payment of high incentive fees
is the strong expectation that year after year they will provide absolute returns. (Note 4)
Hedge funds define themselves as absolute-return strategies and a positive periodic return can be considered an
implicit feature of the contract. This return may also depend on market conditions. For example, if we assume
that the investor has the choice of investing in either a hedge fund or the risk free rate, this minimum return
should be superior to the risk-free rate. Then, for each period, the manager faces a trade-off between increasing
his fees and meeting investors expectations.
Now, let us solve for the optimal trading strategy of the fund from the manager's prospect. Let's denote rb, the
minimum net-of-fees return required by the investor, which in other words can be considered as a benchmark for
the manager. As we mentioned, it may depend on market conditions, and we should have a specific rb for each
period. We assume that rb is known at the beginning of each year and does not change during the year. Another
assumption is that the fund can be liquidated in the event of a very poor performance as in Goetzmann et al.
(2003) and Hodder and Jackwert (2007). Therefore, the manager should always watch the downside risk of the
fund even if its expected return is positive. In our binomial model, a loss can occur with the probability (1-pe).
An optimal behavior would be therefore to ensure that in the case of the occurrence of this state of nature, the
value of the fund remains superior or equal to the liquidation boundary, otherwise the fund will be liquidated
(immediately after the payment of management fees). Let's denote B0, the liquidation boundary at the beginning
of the year. (Note 5)
Another point to consider is the fact that in the three components of the manager's wealth, only the expected
value of incentive fees increases with volatility as we mentioned previously. The expected value of management
fees and that of his ownership in the fund does not change with a greater volatility. On the contrary, they will not
necessarily benefit from a higher volatility because a higher level of risk also increases the magnitude of losses;
and the higher the losses the lower this part of the manager's wealth will be. We then suppose that, according to
his tolerance for risk, the manager sets a limit variance θ² for this part of his wealth.
The manager's optimization problem is then to choose an admissible trading strategy that maximizes his
expected wealth at the end of the evaluation period T, by using
max E WT 
 xe

s.t. E I T   I 0 (1  rbT ) (c1)


 S T , d  B0 (c 2)


 Var aS T  (1  a)mS T    2 (c3)

The first constraint (c1) is related to the minimum net-of-fees return to achieve in order to meet investors
expectation. If the performance of the fund is below this return, the manager exposes the fund to outflows. The
40
second coonstraint (c2) is related to the liquidationn boundary; iif the value oof the fund faalls below B0, this
constraint will be violateed, which will imply the liquuidation of the fund. The thirrd constraint (c (c3) is related to
t the
limit volattility sets by thhe manager forr the part of hiss wealth for whhich a higher vvolatility repreesents a higherr risk,
namely hiis investment in the fund pplus managemeent fees. Som me critical valuues for xe folllow from the three
constraintss:
 (c1) impplies that, depennding on whetther the optionn is expected too be exercised,, we should haave xe ≥ xc1 with
h
  T 
  1
S 0 (1  rbT )  (1  m)(1  r f T )  1k L0 (1  r f T )1  e
2  e


xc1    (12)
1   T 
S 0 (1  m)( e  r f )T  1k S 0 ( e T  r f T )1  e 
2  e 
 
where 1k iss an indicator variable
v such tthat:
1k = k if thhe option is exeercised
1k = 0 otheerwise.
 (c2) impplies that xe ≤ xc2 with
S 0 (1  r f T )  B0 (13)
xc 2 
S 0 ( e T  r f T )
 (c3) impplies that xe ≤ xc3 with


xc3  (14)
 e a  (1  a)m
Then, xc1 rrepresents the lower frontierr for xe, and xc2 and xc3 repreesent its upper frontiers. We can note that xc1 is
positively related to rb, implying
i that the higher invvestors expectaations in term ms of return, thhe higher should be
the investm
ment of the funnd in the riskyy asset. We cann draw the froontiers inside wwhich the mannager can maxiimize
his wealthh without viollating any of the three connstraints. Figuure 1 shows thhese frontiers if we assume e the
following values for thee parameters: μe = 0.08, σe = 0.2, m = 0.022, k = 0.2, rf = 0.01, rb = 0.004, T = 1, a = 30%,
θ = 0.14, and B0 = 50. H0 = 100 and the value of S0 varies from m 100 to 51 coorresponding tto a moneyness L0
varying froom 0 to 49.
Figure 1. Frontiers foor the risky assset weight relaated to constraiints c1, c2 andd c3 and for diffferent levels of
o
m
moneyness
The dottedd lines represeent the frontieer for (c1) and (c3), and thhe full and tinny line, the frrontier for (c2)). xc1
correspondds to the weigght at which the net-of-fees return of thhe fund is equual to the minnimum require ed by
41
investors. We can see thhat xc1 decreasses up to the ppoint at which the option is eexpected to bee exercised without
violating cconstraint (c1).. This correspoonds to a valuee of S0 equal too 88 and a valuue of 12 for L0. Above this point,
p
a lower weeight for the riisky asset impllies a low volaatility of the whhole fund, insuufficient to proovide the minimum
net-of-feess return required by investorrs. So, the mannager should stop to decrease the allocatioon in the risky asset
and xc1 beccomes constannt. For this vallue of xc1, the option is not eexercised and tthe net-of-feess return of the fund
is rb.
Regardingg xc2, we can seee that it decreeases as the mmoneyness L0 inncreases. Thiss suggests that as the value of o the
fund decreeases and gets closer to the liiquidation bouundary, the mannager should rreduce the inveestment in the risky
moneyness. Thee hatched area represents the
asset. As ffor xc3, it remaiins constant annd does not deppend on the m e area
in which thhe manager caan invest in thee risky asset w without violatinng any of the tthree constrainnts. We can see
e that
beyond a certain level of moneynesss it becomes iimpossible to invest withouut violating at least one of these
constraintss. In the presennt example, thiis level of monneyness is about 41, which ccorresponds to a value of S0 equal
e
to 59. Therrefore, above this
t point, it iss impossible too have an optim
mal solution.
The expression of xc3 inn equation (14)) shows that thhe more the mmanager is inveested in the funnd, the less xc33 will
be. It is thhe only frontiier that depennds on the maanager's stake in the fund aand it shows hhow the mana ager's
propensityy of taking riskk will be reducced with his innvolvement inn the fund. Figuure 2 shows thhe values of xc3
c for
values of mmanager stake equal to 10%,, 20% and 30% %.
Figure 2. Fronntiers related too constraint c33 for different values of the m
manager's ownnership a
We can seee how xc3 deccreases with thhe manager's sstake in the fuund. For instannce, for a = 10%, xc3 is allmost
equal to 6,, meaning thatt with a 10% oownership in thhe fund, and aaccording to hiis tolerance foor risk, the mannager
can be com mfortable in innvesting a maximum of six tiimes the value of the fund inn the risky asseet. This represe
ents a
leverage raatio of 5. Whille for a 30% ownership, xc3 = 2.23, which reduces considderably the maaximum investtment
in the riskyy asset.
Dependingg on the managger's stake in tthe fund, the uupper limit forr xe is xc2 or xcc3. However, evven if investin ng xc2
or xc3 in thhe risky asset provides the manager withh the maximum m wealth that he can expecct without violating
constraint 2 or constrainnt 3, it does noot necessarily rrepresent the ooptimal tradingg strategy becaause it constitu
utes a
very risky position. For instance, if thee manager inveests the proporrtion xc2 in the risky asset, thhe value of the fund
will fall too the liquidatioon boundary iff his bet does nnot come true aand the survivval of the fund will be threate
ened.
So, the opptimal investm ment in the riskky asset is below xc2 or xc3 uunless it is necessary. For eexample, suppo ose a
certain ammount of fees (m management fefees + incentivee fees) for whiich the manageer is comfortabble, in other words,
w
for which he can cover his h operating eexpenses and rremain with a certain profit. Then, the prooportion of the fund
that he willl invest in thee risky asset w
will be the one for which he can expect to earn this amoount of fees. Let L us
denote thiss amount v. Too solve for the correspondingg weight in thee risky asset, w we have the following equatio on:
(1  a)mES T   (1  a)kEmaxS T  H T ,0  v (15)
The solutioon is
42
v 1   T 
 mS 0 (1  r f T )  k 1  e  L0 (1  r f T )
(1  a ) 2  e 
 (16)
xv 
1   e T 
 S 0 ( e T  r f T )
mS 0 (  e  r f )T  k 1
2  e 

We can see that the abbove parameteers impact diffferently on xv (the derivatiives of xv witth respect to these
parameterss are presentedd in the appenddix):
 There is a negative reelationship betw ween xv and thhe incentive ffee rate k. A hhigher k impliees a lower xv and
a a
lower k immplies a higherr xv. This is connsistent with C
Carpenter (2000) and shows tthat granting a high-incentivve fee
rate to the manager contrributes to reduucing his appettite for risk
 The sam
me relationship exists betweenn xv and the maanagement feee rate m.
 There is a positive relaationship betw
ween xv and thee level of moneeyness L0, impplying that as tthe option goe
es out
of the monney, the managger should incrrease risk. Carppenter (2000) also finds a sim
milar relationsship.
 There is a negative relaationship betw
ween xv and thee asset under m
management S0. This corroboorates what we e said
in the prevvious section; smaller fundss will have moore incentive tto increase rissk given the im
mpact of the fund's
fu
size on thee manager's compensation.
 Finally, there is a posittive relationshhip between xv the manager'ss stake in the fu
fund a, meaninng that the morre the
manager iss invested in thet fund, the hhigher xv shoulld be in order to earn the saame level of inncentive fees given
g
that they aare paid on a loower proportioon of the fund ((1- a).
Figure 3 shhows the valuees of xv for diffferent levels oof moneyness aand for differeent values of a,, when for example
v = 5. Wee can see how w xv is an inccreasing functiion of a and L0. As the m moneyness incrreases, xv incre eases
unboundeddly.
Figure 3. Values of the

t risky asset weight xv for a = 10%, a = 220% and a = 300% and for diffferent levels of
o
m
moneyness
The relatioonship betweenn xv and a seemms to be contraary to the concclusions of othhers studies whhich find that as
a the
manager's ownership in the fund increeases, his risk ttaking behavioor is reduced. ((Note 6) In facct, it is not the case
because, aas we showed earlier, xc3, wwhich is related to manager's tolerance foor risk, decreasses with a; an nd xc3
prevails uppon xv otherwise, constraint ((c3) is violatedd.
So, even if xv increases with a, its maxximum admisssible value willl be reached eearlier for highher values of a and
as result, the manager'ss ownership iin the fund wwill contributee to reduce hiis risk taking behavior. Th his is
illustrated in figure 4.
43
Figure 4. Values
V of the rrisky asset weiight xv when coonstraint c3 is taken into acccount
Moreover,, as we mentiooned previouslly, the optimaal investment iin the risky assset should nott be superior to t xc2
otherwise, constraint (c22) will be violaated. Then, thee manager willl choose the loower of xv, xc2 and xc3 to inve
est in
the risky aasset. In addittion, xv shouldd be superior or equal to xc1 in order to fill the condiition related to o the
minimum return rb. Therrefore, the optiimal weight foor which all thee conditions arre filled is:
x e*  maxx c1 , min( x v , x c 2 , x c 3 ) (17)
We can theen obtain the optimal

o volatillity of the fundd as follows:
 *  xe* e (18)
The sign oof the relationship between x*e and the paraameters k, m, S0, L0, a holds aalso for σ and these parameters.
Figuree 5. Optimal weight

w of the risky asset for a = 10%, a = 20% and a = 300% and for diff
fferent levels of
m
moneyness
44
Figure 6. O
Optimal volatiility of the fundd for a = 10%,, a = 20% and a = 30% and ffor different leevels of money
yness
Figure 5 aand 6 represennt the optimal proportion off the fund to bbe invested inn the risky assset and the optimal
volatility oof the fund. We
W can see in fiigure 5 that foor low values oof the moneyneess L0, the mannager increase es the
investmennt in the risky asset
a by using leverage as thhe probability oof exercising hhis option-likee contract is higher;
and the higgher his ownerrship in the funnd, the higher the use of leveerage. Howeveer, the manageer starts to decrease
the investmment in the riskky asset for soome levels of m
moneyness andd each of thesee levels dependds on the value e of a,
his ownerrship in the fuund. The highher his ownersship, the earlieer x*e starts too decrease andd for L0 = 11, the
investmennt in the risky asset
a becomes the same whattever the owneership of the m manager. For L0 = 38, the man nager
stop using leverage and starts to increaase the investm ment in the riskk free asset to rreduce the whhole risk of the fund
given that asset value gets
g closer to tthe liquidationn boundary. Beeyond the valuue of L0 = 41, it is impossib ble to
invest withhout violating at least one off the constraintts and it is therrefore impossibble to have an optimal weigh ht for
the risky aasset beyond thhis value.
6. Conclussion
The optionn-like compennsation contracct of hedge funnd managers sserves as an inncentive to supperior performance.
However, this particularr form of manaager compensaation can also bbecome an inccentive to takee more risk in order
o
to end up with an incenntive contract iin the money. The purpose oof this paper iis to evaluate tthe optimal tra
ading
strategy fo
for a hedge fund
fu manager facing different constraintts among otheers, the possiibility of the fund
liquidationn or a minimuum return to deliver in ordder to meet innvestors expecctations. To tthis end, we use u a
binomial mmodel and our analysis docum ments interestiing findings.
First, we ffind that a highh level of risk positively imppacts the expeected manager''s incentive feees at the end of
o the
period butt negatively immpacts the expeected investorss' wealth. Withh a one-year coontract and witth no investme ent in
the fund, the manager's optimal behhavior would be to maxim mize the valuee of his optioon-like contrac ct by
increasing the volatility of the fund innfinitely if he does not have to meet inveestor expectatiions. Howeverr, the
time horizzon of the mannager's contract is more thann one year and he must makee a trade-off between current and
future payoffs because a high volatilityy level increasses the magnituude of potentiaal losses and a decline in the fund
value also implies a declline in his overrall future weaalth.
Among otther interestingg results of ouur model, we ffind that even if the optimaal volatility of the fund incre eases
with the "out of the mooneyness" of thhe option, therre is a certainn point at whicch the optimall volatility starrts to
decrease; and this pointt can vary froom one managger to another given that it is related to ffactors such as a his
ownershipp in the fund annd his tolerancce for risk. Thiis is consistentt with Brown, Goetzmann annd Park (2001) and
Clare and Motson (20099) who docum ment in empiriccal studies thatt it is difficultt to find a general trend in hedge
h
fund manaagers risk-takinng behavior whhen their optioon is OTM.
The resultts of the presennt study suggeest that incenttive contracts, to the extent that they are llong-term-orie ented,
are designned to encouragge good perforrmance. Our conclusions appply not only too hedge funds, but to any forrm of
contract w
where the manaager's compenssation has an iincentive featuure. With a higgh water mark provision, as is i the
case for heedge funds, thhe incentive coontract will be either at the m
money or out oof the money at the beginnin ng of
each year.. Each loss wiill contribute tto increase thee "out of the moneyness" oof the option-llike payoff and d the
greater thee "out of the moneyness",
m thhe lower the exxpected fees off the manager.. In this respecct, the very forrm of
45
hedge fund contracts contributes to lower the manager's appetite for risk because he earns incentive fees only
after recovering past losses. With the absence of a high water mark provision, which is the case in many
instances such as corporate executives compensation contracts, the option is at the money at the beginning of
each year which does not help reduce the agency problem. Nowadays, where the compensation of corporate
executives is increasingly called into question, hedge fund compensation contracts, although not perfect, could
be a good source of inspiration.
Acknowledgements
The author is grateful to seminar participants at the 2009 NFA annual meetings in Niagara-On-The-Lake, the
2010 EFA meetings in Miami, for many helpful comments. I would also like to thank Nicolas Papageorgiou and
Iwan Meier for their helpful comments. I gratefully acknowledge financial support by the Centre de Recherche
en E-Finance (CREF), l'Institut de Finance Mathématique de Montréal (IFM2).
References
Agarwal, V., Daniel, N. D., & Naik, N. (2009) Role of managerial incentives and discretion in hedge fund
performance. Journal of Finance, 64, 1985-2428. http://dx.doi.org/10.1111/j.1540-6261.2009.01499.x
Brown, S. J., Goetzmann, W. N., & Park, J. (2001). Careers and survival: competition and risk in the hedge fund
and CTA industry. Journal of Finance, 56, 1869-1896. http://dx.doi.org/10.1111/0022-1082.00392
Carpenter, J. N. (2000). Does option compensation increase managerial risk appetite? Journal of Finance, 55,
2311-2331. http://dx.doi.org/10.1111/0022-1082.00288
Clare, A., & Motson, N. (2009). Locking in the profit or putting it all on black? An investigation into the
risk-taking behaviour of hedge fund managers. Working paper, The Cass Centre for Asset Management
Research. Retrieved from
http://www.cass.city.ac.uk/__data/assets/pdf_file/0006/69936/Locking-in-the-profits-or-putting-it-all-on-bla
ck.pdf
Goetzmann, W. N., Ingersoll, J. E., & Ross, S. A. (2003). High-water marks and hedge fund management
contracts. Journal of Finance, 58, 1685-1718. http://dx.doi.org/10.1111/1540-6261.00581
Hodder, J. E., & Jackwerth, J. C. (2007). Incentive contracts and hedge fund management. Journal of Financial
and Quantitative Analysis, 42, 811-826. http://dx.doi.org/10.1017/S0022109000003409
Kouwenberg, R., & Ziemba, W. T. (2007). Incentives and Risk Taking in Hedge Funds. Journal of Banking and
Finance, 31(11), 3291-3310. http://dx.doi.org/10.1016/j.jbankfin.2007.04.003
Panageas, S., & Westerfield, M. M. (2009). High-water marks: high risk appetites? Convex compensation, long
horizons, and portfolio choice. Journal of Finance, 64, 1-36.
Wilmott, P. (2006). Paul Wilmott on quantitative finance (2nd ed.). John Wiley & Sons.
Notes
Note 1. For more information on the model, the reader can refer to the book Paul Wilmott On Quantitative
Finance (2nd ed.), John Wiley & Sons, 2006.
Note 2. See appendix for the demonstration. Equation (8) implies that if xeS0(σe√T-rfT) – L0(1+rfT) > 0, the
option is expected to be exercised. While if xeS0(σe√T-rfT) – L0(1+rfT) ≤ 0, the option is not expected to be
exercised.
Note 3. Data from TASS database.
Note 4. Agarwal, Daniel and Naik (2009) find a median lockup period of 1 year and a median redemption delay
of 0.2 year on a merged database formed by CISDM, HFR, MSCI and TASS.
Note 5. It is not obvious to specify a liquidation boundary corresponding to the hedge fund industry because
individual investors may have different rules for liquidation. Goetzmann, Ingersoll and Ross (2003) use different
liquidation boundaries such that the investor will liquidate if the asset value falls by 5% or 20% from his
personal high water mark. Hodder and Jackwerth (2007) assume that the fund is liquidated if the asset value falls
by 50% from the high water mark.
Note 6. For instance, Kouwenberg and Ziemba (2007) or Hodder and Jackwerth (2007).
46
Appendix
A.1. Expected Values and the Derivates with Respect to Volatility
From the binomial model we have:
EST =S0 (1+Te )
With
μ + (1- )
The expected value of the manager's wealth is expressed as follows:
EWT   aEST   (1  a)mEST   kEmaxST  H T ,0
  p e maxx e S T ,u  H 0 (1  r f T ),0 
 aS 0 (1  T )  (1  a ) mS 0 (1  T )  k  
  (1  p e ) maxx e S T , d  H 0 (1  r f T ),0 
 
 p e max x e S 0 ( e T  r f T )  L 0 (1  r f T ), 0
 aS 0 (1   T )  (1  a )  mS 0 (1   T )  k 
 


 
  (1  p e ) max x e S 0 (  e T  r f T )  L 0 (1  r f T ), 0   

where L0 = H0-S0.
With L0 ≥ 0 given that H0 ≥ S0, we have:

(1  p e ) max xe S 0 ( e T  r f T )  L0 (1  r f T ),0  0 
and as result,
 
E WT   aS 0 (1  T )  (1  a ) mS 0 (1  T )  kpe max xe S 0 ( e T  r f T )  L0 (1  r f T ),0 
With p e  1   e T , we obtain:
2 2 e



 xe S 0  e T  (  e  r f )T  

1 
E WT   aS 0 (1  T )  (1  a ) mS 0 (1  T )  1k 
2  L (1  r T )  e

 T xe S 0 r f T  L0 (1  r f T )  
 
e
0 f
   
where 1k is an indicator variable such that:

1k = k if the option is exercised
1k = 0 otherwise.
Then, the derivative of with respect to the volatility of the risky asset is:
EWT  1 
 (1  a)1k  xe S 0 T 

 e T xe S 0 r f T  L0 (1  r f T ) 

 e 2   e2 
By the same way we have:

EI T   (1  a)(1  m) ES T   kEmaxS T  H T ,0
47



 x e S 0  e T  (  e  r f )T  

1 
 (1  a ) (1  m) S 0 (1  T )  1k 
 2  L (1  r T )  e

 T x e S 0 r f T  L0 (1  r f T ) 


e
0 f
   
Then, its derivative of with respect to the volatility of the risky asset is:
E I T  1 
  (1  a )1k  x e S 0 T 

 e T x e S 0 r f T  L0 (1  r f T )  

 e 2   e2 
A.2. Derivatives of xv with Respect to Parameters k, m, L0, S0 and a
1   e T
2 
1
e


 (

1  a )

 m(1  r f T ) L0 (  e  r f )T  S 0 ( e T  r f T )  v ( e T  r f T )
x  
(*) v  0
k  
2
1   T 
S 0 m(  e  r f )T  k 1  e ( e T  r f T )
 2  e 
 
  1   T   
 S 0 (1  r f T )  m (  e  r f )T  k  1  e ( e T  r f T )  
  2   e   
 
  v 1   e T  
  (  e  r f )T   mS 0 (1  r f T )  k 1  L (1  r f T )  
x v   (1  a ) 2   e  0  
(*)  0
m  1   e T  
2
S 0  m (  e  r f )T  k 1  ( T  r f T ) 
 2   e  e 
1   e T
(1  r f T )
k 1
xv 2  e 

(*)  0
L0  1   e T  
S 0 m(  e  r f )T  k 1  ( e T  r f T )
 2   e  
 v 1   T  
  k 1  e L0 (1  r f T )
  e 
x  (1  a ) 2  
(*) v  0
S 0  1   T  
S 02 m(  e  r f )T  k 1  e ( e T  r f T )
2   
  e  
xv 1
(*)  0
a  1   T  
(1  a) S 0 m(  e  r f )T  k 1  e
2 ( e T  r f T )
 2  e 
 
48
ISSN 1916-971X E-ISSN 1916-9728
Behaviour of Bank Share Prices and Their Impact on National Stock

Market Indices:
Comparing Countries at Different Levels of Economic Development
during Recessionary and Non-Recessionary Periods
Lilian W. Komo1 & Isaac K. Ngugi2
1
The Executive Business Centre, Lansdowne Campus, Bournemouth University, United Kingdom
2
The Business School, Talbot Campus, Bournemouth University, United Kingdom
Correspondence: Dr Isaac K. Ngugi, The Business School, Bournemouth University, Talbot Campus, Fern
Barrow, Poole, Dorset, BH12 5BB, UK. Tel: 44-120-2 96-3554. E-mail: [email protected]
Abstract
The performance of banks and their effects on the global economy has been of interest to politicians,
academicians, businesses and the general public especially in connection with the 2008/09 credit crisis. In
particular, the extent to which the bank stock prices affected the national stock market indices in different
countries before and during the crisis is unclear. It is also not clear whether the national stock markets of
countries at different levels of economic development reacted differently to the crisis. This paper contributes to
filling these gaps. It uses regression and correlation analytical techniques in the analysis of the impact and
behaviour respectively. The results suggest that the behaviour of most of the stock market indices was similar
regardless of the level of economic development. The mean stock market indices were statistically significantly
higher before than during the credit crisis. Bank share prices were generally negatively correlated comparing the
period before and during the crisis. Some countries were characterised by a few powerful banks that determined
the course of the respective national stock market index. This has implications on policy and reveals the
intervention points in regulating the performance of stock markets and stabilizing the financial sector.
Keywords: bank shares prices, stock markets, credit crises, economic development
1. Introduction
The 2008/09 financial crisis revealed that the power of globalisation can take a problem from one corner of the
globe to multiple destinations (Chittedi, 2009). The crisis began in the US subprime mortgage sector, causing
house prices to decline, economic activity to reduce and risk aversion to increase. It caused the failure of several
large US based financial firms and had ramifications on household consumption (Naudé, 2009). The crisis
rapidly spread to other parts in the world thereby causing failure of banks in Europe and decline in various stock
indices as well as significant reductions in the market value of equities and commodities (Usman, 2010). The
crisis spread further to other developed countries, as well as to emerging markets through a range of financial
and real sector channels (Stephanou, 2009). The international trade declined and concomitantly the economies
worldwide slowed down (World Bank, 2009). Governments were forced to make enormous financial
commitments to mitigate the situation.
The crisis was caused by factors such as: poor supervision of banks and financial institutions; the extension of
mortgage loans to borrowers at highly concessional terms; and excessive relaxation of fundamental rules and
regulatory requirements for financial institutions (Lunogelo et al., 2010; Diamond and Rajan, 2009). Financial
crises have been found to be related to a currency, a stock market or a banking crisis (Pericoli and Sbracia, 2003).
The financial assets that are considered in this regard include; currencies (AuYong et al., 2004), bonds (Yang,
2005) and stock markets (Syriopoulos, 2007).
Following the 2008/09 crisis, the performance of banks and their effects on the global economy has become of
much interest to politicians, academicians and the general public. Attempts have been made to examine aspects
of the crisis a well as the measurement of systemic risk and regulatory structure (Dwyer, 2012) but there are still
49
many lacunae. The crisis exposed major weaknesses in the extant repertoire of knowledge of the elements that
drive global finance and revealed several knowledge gaps. For example, it is not clear to what extent the bank
stock prices affected the national stock market indices in different countries. Also it is not clear if the national
stock markets in various countries reacted differently to the crisis and if this varied with the levels of economic
development. In line with these gaps, this study has four objectives: a) to examine the behaviour of national
stock market indices across countries at different levels of economic development b) to estimate the impact of
share prices of leading banks on respective national stock market index c) to compare the mean and behaviour of
national stock market indices before (2006-2007) and during (2008-2009) the crisis d) to compare the mean and
behaviour of bank stock prices across countries at different levels of economic development before (2006-2007)
and during (2008-2009) the credit crisis.
The results of this study are relevant to many stakeholders. They are useful to academia in informing the
efficient markets hypothesis - stock markets may be competitive or else they may be controlled by a few
dominant firms. They are relevant to policy makers, managers of financial institutions such as banks and the
regulators of financial markets especially in preparedness for and management of financial crises. The results are
also beneficial to shareholders in informing them about the behaviour of stock markets before and during a
financial crisis, which is relevant for their investment decisions.
2. Theoretical Background
2.1 Significance of Banks to the World Economy
Banks are significant to the world economy and make up a good portion of the equity market. The global
financial sector has nearly $6 trillion in market capitalization, implying that banks account for a significant share
in the global economy. In 2008, the financial stock worldwide, specifically equity market capitalization and
outstanding bonds and loans, was estimated at US$ 175 trillion and this increased to US$ 212 trillion by the end
of 2010 (Roxburgh et al., 2011). Banks are instrumental in providing capital for infrastructure, innovation, job
creation and overall prosperity while playing an integral role in society because they affect the spending of
individual consumers and the growth of entire industries (Cogan, 2008). Banks play a delegated role of
monitoring investments on behalf of investors (Diamond, 1984) and have the capability of reducing liquidity risk
thereby creating investment opportunity (Diamond and Dybvig, 1983).
Stock returns of banks are critical to the future economic growth and prosperity of nations (Cole et al., 2008) as
they are a key conduit of economic resources. It is a well-recognised fact that the development of the financial
sector can make an important contribution to economic growth and poverty reduction. This is especially true in
developing countries, given that their financial sectors are to a large extent not well developed, and without a
well-functioning financial sector, economic development may be constrained (Cogan, 2008).
Stock prices of banks that are listed in the domestic stock exchanges reflect the performance of the banking
sector in that country (Fariborz and Qiongbing, 2009). Furthermore, most stock market indicators tend to be
highly correlated with banking sector development (Demirguc-Kunt and Levine, 1996). Also, volatility in the
banking industry can be used to indicate the stability of a nation’s banking sector performance (Fariborz and
Qiongbing, 2009). Yartey and Adjasi (2007) establish that a percentage point increase in banking sector
development increases stock market development in Africa by 0.59 percentage point, therefore in turn affecting
macroeconomic stability and economic development.
2.2 Relevance of Stock Markets
Stock markets are relevant at least in: development of financial markets; promoting economic growth; and
enabling fair resource custodianship and allocation. It is through stock markets that many firms are able to
acquire capital quickly and efficiently (Kumar, 1984). The acquired capital is then invested or rather directed
into profitable projects, consequently promoting sustainable investment growth. Indeed, stock market activities
have been found to be positively correlated with investment (Tobin, 1969; Furstenberg, 1977).
The stock market accelerates economic growth by providing a boost to domestic savings and increasing the
quantity and the quality of investment (Singh, 1997). It also encourages savings by providing individuals with an
additional financial instrument that may better meet their risk preferences and liquidity needs (Levine and Zervos,
1998).
Stock markets have been found to have an influence on corporate governance (Abor et al., 2010). This is
particularly by moderating the interests of shareholders vis-à-vis those of managers (ibid). Perhaps more benefits
would be realized if markets were ideal, whereby prices provide accurate signals for resource allocation where
firms can make production-investment decisions, and investors can choose among the securities that represent
50
ownership of firms’ activities under the assumption that securities prices at any time ‘fully reflect’ all available
information (Beechey et al., 2000). Stock prices are related to the availability and cost of equity finance and they
influence firms’ investment decisions (Baker et al., 2003). Comparably, increases in the stock prices reflect
availability of equity finance at relatively lower cost.
2.3 Stock Market Indices
A stock market index is a mean price for a group of stocks especially those larger in quantity. The stocks are
usually on a particular stock exchange or across an investing sector. Indices are computed from stocks that have
something in common. For instance, such stocks could be from the same exchange or from the same industry.
Stock indices are indicators of the economic health of the particular industry that they represent (Vinod, 2009).
Vinod (2009) goes on to state that a stock market index tracks the performance of a specific basket of stocks
over a given day or a given period of time. It may have many specific uses to different participants in the
financial market place. One basic use of an index is in computing the total risk and return of the aggregate
market portfolio for a given period of time to set a benchmark for evaluating the performance of actively
managed individual portfolios. The study reckons that in some cases, the stock index is used as a proxy for
market sentiments and investor confidence in the capital markets.
2.4 Stock Prices
Stock prices basically signify the perceived value of the investment or the company that they represent. They
reflect the marginal productivity of capital and their increase consequently implies an increase in this
productivity. In turn, an increase in the marginal productivity of capital is directly related to an increase in
investment activities, and hence the positive correlation between changes in stock prices and investment growth.
Changes in prices of stocks form an important component of variation in the market value of capital (Fama,
1981). Thus, there is a causal relationship between changes in the prices and in the market value of capital and
investment. Since an increase in prices of stock is associated with an expansion in investment, stock markets
have been found to significantly predict investment (Barro, 1990).
3. Methodology
3.1 Sampling and Data Collection
The data for this research comprises of the national stock market indices of countries at the three levels of
economic development and bank stock prices of the leading banks listed in the respective stock markets.
Purposively, three countries were selected from each category. Due to the author's particular interest in
understanding more about Africa, the developing countries were selected only from the African continent. From
each country, leading banks that are listed in the national stock market exchange were selected depending on
level of capitalization as at the year 2010 and availability of data (Table 1). Leading banks registered at least
within the five years (2006-2010) were selected so as to capture data before (2006-2007) and during (2008-2009)
the credit crisis.
Data were extracted from Thomson Analytics, in US dollars. The Thomson Analytics database is professional,
reliable and easily accessible compared to gathering the data directly from the individual stock markets.
Extracting data from the individual stock markets is more expensive and time consuming.
The collected data comprises of national stock market indices and selected banks’ share prices for the period
2006-2009. This period is relevant because it allows a comparative analysis for the period before the credit crisis
(2006-2007) and during the crisis (2008-2009). The data included banks’ stock market closing price and monthly
price close data. The monthly price close data were for the stock indices. The use of monthly excess returns is
justified considering that examination of co-movements of short-term returns, that is, daily or weekly returns,
may understate the degree of integration due to non-synchronous trading and non-overlapping trading hours
across different markets (Karolyi and Stulz, 1996). Furthermore, concerning the potential long-run nature of
convergence and the co-integration analysis for common trends, it has been found that it is less the sample
interval than the time span which plays an important role in detecting co-dependencies (ibid).
51
Table 1. Selected countries, stock markets and banks

Level of Country Stock market Stock market index Banks
economic
development
Developing Kenya Nairobi Stock Exchange NSE Barclays Bank of Kenya Ltd, Kenya
countries Commercial Bank Ltd, Standard
Chartered Bank Ltd., NIC Bank Ltd, and
National Bank of Kenya Ltd
Nigeria Nigerian stock exchange S&P/IFCG First Bank of Nigeria Plc, Access Bank
(NSE) plc, Guaranty Trust Bank Plc, Stanbic
IBTC Bank Plc, United Bank for Africa
(UBA) Plc.
Egypt Egyptian stock exchange The Egypt Morgan Stanley Commercial International Bank Egypt,
Capital International (MSCI) Egyptian Gulf Bank, Export Development
Bank of Egypt and Faisal Islamic Bank of
Egypt.
Emerging China Shanghai Stock Exchange Shanghai Stock Exchange Bank of China, China Citic Bank, China
countries (SSE) Composite Index Construction Bank, China Merchants
bank and Industrial & Commercial Bank.
Brazil BM&FBOVESPA BOVESPA Banco Bradesco, Banco de Brasil, Itau
Unibanco and Santander Brazil.
India Bombay Stock Exchange BSE National Index HDFC Bank Limited, HSBC India, ICICI
Bank Limited, Punjab National Bank and
the State Bank of India.
Developed United London Stock Exchange Financial Times Stock Barclays, HSBC, Lloyds, Royal bank of
countries Kingdom Exchange (FTSE) 100 Share Scotland and Standard Chartered.
Index (Footsie)
Australia Australian Stock Exchange ASX index Commonwealth Bank, Westpac, ANZ,
market National Australia Bank and Bendigo and
Adelaide Bank.
United NYSE NYSE Composite Index Bank of America Corporation, HSBC
States of USA Inc., J. P. Morgan Chase & Co.,
America Bank of New York Mellon Corporation
and Wells Fargo.
3.2 Data Analysis

In social science research, statistical techniques are considered a major tool for data analysis (Nachmias and
Nachmias, 2008). In this study, SPSS is the main software that was used in data analysis and this largely
involved regression and correlation analysis. The key variables in this study are the national stock market indices
and the monthly banks’ share prices. For analysis of the behaviour of national stock market indices across
countries at different levels of development, correlation analysis was undertaken. To estimate the impact of share
prices of leading banks on respective national stock market indices, regression analysis was carried out.
Statistical tests were conducted to test the significance of the mean differences between the periods before
(2006-2007) and during (2008-2009) the global financial crisis.
4. Results and Discussion
4.1 Behaviour of National Stock Market Indices across Countries at Different Levels of Development
In this section, the results and discussion with respect to the first objective of this study are presented. The
behaviour of national stock markets across countries at different levels of economic development is examined.
This was achieved by analysing the correlation between the national stock market indices across the countries
during the four year period, 2006-2009 (Table 2). Pearson's correlation coefficient was estimated and this is
appropriate in measuring the strength of the linear relationship between two variables (Burns and Burns, 2008;
Stockburger, 1996), hence assessing herewith the similarity in the movements of the stock market indices across
countries and testing whether this was statistically significant or not (Table 2).
52
Table 2. Correlations between indices across stock markets in different countries

NYSE
Brazil Shanghai MSCI India FTSE 100 in S&P IFCG D
Index NYSEALL ASX NSE
index All Share Egypt BSE 100 US Dollars Nigeria in USD
COMPOSITE
NYSE Nyseall Pearson 1 .892** .241* .432** .675** .556** .992** .928** .711**
Composite correlation
Sig. .000 .040 .001 .000 .000 .000 .000 .000
ASX Pearson .892** 1 .628** .716** .829** .829** .860** .810** .719**
correlation
Sig. .000 .000 .000 .000 .000 .000 .000 .000
Brazil index Pearson .241* .628** 1 .697** .722** .831** .164 .215 .391**
correlation
Sig. .040 .000 .000 .000 .000 .119 .059 .002
Shanghai-All- Pearson .432** .716** .697** 1 .624** .845** .403** .391** .573**
Share correlation
Sig. .001 .000 .000 .000 .000 .001 .002 .000
MSCI Egypt Pearson .675** .829** .722** .624** 1 .785** .624** .695** .819**
correlation
Sig. .000 .000 .000 .000 .000 .000 .000 .000
IndiaBSE 100 Pearson .556** .829** .831** .845** .785** 1 .498** .493** .547**
correlation
Sig. .000 .000 .000 .000 .000 .000 .000 .000
FTSE 100 US Pearson .992** .860** .164 .403** .624** .498** 1 .922** .695**
Dollars correlation
Sig. .000 .000 .119 .001 .000 .000 .000 .000
NSE Pearson .928** .810** .215 .391** .695** .493** .922** 1 .785**
correlation
Sig. .000 .000 .059 .002 .000 .000 .000 .000
S&P IFCG D Pearson .711** .719** .391** .573** .819** .547** .695** .785** 1
Nigeria in USD correlation
Sig. .000 .000 .002 .000 .000 .000 .000 .000
Notes: **. Correlation is significant at the 0.01 level (1-tailed). *. Correlation is significant at the 0.05 level (1-tailed). N = 54.
The results in Table 2 show that the Pearson correlation coefficient of all the indices is positive. Six out of the
nine stock market indices are significantly positively correlated. However, in two instances the correlation was
not significant. First is the behaviour of the Nairobi Stock Exchange index and the Brazilian Stock Market Index.
The other instance is between the FTSE 100 and the Brazilian Stock Market Index. The Brazilian stock market
index had the highest index throughout the period of analysis. The three countries that had the correlation
coefficient not significant (Kenya, Brazil and UK) are at different levels of economic development. As expected,
the highest positive correlation amongst the stock market indices was between America and United Kingdom,
both of which are developed countries. In other words, FTSE 100 and the NYSE were the most highly correlated
with a 0.992 Pearson correlation coefficient.
Amongst the developing (African) and developed countries, the Kenyan and American stock market indices
showed the the highest correlation, with a Pearson correlation coefficient of 0.928. On the other hand, between
the developing and emerging countries, Egypt and India had the highest Pearson correlation coefficient, at 0.785.
Considering the developed and emerging economies, Australia and India had the highest Pearson correlation
coefficient, estimated at 0.829.
From the results in Table 2, it is evident that all the stock market indices were affected by the global financial
crisis. All the indices were significantly positively correlated at least with the USA index. There is a convergence
between the findings and theory in that, regardless of the country’s level of development, there was a correlation
in behaviour in most indices considering the pre and during the global financial crisis periods. This is consistent
with Efficient Market Hypothesis since it may be an indication that the stock market indices were reacting to the
news of a failing global economy. Furthermore, during a crisis, markets tend to strongly influence each other
(Malliaris and Urrutia, 1992; Arshanapalli and Doukas, 1993; Gklezakou and Mylonakis, 2009).
53
From the analysis (Table 2), it is a surprise that the stock market indices of Kenya (NSE) and UK (FTSE 100)
had the second highest correlation (the highest was between NYSE and FTSE). This contradicts the findings of
some studies that suggest that the less developed markets are relatively isolated from the capital markets of other
countries and they have a relatively low correlation with markets in the developed countries (Chien-Chiang et al.,
2010). However, the results of our study generally show that the stock market indices exhibited the same
efficient market hypothesis regardless of the country's level of economic development. This could be an
indication of how increasingly the world had become globalised and effects from one part of the world can easily
diffuse to the other ends. Furthermore, crises seem to create excessive correlations between countries.
Researchers call this contagion. A shock in one market leads investors to withdraw funds from other markets
because of irrational fears and thus leads to unusually high co-movements of asset prices (Dutt and Mihov,
2008).
4.2 Impact of Share Prices of Leading Banks on Respective National Stock Market Indices
In this section, the results and discussion with respect to the second objective of this study are presented. The
impacts of share prices of leading banks on respective national stock market index are estimated. This was done
on per country basis and involved estimation of regression equations with the national stock market index as the
dependent variable and the leading banks’ share prices as the independent variables.
4.2.1 Impact of Developed Countries Banks’ Share Prices on the Respective Stock Market Index
Equations 2, 4 and 6 shows the estimated regression models that illustrate the impact of developed countries
banks' share prices on the respective stock market index.
Australia
Specified regression model:
YAU STRALIA= βO +βANZ PANZ + βBENPBEN +βCBA PCBA + βNABPNAB + βWBCPWBC +ε (1)
Results:
YAUSTRALIA= - 460.502 -128.101 ANZ-AU + 122.500 BEN-AU + 92.235 NAB-AU + 186.718 WBC-AU + ε (2)
Four out of the five Australian banks significantly influenced the ASX index. Amongst them, only ANZ-AU
influenced negatively. CBA-AU was the only bank that did not have a significant impact on the Australian Stock
Exchange. WBC-AU had the greatest impact (β=187). The estimated model explains 94% of the variations in the
index, as indicated byR Square.
USA
YUSA = βO +βBoA PBoA + βBNYM PBNYM +βHSBCPHSBC + βJPM-NPJPM-N + βWFPWF+ε (3)
Results:
YUSA = -388.257 + 148.893BNYM + 86.093HSBC-USA + ε (4)
Only two banks were found to significantly impact on the NYSE index at 95% confidence interval, namely the
Bank-NY Mellon (ρ <.000) and HSBC USA (ρ <.000), with the Bank NY Mellon having a greater impact as
evidenced by the respective coefficient (β=149). The estimated model explains 94% of the variations in the
national stock market index, as indicated by the R Square. The share prices of the other three banks (BoA,
JPM-N and WF) did not have a significant impact on the NSE index at 95% confidence interval.
UK
YK = βO +βBARC-LNPBARC-LN + βHSBC-LN PHSBC-LN + βLLOY-LNPLLOY-LN+βRBS-LNPRBS-LN
+ βSTAN-LNPSTAN-LN +ε (5)
Results:
YUK=25.535-1.687 BARC-LN + 1.185HSBC-LN +5.851LLOY-LN + 1.864 STAN-LN + ε (6)
Four banks were found to impact significantly on the FTSE 100 index at 95% confidence interval, namely the
BARC-LN (ρ<.041), HSBC-LN (p<.014), LLOY-LN (p<.008) and STAN-LN (ρ<.000). However, Barclays
affected the index negatively while the rest affected it positively, as evidenced by the respective signs of the
coefficients. The estimates of the coefficients indicate that the share prices of LLOY-LN bank had a greater
54
impact on the FTSE 100 index than the shares of HSBC, Barclays and Standard Chartered. The estimated model
explains 98% of the variations in the national index. The share prices of RBS did not have a statistically
significant impact on the FTSE 100 at 95% confidence interval.
4.2.2 Impact of Emerging Countries Banks’ Share Prices on the Respective National Stock Index
Equations 8, 10 and 12 are the estimated regression models showing the impact of emerging countries banks'
share prices on the respective stock market index.
Brazil
YK = βO +β ITAU UNIP ITAU UNI + β SAN-BR P SAN-BR + β BBAS3-BRP BBAS3-BR + β BBDC4-BR P BBDC4-BR +ε (7)
Results:
YBRAZIL= -3495.049 + 913.868 ITAU UNIBANCO + 1418.830BBDC4-BR+ ε (8)
Two banks impacted significantly on the Brazil national stock market index at 95% confidence interval. These
were ITAU UNIBANCO (ρ<.002) and BBDC4-BR (ρ<.000). Both affected the index positively. The
coefficients indicate that the share prices of BBDC4-BR bank had a greater impact on the Brazilian national
index than the shares of ITAU UNIBANCO. The estimated model explains 96% of the variations in the index.
The share prices of SAN-BR and BBAS3-BR banks did not have a significant impact on the Brazilian index at
95% confidence interval.
China
YK = βO +βBoC PBoC + βCCBPCCB + βCConBPCConB + βCMB PCMB + βICB PICB+ε (9)
Results:
YCHINA= -29.013 + 194.441 CHINA CITIC BANK + 67.510 CHINA MERCHANTS BANK + 349.809 Bank of BoC +ε (10)
Three banks were found to significantly impact on the Shanghai All Share index at 95% confidence interval.
These were the China Citic Bank (ρ<.001), China Merchants Bank (ρ<.004) and Bank of China BoC (ρ<.049).
All the three banks affected the Shanghai All Share index positively. The coefficients indicate that the share
prices of Bank of China had the greatest impact on the index. The estimated model explains 97% of the
variations in the index. The share prices of the other banks (China Construction Bank and Industrial &
Commercial Bank) did not have a significant impact on the Shanghai All Share index at 95% confidence
interval.
India
YK = βO +β HDFC-BYP HDFC-BY + β HSBC INDIAP HSBC INDIA + β ICICI-BY P ICICI-BY + β PNB-BYP PNB-BY
+ β SBI-BY P SBI-BY +ε (11)
Results:
Y INDIA=28.349 + 3.315 HDFC-BY +4.258 HSBC INDIA +3.369 ICICI-BY -3.929PNB-BY + ε (12)
Four out of five banks were found to impact significantly on the India BSE 100 index at 95% confidence interval.
Unlike the other banks (ICICI-BY, HSBC INDIA and HDFC-BY), the share prices of PNB-BY bank had a
negative impact on the India BSE 100 index. The estimated model is strong as it explains 97% of the variations
in the index.
4.2.3 Impact of Developing Countries Banks’ Share Prices on the National Stock Index
Equations 14, 16 and 18 shows the estimated regression models that reflects the impact of developing countries
banks' share prices on the national stock market indices.
Nigeria
YK = βO + βACCESSBANK-LAPACCESSBANK-LA + βFIRSTBANK-LAPFIRSTBANK-LA + βGUARANTY-LAPGUARANTY-LA + βStanbic
IBTC-LAPStanbic IBTC-LA + βUBA-LAPUBA-LA +ε (13)
55
Results:
YNIGERIA= 51.142 + 357.859 FIRSTBANK-LA + 1322.847 STANBIC IBTC-LA + 374.759 UBA-LA+ ε (14)
Three banks out of the five were found to impact significantly on the Nigeria Stock Exchage index at 95%
confidence interval, namely the FIRSTBANK-LA (ρ<.000), Stanbic IBTC-LA (ρ<.001) and UBA-LA (ρ<.029).
The coefficients indicate that the share prices of Stanbic IBTC-LA bank had the greatest impact on the national
index. The estimated model explains 95% of the variations in the index. The share prices of the other two banks
(ACCESSBANK-LA and GUARANTY-LA) did not have a significant impact on the index at 95% confidence
interval although they had positive coefficients.
Kenya
YK = βO +βKCBPKCB + βBBKPBBK + βNBKPNBK + βNICPNIC + βSCBKPSCBK+ε (15)
Results:
YK = -14.604 - 47.352PNIC + 29.273PSCBK+ε (16)
Only two banks out of the five were found to impact significantly on the Nairobi Stock Exchange index at 95%
confidence interval. These were the NIC (ρ<.001) and SCBK (ρ<.000). Surprisingly, NIC affected the index
negatively while the SCBK affected it positively, as evidenced by the respective signs of the coefficients
(βNIC=.-47.352, βSCBK = 29.273). As shown by the values of the coefficients, the share prices of NIC bank had a
greater impact on the national index than the shares of SCBK. The estimated model explains 87% of the
variations in the index. The share prices of the other three banks (BBK, KCB and NBK) did not have a
significant impact on the national index at 95% confidence interval although they had positive coefficients.
Egypt
YK = βO +βCIBE PCIBE + βEGBE-CIPEGBE-CI +βEDBEPEDBE + βFIBPFIB + ε (17)
Results:
YEGYPT= 54.092 +15.734 CIBE + 25.325 EDBE+ 11.193 FIB + ε (18)
In Egypt, three out of the four banks influenced significantly the national index. These were; Commercial
International Bank Egypt (p<.002), Export development bank (p<.022) and the Faisal Islamic Bank (p<.003).
Both banks affected the stock market positively. The coefficients show that the share prices of CIBE had a
greater impact (β=16) on the stock market index than the share prices of FIB (β=11). The estimated model
explains 86% of the variations in the index. On the other hand, the share prices of the Egyptian gulf bank did not
have a significant impact on the Egyptian stock market index at 95% confidence interval although they had
positive coefficients.
In general, given the share prices of the banks whose coefficient were significant, the estimated models can be
used to predict the respective national stock market index. The countries that had most of the banks in their
chosen samples having a statistically significant impact on their stock market indices include Australia, United
Kingdom and India. These countries had four out of five banks significantly affecting their national stock market
indices. Closely following were Egypt and Nigeria, both with three banks affecting the index of their respective
stock markets. On the other hand, America, China, Brazil and Kenya all had only two banks whose share prices
statistically significantly affected the stock market indices.
4.3 Comparison of the National Stock Market Indices Prior and during the Credit Crisis
In this section, the results and discussion with respect to the third objective of this study are presented. This
relates to the comparison of the national stock market indices before (2006-2007) and during (2008-2009) the
credit crisis.
4.3.1 Mean Differences of the National Stock Market Indices before and during the Crisis
In Table 3, a paired sample t-test was carried out to test the mean differences between stock market indices
before and during credit crisis. The table shows a paired sample t-test value associated with the p-value and other
statistics. The results reveal that NYSE, Brazil index, ASX, NSE and FTSE all had a statistically significant
mean difference at 95% confidence interval comparing the periods before and during the credit crisis. This
implies that these indices were most affected in responding to the news of the credit crisis and therefore they
serve as empirical evidence that the efficient market hypothesis discussed in the literature review chapter is still
valid.
56
Table 3. Paired samples statistics showing the mean difference between national stock market indices and
average bank share prices before and during the crisis
Stock market indices Average bank share prices
Mean Std. Deviation Sig. df Mean Std. Deviation Sig.
Country difference (2-tailed) difference (2-tailed)
USA 2009.790* 2058.666 0.000 23 -11.516* 7.516 0.000
Australia 767.415* 1585.640 0.026 23 -5.552* 6.365 0.000
UK -33.532* 30.956 0.000 23 -5.296* 3.756 0.000
Brazil -6476.536* 10882.254 0.008 23 -0.153 3.452 0.830
India 5.769 66.678 0.676 23 1.225 6.594 0.372
China -17.522 252.168 0.737 23 0.088 0.523 0.418
Egypt -4.476 121.427 0.858 23 0.052 2.210 0.920
Kenya 18.505* 27.170 0.003 23 -0.014 0.464 0.882
Nigeria -23.455 356.511 0.750 23 -0.057 0.143 0.065
*Significant at 0.05 level of significance.
4.3.2 Behaviour of the National Indices before and during Credit Crises
Correlation analysis was carried out in order to identify if there was any statistically significant change in
behaviour of the national stock market indices before and during the credit crisis. Table 4 indicates that the stock
market indices for four countries out of the nine had statistically significant correlation for the period before and
during the crisis. All the correlations were negative except that of Brazil and India, probably an indication that
the credit crises spread at differential times across different countries. Else, this could be an indication that some
countries were hit harder than others. The two positively correlated coefficients were not statistically significant
and hence the observed phenomenon may have been a matter of chance. The correlation for the indices before
and after crisis was not statistically significant for the countries in the emerging markets.
Table 4. Paired samples correlation analysis of stock market indices and mean bank share prices before and
during the crisis
Correlation of stock market indices before and Correlation of mean bank stock prices before
during the crisis and during the crisis
Country N Pearson correlation Sig. Pearson correlation Sig.
coefficient coefficient
USA 24 -.668* .000 -.724* .000
Australia 24 -.316 .133 -.007 .976
UK 24 -.702* .000 -.715* .000
Brazil 24 .115 .591 .147 .492
India 24 .041 .848 .350 .094
China 24 -.040 .853 -.087 .687
Egypt 24 -.349 .095 -.481* .037
Kenya 24 -.909* .000 -.870* .000
Nigeria 24 -.827* .000 -.136 .525
Note: *Significant at 0.05 level of significance
Those indices that were statistically significant were negatively correlated. This implies that during the period
before crisis, they moved in one direction and for the period during the crisis, they moved in the opposite
direction. Comparing the two periods, the statistically significant correlated indices were NYSE (USA), NSE
(Kenya), FTSE (UK) and S&P IFCG (Nigeria). The stock market index with the strongest correlation was the
NSE with a correlation of -.909 while the index with the weakest significant correlation was NYSE (-.668). This
implies that the Nairobi Stock Exchange had the biggest change and appears to have been affected more than all
the other national stock market indices in this sample. This is extraordinary considering that Kenya is a
developing country that is quite distanced from the origin or epicentre of the credit crisis for instance in terms of
economic growth, industrialization and overall economic integration. This is however consistent with Mwega's
(2010) findings that Kenya’s stock exchange was the worst hit by the credit crisis considering the African
57
countries, with the NSE 20-Share Index slumping by 35% in 2008. Kenya was followed by Nigeria and
Mauritius which are countries that for long have liberalised their capital markets.
Given that five out of nine stock market indices had statistically significant mean difference and the behaviour of
four indices were statistically different comparing the pre and during the recessionary periods, this confirms the
findings of Modi et al. (2010) that the increasing integration among the national stock exchanges is an indication
of the ease of transmission of financial instabilities from international to domestic markets (financial contagion).
They explain that despite the fact that the poorest countries in the developing world, particularly in Africa, do
not have sophisticated financial markets, thus are not susceptible to direct financial contagion, does not prevent
the dangers arising from contagion. This is because the overall collapse in confidence in the financial system
world-wide raises borrowing costs sharply, cuts back revenues and threatens the solvency of domestic financial
systems even in low-income countries that may be poorly integrated into the international market. Moreover, the
banking sector in some developing nations is largely under foreign ownership. This may imply that decisions
taken by foreign banks to withdraw credit might compound difficulties caused by deleveraging especially as
banks sharply curtail lending activities and focus on consolidating the financial accounts of the parent company
(Mold et al., 2009).
4.4 Comparison of the Bank Stock Prices Reactions across Countries at Different Levels of Development before
and during the Credit Crisis
In this section, the results and discussion with respect to the fourth objective of this study are presented. The
hypothesis that there is no significant difference in the mean and behaviour of bank stock prices across countries
at different levels of development before (2006-2007) and during (2008-2009) the credit crisis is tested.
4.4.1 Mean Differences of Average Bank Share Prices before and after Credit Crises
For the bank share prices, data was taken for the 2006-2009 period. The average share prices, considering the
banks in our sample, for each month per country were calculated. A comparison between the periods 2006-2007
and 2008-2009 was done using a paired samples t-test (Table 3). Generally, the average share prices of most of
the banks reduced during the credit crisis compared to the positions held before the crisis. The developed
countries (Australia, America and United Kingdom) had the most notable changes as their bank stock price
averages reduced the most compared to the other countries and the mean difference in each case was
significantly different.
4.4.2 Behaviour of Average Bank Share Prices Prior and during the Recessionary Period
Table 4 presents the paired samples correlations of average bank stock prices across countries at different levels
of economic development before (2006-2007) and during (2008-2009) the financial crisis. The statistics indicate
that four out of nine average bank share prices had statistically significant correlation for the period before and
during the crisis. Those that were statistically significant were negatively correlated. The statistically significant
correlated average bank share prices included those of the banks in USA, UK, Kenya and Egypt. The country
with the strongest correlation in the average share prices for the two periods was Kenya with a correlation of
-.870 while that with the weakest correlation was U.K (-.715). The banks share prices of the developed and
developing countries were particularly significantly affected by the crisis. The global financial crisis stemmed
from the US-based transnational financial institutions and spread quickly beyond the US. The credit crisis first
spread to other developed countries and further to other regions including the developed countries (Roach,
2009).
5. Conclusions and Recommendations
This study achieved the four objectives set forth in the introduction. The results suggest that, the behaviour of
most of the stock market indices was similar regardless of the level of economic development. Some countries
were characterised by only a few powerful banks that determined the course of the national stock index. In other
words, only a few banks had significant impact on the national stock market index and the estimated models
were a good fit, each explaining more than 80% of the variations in the index. This reveals the intervention
points if policy makers were to stabilise the stock markets. National stock market regulators could set an early
warning department for detecting the early signs in the movement of the indices and establishing the cause. If the
movement is negative and caused by movement in shares of one or a few organisations, then corrective actions
could be targeted to such companies. A kitty could be established for purchasing shares of the affected
companies as a corrective measure and dispose them once stability is attained. The stock market index and
average bank share prices for most countries had statistically significant correlation before and during the
2008/09 credit crisis but moving in opposite direction.
58
It is evident from the results that regardless of the level of development of the countries, they all had some
reaction to the news of a looming credit crisis. Following the spread of the financial crisis to different parts of the
globe, the bank-share prices and the stock market indices registered changes to varying degrees. These responses
to the global crisis signify efficiency of the stock markets and this is consistent with the efficient market
hypothesis. The knowledge and measurement of the level of the efficiency of the stock market is essential to
policy makers, investors, and other major players that are involved in ensuring long-term real capital in a country.
As Alam and Uddin, (2009) note, across the globe, a mature level of the stock market efficiency is perceived as a
barometer of the economic health and prospect of a nation. It is also a reflection of the confidence of both the
domestic and global investor.
There is a need to reform the existing international financial architecture in order to reduce the likelihood of
financial crises and to manage them better. These reforms should be characterised by a full participation and
representation of countries at all levels of economic development. Despite the challenges posed by financial
globalization, it would be useful to design and implement a new approach to capital regulation and supervision if
future stability of the global financial system is to be guaranteed. This concurs with Moshirian's (2008) call for
stronger and more effective international institutions that are capable of monitoring the activities of international
financial institutions. Furthermore, strengthening capital restrictions and official supervisory powers can improve
the efficient operations of banks (Chortareas et al 2012).
Usman (2010) reckons that the incidences of financial crises have put the international financial system in
question. Perhaps the structures and principles of Islamic finance might remedy the situation, or probably help to
avert a global financial crisis in future. This financial system offers more ethical and efficient alternative to the
conventional interest-based financial system. It deters interest and introduces in its place the principle of risk and
reward-sharing. It also tends to daunt short selling because one may only sell something that he/she owns (ibid).
The system is considered a value-based one and which primarily aims at enhancing moral and material wellbeing
of the individual as well as of the society as a whole (Ahmed, 1994). It may be worthwhile to borrow some
principles about the frameworks that are used by this system to uphold ethical business practices in finance. With
the innovation of financial products still to continue, is it possible to learn and maybe use the technicalities that
are employed in coming up with ethically sound financial products and contracts. The financial industry could
borrow a leaf from the good qualities of Islamic finance as it seeks to perfect ethical business practices in the
global financial markets. Besides, financial institutions should always recognise that they have fiduciary
responsibility to act in the best interests of the consumer, while regulation and supervision of financial markets
should enhance trust in financial institutions and in the financial system.
6. Limitations and Suggestions for Further Research
Some banks did not have complete data for the entire period. For instance in the case of Brazil, data for the bank
Banco Santander was missing for the period between January 2006 and February 2007, whereas for Egypt, all
the four banks had some data missing between the years 2006-2008. This may cause a slight bias. This study is
limited to only three countries for each level of economic development. This means that there is scope in
enlarging this to a larger sample. In addition, non-probability sampling was used to select the countries and stock
exchanges from which data was extracted, therefore the objectivity and representativeness of the study may have
been compromised. Further research could adopt alternative sampling techniques. In some cases (e.g. Australian),
some banks were found to significantly influence the national index positively (despite the recession) while
others influenced negatively. It would be interesting to undertake in-depth case studies focused on such
unexpected or exemplar cases and to draw lessons.
References
Abor, J., Adjasi, C. K. D., Bokpin, G. A., & Osei, K. A. (2010). Do Emerging Financial Markets Matter in
Investment Opportunity Set? A Dynamic Panel Analysis. Journal of Money, Investment and Banking, 4.
Adjasi, C. K. D., & Biekpe, N. (2009). Do Stock Markets Matter in Investment Growth in Africa? The Journal of
Developing Areas, 43, 109-120. http://dx.doi.org/10.1353/jda.0.0032
Ahmed, Z. (1994). Islamic Banking: State of the Art. Islamic Economic Studies, 2, 1-34.
Alam, M. M., & Uddin, G. S. (2009). Relationship between Interest Rate and Stock Price: Empirical Evidence
from Developed and Developing Countries. International Journal of Business and Management, 4.
Anmar, P., & Beer, J. D. (2010). Comparing the South African Stock Market's Response to Two Periods of
Distinct Instability – the 1997/8 Asian and Russian Crisis and the Recent Global Financial Crisis. Retrieved
from
59
http://ec.ut.ee/eaces2010/artiklid/Pretorius%20de%20Beer-Comparing%20the%20South%20African.pdf
Arshanapalli, B., Doukas, J., & Lang, L. (1993). International Stock Market Linkages: Evidence from the Pre-
and Post-October 1987 Period. Journal of Banking and Finance, 17, 193-208.
http://dx.doi.org/10.1016/0378-4266(93)90088-U
Auyong, H., Gan, C., & Treepongkaruna, S. (2004). Cointegration and Causality in the Asian and Emerging
Foreign Exchange Markets: Evidence from the 1990s Financial Crises. International Review of Financial
Analysis, 13, 479-515. http://dx.doi.org/10.1016/j.irfa.2004.02.024
Barro, R. J. (1990). The Stock Market and Investment. Review of Financial Studies, 3, 115-131.
http://dx.doi.org/10.1093/rfs/3.1.115
Beechey, M., Gruen, D., & Vickery, J. (2000). The Efficient Market Hypothesis: A Survey. In Economic
Research Department. R. B. O. A. (Eds.). Retrived from: http://web.cenet.org.cn/upfile/93098.pdf
Boshoff, W. H. (2006). The Transmission of Foreign Financial Crises to South Africa: A Firm-Level Study.
Journal for Studies in Economics and Econometrics, 30, 61-85.
Burns, R. B., & Burns, R. A. (2008). Business Research Methods and Statistics Using Spss. London, SAGE.
Chien-Chiang, L., Jun-De, L., & Chi-Chuan, L. (2010). Stock Prices and the Efficient Market Hypothesis:
Evidence from a Panel Stationary Test with Structural Breaks. Japan and the World Economy, 22, 49-58.
http://dx.doi.org/10.1016/j.japwor.2009.04.002
Chittedi, K. R. (2009). Global Stock Markets Development and Integration: With Special Reference to Bric
Countries. MPRA Paper. Retrieved from
http://mpra.ub.uni-muenchen.de/18602/1/MPRA_paper_18602.pdf
Chortareas, G., Girardone, C., & Ventouri, A. (2012). Bank supervision, regulation, and efficiency: Evidence
from the European Union. Journal of Financial Stability, 8, 292-302.
http://dx.doi.org/10.1016/j.jfs.2011.12.001
Cogan, A. (2008). The Cluster Approach to Economic Growth. Business Growth Initiative Technical Brief No. 7.
ed. Washington, D.C, Weidemann Associates, Inc. Retrieved from: http://tinyurl.com/3jekqj
Cole, R., Moshirian, F., & Wu, Q. (2008). Bank Stock Return and Economic Growth. Journal of Banking &
Finance, 32, 995-1007. http://dx.doi.org/10.1016/j.jbankfin.2007.07.006
Demirguc-Kunt, A., & Levine, R. (1996). Stock Markets, Corporate Finance and Economic Growth: An
Overview. The World Bank Economic Review, 10, 223-239. http://dx.doi.org/10.1093/wber/10.2.223
Diamond, D. (1984). Financial Intermediation and Delegated Monitoring. Review of Economic Studies, 51,
Diamond, D., & Dybvig, P. (1983). Bank Runs, Deposit Insurance and Liquidity. Journal of Political Economy,
91, 401-419. http://dx.doi.org/10.1086/261155
Diamond, D., & Rajan, R. (2009). The Credit Crisis: Conjectures About Causes and Remedies. AER Papers and
Proceedings. forthcoming. http://dx.doi.org/10.1257/aer.99.2.606
Dutt, P., & Mihov, I. (2008). Stock Market Co-movements and Industrial Structure.
Dwyer, G., Tabak, B., & Vilmunen, J. (2012). The financial crisis of 2008, credit markets and effects on
developed and emerging economies. Journal of Financial Stability, 8, 135-137.
http://dx.doi.org/10.1016/j.jfs.2012.03.001
Fama, E. P. (1981). Stock Returns, Real Activity, Inflation, and Money. American Economlc Review, 71,
545-565.
Fariborz, M., & Qiongbing, W. (2009). Banking Industry Volatility and Banking Crises. International Financial
Markets, Institutions and Money, 19, 351-370. http://dx.doi.org/10.1016/j.intfin.2008.02.002
Furstenberg, V. (1977). Corporate Investment: Does Market Valuation Matter in the Aggregate? Brookings
Papers on Economic Activity, 2.
Gklezakou, T., & Mylonakis, J. (2009). Interdependence of the Developing Stock Markets, before and During
the Economic Crisis: The Case of South Europe. Journal of Money, Investment and Banking, 70-78.
Karolyi, G. A., & Stulz, R. M. (1996). Why Do Markets Move Together? An Investigation of Us-Japan Stock
Return Comovements. Journal of Finance, 51, 951-986.
60
http://dx.doi.org/10.1111/j.1540-6261.1996.tb02713.x
Levine, R., & Zervos, S. (1998). Stock Markets Banks and Economic Growth. American Economic Review, 88,
537-58.
Lunogelo, H. B., Mbilinyi, A., & Hangi, M. (2010). Global Financial Crisis Discussion Series Paper 20:
Tanzania Phase 21. Global Financial Crisis Discussion Series. London, Overseas Development Institute.
Retrieved from: http://www.odi.org.uk/resources/download/4738.pdf
Malliaris, A. G., & Urrutia, J. L. (1992). The International Crash of October 1987: Causality Tests. Journal of
Financial and Quantitative Analysis, 27, 353-364. http://dx.doi.org/10.2307/2331324
Modi, A. G., Patel, B. K., & Patel, N. R. (2010). The Study on Co-Movement of Selected Stock Markets.
International Research Journal of Finance and Economics. Retrieved from
http://www.eurojournals.com/IRJFE_47_15.pdf
Mold, A., Paulo, S., & Prizzon, A. (2009). Taking Stock of the Credit Crunch: Implications for Development
Finance and Global Governance. OECD Development Centre. http://dx.doi.org/10.1787/224377364587
Moshirian, F. (2008). Financial Services in an Increasingly Integrated Global Financial Market. Journal of
Banking and Finance, 32, 2288-2292. http://dx.doi.org/10.1016/j.jbankfin.2008.03.003
Mwega, F. M. (2010). Global Financial Crisis: Kenya Phase 2 Discussion Series. London, Overseas
Development Institute. Retrieved from: http://www.odi.org.uk/resources/download/3312.pdf
Nachmias, C. F., & Nachmias, D. (2008). Research Methods in the Social Sciences. New York, Worth.
Naudé, W. (2009). Fallacies About the Global Financial Crisis Harms Recovery in the Poorest Countries.
Retrieved from
https://www.cesifogroup.de/pls/guestci/download/C-ESifo%20Forum%202009/CESifo%20Forum%204/20
09/forum4-09-focus1.pdf
Pericoli, M., & Sbracia, M. (2003). A Primer on Financial Contagion. Journal of Economic Surveys, 17, 571-608.
http://dx.doi.org/10.1111/1467-6419.00205
Roach, S. S. (2009). A Lethal Shakeout. Foreign Policy. Washington, DC. Retrieved from
http://www.foreignpolicy.com/story/cms.php?story_id=4592&.
Roxburgh, C., Lund, S., & Piotrowski, J. (2011). Mapping global capital markets 2011. McKinsey Global
Institute.
Singh, A. (1997). Financial Liberalization, Stock Markets, and Economic Development. The Economic Journal,
107, 771-782. http://dx.doi.org/10.1111/j.1468-0297.1997.tb00042.x
Smith, M. J. (1998). Social Science in Question. London: Sage.
Stephanou, C. (2009). Dealing with the Crisis: Taking Stock of the Global Policy Response Crisis Response
Policy Brief 1. Retrieved from http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1427416
Stockburger, D. W. (1996.) Introductory Statistics: Concepts, Models, and Applications. Missouri State
University.
Syriopoulos, T. (2007). Dynamic Linkages between Emerging European and Developed Stock Markets: Has the
Emu Any Impact? International Review of Financial Analysis, 16, 41-60.
http://dx.doi.org/10.1016/j.irfa.2005.02.003
Tobin, J. (1969). A General Equilibrium Approach to Monetary Theory. Journal of Money Credit and Banking,
1, 15-29. http://dx.doi.org/10.2307/1991374
Usman, M. (2010). Global Financial Crisis: Its Impact on Developing Countries and Lessons for Pakistan. IPRI
Journal, 1, 93-118.
Vinod, T. (2009). Importance of Stock Market and How Stock Market Is Important for Countries Economy.
Retrieved from http://www.sharetipsinfo.com/economy-stock-market.html
World Bank. (2009). The Global Financial Crisis. Washington, DC, World Bank.
Yang, J. (2005). Government Bank Market Linkages: Evidence from Europe. Applied Financial Economics, 15,
Yartey, C. A., & Adjasi, C. K. D. (2007). Stock Market Development in Sub-Saharan Africa: Critical Issues and
Challenges. IMF Working Paper African Department. Retrieved from http://ssrn.com/abstract=1094214
61
ISSN 1916-971X E-ISSN 1916-9728
The Determinants of Informal Capital in the Financing of Small Firms

at Start-Up: An Ethnic Comparison of Small Firms in Sweden
Darush Yazdanfar1 & Saeid Abbasian2
1
Department of Social Science, Unit of Business Administration, Mid Sweden University, Sweden
2
Department of Social Science, Unit of Tourism Science & ETOUR, Mid Sweden University, Sweden
Correspondence: Saeid Abbasian, Department of Social Science, Unit of Tourism Science & ETOUR, Mid
Sweden University, Sweden. E-mail: [email protected]
Abstract
This study explains empirically the differences in the use of informal financing between native- and
immigrant-owned small businesses in terms of ethnicity and other relevant variables. A sample of 2814 native-
and immigrant-owned small businesses, consisting of a unique database gathered, was analysed and several
univariate and multivariate methods employed. The results suggest that ethnicity is a significant explanatory
variable and an important factor in informal capital access in the start-up stage in terms of loans from family
members and friends. Moreover, the other independent variables, namely gender, age, experience of starting
businesses, the amount of start capital, and firm size, affect loans from family members, whereas loans from
friends are influenced by age, size, and industry affiliation. Since knowledge about informal capital determinants
is limited, the results of this study add to our understanding of the variables that explain the financing behaviour
of small businesses at start-up.
Keywords: small business finance, ethnicity, informal capital sources, start-up stage, bank loan
1. Introduction
Many problems, including financing, are common for both native-born and immigrant owners in the small
business sector in western countries regardless of gender or ethnicity (Smallbone, Bertotti, & Ekanem, 2005).
Nevertheless, it is supposed that immigrants compared with native-born owners have considerable difficulties
raising start-up capital (Ram, Theodorakopoulos, & Jones, 2008; Ramangalahy, Brenner, Menzies, & Filion,
2002).
According to Irwin and Scott (2010), ethnicity and gender are two significant variables that seem to be barriers
to acquiring bank funding in the start-up phase. In this paper, we examine whether variables including ethnicity
and gender influence access to financial sources, especially the role of informal capital in financing small firms
in Sweden. By informal capital, we mean all financial support that is received in other ways than the formal
route (i.e. banks and governmental authorities). This includes all loans, grants, and contributions by family
members, friends, and relatives (Raijman & Tienda, 2003; Piperopoulos, 2010; Rezaei, 2007).
2. Previous Research
Despite the general perception that owners of small firms have problems with funding (Berger & Udell, 1995), it
seems that both ethnicity and gender still matters, as several previous studies have used them as analytical tools.
It is possible to correlate the significance of informal capital with numerous hypotheses and variables. Altinay
and Altinay (2006), for example, suggested four general hypotheses and 22 variables such as ethnicity, age of
owner, firm characteristics, and owner’s education and language skills to find any possible correlation between
the growth of small firms and these variables.
3. Ethnicity
Immigrants in general have less income than do native-born inhabitants (Cobb-Clarck & Hildebrand, 2002). In
addition, immigrant entrepreneurs show less likelihood than do their native counterparts to gain financial capital
from mainstream credit institutions and banks (Ram & Jones, 1998; Ram, Smallbone, & Deaknis, 2002; Bruder,
Neuberger, & Räthke-Döppner, 2011). Further, different groups of immigrant entrepreneurs are treated by banks
62
in different ways (Deakins, Ishaq, Smallbone, Whittam, & Wyper, 2007), which means that the chance of taking
out a bank loan is not equal for all immigrant groups/individuals. Moreover, the reliance on informal capital also
varies by ethnic group (Basu & Parker, 2001). In particular, immigrant-owned firms with no track records or
mainstream networks in the host society face larger difficulties gaining funding from banks (Ofori, 2005;
Cavalluzzo & Wolken, 2005). Therefore, immigrant entrepreneurs mobilise their financial resources mostly
through their own savings and ethnic informal sources rather than by formal mainstream networks (Bond &
Townsend, 1999; Huck, Rhine, Bond, & Townsend, 1999).
4. Gender
Gender is a major factor in the likelihood to gaining informal capital. According to previous research, women
face more problems than do men in raising venture capital or taking out bank loans (Greene, Brush, Hart, &
Saparito, 2001; Brush, Carter, Gatewood, Greene, & Hart, 2004; Fuller-Love, Lim, & Akehurst, 2006). Some
research does not exclude the impact of existing general perceptions on different gender-based behavioural
patterns (Robb & Wolken, 2002; Huang & Kisgen, 2008) that might affect banks’ attitudes towards women and
men.
Other authors (e.g. Fay & Williams, 1993) discuss gender-based clichés combined with discrimination from
financial institutions. Native-born female entrepreneurs are more often refused credit by banks and they pay
more interest than their male counterparts do (Verheul, van Stel, & Thurik, 2004; Muravyev, van Stel, & Thurik,
2007). By contrast, they have a significantly greater likelihood than do immigrants of obtaining a bank loan (see,
for example, Blanchflower, Levine, and Zimmerman, 2003). Women often start businesses with fewer resources
compared with men (Carter & Allen, 1997) and they often have fewer external sources of capital than men do
(Brophy, 1989; Brush, 1992). This might be mostly due to discouragement (Kon & Storey, 2003) or concern
about being refused financing (Marlow & Carter, 2006). The situation becomes even worse for female
immigrants as they have less access to financial resources compared with both native-born women entrepreneurs
and male counterparts from their own ethnic groups (Maltay, Scott, & Whittam, 2011; Davidson, Fielden, &
Omar, 2010; Essers & Benschop, 2007; Shanmuganthan, Spinder, Stone, & Foss, 2003; Schrover, Van Der Leun,
& Quispel, 2007).
5. Firm Characteristics
Firm characteristics (e.g. size, legal form, industry affiliation/sector, age) are also considered to play a role in the
differences in access to start-up capital among all business owners. Businesses owned by women and immigrants
are generally smaller and younger than are male-owned businesses (Devine, 1994; Coleman, 1998; Hussain,
Matlay, & Scott, 2008). Blanchflower et al. (2003) and Cavalluzzo, Cavalluzzo, and Wolken (2002) showed that
firm age, size, turnover, previous bankruptcy and creditworthiness, location, and industry affiliation might all
play a role in favouring male and female native-born entrepreneurs over male and female minority entrepreneurs
as well as favouring male native-born entrepreneurs against female native-born entrepreneurs.
Loscocco, Robinson, Hall, and Allen (1991) explained that female entrepreneurs have a disadvantage in
accessing finance through a combination of the firm’s and the owner’s characteristics such as firm size (turnover)
and income, lack of experience, and their concentration in the least profitable industries. Other studies have also
shown that businesses owned by women and immigrants in contrast to those owned by male native-born
entrepreneurs are highly concentrated to the services sector (Kalleberg & Leicht, 1991; Hedberg, 2009). Firms
operating in these traditional sectors rely more on personal savings and loans from family and friends, and less
on bank loans (Ram, Smallbone, Deakins, & Jones, 2003). Further, the legal form of the business has been used
to analyse differences (Bressler & Wiseman, 2011; Papadaki & Chami, 2002).
6. Owner Characteristics
Owner characteristics have been examined to explain differences between small business owners’ access to
start-up capital. Human capital, especially educational level and language ability, have been examined as impact
factors on immigrant business owners’ chances of gaining bank loans compared with indigenous business
owners (Rowley, 2004; Toussaint-Comeau, 2005; Altinay & Altinay, 2008). Bates’ (1997) study of Korean and
Chinese entrepreneurs in the US confirmed that highly educated owners had much better access to bank loans
than did less educated, who in turn mostly relied on their families and friends. Collins and Low (2010) reported
that variables such as education and language ability play an important role in Asian women’s
self-entrepreneurship and in their success in Australia. Carter, Brush, Greene, Gatewood, and Harts (2003)
suggested that there is a strong relation between the higher education of women and their odds of receiving
external financing.
63
The age of the owner has also been used as a possible factor that affects access to financial resources in several
previous studies (e.g. Dhaliwal & Kangis, 2006). Further, the owner’s previous work experiences,, including
previous experience of businesses plus outside work experience have also been used by several authors as
analysis tools. Studies, for example, show that male and female entrepreneurs differ in previous experiences and
professions, especially in prior entrepreneurial experiences (Brush, 1992; Verheul & Thurik, 2001; Fischer,
Reuber, & Dyke, 1993). Papadaki and Chami (2002) counted 13 impact factors on the growth of small
businesses, including outside work experience, previous businesses in the family, knowledge of the current
industry, and previous ownership. Moreover, Yazdanfar and Jahandar (2011) showed that owners’ previous
experiences of starting up a business and outside work experience influence the use of external capital in the
start-up stage.
Various authors emphasise the owner’s own savings and how this affects financing or success in different terms.
Although Kushnirovich and Heilbrunn (2008) emphasised access to own start-up capital, Basu (1998) found a
strong relation between own savings (personal capital) plus the share of informal capital in the start-up, on one
hand, and the success of the business, on the other hand. Cavalluzzo and Wolken (2005) suggested that personal
wealth and credit history are possible explanations for the differences in receiving bank loans.
Other previous studies point to the cultural gap as an additional obstacle for ethnic groups when they apply for
bank loans (Osili & Paulson, 2005, 2006; Alesina & La Ferrara, 2005). This induces immigrant-owned firms to
prefer internal and informal capital sources to external and informal ones (Barrett, Jones, McEvoy, &
McGoldrick, 2002; Ram et al., 2003). Culture and cultural proximity can also be a reason why some immigrant
groups are favoured or disfavoured by credit institutions (Albareto & Mistrulli 2011; Estapé-Dubreuil &
Torreguitart-Mirada, 2010).
7. Hypotheses
Based on previous studies and our available data, the following hypotheses have been formulated. Unlike
previous studies, the current research both uses a larger number of independent variables and has a larger
sample.
Hypothesis 1: Ethnic background is significantly related to attitude towards informal capital in terms of loans
from family members and friends in the start-up stage.
Hypothesis 2: Gender is significantly related to attitude towards informal capital in terms of loans from family
members and friends in the start-up stage.
Hypothesis 3: Owner age is significantly related to attitude towards informal capital in terms of loans from
family members and friends in the start-up stage.
Hypothesis 4: Previous experience is significantly related to attitude towards informal capital in terms of loans
from family members and friends in the start-up stage.
Hypothesis 5: Outside work experience significantly influences attitude towards informal capital in terms of
loans from family members and friends in the start-up stage.
Hypothesis 6: Educational level is significantly related to attitude towards informal capital in terms of loans from
family members and friends in the start-up stage.
Hypothesis 7: Access to personal start-up capital is significantly related to attitude towards informal capital in
terms of loans from family members and friends in the start-up stage.
Hypothesis 8: Firm size, in terms of number of employees, is significantly related to attitude towards informal
capital in terms of loans from family members and friends in the start-up stage.
Hypothesis 9: The legal form of the firm is significantly related to attitude towards informal capital in terms of
Hypothesis 10: A firm’s industrial sector is significantly related to attitude towards informal capital in terms of
8. Data Sources and Model Specification
8.1 Data Sources
The novel firm-level panel database used in this study was gathered by the Swedish Small Business Forum in the
autumn of 2008. This data sample consists of all available small firms incorporated between 2005 and 2008 in
four regions located in south-east Sweden. The preliminary sample covers 2832 firms. Firms with missing ethnic
64
background variables were removed from the sample, leaving a final sample of 2510 native- and 304
immigrant-owned active small firms.
8.2 Model Specification
To investigate the hypotheses formulated earlier, a number of univariate analyses – including analysis of
variance (ANOVA) – were employed as a first step to explore the differences between native- and
immigrant-owned firms in access to informal start-up capital. Subsequently, a binary logistic regression model
was implemented to identify the variables that can distinguish between Swedish native- and immigrant-owned
firms with regard to patterns of informal capital acquisition in the start-up stage.
The value of the dependent variables in the model can vary between 0 and 1. Those sample firms that have used
a loan from family members or friends take a value of 1, otherwise 0 (in other words, used = 1 and not used = 0).
The underlying estimation equation in the binary logistic regression model can be shown as follows:
∗
= 0+1(X1)+2(X2)+3(X3)+4(X4) +5(X5)+6(X6)+7(X7)+ 8(X8)+ 9(X9)+
where:
= Binary variable if the firm has used a certain type of funding source 1; otherwise 0
0= Constant
The dependent variable in this study is informal capital in terms of loans from family members and friends.
X1: Ethnicity (native: 1, immigrant 2)
X1: Gender (female: 1, male: 2)
X3: Owner’s age, natural logarithm of the age
X4: Earlier experience of starting up a business (Yes: 1, No: 0)
X5: Outside work experience (Yes: 1, No: 0)
X6: University education (Yes: 1, No: 0)
X7: The owners' personal start-up capital, natural logarithm of amount
X8: Number of employees as a proxy of firm size, natural logarithm of the number of employees.
X9: Dummy variable for the legal form of firm has been coded as, less formalised legal form: sole
proprietorships (1), trading partnerships as (2), and the most formalised legal form limited liability companies as
(3).
X10: Dummy variable for industry has been coded: manufacturing sector with most physical capital (1),
electronic or metal construction industries (2), transport (3), retail and wholesale (4), consulting and other
services (5), and restaurants and related market segments (6). random error
9. Empirical Results
9.1 Descriptive Analysis
Descriptive statistics are used to report the key characteristics of the sample in general as well as separately for
native- and immigrant-owned firms. The statistics contain the number of employees, amount of start-up capital,
age of owner, earlier experience of starting up a business, outside work experience, university education, legal
form, and industry affiliation.
As clearly shown by Table 1, there was no significant difference between the age of native- and immigrant-born
owners of firms (F = 3.37, p = 0.623). The average age of business owners was 42.5 years. On the question of
whether owners had ever started or owned other businesses, around 35 per cent of respondents answered YES. In
addition, more than 40 per cent of respondents expressed that they had outside work experience. However, there
was no statistically significant difference between native and immigrant owners in terms of outside work
experience (F = 2.66, p = 10.36). Whereas 24.3 per cent of respondents had a secondary education, only 15.4 per
cent had a university education. Native business owners had an almost similar educational level to that of
immigrant owners, as the average university education levels did not differ significantly. Similarly, the average
ages of native and immigrant owners did not differ significantly (F = 0.7852, p = 37.62).
65
Table 1. Mean, standard deviation, and number of variables included in the pooled sample
Age Started businesses Outside University Number of Start-up Legal Industry
before work experience education employees capital form affiliation
Natives Mean 42.650 1.655 1.576 1.819 1.966 3.840 20.390 3.249
Std. Deviation 12.584 0.479 0.498 0.424 2.957 2.110 16.458 1.422
N 2510 2510 2510 2510 2510 2510 2510 2510
Immigrants Mean 41.25 1.61 1.63 1.80 3.24 4.00 19.91 3.44
Std. Deviation 11.95 0.51 0.50 0.43 14.01 2.00 15.79 1.62
N 304 304 304 304 304 304 304 304
Total Mean 42.50 1.65 1.58 1.82 2.10 3.86 20.34 3.27
Std. Deviation 12.5228 0.4823 0.4985 0.4240 5.3938 2.0987 16.3851 1.4451
N 2814 2814 2814 2814 2814 2814 2814 2814

ANOVA F 3.3786 2.5011 2.6606 0.7852 15.2834 1.4954 0.2344 4.9584
Sig. 0.07 0.11 0.10 0.38 0.000 0.22 0.63 0.03
Welch test 0.056 0.132 0.104 0.379 0.617 0.114 0.204 0.045
Levene Statistic 0.028 0.001 0.005 0.139 0.100 0.000 0.039 0.000
Levene statistic: test of homogeneity at the 0.05% level; Welch: robust tests of equality of means between the two groups at the 0.01% level.
ANOVA at 0.05%.
The statistical analysis of our sample indicated that firms are in general young micro firms established between
2005 and 2008. Interestingly, the size of native- and immigrant-owned firms in terms of number of employees is
significantly different (F test = 15.25, p = 0.000); immigrant-owned firms have more employees than do
native-owned firms (F test = 15.2, p = 0.000). Moreover, with regard to legal form, around 70 per cent of all
sample firms are incorporated as sole proprietorships. The rest are incorporated as limited liability companies
(22%) and trading partnerships (8%). The ANOVA results show that there is no significant difference between
native- and immigrant-owned firms in terms of legal form (F = 0.23, p = 62.83).
Concerning industrial affiliation, 57 per cent of all firms are active in the services sector, 18 per cent in the
manufacturing sector, 12 per cent as retail firms, and the remaining 13 per cent are affiliated to other industries
including transport, construction, restaurants, and consulting. The results confirm the significant difference in the
mean scores of industry affiliation of native and immigrant-owned firms (F = 4.95, p= 2.65).
9.2 Sources of Start-Up Funds by Ethnic Background
The descriptive statistics of differences between native- and immigrant-owned firms with regards to informal
financing (i.e. loans from family members and friends) at start-up are shown in Table 2. The results demonstrate
the percentage of native- and immigrant-owned firms, respectively, which obtained capital from family members
(25–37%) and friends (4–20%).
The majority of firms classified loans from family members as a first financing alternative (26%) and loans from
friends as a second option (6%). Moreover, the sample is relatively characterised by a low standard deviation for
both native and immigrant groups (Table 2).
Table 2. Sources of start-up funds by ethnic background

Family members Natives 75% 25% 2,510 0.43
Immigrants 63% 37% 304 0.48
Total 74% 26% 2,814 0.44
Friends Natives 96% 4% 2,510 0.19
Immigrants 80% 20% 304 0.40
Total 94% 6% 2,814 0.23
66
Table 3 displays the mean, standard deviation and number of observation of dependent variables (Yes = 1 and
No = 0 answers), showing the importance of loans from family members and friends as financing sources of
firms in the start-up stage. The results show again that a higher proportion of immigrant-owned than
native-owned firms use loans from family members or friends to start their firms.
Table 3. Mean, standard deviation and number of dependent variables

National Family members Friends
Natives Mean 1.250 1.039
Std. Deviation 0.43 0.19
N 2,510 2,510
Immigrants Mean 1.37 1.20
N 304 304
Total Mean 1.26 1.06
N 2,814 2,814
9.3 Test of Statistical Significance: ANOVA Test

In this section, a significance test is performed to study which variables are statistically significant to explain the
differences between the external financing of native and immigrant groups. The null hypothesis is there is no
significant difference between native and immigrant groups concerning the use of different kinds of external
capital in the start-up stage. As presented in Table 4, the results of the ANOVA show, on one hand, that
statistically significant differences are found between native and immigrant groups regarding loans from family
members and friends at the 5 per cent level. On the other hand, the differences concerning other sources, namely
venture capital funding, government grants, funding from risk-capital companies, and other sources, are not
statistically significant.
The validity tests of ANOVA, including Welch’s test, Levene statistic, the and Jarque–Bera test (J-B test),
confirm the robustness of the results. To investigate further, logistic regression analyses were carried out.
Table 4. ANOVA, Levene, and Welch tests of the dependent variables

Sum of squares df Mean Square F Sig. Welch Levene J-B
test statistic p-value
Loans from Between groups 3.60673 1 3.60673 18.75236 0.000** 0.000 0.000 0.000
family Within groups 540.8453 2812 0.192335
members
Total 544.452 2813
Loans from Between groups 6.762878 1 6.762878 132.7525 0.000** 0.000 0.000 0.000
friends
Within groups 143.2531 2812 0.050943
Total 150.016 2813
Total 286.6951 2813
Notes: **Coefficients are significant at the 0.05 level; Levene statistic – test of homogeneity at the 0.05% level; Welch t test – robust tests of
equality of means between the groups and J-B p-value of means at the 0.05% level.
9.4 Logistic Regression Analyses

This section is based on the outcomes of four logistic regression analyses. It presents the empirical results,
research findings, and diagnostic validation tests.
9.4.1 Loans from Family Members
In the first step, all the independent variables were entered into the binary logistic regression model. However,
the independent variables in the final model are those that are most significant in explaining the changes in the
dependent variable and non-significant variables have been removed from the model. As shown by the result of
67
the first stepwise logistic regression, reported in Table 5, the significant variables that affect loans from family
members are owners’ ethnicity, gender, age, previous experience of starting a business, access to start-up capital,
and firm size. These results support hypotheses 1, 2, 3, 4, 7, and 8. However, the results provide no support for
the other hypotheses. Comparing the odds ratios (Exp (B)) of predictors, it can be confirmed that the change in
the dependent variable (loans from family members) is largely explained by ethnicity (Exp B:1,699), previous
experience of starting a company (Exp B:1,36), and firm size (Exp B:1,288). The size of the owner's personal
start-up capital (Exp B:1,10), gender (Exp B:0,624), and age (Exp B:0,348) also play a role in explaining
changes in loans from family members.
The validity of the model was examined by several diagnostic tests including the Omnibus test and Hosmer and
Lemeshow. These tests yielded a Chi-square value (χ2=60,4 p = 0.000<0,05) that showed that the explanatory
variables included in the model significantly affect the independent variable (loans from family members).
Moreover, the model has a Hosmer–Lemeshow goodness of fit (χ2=9.8, p = 0,27>0.05) and overall correctly
classified 73.7 per cent of the cases included in the sample. The results of the Cox and Snell tests and
Nagelkerke R squared also indicate that more than 21 and 31 per cent of the variation in the dependent variable
is explained by the set of independent variables. The Wald statistics provide further evidence of the significance
of the coefficients in the estimated model.
Table 5. First and second stepwise logistic regression analyses: dependent variables: loans from family members,
loans from friends
Loans from family members B S.E. Wald Df Sig. Exp(B)
Ethnicity 0,53034 0,12998 16,6471 1 0,00005 1,69951
Gender -0,47026 0,08323 31,9194 1 0,00000 0,62484
Age -1,05425 0,34309 9,44184 1 0,00212 0,34845
Experience of starting businesses before 0,31282 0,09874 10,0359 1 0,00154 1,36728
Start-up capital 0,10037 0,02235 20,1700 1 0,00001 1,10559
Size 0,25371 0,11952 4,50558 1 0,03378 1,28880
Constant -0,13195 0,66958 0,03883 1 0,84378 0,87638
Chi-square Df Sig. Classification accuracy 73.83
Omnibus tests 95,2288 6 0.0000 Cox and Snell tests 0,0332749
Hosmer and Lemeshow test 19,3383 8 0,01315 Nagelkerke R squared 0,0486744
Wald 582,390 1 0.0000 -2 Log likelihood 3143,166
Loans from friends B S.E. Wald Df Sig. Exp(B)
Ethnicity 1,72553 0,17946 92,4465 1 0,0000 5,61554
Age -2,20025 0,64121 11,7744 1 0,0006 0,11077
Size 0,44002 0,16775 6,88005 1 0,0087 1,55274
Industry 0,16188 0,05799 7,79230 1 0,0052 1,17572
Constant -2,05317 1,06604 3,70939 1 0,0541 0,12832
Chi-square Df Sig. Classification accuracy 94,3496802
Omnibus tests 113,5314 4 0.0000 Cox and Snell tests 0,03954216
Hosmer and Lemeshow test 10,87728 8 0,20875 Nagelkerke R squared 0,11221046

Wald 1189,01 1 0.0000 -2 Log likelihood 1109,070a
Notes: All regression coefficients are significant at the 0.05 level of significance (p = 0.000).
9.4.2 Loans from Friends

Table 5 reports the results of the second binary logistic regression model testing the hypotheses formulated in
section 1.1. As the findings show, in agreement with hypotheses 1, 2, 8, and 10, four independent variables –
ethnicity, owners’ age, firm size, and industry affiliation – are found to be significant at the 5 per cent level.
However, the results provide no evidence supporting the other hypotheses regarding the impact of other
explanatory variables on loans from friends. Thus, ethnicity, firm size, and industry affiliation influence the
variable loans from friends positively, while owners’ age is negatively related in this respect.
68
As shown by the diagnostic tests, the Omnibus (χ2 = 113.5, p=0.0000 < 0.05), Hosmer and Lemeshow (χ2 =
10.877, p = 0.20875 > 0.05), and Wald (χ2 = 1189, p = 0.000 < 0.05) tests all confirm the statistical significance
of the results. Moreover, the estimated model explains between 39.5 and 22.1 per cent of the variance (Cox and
Snell R2 and Nagelkerke, respectively), confirming an association between the dependent and independent
variables. The classification accuracy ratio of the model, which compares the prognosticated in relation to the
observed results, implies very good classification power of the variables (94.3).
To sum up, the explanatory variable ethnicity has a positive impact on the acquisition of informal capital in
terms of loans from family members and friends. In similarity with Irwin and Scott’s (2010) results we found
that ethnicity is significant barrier to receiving bank funding in the start-up phase, whilst in contrast to their
results we found that also gender constitutes barrier. Such barriers in turn motivate immigrant and female
entrepreneurs to borrow from the most accessible source, namely family members. In addition, the impact of
gender on loans from family members is negative, indicating that male owners tend to rely less on this financing
source than do their female counterparts. However, gender has no influence on loans from friends. The findings
also show that age has a negative effect on loans from family members and friends. This is different from
previous research (see Dhaliwal and Kangis, 2006). Whereas the experience of starting up businesses before and
the amount of start-up capital positively influence loans from family members, they have no effect on loans from
friends. Industry affiliation, which is positively related to loans from family members, has no influence on loans
from friends. Finally, the size of firms is positively related to both types of informal capital. Surprisingly, and in
contrast to the common perception in research, immigrant-owned firms in this sample are larger than are those
owned by native- born firms.
10. Conclusion
The increasing importance of entrepreneurship in job creation among immigrants has been acknowledged in the
literature. Given that the availability of capital plays a vital role in starting a business, this study focuses on the
role of ethnicity and other relevant variables in using informal financing in the start-up stage. One of the main
findings of the study is that owners’ ethnicity is significant in explaining differences between native- and
immigrant-owned firms in terms of the use of informal capital. Thus, immigrant-owned firms tend to rely more
on loans from family members and friends than do native-owned firms.
Furthermore, the findings provide evidence that immigrant-owned firms use more loans from family members
and friends than do native-owned firms. Gender has no association with loans from friends. Since age has a
negative effect on loans from family members and friends, older firm owners tend to use informal capital more
often compared with younger counterparts. In addition, whereas the impact of experience of starting businesses
earlier and the amount of start-up capital on loans from family members is significant and positive, these also
influence loans from friends. Moreover, industry affiliation, which is positively related to loans from family
members, is related to loans from friends. Finally, firm size is positively related to both types of informal capital.
References
Albareto, G., & Mistrulli, P. E. (2011). Bridging the gap between migrants and the banking system. Working
paper 794, Banca D’Italia Eurosistema.
Alesina, A., & La Ferrara, E. (2005). Ethnic Diversity and Economic Performance. Journal of Economic
Literature, 43(3), 762-800. http://dx.doi.org/10.1257/002205105774431243
Altinay, L., & Altinay, E. (2006). Determinants of Ethnic Minority Entrepreneurial Growth in the Catering
Sector. The Service Industries Journal, 26(2), 203-221. http://dx.doi.org/10.1080/02642060500369354
Altinay, L., & Altinay, E. (2008). Factors influencing business growth: the rise of Turkish entrepreneurship in
the UK. International Journal of Entrepreneurial Behaviour & Research, 14(1), 24-46.
Andersson, P. (2006). Four Essays on Self-Employment. Swedish Institute for Social Research, Stockholm
University: Stockholm.
Barrett, G., Jones, T., McEvoy, D., & McGoldrick, C. (2002). The economic embeddedness of immigrant
enterprise in Britain. International Journal of Entrepreneurial Behaviour & Research, 8(1/2), 11-31.
Basu, A. (1998). An exploration of entrepreneurial activity among Asian small businesses in Britain. Small
Business Economics, 10(4), 313-326. http://dx.doi.org/10.1023/A:1007956009913
69
Basu, A., & Parker, S. C. (2001). Family finance and new business start-ups. Oxford Bulletin of economics and
Statistics, 63(3), 333-358. http://dx.doi.org/10.1111/1468-0084.00224
Bates, T. (1997). Financing small business creation: The case of Chinese and Korean immigrant entrepreneurs.
Journal of Business Venturing, 12(2), 109-124. http://dx.doi.org/10.1016/S0883-9026(96)00054-7
Berger, A. N., & Udell, F. G. (1995). Relationship lending and lines of credit in small firm finance. Journal of
Business, 68(3), 351-381. http://dx.doi.org/10.1086/296668
Blanchflower, D. G., Levine, P. B., & Zimmerman, D. J. (2003). Discrimination in the small-business credit
market. The Review of Economics and Statistics, 85(4), 930-943.
Bond, P., & Townsend, R. (1999). Formal and informal financing in a Chicago ethnic neighbourhood. Economic
Perspectives (II), Federal Reserve Bank of Chicago, pp. 3-27.
Bressler, M. S., & Wiseman, M. (2011). Understanding Asian-American businesses and their role in economic
growth and development. Research in Business and Economics Journal, 4, 10. Retrieved from
http://www.aabri.com/rbej.html.
Brophy, D. (1989). Financing women owned entrepreneurial firms. In Hagan, O., Rivchun, C., & Sexton, D.
(Eds.), Women Owned Businesses (pp. 55-76). New York: Praegar.
Bruder, J., Neuberger, D., & Räthke-Döppner, S. (2011). Financial constraints of ethnic entrepreneurship:
evidence from Germany. International Journal of Entrepreneurial Behaviour & Research, 17(3), 296-313.
Brush, C. (1992). Research on women business owners: past trends, a new perspective and future directions.
Entrepreneurship Theory and Practice, 16(1), 5-30.
Brush, C., Carter, N. M., Gatewood, E., Greene, P. G., & Hart, M. M. (2004). Clearing the Hurdles: Women
Building High-Growth Businesses. FT/Prentice Hall. Upper Saddle River, NJ.
Carter, N. M., & Allen, K. R. (1997). Size determinants of women-owned businesses: choice or barriers to
resources? Entrepreneurship and Regional Development, 9(3), 211-220.
Carter, N. M., Brush, C. G., Greene, P. G., Gatewood, E., & Harts, M. M. (2003). Women entrepreneurs who
break through to equity financing: the influence of human, social and financial capital. Venture Capital,
Cavalluzzo, K. S., & Wolken, J. D. (2005). Small business loan turndowns, personal wealth and discrimination.
Journal of Business, 78(6), 2153-2178. http://dx.doi.org/10.1086/497045
Cavalluzzo, K. S., Cavalluzzo, L. C., & Wolken, J. D. (2002). Small business financing, and discrimination:
Evidence from a new survey. The Journal of Business, 75(4), 641-679. http://dx.doi.org/10.1086/341638
Cobb-Clarck, D., & Hildebrand, V. (2002). The wealth and Asset Holdings of US and Foreign Born Households:
Evidence from SIPP data. Discussion Paper 674 from the Institute for the study of Labor: Bonn, Germany.
December 2002. Retrieved may 2012 from: www.iza.org.
Coleman, S. (1998). Access to capital: a comparison of men and women-owned small businesses. Paper
presented at The Babson-Kauffman Entrepreneurship Research Conference, 21 May, Gent: Belgium.
Collins, J., & Low, A. (2010). Asian female immigrant entrepreneurs in small and medium-sized businesses in
Australia. Entrepreneurship & Regional Development, 22(1), 97-111.
Davidson, M. J., Fielden, S. L., & Omar, A. (2010). Black, Asian and Minority Ethnic female business owners:
Discrimination and social support. International Journal of Entrepreneurial Behaviour & Research, 16(1),
Deakins, D., Ishaq, M., Smallbone, D., Whittam, G., & Wyper, J. (2007). Ethnic minority business in Scotland
and the role of social capital. International Small Business Journal, 25(3), 307-326.
Devine, T. J. (1994). Characteristics of self-employed women in the United States. Monthly Labour Review,
117(3), 20-34.
70
Dhaliwal, S., & Kangis, P. (2006). Asians in the UK: gender, generations and enterprises. Equal Opportunities
International, 25(2), 92-108. http://dx.doi.org/10.1108/02610150610679529
Essers, C., & Benschop, Y. (2007). Enterprising identities: Female entrepreneurs of Moroccan and Turkish
origin in the Netherlands. Organization Studies, 28(1), 49-69. http://dx.doi.org/10.1177/0170840607068256
Estapé-Dubreuil, G., & Torreguitart-Mirada, C. (2010). Microfinance and gender consideration in developed
countries: the case of Catalonia. Management Research Review, 33(12), 1140-1157.
Fay, M., & Williams, L. (1993). Gender bias and the availability of business loans. Journal of Business
Venturing, 8(4), 363-377. http://dx.doi.org/10.1016/0883-9026(93)90005-P
Fischer, E., Reuber, R., & Dyke, L. (1993). A theoretical overview and extension of research on sex, gender, and
entrepreneurship. Journal of Business Venturing, 8(2), 151-168.
http://dx.doi.org/10.1016/0883-9026(93)90017-Y
Fuller-Love, N., Lim, L., & Akehurst, G. (2006). Guest editorial: Female and ethnic minority entrepreneurship.
International Entrepreneurship and Management Journal, 2(4), 429-439.
http://dx.doi.org/10.1007/s11365-006-0007-y
Greene, P. G., Brush, C. G., Hart, M. M., & Saparito, P. (2001). Patterns of venture capital funding: is gender a
factor? Venture Capital: An International Journal of Entrepreneurial Finance, 3(1), 63-83.
Hedberg, C. (2009). Intersections of Immigrant Status and Gender in the Swedish Entrepreneurial Landscape.
Working Paper 8, Stockholm University Linnaeus Center for Integration Studies (SULCIS).
Huang, J., & Kisgen, D. J. (2008). Gender and corporate finance. Paper presented at USC FBE Finance Seminar,
October. Los Angeles, CA.
Huck, P., Rhine, S., Bond, P., & Townsend, R. (1999). Small business finance in two Chicago Minority
neighborhoods. Economic Perspectives, 23(2), 46-62.
Hussain, J., Matlay, H., & Scott, J. (2008). Financial education in small ethnic minority businesses in the UK.
Education and Training, 50(8/9), 737-747. http://dx.doi.org/10.1108/00400910810917109
Irwin, D., & Scott, J. M. (2010). Barriers faced by SMEs in raising bank finance. International Journal of
Entrepreneurial Behavior & Research, 16(3), 245-259. http://dx.doi.org/10.1108/13552551011042816
Kalleberg, A. L., & Leicht, K. T. (1991). Gender and organizational performance: Determinants of small
business survival and success. Academy of Management Journal, 34(1), 136-161.
Kon, Y., & Storey, D. J. (2003). A theory of discouraged borrowers. Small Business Economics, 21(1), 37-49.
http://dx.doi.org/10.1023/A:1024447603600
Kushnirovich, N., & Heilbrunn, S. (2008). Financial funding of immigrant businesses. Journal of Developmental
Entrepreneurship, 13(2), 167-184. http://dx.doi.org/10.1142/S1084946708000910
Loscocco, K. A., Robinson, J., Hall, R. H., & Allen, J. K. (1991). Gender and small business success: An Inquiry
into women's relative disadvantage. Social Forces, 70(1), 65-85.
Maltay, H., Scott, M., & Whittam, G. (2011). Ethnic entrepreneurship in reverse in the UK: Is there gender bias
in access to finance for South Asian women entrepreneurs? Paper presented at the 56th Annual ISCB World
Conference, Stockholm: Sweden.
Marlow, S., & Carter, S. (2006). If you don’t ask you don’t get! Women, self-employment and finance. Paper
presented to Warwick Business School Small Firms Finance Conference, May, Coventry, UK.
Muravyev, A., Schäfer, D., & Talavera, O. (2007). Entrepreneurs’ gender and financial constraints: Evidence
from international data. Working paper from German Institute for Economic Research (DIW) and European
University Institute. July 16, 2007: Berlin: Germany.
Ofori, S. K. A. (2005). Financing the UK “Immigrant -entrepreneurs” sector: Policy, Practices and Opportunity.
Working paper from The Durham Law School and Durham Business School: Durham University.
Osili, U. O., & Paulson, A. (2005). Institutions and Financial Development: Evidence from International
Migrants in the U.S. Federal Reserve Bank of Chicago. Working Paper No. 2004-19.
71
Osili, U. O., & Paulson, A. (2006). What Can We Learn about Financial Access from U.S. Immigrants? Federal
Reserve Bank of Chicago, Working Paper No. 2006-25.
Papadaki, E., & Chami, B. (2002). Growth Determinants of Micro-businesses in Canada. Small Business Policy
Branch Industry, Canada.
Piperopoulos, P. (2010). Ethnic minority businesses and immigrant entrepreneurship in Greece. Journal of Small
Business and Enterprise Development, 17(1), 139-158. http://dx.doi.org/10.1108/14626001011019170
Raijman, R., & Tienda, M. (2003). Ethnic foundations of economic transactions: Mexican and Korean immigrant
entrepreneurs in Chicago. Ethnic and Racial Studies, 26(5), 783-801.
Ram, M., & Jones, T. (1998). Ethnic Minorities in business. Small Business Research Trust, Report January.
Open University Business School: UK. http://dx.doi.org/10.1080/1369183032000123440
Ram, M., Smallbone, D., & Deaknis, D. (2002). Ethnic Minority Business in UK: Access to Finance and
Business Support. London: British Bankers Association. http://dx.doi.org/10.1177/0950017008093479
Ram, M., Smallbone, D., Deakins, D., & Jones, T. (2003). Banking on break out: Finance and the development
of Ethnic Minority Businesses. Journal of Ethnic and Migration Studies, 29(4), 663-681.
Ram, M., Theodorakopoulos, N., & Jones, T. (2008). Forms of capital, mixed embeddedness and Somali
enterprise. Work, employment and society, 22(3), 427-446.
Ramangalahy, C., Brenner, G. A., Menzies, T. V., & Filion, L. J. (2002). Ethnic enterprise Start up stage: An
Empirical survey among the Chinese, Italian and Sikh entrepreneurial communities. Working paper no 15,
at McGill Conference on International entrepreneurship: Researching New frontiers, September 13th to 16th,
2002, McGill University, Montreal, Canada.
Rezaei, S. (2007). Breaking out: The dynamics of immigrant owned businesses. Journal of Social Science, 3(2),
94-15.
Robb, A., & Wolken, J. D. (2002). Firm, owner, and financing characteristics: differences between female- and
male-owned small businesses. Federal Reserve System Research Paper Series - FEDS Papers, Working
paper No. 18, March. Page nos: 28. Washington: USA.
Rowley, T. (2004). Entrepreneurship Means Adaptation. Retrieved from http://www.matr.net/print-12097.html
(accessed 22 October 2012).
Schrover, M., Van Der Leun, J., & Quispel, C. (2007). Niches, labour market segregation, ethnicity and gender.
Journal of Ethnic and Migration Studies, 33(4), 529-540. http://dx.doi.org/10.1080/13691830701265404
Shanmuganthan, P., Spinder, D., Stone, M., & Foss, B. (2003). Does ethnic focus change how banks should
implement customer relationship management? Journal of Financial Services Marketing, 8(1), 49-62.
http://dx.doi.org/10.1057/palgrave.fsm.4770106
Smallbone, D., Bertotti, M., & Ekanem, I. (2005). Diversification in ethnic minority business- the case of Asians
in London’s creative industries. Journal of Small Business and Enterprise Development, 12(1), 41-56.
Toussaint-Comeau, M. (2005). Self-employed immigrants: An analysis of recent data. Chicago Fed Letter, No
213. April, 4 pages.
Verheul, I., & Thurik, R. (2001). Start-up capital: Does gender matter? Small Business Economics, 16(4),
329-346. http://dx.doi.org/10.1023/A:1011178629240
Verheul, I., van Stel, A., & Thurik, R. (2004). Explaining female and male entrepreneurship across 29 countries.
Discussion Paper nr 0804 on Entrepreneurship, Growth and Public Policy. Max Planck Institutet for
Research into Economic Systems Group Entrepreneurship, Growth and Public Policy, Jena: Germany. Page
nos: 32.
Yazdanfar, D., & Jahandar, S. (2011). Acquisition of external capital at start-up stage: Differences between
Swedish female- and male-owned firms. International journal of Entrepreneurship and Small Business,
15(4), 435-451. http://dx.doi.org/10.1504/IJESB.2012.046474
72
ISSN 1916-971X E-ISSN 1916-9728
Examining the Value of Money in England over the Long Term

(1259-2009)
Adam Abdullah1
1
Faculty of Business Management and Accountancy, Universiti Sultan Zainal Abidin, Kuala Terengganu,
Malaysia
Correspondence: Adam Abdullah, Faculty of Business Management and Accountancy, University Sultan Zainal
Abidin, Gong Badak Campus, 21300 Kuala Terengganu, Terengganu, Malaysia. Tel: 60-9-668-7705. E-mail:
[email protected]
Received: January 20, 2013 Accepted: January 28, 2013 Online Published: February 22, 2013
Abstract
This paper seeks to analyze the value and purchasing power of money in England over 750 years, during which
time the pound has lost almost all of its value. This paper proposes a suitable methodology of measuring nominal
inflation in England over the long term, by constructing a wholesale commodity price index in order to
determine the purchasing power of money, being the inverse of prices. The index adjusts indices of the value of
gold and silver, to determine the purchasing power of gold, silver, and also real prices, thereby providing a
suitable technique to measure the store of value function of money. In so doing, we discover that a monetary
theory of value accurately interprets the empirical evidence.
Keywords: gold, silver and price indices, value and purchasing power of money
1. Introduction
Monetary economics provides a framework of analysis of the functions of money to measure the effects on
monetary systems and economies. In England in 1259, the official mint price of gold was 17s 9.3d per fine troy
ounce or £0.88875/FTO (Feavearyear, 1932, p. 347), whereas the average market price of gold in 2009 was
£621.59/oz, hence £1 in 1259 has sunk to £0.00143 (0.88875 / 621.59), or 1/7th of one penny in 2009. Whilst the
unit of account remained the same, the monetary standards varied. Hence, this paper involves a systematic
investigation of historical empirical evidence and statistics, to establish which medium of exchange maintains its
store of value. The purchasing power of money (PPM) measures the erosion of the real value of money due to
inflation, and conceptually is therefore the inverse of the price level (Rothbard, 1983, p. 30). Yet the value of
money (VM), or what a currency exchanges for a given weight of precious metal, is distinct from the purchasing
power of money, or the ratio of exchange between money and commodities (Ricardo, 2004, p. 90). By analyzing
the VM, the PPM and prices in England from 1259-2009, a monetary theory of value uncovers their causal
significance from the perspective of monetary theory, monetary economics and economic history.
The dispersion of prices requires a statistically convenient definition of the movements in the general level of
prices in the form of a price index number, and its reciprocal indicates the purchasing power of money (Fisher,
1911, p. 184). The change in the general price level of all commodities shows the relative value of money
(Warren, 1935, p. 10). Specifically, the purchasing power of gold (PPG) or the purchasing power of silver (PPS)
involves how much can be purchased with the proceeds of the sale of an amount of gold or silver. Statistically,
this requires the construction of an index for the price of bullion divided by an index of the price for goods and
services. It requires the construction of a unified series of commodity prices, which together with a unified price
of bullion series will be able to capture changes in purchasing power in order to analyze how changes in
monetary systems relate to the gyrations in purchasing power and how this has affected prices and economic
well being. The construction of a price index has the purpose of measuring variations in the purchasing power of
money, which requires an explanation of the methodology associated with the construction of a suitable price
index, including types of commodities, the treatment of missing values, the type of mean and the compilation of
relevant data.
An investigation into the value of money and prices over a long period of time, presumes both comparability and
continuity, indeed price history is “a study not of isolated facts but of relations; comparison is its essence. This
73
makes it necessary to make as sure as we can in each case that, in comparing prices at different times, we are
comparing like with like” (Beveridge, 1965, p. xxvi). In terms of purchasing power of money (in the context of
gold and silver), when we contemplate nominal and real prices, what we are really assessing is whether prices
take into account the value of money: Beveridge (1965) clarified that if we mean by this,
changes in the value of money in relation to commodities, the answer is that the course of prices, as recorded in
contemporary money, is itself the record and measure of such changes…[if] the change in the value of money is
meant a change in the silver or gold content of the currency, the answer is that [the compilation of historical
prices], gives the means of converting all such prices into bullion equivalents… The main purpose of price
history - [the] comparison of the values placed on different articles or services at different times as a guide to
changes of economic structure or conditions – can be achieved without this conversion. The use of bullion
equivalents is…for the special purposes of relating prices to the supply of precious metals or to the currency
policy (p. xlvii).
The prices that Beveridge (Note 1) was referring were expressed in terms of the British pound (GBP), which as a
monetary unit of account has been defined in silver, gold and now debt organized into paper (fiat money). Those
currencies, which provide the best continuity, are of the most interest to price historians, such as the British
pound (GBP). In terms of the functions of money, whatever the diminution of the bullion content, for long
periods of time, the GBP has never ceased to be used as a medium of exchange, as a unit of account and as a
standard of deferred payment for the settlement of a debt – however, what changed was the store of value, and
this function of money is measured by the changes in the value and purchasing power of money.
Of particular relevance is the historical data and methodology compiled by Jastram on the purchasing power of
gold and silver in England and the United States between 1560-1979. Roy W. Jastram (1915-1991) was an
American economist, whom wrote The Golden Constant (1977) and Silver: The Restless Metal (1981), which
were landmark works on the empirical study of gold and silver over the long term, but in his concluding remarks,
the intent was to establish that these metals had become de-monetized and highly manipulated commodities, and
in his opinion, there was no possibility of a return to a true gold standard, and yet believes that a monetary
system should be “managed” (Jastram, 1977, pp. 181-5). Indeed, Jastram was by no means a ‘gold bug’ since he
stated that “we are accustomed to thinking of gold itself as money. It is not…” (Jastram, 1977, p. 73), but did
admit that gold maintains its purchasing power over long periods of time and concluded that gold was a good
hedge against deflation but not inflation (Jastram, 1977, p. 132). Jastram also largely dismissed silver as having
been particularly subjected to manipulation over time, “We have escaped from clipped, debased and manipulated
coinage into manipulated, debased and politicized paper – not much of an exchange. World stability rests on
money” (Jastram, 1981, p. 158).
Jastram commenced his analysis from 1560 up to 1790 with data derived from the Mercantile Era [1550-1830]
of Beveridge’s Prices and Wages in England (1965), before the advent of well-organized indices, which Jastram
links into his series. However, Jastram ignores earlier price data available from Rogers (1866-1892), which
allows research to be extended back to 1259 capturing a period of very low and stable prices up until the Great
Debasement in the mid 16th century, which significantly increased prices. Furthermore, writing in the late 1970s,
Jastram could not have included the very inflationary 30 years since his books were published. Jill Leyland
updated Jastram’s research up to 2007, and was published by the World Gold Council in 2009, however, there is
still no price analysis prior 1560. Jastram adopted a wholesale price series and a simple geometric mean that
facilitates the adjustment of the base year without distorting the data. Jastram argues his un-weighted wholesale
commodity price index would arguably better represent the level of trade meaningful for the purchasing power
of a precious metal, as opposed to a cost-of-living index of what a wage-earner typically spends on a market
basket of retail goods and services that a family consumes (Jastram 1977, p. 61). It therefore includes a much
wider set of wholesale commodity prices generally for the period 1560-1790 covering (for example in 1700) up
to 16 categories and 72 commodities, as compared to a retail price index developed by Brown and Hopkins (BHI,
1956, 1981, p. 28-31), which although does commence at a much earlier date from 1264-1954, adopted a
weighted arithmetic mean index for retail prices that was designed as a cost-of-living index for worker’s families
and confined to consumer goods only, comprising 6 categories and 17 commodities.
O’Donoghue and Goulding constructed a composite price index in Consumer Price Inflation Since 1750, which
evaluates the purchasing power of money from 1750 to 2003 (ONS-CPI, 2004, pp. 38-46), but other than modern
retail price data from the ONS, is largely based on the Brown-Hopkins Index. Another interesting data set on
English agricultural prices was published by Clark (2004, pp. 41-124), The Price History of English Agriculture
1209-1914, involved a weighted-geometric mean for agricultural prices involving 26 commodities that were also
presented in arable, pasture and wood product sub-series, in order to assess movements in these series in relation
74
to agrarian history. At first glance, a weighted-geometric mean would statistically provide the most accurate
index according to Fisher (1922, p. 360). A long-term index of this nature could easily be linked into the Jastram
index without the trouble of constructing a new unweighted geometric mean for the same period, though Clarks’
farm price index at times understates prices given the output weights attached to the construction of the index.
The weights alter by period and by commodity, and some are based “on consumption of workers in mid 19th
century England”, admitting to “uncertainties” relating to farm output prior this time and “speculative” arable
and pasture outputs (Clark, 2004, pp. 44-45,56). Fisher added one proviso that “if the simple weighting does not
happen to be too erratic, the geometric is the best formula” (Fisher, 1922, p. 212), and Keynes stipulated that
constructing a consumption index for the purposes of assessing purchasing power is appropriate provided that
“actual consumption furnishes us with our standard [of weighting]” (Keynes, 1930, Vol.1, p. 54), but in reality,
“no such consumption data exists” (Jastram, 1977, p. 67). In essence, we concur with Jastram’s views and
expand on the methodology below as to why we also have chosen the simple geometric mean to construct a
wholesale price index in determining the purchasing power of money (PPM), the PPG, the PPS, and other than
prices expressed in GBP, also determine prices expressed in bullion.
This paper is organized into four sections. The first section details the methodology in compiling indices for the
price of gold and silver in England over the long term. The second section details the methodology in
constructing a wholesale commodity price index from 1259-1560, including the selection of price series, the
treatment of missing values, an analysis of the type of mean adopted and an explanation concerning the
compilation of the entire index, by linking up with other well-organized indices from 1560-2009. The third
section presents an analysis of our findings, together with an interpretation of a monetary theory of value, whilst
the fourth section provides a brief summary and concluding remarks.
2. Indices for the Price of Gold and Silver
The main reason for choosing to examine the prices of gold, silver and commodities through the GBP is that,
whilst the English changed their system of weight from tower to troy pounds in 1526, the monetary unit of
account has not changed at all, which provides for a consistent period of study over the long term (Note 2). For
the price of gold and silver, the average annual prices of prevailing fineness, converted into troy ounces, quoted
in order of preference (1) by the market, (2) by the bank’s buying price, (3) by the mint. The mint price was the
price fixed by law at which the mint could buy bar gold or silver. Until 1526 when the troy measure was
introduced into England, the price was in tower lbs (the tower pound was 15/16 lighter than a troy lb of 5760
troy grains), and for the purposes of comparison, has been converted into troy ounces.
The market price represents the price of gold or silver freely arrived at between parties, whereas the Bank of
England’s and the mint’s buying price represents the posted price at which these institutions stand ready to
purchase bullion even if there are no takers. In reality there was no meaningful private market price beyond the
mint price until the advent of the Bank of England in 1694, nor was there any meaningful market beyond the
bank's selling/buying price after 1694. The market price typically fluctuated between the mint price, which was
the bank's selling price, and the bank's buying price. Following the Bank Act of 1844, the market price of gold
traded within a very narrow band between the bank's selling price of £3 17s 10.5d and the bank’s buying price of
£3 17s 9d. By the time England was firmly established on a de jure gold standard in 1816, the Bank was
effectively controlling the market price of gold (Frankel, 1953, p. 36). Historically, English currency was silver
and was typically weighed rather than accepted by tale (Rogers, 1866, Vol.1, p. 175). Soon after the Bank was
established, England went on to a de facto gold standard from 1717, but already a market price for silver had
developed from the late eighteenth century, with regular quotes for “Pieces of Eight” or Spanish dollar (peso or
piastre), a very popular trade coin, that was not only adopted by the Americans (the dollar) but also by the
Ottomans (kurus) in 1703. Sources for English prices of gold and silver are presented in tables 1 and 2.
75
Table 1. Compilations of the English price of gold, 1257-2009

Source Period Frequency Denomination
Mint Price1 1257-1692 annual mint price s, d per FTO
Houghton, Castaing’s,
1693-1789 annual market price £ / STO .917 gold3
Lloyd’s List2
Parliamentary Papers 18194 1790-1818 annual market price £ / STO .917 gold
Lutyens’ Course of the Exchange5 1819-1832 annual market price6 £ / STO .917 gold
Sharps, Pixley Ltd London7 1833-1967 annual market price £ / STO .917 gold to 1918 & .995 ex 1919
LBMA8 1968-2009 annual market price £ / FTO of 995 gold
Notes: 1 Feavearyear (1932, pp. 347-8), Jastram (2009, pp. 318-330).
2
John Houghton’s A Collection for Improvement of Husbandry and Trade quoted gold and silver prices from 1693-1697 (Li, 1963, pp.
10-11). Castaing’s The Course of the Exchange and Other Things included the market price per oz of standard 22k gold in £s and Spanish
pieces of eight silver in pennies. It was first published in 1698 and became the forerunner of the London Stock Exchange’s Daily Official List.
The Lloyds News was established in 1696 by Edward Lloyd, and was the forerunner of the Lloyds List, which was published from 1734, and
also quoted twice-weekly prices including gold and silver from 1735. For the period 1700-1717 the Bank’s buying price was adopted since
no market prices were quoted for these years in Castaing’s (Jastram 2009, pp. 318-330).
3
Standard gold (or ‘crown’) gold was 917 gold or 91.7% pure (22 karat).
4
Jastram (2009, pp. 318-330)
5
Jastram (2009, pp. 318-330)
6
Rothschilds became the exclusive bullion broker to the Bank of England in 1825 (Jastram, 1977, p. 24). The first bullion brokers were
Mocatta and Goldsmid with Moses Mocatta starting in 1671.
7
Jastram (2009, pp. 318-330); Shrigley (1935, pp. 90-92) quotes market gold:silver ratios from 1687-1932, actual gold prices from
1870-1932, actual silver prices from 1833-1933; source for both Jastram & Shrigley is Sharps, Pixley (Pixley & Abell).
8
London Bullion Market Association gold price data is from The London Gold Market Fixing Ltd.
Table 2. Compilations of the English price of silver, 1257-2009

Source Period Frequency Denomination
Mint Price1 1257-1692 annual mint price d per STO (92.5%)
Houghton, Castaing’s, Lloyd’s List2 1693-1810 annual market price £, s, d per STO
Parliamentary Papers 18193 1811-1818 annual market price £, s, d per STO
Lutyens’ Course of the Exchange4 1819-1832 annual market price £, s, d per STO
Sharps, Pixley Ltd London5 1833-1967 annual market price £, s, d / STO to 1944, & .999 silver ex 1945
LBMA6 1968-2009 annual market price £ decimal from 1971
Notes: 1 Feavearyear (1932, p. 346) has exact prices. Jastram (1981:164-188) also carries mint prices per standard troy oz (sterling silver of
0.925% pure) with data provided for Mocatta Metals Corporation, providing tower and troy prices with metric equivalent to 4 decimals.
2
Houghton silver prices from 1693-1697 (Li, 1963, pp. 10-11). Castaing’s original folios available from the Guildhall Library and Jastram
(1981, pp. 164-188) to 4 decimal places. For the period 1700-1717 the Bank’s buying price was adopted since no market prices were quoted
for these years in Castaing’s.
3
Jastram (1981, pp. 164-188)
4
Jastram (1981, pp. 164-188)
5
Jastram (1981, pp. 164-188); Shrigley (1935, table at end of book) quotes market gold:silver ratios from 1687-1932, actual gold prices from
1870-1932, actual silver prices from 1833-1933; source for both Jastram & Shrigley is Sharps, Pixley (Pixley & Abell).
6
LBMA silver price data is from London Silver Market Fixing Ltd.
3. Price Series, Missing Values, Type of Mean, Index Compilation

Other than the construction of gold and silver price indices, we can construct a suitable index for English
commodity prices between 1259-1560, to tie-in with existing indices constructed thereafter, as discussed below.
The primary source for actual prices is Thorold Rogers (Note 3) (1866-1892). In order to assure breadth of
coverage, the following groupings, established by the International Scientific Committee on Price History for the
study of historical prices (Note 4), have been adopted (table 3):
76
Table 3. Grouping of articles for price history

I Grain and Other Crops XI Light, Fuels, etc.
II Grain Products XII Textiles
III Livestock, Meat and Poultry XIII Hides and Skins, etc.
IV Dairy Products, Fats, etc. XIV Building Materials
V Fish XV Metals
VI Vegetables XVI Chemicals and Miscellaneous
VII Fruit XVII Other Articles of Common Use
VIII Sugar, Spices, etc.
IX Miscellaneous Foods
X Drinks
Our analysis accordingly compiled a number of price series that reflected both a credible width and depth to
English prices involving 31 items reflected in all 17 groups (listed in table 4 below).
In order to avoid distortion of the overall price index, a minor problem arose regarding the methodology of
determining missing values. Our general precedent for estimating missing values being derived from available
values of other products are Pamuk (2002) and Clark (2004): “In cases where the prices of one or more of these
items were not available for a given year, missing values were estimated by an algorithm that applied regression
techniques to the available values” (Pamuk, 2002, p. 297). “...Where individual price quotes were missing, the
index was interpolated using the prices of other products” (Clark 2004, p. 46). For a more detailed methodology
of interpolation and extrapolation our precedent is Allen (2001), “...There were many gaps in the underlying
price series. Generally, these were filled by interpolation. As a result, year-to-year fluctuations in the price level
are damped, but the general trends and relative levels...are preserved” (Allen 2001, p. 420), and also Abel
(1980),
In years when there are oat prices but none for rye, the three preceding and the three following rye prices are
expressed as a percentage of the corresponding oat prices. Subsequently, the oat price in question is multiplied
by the index number thus obtained, and the resulting value inserted into the series for the year in which the price
of rye is missing. If the price of rye in the years 1,2,3,4,5,6,7,8, etc. is called A,B,C,D,E,F,G,H, etc. and the oat
prices a,b,c,d,e,f,g,h, etc., the missing rye price E in year 5 can be determined by the following equation.
BC  D F G H
e  E
bcd  f  g h
The rye price obtained by this method is put between brackets in [the] table. (p. 301).
Whilst Abel's equation provides for a more responsive regression-like technique, where a more careful analysis
is ultimately desired during periods of marked inflation or deflation, we have therefore adapted a more sensitive
formula by limiting the averaging of available values to one year, or in the case of longer periods of missing
values, the indexing of the current over the preceding year. Of course, nothing can better an original observation,
but we found that these revised techniques work well to nonetheless generate a more realistic movement in
prices. The ability to index values from other items is facilitated with a larger collection of price series.
The only remaining issue was to establish from which other price series missing values were to be computed,
and this was a subjective, albeit logical, exercise in comparison. Estimated values derived from associated
relationships between arable, pasture, meat, fish, dairy, other farm products, drink, wood, fuel, candles, textiles,
light industry, with sources of data, are ultimately captured together with observed price data, and our indicative
guide on this was Allen (2001, App.1, pp. 435-441). Average annual prices were decimalized from original
observations quoted, for example in £, s, and d, in order to facilitate the calculation of missing values, being
typically derived from a related genus.
With respect to the construction of index numbers, in considering a stochastic approach to measuring inflation
and the purchasing power of money, analysis is required on what type of index number to adopt, and what
mathematic average to apply. Fisher examined index numbers in The Purchasing Power of Money (1911) and in
The Making of Index Numbers (1922). In assessing what index number was best, he concluded that, “if the
simple weighting does not happen to be too erratic, the geometric is the best formula” (Fisher, 1922, p. 212).
Fisher went on to specify the “ideal” formula, for use in his quantity theory of money and equation of exchange,
MV=PT, where T is the volume of trade as represented by Q or the quantity of goods and services purchased by
prices (P). An index number of P implies an index number of T, or an index number of trade (in the form of
77
quantities of goods and services), thus P1T1 = ∑P1Q1 and so Fisher’s “ideal” formula (Fisher, 1911, p. 418; 1922,
pp. 197, 241-2) echoes Walsh’s (1901, p. 429) geometric cross (mean) of the Laspèyre (with base year quantities
as weights) and Paasche (with current year quantities as weights) formulae, or LP  as represented in the
following equations for prices (1.0) and for quantities (2.0).
 P1Q0  P1Q1
for prices (price index) . (1)
 P 0Q0  P 0Q1
 Q1 P0  Q1 P1
for quantities (quantity index) . (2)
 Q 0 P 0  Q 0 P1
However, our analysis involves compiling historical prices without consumption data necessary for a standard of
weighting. The unweighted arithmetic ratio-of-aggregates (∑P1/∑P0) type of index has a heavy inherent bias
(Jastram, 1977, p. 68). The best measure for the average ratio of change in prices is the geometric mean
(Mitchell, 1938, p. 76), and an unbiased simple index will in practice result only in small differences in
comparison with a carefully constructed weighted index (Fisher, 1922, p. 445). Instead of taking the arithmetic
mean by adding a set of price ratios together and dividing n into the result, we can construct a simple geometric
mean of a set of price-ratios by multiplying them together and taking the nth root of their product, with n being
the number of commodities included (3.0),
n P1 . P 1 ... (3)
P0 P 0
This is also known as the Jevons Index (Jevons, 1883, p. 332), and is identical to the unweighted geometric mean
prices (4.0) (ILO-PPI, 2004, p. 217).
1/ n
 P1   ( P1)1/ n (4)
PJ    
 P0   ( P 0)1/ n
Mitchell in The Making and Using of Index Numbers (1938) attributes three advantages to the adoption of a
geometric mean:
First, unlike the arithmetic mean, [the geometric mean] is not in danger of distortion from the asymmetrical
distribution of price fluctuations…If, for example, one commodity rose tenfold in price and another commodity
fell to one-tenth of the old price, the arithmetic mean would show and average rise of 505 per cent (1,000+10) / 2,
while the geometric mean would show no change in the average, since 1,000 x10   100 …The second merit
claimed for the geometric mean is that they can be shifted from one base period to the other without producing
results that seem to be inconsistent…A third advantage of the geometric means is that they are likely to be nearer
the modes of distributions, which they represent than are arithmetic means (pp. 69-71).
We can adopt the geometric type of index from the 13th to the mid 19th century, and link up with similar other
well-organized indices from the end of the 18th century that capture not just agricultural prices, but also
manufacturing goods associated with the industrial revolution, and the development of an increasingly urbanized
and industrialized economy together with an increasing service component, requiring a more detailed index
construction from the mid 19th century. By relying on compilations of prices data or indices published directly
from their authors or official publications, and providing full disclosure as to where and how the data was
obtained and compiled, we may ascertain the reliability and credibility of the underlying source material. In this
regard, table 4 presents a chronology for the compilation of English prices, that builds on the methodology of
Jastram, by applying the same approach he took for the period 1560-1790 in constructing a geometric index of
wholesale commodities from Beveridge’s data (Note 5), and construct a similar geometric index based on
Rogers’ data, for an earlier period from 1259-1560. At the other end of the spectrum, we can extend the overall
index from 1979 to 2009 with updated producer price index (PPI) data from the Office of National Statistics.
78
Table 4. Compilations of English commodity prices, 1259-2009

Source Period Time Base or Currency Price Measure
Rogers (1866-92)1 1259-1560 s, d contract, wholesale
Jastram-Beveridge Index2 1560-1790 1930 wholesale
Gayer-Rostow-Schwartz Index3 1790-1850 1821-1825 wholesale
Sauerbeck-Statist Index4 1850-1965 1867-77 wholesale
Central Statistical Office WPI5 1965-1979 1930 wholesale
Office of National Statistics PPI6 1979-2009 2005 wholesale
1
Notes: Rogers (1866-1892) compiled price data from 1259-1793
2
Jastram (1977, pp. 30-33), Beveridge (1965) provided price data from 1550-1830 and commodity price-ratios, which were individually on a
base average of 1720-1744=100 (Beveridge, 1965, pp. 682-741, and folding pages). Jastram’s index, of up to 72 commodities, was of the
geometric type with the base year at 1930, since the Board of Trade once used 1930 as a base year (Jastram, 1977, pp. 72, 203).
Jastram-Beveridge data was used until 1790 although for 1692-1702, a price index by Li (1963, p. 9) of weekly prices of 13 commodities
from Houghton was adopted in lieu of Jastram.
3
BHS (1962, p. 470), Jastram (1977, pp. 190-193). The Jastram-Beveridge Index was chained to the geometric Gayer-Rostow-Schwartz
Index (1953) of 76 commodities, by linking-up in the overlap year of 1790 though division, and it too was of the geometric type, for the
period 1790-1850.
4
BHS (1962, p. 474), BHS (1971, p. 187), Jastram (1977, pp. 194-198). The Gayer-Rostow-Schwartz Index is spliced into the
Sauerbeck-Statist Index (1886), for the period of 1850-1965, which is a simple arithmetic index of 45 prices series for 38 commodities. The
Board of Trade also produced a weighted arithmetic wholesale price index from 1871-1955, but instead the Sauerbeck-Statist Index is
preferred since it was published from 1846-1965.
5
BHS (1971, p. 187), Jastram (1981, pp. 196-199). No data during WWII with the CSO’s Index Number of Wholesale Price but continues
from 1946-1979. Instead of adopting revised ONS PPI data from 1931-2007 in Jastram (2009, pp. 283-296), the Sauerbeck-Statist Index was
continued until 1965 and chained with the original CSO index in Jastram (1981) from 1965-1979.
6
The Central Statistics Office was merged into the Office for National Statistics in 1996, and the Index Number of Wholesale Price was
renamed the Producers Price Output Series (PPI) in 1978. The ONS PPI was linked to the CSO WPI from 1979-2009. The CSO/ONS adopt
the Laspèyres arithmetic quantity-weighted index.
Statistical analysis involves an overall population rather than a sample of gold, silver and price data, in order to
achieve an actual rather than a probable outcome. The correlation coefficient for a population is applied when
measuring the strength of dependence between two variables, such as, correlating the nominal PG and CP over
the entire period of analysis to determine the impact of the diminution of monetary value upon commodity prices.
Correlation does not reveal causality, and another way of assessing performance is to present a summary analysis
of long term stability and short term volatility on prices thereby examining the related monetary systems and
rank their performance (Mueller, 2010, pp. 329-331, 441). Recognizing this says as much for the performance of
the relevant monetary authorities as for the medium of exchange itself, we may adopt two measures of
comparison: (1) long term price stability measured by the average absolute annual change in the index for
wholesale prices over the periods mentioned, and (2) short term volatility, measured by the standard deviation of
annual wholesale price changes during those periods, derived from the population standard deviation (5.0),
( x  x ) 2 (5)

n
Weighing stability and volatility equally, we can then present a stability rank for each monetary system and also
provide some additional commentary. Ultimately, by statistically determining a full population of commodity
prices in real terms, in terms of gold and silver, we will achieve the desired analysis in explaining actual
causality in the relationship between a loss of intrinsic value and higher nominal commodity prices.
4. Findings
In the preceding section we have carefully provided a suitable methodology for the construction of indices for
the prices of gold (PG), silver (PS) and commodity prices (CP). Additionally, we will present our findings
involving the PPG, the PPS, the British pound adjusted by wholesale commodities (PPBw) to reflect the
purchasing power of money (PPM), and the pound as a unit of account adjusted by the price of gold (GBPg) and
the price of silver (GBPs). We will then present wholesale commodity prices expressed in terms of gold (CPgc)
and silver (CPsc), in order to determine the reason for the devaluation of the GBP as a unit of account, whether
as specie or as fiat money. Finally, we will interpret our findings in the context of a monetary theory of value,
and relate real prices with the theory of value in exchange. The PG, CP and PPG are presented in figure 1, and
the PS, CP and PPS are presented in figure 2. Clearly, in both figures, we may immediately observe that the PPG
79
holds its store of value over the long term. We may also observe the positive relationship between gold and
prices, and the correlation coefficient between PG and CP is 0.93. The PG (or PS) is a hedge against commodity
price deflation when the currency is expressed as a fixed weight of precious metal, but when the currency is not
fixed and is free to discover its rate of exchange with a fixed weight of precious metal, such as under the fiat
standard, then PG (or PS) is a hedge against price inflation.
INDICES OF THE PRICE OF GOLD (PG), WHOLESALE COMMODITY PRICES (CP)

AND THE PURCHASING POWER OF GOLD (PPG), ENGLAND, 1259-2009 (1560=1)
(logarithmic scale)
1,000.00
100.00
10.00
1.00
0.10
0.01
1259 1309 1359 1409 1459 1509 1559 1609 1659 1709 1759 1809 1859 1909 1959 2009
PG CP PPG
Sources: (PG) - MP, HCL, PP, L, SP, LBMA, (CP) - R, BHS, JB, GRS, SS, CSO, ONS
Figure 1. Indices of the Price of gold, wholesale commodity prices, and the purchasing power of gold, England,
1259-2009
INDICES OF THE PRICE OF SILVER (PS), WHOLESALE COMMODITY PRICES (CP)

THE PURCHASING POWER OF SILVER (PPS), ENGLAND, 1259-2009 (1560=1
(logarithmic scale)
1,000.00
100.00
10.00
1.00
0.10
0.01
1259 1309 1359 1409 1459 1509 1559 1609 1659 1709 1759 1809 1859 1909 1959 2009
PS CP PPS
Sources: (PS) - MP, HCL, PP, L, SP, LBMA, (CP) - R, BHS, JB, GRS, SS, CSO
Figure 2. Indices of the price of silver, wholesale commodity prices, and the purchasing power of silver, England,
1259-2009
If we moved the base reference year from 1560 to 1259, without creating any distortion given that we have
adopted the geometric mean, co-movement is very close prior the Great Debasement (1542-1551), and equally if
we move the base reference year to say 1930, the same close movement is apparent under the gold standard. In
other words, devaluation is not the exclusive preserve of fiat money, but manifestly occurred as a result of
debasement by reducing the gold and silver content in coinage during the reigns of Henry VIII and Edward IV
and referred to as the Great Debasement. Hence, our choice of 1560 for the base reference year in figures 1 and 2
is to retain the structure of the co-movement between prices, precious metals and their purchasing power.
However, we need to adjust the GBP as a unit of account both in terms of wholesale commodity prices and in
terms of both gold and silver, in order to better evaluate the value and purchasing power of money in England
over the long term. Accordingly, in figure 3 we present the purchasing power of gold (top line), the value
(middle) and the purchasing power of the pound (bottom) over the long term, with the base reference year at
1259.
80
INDICES OF THE VALUE OF THE ?IN TERMS OF GOLD (GBPg), THE PURCHASING
POWER OF THE ?(PPBw) AND OF GOLD (PPG), ENGLAND, 1259-2009 (1259=1)
(logarithmic scale)
10.000
1.000
0.100
0.010
0.001
1259 1309 1359 1409 1459 1509 1559 1609 1659 1709 1759 1809 1859 1909 1959 2009
GBPg PPBw PPG
Sources: (GBPg) - MP, HCL, PP, L, SP, LBMA, (PPBw) - R, BHS, JB, GRS, SS, CSO, ONS
Figure 3. Indices of the value of the GBP in terms of gold, the purchasing power of the British Pound, and the
purchasing power of gold, England, 1259-2009
As England moved off the gold standard (1931) and the gold exchange standard (1971), the loss of value and
purchasing power is very evident post 1971, with the GBP seemingly falling off a monetary cliff, in terms of
value (GBPg) and purchasing power (PPBw). The reciprocal of the PPBw is nominal inflation (CP), being
wholesale commodity prices expressed in pounds. The GBPg is identical to the index of the market value of the
pound expressed in grams of pure gold and thus represents the intrinsic value and gold content (GC) of the
pound. The reciprocal of GC is the index of the price of gold (PG). We conclude from figure 3 that the loss in the
value of the pound has affected its loss in purchasing power. Meanwhile, we also present our findings in
connection with silver in figure 4. England was essentially on a silver standard before it moved on to a de facto
gold standard in 1717 and a de jure gold standard in 1816. Moreover, following the de-monetization of silver in
the U.S. in 1873, the international price of silver subsequently declined, and thus we cannot perhaps read too
much into the value of the pound when adjusted for silver (GBPs) when the medium of exchange was operating
under a gold or fiat standard, but we can be more complimentary about silver prior the Great Debasement
(1542-1551), given the prevailing monetary system was specifically a monometallic silver commodity standard,
and revealed that it was no less stable than the de facto gold standard that existed after 1717.
INDICES OF THE VALUE OF THE ?IN TERMS OF SILVER (GBPs), THE PURCHASING
POWER OF THE ?(PPBw) AND OF SILVER (PPS), ENGLAND, 1259-2009 (1259=1)
(logarithmic scale)
10.000
1.000
0.100
0.010
0.001
1259 1309 1359 1409 1459 1509 1559 1609 1659 1709 1759 1809 1859 1909 1959 2009
GBPs PPBw PPS
Sources: (GBPs) - MP, HCL, PP, L, SP, LBMA, (PPBw) - R, BHS, JB, GRS, SS, CSO, ONS
Figure 4. Indices of the value of the GBP in terms of silver, the purchasing power of the British Pound, and the
purchasing power of silver, England, 1259-2009
Notwithstanding, the affects of the Great Debasement, the gold standard and finally the de-monetization of silver,
the PPS over the long term remains reasonably constant, and to appreciate the implications of this more clearly,
we now turn to exploring the relationships between nominal and real prices. To facilitate our analysis, we present
the index of nominal wholesale commodity prices in figure 5.
81
INDEX OF WHOLESALE COMMODITY PRICES (CP

ENGLAND, 1259-2009, (1259=1)
(logarithmic scale)
1,000
100
10
0
1259 1309 1359 1409 1459 1509 1559 1609 1659 1709 1759 1809 1859 1909 1959 2009
Sources: WPI - R, BHS, JB, GRS, SS, CSO, ON
Figure 5. Index of wholesale commodity prices, England, 1259-2009
In figure 5, we notice that from 1259-2009, over a period of 750 years, nominal prices in England, expressed in
pounds increased about 550 times. With the index at 1.00 in 1259, by the outbreak of WWI in 1914 it had risen
to 10; off the gold standard after WWI in 1920 it tripled to 30 before declining to 15 on resumption of the gold
standard in 1925; upon England’s final exit from the post-war gold standard it was back at 10; and off the gold
standard after WWII in 1945 it had doubled to 20; and upon the collapse of the gold exchange standard in 1971 it
stood at 64; but without any recourse to gold, it rose to 550 by 2009. In order to determine the extent of
devaluation, we need to express commodity prices in terms of gold (or silver) and obtain real prices. We know
that the reciprocal of the PG, being the index of the market value of the unit of account for a fixed weight of gold,
or the money value of gold, is the GC, being the index of the market value of gold in grams of pure gold per unit
of account, or the gold value of money (the intrinsic value of a unit of account, such as the pound). To express
prices in terms of gold, we multiply the index of commodity prices (CP) by the index of the gold value of money
(GC), hence, CPgc = CP x GC. Similarly, to obtain real prices in terms of silver, CPsc = CP x SC. We present
CPgc in figure 6 and CPsc in figure 7.
WHOLESALE COMMODITY PRICE INDEX IN GRAMS OF PURE GOLD (CPgc),

ENGLAND, 1259-2009 (1259=1)
(linear scale)
5.0
4.5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
1259 1309 1359 1409 1459 1509 1559 1609 1659 1709 1759 1809 1859 1909 1959 2009
CPgc 20 Year Moving Average
Figure 6. Commodity prices expressed in terms of gold, England, 1259-2009
82
WHOLESALE COMMODITY PRICE INDEX IN GRAMS OF PURE SILVER (CPsc),

ENGLAND, 1259-2009 (1259=1)
(linear scale)
14.0
12.0
10.0
8.0
6.0
4.0
2.0
0.0
1259 1309 1359 1409 1459 1509 1559 1609 1659 1709 1759 1809 1859 1909 1959 2009
CPsc 20 Year Moving Average
Sources: (PS) - MP, HCL, PP, L, SP, LBMA, (CP) - R, BHS, JB, GRS, SS, CSO, ONS
Figure 7. Commodity prices expressed in terms of silver, England, 1259-2009
When we express nominal CP in bullion equivalent prices, in gold (CPgc), we note that over the same period and
utilizing a 20 year moving average, it is quite remarkable that prices stayed within the relatively narrow range of
1.0 to 3.5 for gold and whilst there were medium term movements, these occurred around a long-term trend that
rose only modestly, hence debasement or devaluation was the most important determinant of English prices over
the long-term. For silver (CPsc) the 20 year moving average range was 1.0 to 3.6 up to 1717 before England
went onto a de facto gold standard, rising to 5.0 in 1816 when England went on the gold standard de jure, before
deteriorating to 13.5 after the silver was de-monetized in 1873 and fiat money imposed itself in the 20th century –
that is still a lot better than 550 in nominal terms. What is very impressive is just how stable silver was prior to
the Great Debasement during a silver commodity standard, when prices remained within a very narrow range of
1.0 to 1.4 and gold remained within 1.0 to 1.2 implying that the silver coinage in circulation was deteriorating
faster than the nominal price of gold.
We complete our analysis, by presenting in figures 8 and 9, the bullion equivalents of prices in gold (CPgc) and
silver (CPsc), so that CPgc = CP x GC, and CPsc = CP x SC. Their reciprocal, as with PPM = 1/P, expresses the
PPG or PPS, so that PPG = 1/CPsc, and the PPS = 1/CPsc.
WHOLESALE COMMODITY PRICE IN GRAMS OF PURE GOLD (CPgc) AND THE

PURCHASING POWER OF GOLD (PPG), ENGLAND, 1259-2009 (1259=1)
(logarithmic scale)
100.0
10.0
1.0
0.1
1259 1309 1359 1409 1459 1509 1559 1609 1659 1709 1759 1809 1859 1909 1959 2009
CPgc PPG
Figure 8. Commodity prices expressed in terms of gold and the purchasing power of gold, England, 1259-2009
83
WHOLESALE COMMODITY PRICE INDEX IN GRAMS OF PURE SILVER (CPsc) AND

THE
PURCHASING POWER OF SILVER (PPS), ENGLAND, 1259-2009 (1259=1)
100.0
10.0
1.0
0.1
0.0
1259 1309 1359 1409 1459 1509 1559 1609 1659 1709 1759 1809 1859 1909 1959 2009
CPsc PPS
Sources: (PS) - MP, HCL, PP, L, SP, LBMA, (CP) - R, BHS, JB, GRS, SS, CSO, ONS
Figure 9. Commodity prices expressed in terms of silver and the purchasing power of silver, England, 1259-2009
In other words, CPpg or CPgc are real prices that revolve around a long-term trend that increases only modestly,
but revolves around a constant of 1.00, had the VM been maintained, rising and falling with the natural supply
and demand of commodities, even when we consider the subsequent structural economic adjustments in trade
that took place as a result of serial monetary distortion. By clearly determining that the increase in nominal
prices was caused by debasement and devaluation, we can say that the reduction in the value of money, as a
result of an increase in the excessive supply of money in relation to demand resulted in an increase in the price
level, hence price inflation is the effect and not the cause. This takes into account, both the quality and quantity
of money, for it takes into account value as well as demand and not just supply, as opposed to the quantity theory,
which largely ignores both value and demand, in favour of a reliance based upon the supply of money alone.
Hence, quantity theorists would encourage monetary policy to target inflation by altering the quantity of money,
given that they argue that there is a predictable relationship between the supply of money (M) and prices (P),
where velocity (V) is constant and real output (Y) is given, such that MV = PY, as derived from Irving Fisher’s
equation of exchange (1911). Pigou’s Cambridge approach (1917) and Freidman’s monetarist re-statement (1956)
are variants on the same theme, since they hold that the money demand function, being the reciprocal of V, is
stable.
Furthermore, a monetary theory of value is also distinct from the mercantilist purchasing power theory, which
has evolved into the effective demand of Keynes. The common thread that runs through mercantilism is the
notion that, the authorities should undertake an active role in the intervention and manipulation of money, to
facilitate a favourable balance of payments in international trade, that will draw precious metal (money) into, and
thereby stimulate, the domestic economy. The increase in the quantity of money would increase purchasing
power and increasing the demand for labour and production. Since investment fluctuates with the interest rate, a
low interest rate will increase employment: and in lieu of an easy money policy for private investment, then
direct government deficit-spending would guarantee full employment (Hahn, 2007, pp. 107, 139, 166).
Keynesians would encourage monetary policy to alter interest rates in order to manage the quantity or supply of
money. However, Keynes’ admission that, “the primary effect of a change in the quantity of money on the
quantity of effective demand is through its influence on the rate of interest” (Keynes, 2007, p. 298), would imply
that, “by effective demand Keynes seems to mean little more than total monetary demand; therefore doubling the
quantity of money, say, directly doubles the effective demand because the two terms practically mean the same
thing” (Hazlitt, 2007, p. 299). In reality, we have a simple quantity theory buried within Keynesian economics.
Our empirical data shows that absent of backing by a precious metal that would anchor fiat money to intrinsic
value, the classical quantity theory and the Keynesian purchasing power theory, have both been responsible for
influencing monetary policy that has led to the same macroeconomic result under the fiat standard since 1971, as
observed in figures 3 and 4: an exponential decay in the value of money, due to excessive debt and money
creation in relation to demand, the effect of which has led to an exponential increase in nominal prices, and when
corrected for this loss of value, real prices are low and constant.
Our analysis equally applies to the more notable periods involving higher prices of precious metals and
commodities during the Great Debasement (1542-1551), and during the Bank Restriction Period (1797-1821),
when England went off the gold standard for 24 years, due to the Bank of England suspending the redemption of
bank notes for gold coins. During the Great Debasement, the PS increased by 64% from 1542-1551, and
immediately following 1551, the average CP for the five years from 1552-1556, also increased by 64%. With
84
differences in domestic and foreign bimetallic ratios, Gresham’s Law saw gold exported and silver imported for
the domestic production of debased silver coins (Gould, 1970, p. 30). Thus, we have a decrease in the value of
money (reflected in a higher PS), as a result of an excessive increase in the supply of debased silver coinage, in
relation to demand, the effect of which was an increase in prices.
During the Bank Restriction Period (1797-1821), the un-backed issuance of paper money and the subsequent
high price of gold, was condemned by the Bullion Committee Report in 1810, as cited by Cannan (1925).
There is at present an excess in the paper circulation of this country, of which the most unequivocal symptom is
the very high price of bullion… that this excess is to be ascribed to the want of a sufficient check and control in
the issues of paper from the Bank of England; and originally, to the suspension of cash payments, which
removed the natural and true control (p. 66).
Although Ricardo would never advise a government to restore a currency to par that had depreciated in value by
30% (Ricardo, 1951, pp. 71-74), the gold standard was established de jure in 1816 (whilst payments were still
suspended), and the old parity was chosen rather than a new parity reflecting a higher market price during the
suspension, such that the PG fell from £5.35 in 1815 to the original mint price of £3.89375 (£3 17s 10.5d) in
1821, a decline of 27% that increased the value of the pound and reduced prices. Here, we have an increase in
the value of money, as a result of a decline in the volume of paper money, in relation to demand, the effect of
which was a decline in prices.
Our findings have gone much further than merely presenting the correlation between the price of bullion and the
price of commodities (whether over the short or long term). Whilst the correlation between the PG and CP in
England between 1259-2009 is 0.93, this certainly indicates that a decrease in the value of money (reflected in a
higher price of gold) is strongly correlated with higher nominal prices over the long term, but it does not
however reveal a great deal regarding causality, or indeed, how to resolve the loss of value in the medium of
exchange and the impact upon prices. The findings have explained with empirical evidence and through statistics,
both the impact and adjustment needed with regard to the value of money and its purchasing power and the
associated effect upon the general price level. Having expressed prices in real terms in terms of gold and silver,
when multiplying real prices by the PPG or PPS we obtain unity: CPgc x PPG (or CPsc x PPS) = 1. This
confirms that price inflation or deflation, are purely a monetary phenomenon, and any diminution of value is
proportional to the increase in nominal prices. With a stable currency, real prices (CPgc, CPsc) and real money
(PPG, PPS) may vary in the short term, but revolve around a long-term secular trend that is constant. When we
apply the criteria of long-term price stability and short-term volatility (table 5), we notice a highly stable
international gold standard on one hand, and a largely unstable inter-war gold-exchange standard on the other.
Table 5. Wholesale price index, England: Long-term stability and short-term volatility
Long-Term Short-Term Stability
Period Monetary System Comments
Stability Volatility Rank
1260-1509 domestic silver commodity standard 0.5% 4.8% 3 silver commodity standard
(upto the reign of Henry VIII)
1509-1717 domestic silver commodity standard 1.3% 6.8% 4 silver commodity standard
(includes Great Debasement 1542-1551
& Great Recoinage 1696-1699)
1717-1821 domestic gold commodity standard (de 0.3% 4.9% 2 fractional reserve gold standard
facto from 1717; includes Bank
restriction period 1797-1821)
1821-1914 domestic & international gold standard -0.2% 4.1% 1 fractional reserve gold standard
(de jure from 1816)
1914-1944 inter-war gold standard (U.K. off gold 3.2% 10.6% 7 less than 25% on a gold standard: early
standard 1914-1925 and from 1931) 1920s inflation & 1930s deflation
1944-1971 Bretton Woods gold-dollar standard 4.8% 6.4% 5 fractional reserve gold exchange
standard
1971-2009 international fiat paper standard 6.1% 6.0% 6 fractional reserve debt as money
With regard to the international gold standard, there was a long term secular decline in prices following the
de-monetization of silver, and absent of debasement, the domestic silver commodity standards would have
85
performed better. With regard to the inter-war gold-exchange standard, England was only on the gold standard
for less than 25% of the time during this period. Volatility was more obvious during the breakdown of monetary
order during the 20th century and the worst performing standard in terms of long-term stability was the
exponential decay of the purchasing power of the modern fiat paper standard following an annual increase in
prices of about 6.1% from 1971, which may be contrasted with adjusting nominal prices for the value of money
in terms of gold (CPgc) throughout the entire period from 1259-2009, with an average annual change of only
0.4%, and thus proves to be highly stable over the long term.
5. Conclusion
By examining the value of a currency in terms of its’ rate of exchange with gold or silver, not only are we able to
detect the underlying effectiveness of its’ purchasing power, but we also obtain a very clear sense of its’ worth as
a measure and store of value. By carefully constructing indices for the prices of gold, silver and commodities in
England over the long term, our findings have demonstrated that the purchasing power of gold and silver are
stable over the long term. We also found that debasement and devaluation have been the primary cause in the
decline of the value and purchasing power of money (the pound), the effect of which was an increase in nominal
prices. Therefore, prices increase inversely with the bullion content of currency, or in proportion to the rate of
debasement or devaluation. By correcting the loss of the value of money, and expressing the prices in bullion at
the rate of intrinsic decay, we discover not only constant prices expressed in either gold or silver, but through its
reciprocal, we also discover a constant purchasing power of gold or silver.
As mentioned in the introduction, there is a difference between the value of money and what money can
purchase at any given point in time. Changes in the purchasing power of money can originate from the supply
and demand of money, or from the supply and demand of commodities. Inflation and deflation are purely
monetary phenomena, and price changes due to monetary distortions are distinct from non-monetary reasons.
Value in exchange is measured in terms of prices. The price of a commodity is affected by the underlying value
of money (VM). The price of a commodity may be expressed as the ratio of the supply and demand of that
commodity, and the demand and supply of money. The price of a commodity is thus the ratio of two values, each
determined by its’ own respective supply and demand. Debasement and devaluation involves a decline in the
VM, caused by an excessive supply of money in relation to demand, the effect of which is an increase in the
price level.
Demand for a commodity

Supply of a commodity
Price of a commodity = (6)
Demand for money
Supply of money
When the VM reflected in the denominator in (6.0) is constant (VM = 1) then the value of the numerator would
be constant over the long term. Indeed, it is worth recalling for “a currency, to be perfect, [it] should be
absolutely invariable in value” (Ricardo, 1816, pp. 13-14). The management of the value and purchasing power
of money implies an ability to provide stable prices, and yet prices are supposed to change, in order to efficiently
transmit information and organize an economy. In terms of monetary policy, we should not target stable prices
(which have increased exponentially under the fiat standard), by devaluing money and its purchasing power, but
we should target a stable currency, absent of government intervention and manipulation, which in turn implies a
stable purchasing power and thus stable prices. Moreover, wealth is transferred to the issuer as a result of the
over-issuance of currency, since wealth is confiscated through inflation: “the decrease in purchasing power
incurred by holders of money due to inflation imparts gains to the issuers of money” (Reubling, 1975, p. 22).
Well before his appointment as Chairman of the Federal Reserve, as an economist, Alan Greenspan (1966)
mused over a gold standard, equating it with economic freedom, although his sentiments equally apply to a gold
or silver commodity standard, in the absence of which,
There is no way to protect savings from confiscation through inflation. There is no safe store of value…no way
for the owners of wealth to protect themselves…This is the shabby secret…Deficit spending is simply a scheme
for the confiscation of wealth. Gold [or silver] stands in the way of this insidious process. It stands as a protector
of property rights.
References
Abel, W. (2006). Agricultural Fluctuations in Europe from the Thirteenth to the Twentieth Centuries. (Olive
Ordish, Trans.). London: Routledge.
86
Allen, R. C. (2001). The Great Divergence in European Wages and Prices from the Middle Ages to the First
World War. Explorations in Economic History, 38(2), 411-447. http://dx.doi.org/10.1006/exeh.2001.0775
Beveridge, W. (1965). Prices and Wages in England from the Twelfth to the Nineteenth Century, 1. London:
Frank Cass & Co. Ltd.
Brown, H. P., & Hopkins, S. V. (1956, November). Seven Centuries of the Prices of Consumables, Compared
with Builders’ Wage-Rates, Economica, 23, 296-314, Brown-Hopkins Index reprinted in Carus-Wilson.
(1962), 2, 179-196, and with additional appendix in Brown, H.P., Hopkins, S.V. (1981). A Perspective of
Wages and Prices, 13-59. London: Methuen.
Cannan, E. (Ed.). (1925). The Paper Pound of 1791-1821, A Reprint of The Bullion Report (2nd ed.). London:
P.S. King & Son.
Central Statistics Office. see ONS.
Clark, G. (2004). The Price History of English Agriculture 1209-1914. In Field, A. J. (Ed.), Research in
Economic History (Volume 22, pp. 41-124). London: Elsevier.
Feavearyear, A. E. (1932). The Pound Sterling, A History of English Money. London: Oxford University Press.
Fenton, R. E. (2010). International Monetary Conference, held in compliance with the invitation extended to
certain governments of Europe by the government of the United States, in pursuance to the second section
of the Act of Congress of February 28 1878, in Paris in August 1878, under the auspices of the Ministry of
Foreign Affairs of the Republic of France, Senate Executive Document No.58, 45th Congress, 3rd Session,
Washington, DC: Government Printing Office, 1879, re-printed by Nabu Press.
Fisher, I. (1911). The Purchasing Power of Money. New York: Macmillan.
Fisher, I. (1922). The Making of Index Numbers: A Study Of Their Varieties, Tests, and Reliability. Boston:
Houghton Mifflin Company, reprinted (2008) Whitefish MT: Kessinger Publishing.
Francis, J. H. (1862). A History of the Bank of England, Its Times and Traditions from 1694 to 1844 (1st ed.).
Homans, I. S. (Ed.). New York: Banker’s Magazine.
Frankel, H. (1953). The Price of Gold and the Purchasing Power of the Pound Sterling. Journal of the Royal
Statistical Society, 116(1), 35-46.
Friedman, M. (1956). The Quantity Theory of Money: A Restatement, in (1969), The Optimum Quantity of
Money and Other Essays, 51-67. New Brunswick, NJ: Transaction Publishers.
Gayer, A. D., Rostow, W. W., Schwartz, A. J., & Frank, I. (1953). The Growth and Fluctuation of the British
Economy, 1, 468-70. Oxford: Clarendon Pres.
Gould, J. D. (1970). The Great Debasement, Currency and the Economy in Mid-Tudor England. Oxford: Oxford
University Press.
Greenspan, A. (1966). Gold and Economic Freedom. The Objectivist Newsletter.
Hahn, L. A. (2007). The Economics of Illusion, originally delivered as a lecture in Zurich along with an abridged
version published in German in September 1947, partly translated into English as The Easy Money Policy –
The End of an Illusion in December 1947, fully translated into English and published in the U.S. in 1949,
republished with introduction by Henry Hazlitt, Auburn: Ludwig von Mises Institute.
Hazlitt, H. (2007). The Failure of the New Economics. Auburn: Ludwig von Mises Institute.
Houghton, J., (A Collection for Improvement of Husbandry and Trade), Castaing, J., (The Course of the
Exchange and Other Things), Lloyds List, see Li (1963), Guildhall Library (U.K. Data Archive), Jastram
(1981, 2009).
ILO-PPI. (2004). Producer Price Index Manual: Theory and Practice. Geneva: International Labour Office
Jastram, R.W. (1977). The Golden Constant, The English and American Experience 1560-1976. originally
published New York: John Wiley & Sons, published by Leyland, J. (2009), with updated material
1560-2007. London: Edward Elgar Publishing.
Jastram, R. W. (1981). Silver, The Restless Metal. New York: John Wily & Sons.
Jevons, W. S. (1883). Money and the Mechanism of Exchange. originally published (1875) London: Henry S.
King & Co., published New York: D.Appleton and Company.
87
Keynes J. M. (2007). The General Theory of Employment, Interest and Money, originally published in 1936,
re-published for the Royal Economic Society, Basingstoke: Palgrave Macmillan.
London Bullion Metals Association. (LBMA). Retrieved from 19 January 2013,
http://www.lbma.org.uk/pages/index.cfm?page_id=6&title=statistics.
Lutyens (Course of the Exchange), see Jastram (1981, 2009).
MacLeod, H. D. (1883). The Theory and Practice of Banking (4th ed.). London: Longmans, Green & Co.
Mint Price. (1932). see Feavearyear.
Mitchell, B. R., & Deane, P. (1962). Abstract of British Historical Statistics. Cambridge: Cambridge University
Press, updated by Mitchell, B.R., and Jones, H.G., 1971, Second Abstract of British Historical Statistics.
Cambridge: Cambridge University Press.
Mitchell, W. C. (1965). The Making and Using of Index Numbers. New York: Augustus M. Kelley.
Mueller, J. D. (2010). Redeeming Economics, Rediscovering the Missing Element. Wilmington, Delaware: ISI
Books.
O’Donoghue, J., & Goulding, L. (2004, March). Consumer Price Inflation Since 1750. published a Composite
Price Index 1750-2003, in Economic Trends 604, 38-46. Office for National Statistics.
Office for National Statistics. [ONS]. retrieved from 19 January 2013 http://www.statistics.gov.uk/
Pamuk, S., & Ozmucur, S. (2002, June). Real Wages and Standards of Living in the Ottoman Empire, 1489-1914.
The Journal of Economic History, 62(2), 293-321. Cambridge University Press.
Pigou, A. C. (1917). The Value of Money. Quarterly Journal of Economics, 32(1), 38-65.
Ricardo, D. (1816). Proposals for an economical and secure currency: With observations on the profits of the
Bank of England, as they regard the public and the proprietors of bank stock (2nd ed.). London: John
Murray.
Ricardo, D. (1951). “Letter to Wheatley” dated 18 September 1821. In P. Sraffa (Ed.), Works and
Correspondence of David Ricardo, 9, 71-74. Cambridge: Cambridge University Press.
Ricardo, D. (2004). The Principles of Political Economy and Taxation. London: J. M. Dent & Sons.
Rogers, J. E. T. (1866-1892). A History of Agriculture and Prices in England, From the Year After the Oxford
Parliament (1259) to the Commencement of the Continental War (1793), Vols.1,2 (1866), Vols.3,4 (1882),
Vols.5,6 (1887), Vol.7 (parts I and II, 1892), originally published Oxford: Clarendon Press, re-printed
Boston: Adamant Media Corporation, 2005.
Rothbard, M. N. (1983). The Mystery of Banking. New York: Richardson & Snyder.
Ruebling, C. E. (1975, February). Financing Government Through Monetary Expansion and Inflation. Federal
Reserve Bank of St. Louis Review, 15-23.
Sauerbeck, A. (1886, September). Prices of Commodities and the Precious Metals. Journal of the Statistical
Society of London, 49(3), 581-648. published by Blackwell Publishing for the Royal Statistical Society, and
continued annually thereafter in the same source by Sauerbeck, A., and subsequently by the editor of The
Statist, reproduced in BHS 1962 p. 474, and by Jastram 1977 pp. 194-198.
Sharps, Pixley Ltd, London, see Shrigley. (1935). Jastram (1981, 2009).
Shrigley, I. M. (Ed.). (1935). The Price of Gold, documents illustrating the statutory control through the Bank of
England of the market price of gold, 1694-1931. London: P. S. King & Son.
Walsh, C. M. (1901). The Measurement of General Exchange-Value, London: Macmillan.
Warren, G. F., & Pearson, F. A. (1935). Gold and Prices. New York: John Wiley & Sons.
Notes
Note 1. William Beveridge (1879-1963) was an English economist, whom compiled a collection of prices series
for nearly 170 commodities in his Prices and Wages in England from the Twelfth to the Nineteenth Century, from
about 1550-1830 and was originally published in 1939.
Note 2. The currency was £1 = 20 shillings (s), and 1s = 12 pennies (d), hence £1 = 240d. The tower pound =
7,200 wheat grains, 1 tower lb = 12 tower ozs, 1 tower oz = 600 wheat grains, and 1 tower pennyweight (pwt or
dwt) = 30 wheat grains. Hence 1 tower lb = 240 dwt (7,200 / 30). The troy pound = 5,760 barley grains, 1 troy lb
88
= 12 troy ozs, 1 troy oz = 480 barley grains and 1 troy dwt = 24 barley grains. Hence 1 troy lb = 240 dwt (5,760 /
24). The tower lb was 15/16 lighter than the troy lb: thus, 1 tower lb of 7,200 tower grains = 5,400 troy grains,
and 1 tower oz = 450 troy grains and 1 tower dwt = 22.5 troy grains. Since, 1 barley grain = 0.06479891g, then 1
troy oz = 31.1034768 grams or 1g = 0.0321507 troy ozs, so that 1kg = 32.1507 troy ozs and I MT = 32,150.7
troy ozs.
Note 3. Thorold Rogers (1823-1890) was an English economist whom compiled prices of commodities from
1259-1793 in A History of Agriculture and Prices in England from 1259 to 1793, originally published 1866–
1892.
Note 4. Beveridge (1965, p. lx), cf. Jastram (1977, p. 62).
Note 5. The Jastram-Beveridge price index covered (for example in 1700) up to 16 categories and 72
commodities.
89
ISSN 1916-971X E-ISSN 1916-9728
Consequential Effects of Budget Deficit on Economic Growth:

Empirical Evidence from Ghana
Samuel Antwi1, 2, Xicang Zhao1 & Ebenezer Fiifi Emire Atta Mills3
1
School of Finance and Economics, Jiangsu University, Jiangsu, China
2
Koforidua Polytechnic, P.O.Box KF981, Koforidua, Ghana
3
School of Business Administration, Jiangsu University, Jiangsu, China
Correspondence: Samuel Antwi, School of Finance and Economics, Jiangsu University, 301 Xuefu Road,
Zhenjiang, Jiangsu, China. E-mail:[email protected]
Abstract
The study evaluates budget deficit sustainability of Ghana between 1960 and 2010 using the present value
budget constraint approach. By applying annual time series data, the ADF and PP tests for unit root rejected the
null hypothesis at 1 percent significance level after first difference. Hence, both government expenditure and
revenue of Ghana are stationary and integrated of order one. The Granger causality test supported a
bi-directional causation such that both expenditure and revenue of Ghana have temporal precedence over each
other. This means past and present values of government revenue provide important information to forecast
future values of expenditure. The test for cointegration favored the sustainability of budget deficit of Ghana at 10
percent significance level in the strong sense. In this case, government can continue to service its past
accumulated deficits without large future correction to the balance of income and expenditure. Again, the study
achieved the conventional negative sign of the speed of adjustment to long run equilibrium following shocks to
the system at 5 percent significance level.
Keywords: Ghana, budget deficit, government revenue, government expenditure, cointegration
1. Introduction
Budget deficit arises when the demand for government expenditure far exceeds government revenue that needs
to be financed by net lending. For the economy of Ghana, there has been persistent tendency towards budget
deficit since independence as a result of ever expanding government expenditure, inadequate revenue generation
capacity of government and increasing debt levels [Pomeyie, 2001: 162]. For instance, the deficit-GDP of Ghana
increased from 7.8 percent in 2005 to 8.1 percent in 2006 and 9.6 percent in 2007 and 14.5 percent in 2008 [IEA,
2008: 23-29]. As the economy of Ghana grows, policy makers have been concerned with the extent to which the
budget deficit is sustainable. For most years, government expenditure has exceeded government revenue in
Ghana leading to deficits on the budget. Expenditure has been rising steadily due to increase demand for
infrastructure and payment of interest on debt. For instance, total expenditure to GDP increased from 31.62
percent in 2005 to 33.71 percent in 2006 and 35.9 percent in 2007 [Bank of Ghana, 2007: 2]. Yet, it is important
that the government of Ghana run some fiscal deficits in order to stimulate economic growth by building up
enough capital stock. This would place the economy on its steady state growth path so that debt can be issued to
cover the deficits and repaid in the future [Xiomara and Greenidge, 2003: 2-3]. Contrary, the various sources of
Ghana’s revenue have become highly inadequate due to narrow tax base, high rate of tax evasion and corruption
in the revenue collection agencies. This has led to over spending by the government which tends to create deficit
on the budget. For instance, total government revenue to GDP increased marginally from 23.87 percent in 2005
to 24.1 percent 2006 but declined to 23.6 percent in 2007 [Bank of Ghana, 2007]. The persistent deficit in Ghana
means that the debt level and its servicing will continue to grow without limit unless constrained. This may lead
to explosion of the debt-GDP ratio of Ghana due to higher interest payment such that large and costly adjustment
must be made in future to correct the budget. In this regard, it is important to examine the problem of study.This
study aims at evaluating the sustainability of budget deficit in Ghana between 1960 and 2010.
90
2.1 Government Budget Deficit
A deficit policy plays a vital role in assisting countries achieve macroeconomic stability, poverty reduction,
income redistribution and sustainable growth. For this reason, most governments use the budget as effective tool
in achieving their economic objectives. This means that large and accumulating budget deficit may not
necessarily be a bad policy objective if such deficits are effectively utilized to enhance economic growth. It is in
line with this that an appropriate operational definition and measure of budget deficit must be clearly stated.
Otherwise, the occurrence of large nominal budget deficit may be misleading depending on the operational
measure adopted by a particular country.
2.2 Causes and Determinants of Budget Deficit Growth
In general, changes in budget deficit is attributed to changes in government spending or tax revenue or both.
Government receives revenue in its daily transactions and on capital items in the form of taxes and interests. On
the other hand, government pays for daily activities and capital items such as administrative expenses, loans and
grants. Thus, budget deficit increases when government spending persistently exceeds its revenue. If expenditure
continue to mount up throughout the years whereas revenues especially taxes are poorly collected, it widens the
budget deficit position of the country. In this case, the accumulated value of past deficit creates increase debts
which must be financed together with the accompanying interest payments.
2.3 Influence of Revenue-Expenditure Relationship on Budget Deficit
A method of determining sustainable budget deficit is to check whether government revenue and expenditure are
cointegrated. This implies that there may be significant long-term economic relationship between these two
variables. There are four hypotheses that examine the influence of revenue and expenditure on budget deficit.
The tax-spend hypothesis postulates that raising taxes in an attempt to reduce deficit also causes expenditure to
rise. It means that government raises tax revenue ahead of engaging in new expenditure. Contrary, the spend-tax
hypothesis predicts that government initially incurs expenditure and then increases tax revenue to finance the
deficit.
2.4 Sustainability of Budget Deficit
A sound fiscal policy is mandatory for macroeconomic stability and sustainable growth which is a major goal of
most emerging market countries such as Ghana. Yet, the size of budget deficit and ways of financing it
determine the fiscal constraint of the country in the long term. In this sense, sustainable budget deficit becomes
an important factor for which government authorities should pay particular attention [Kustepeli and Onel, 2004:
1-2]. The government‘s ability to borrow is constrained by the size of its permanent income just like an
individual, even if it remains in authority infinitely. This implies that whatever debt it accumulates has to be
repaid in the future.
2.5 Dimensions of Budget Deficit Sustainability
The controversies regarding conditions of intertemporal budget constraint and the shift to long-term horizon has
expanded the way governments and international organizations think about budget deficit sustainability.
Although it has retained its original meaning as a measure of the solvency of government, it has acquired several
dimensions in relation to governments that have no difficulty in meeting present obligations. Current
sustainability analysis focuses on fiscal conditions that may retard economic growth, increase tax burdens or
transfer significant costs to future taxpayers. These dimensions reflect concerns that governments accumulate
long-term liabilities that do not appear in current budgets but may disadvantage future generations when they are
due [Díaz, Izquierdo and Ugo, 2004: 6].
2.6 Determinants of Budget Deficit Growth
A model involving variation in inflation, government expenditure during wartime, cyclical fluctuation in output
during economic boom and recession in the postwar period was tested if it differs significantly from those during
the world wars in the Swiss federal state. The estimate showed some cyclical fluctuation in the world war
periods. This supports the assertion that significant determinant of budget deficit is increase in state expenditure
during wartime. In this case, civilian expenditure was reduced and/or taxes increased to finance military
expenditure during the war [Gebhard and Silika, 2006: 18-21]. In Ghana, changes in inflation, interest rate and
real GDP have reacted negatively to changes in budget deficit. For instance, high inflation in 1983 caused budget
deficit to increase by 35.8 percent due to decline in direct tax revenue. Also, changes in real interest rate
increased budget deficit by 11.3 percent of GDP in 1984. Again, high wage bill increased the deficit by 2.5
91
percent in 1985. Thus, changes in macroeconomic variables have had strong impact on the fiscal deficit in Ghana.
However, these effects have become less pronounced over the past years as the Ghanaian economy has grown
more stable [Wetzel and Roumeen, 1991: 48].
2.7 Effects of Budget Deficit Growth and Economic Sustainability
Financing of budget deficit in Ghana have had diverse macroeconomic burden on the economy. Central bank
financing for instance, have expanded the monetary base and money supply. According to Wetzel and Roumeen
[1991: 48], it distorted the distinction between monetary and fiscal policies whereas the sale of domestic bond
increased interest rate. This led to increase in the net domestic financing from 0.49 percent of GDP in 2004 to
4.15 percent in 2006. As a result, money supply (currency and deposits) increased from 6.84 percent in 2005 to
34.4 percent in 2006. However, domestic interest and bank rate reduced due to low demand for bonds [ISSER,
2007: 30, 60, 65]. Also, Ghana has accumulated large external debt and so borrow externally only on short term
bases at high interest rate. This is because foreign financing raises the cost of servicing external debt. For this
reason, Ghana’s access to external borrowing prior to 1984 had been limited, ranging between -0.74 and 1.62
percent of GDP. In recent times however, debt levels have been falling. External debt fell from 72.5 percent in
2004 to 26.9 percent in 2006 with debt service to GDP reducing from 6.8 percent in 2004 to 6.0 percent in 2006.
[Wetzel and Roumeen, 1991: 48; ISSER, 2007: 50, 101].
3. Methodology
3.1 Variables of the Study
The main variables employed in the study include government revenue, government expenditure and real gross
domestic product (GDP).
1) Government revenue includes all amounts of money or income received from sources outside the government
entity. This includes taxes, loan repayments, direct income, interests and grants obtained locally and externally.
2) Government expenditure involves spending by government authorities on goods and services. This includes
spending on road maintenance, health, administration and security. It also includes subsidies, grants and debt
servicing.
3) Real gross domestic product is the measure of changes in physical output in an economy between different
time periods by valuing all goods and services at the same constant price. Thus, it measures the value of final
goods and services produced in a given year when valued at constant prices [Parkin and Bade, 2003].
These variables are relevant to the study because changes in government spending or revenue create changes in
fiscal deficits. For instance, rapid increase in government expenditure coupled with shortfalls in tax revenue will
persistently create budget deficit.
3.2 Model Specification, Estimation and Tests
This study adopts the present value budget constraint model used by Kustepeli and Onel [2004: 7-8] in
specifying a model for budget deficit sustainability of Ghana. It is the most commonly used model in the study
of budget deficit sustainability base on intertemporal budget constraint (IBC) of government [Gebhard and Silika,
2006: 5]. The model begins with government budget constraint expressed as follows:
Gt  (1  rt )  t 1  Rt   t (1)
Where money finance ( M t  M t 1 ) is assumed to be negligible and t B is stock of government debt, Rt is

government revenue Gt is government expenditure excluding interest payments and rt is real interest rate. The
expression shows that the public sector finances its fiscal deficit through borrowing. Hence, it indicates the
sources and uses of public sector funds. It is therefore suitable for this study since Ghana finances its fiscal
deficits mainly through borrowing [ISSER, 1994: 30].
Equation 1 is then solved recursively by including subsequent and infinite periods. This generates the present
value budget constraint (PVBC) expressed as follows:
x 1 x 1

 1   1 
t 1     ( Rt  x  Et  x )  lim   t  x (2)
x  0  1  rt 
x  1  r
 t 
92
Where Et  Gt  (rt  r )  t 1 , and the interest rate rt is assumed to be stationary around a mean, r . Theexpression
means that the stock of government debt must be equal to the present value of primary budget surpluses with no
issue of new debt to finance the deficit.
Yet, intertemporal sustainability requires that the no Ponzi game (NPG) condition holds.
Therefore, the PVBC is reduced to NPG condition represented in equation 3;
x 1
 1  (3)
lim   t  x  0
x  1  r
 t 
This means that present discounted value of all future debt balances must sum up to zero because if lenders
behave optimally and rationally, government must pay off its debt at some point in time. Thus, government
cannot continue to roll over its debt perpetually into the future since lenders are rational.
In this study, we test the cointegration between government revenue ( Rt ) and government expenditure ( Gt ) in
the following testable regression function:
Rt     Gt  t (4)
Where α is a constant which shows the degree of drift in the parameters, t is the error term whilst β, shows
the extent to which changes in government expenditure affects the value of government revenue [Bajo-Rubio,
Díaz-Roldán and Esteve, 2005: 2-4]
3.3 Data Sources
The study employed annual time series data on government revenue-GDP and government expenditure-GDP of
Ghana for the period 1960 to 2010. The data was obtained from World Bank development indicators’ CD-ROM
and the state of the Ghanaian economy.
3.4 Data Analysis
In order to determine the causal link between government expenditure-GDP and revenue-GDP series of Ghana,
we apply the Granger causality test as stated in the following expression below:
n n
ln EXGt   0   a1i ln EXGt i   a2i ln REVGt i  1t (5)
i 1 i 1
m m
ln REVGt  0   1i ln REVGt k    2 k ln REVGt k   2t (6)
k 1 k 1
Where ln EXGt and ln REVGt are log of government expenditure-GDP and log of government revenue-GDP at
time t;  0 and  0 are intercepts, a1i and 1i are slope coefficients of own lagged values, a2i and  2k are slope
coefficients of lagged values of other variables in equation 5 and 6 respectively whilst ε is error term. In
estimating the causal link between government expenditure-GDP and revenue-GDP series of Ghana, the null
hypothesis of no Granger causality is tested against the alternative of Granger causality. In this case, the null
 
hypothesis is not rejected if a2i and  2k are statistically equal to zero simultaneously.
In applying the Engle-Granger test, the long run cointegration relation between government expenditure-GDP
and government revenue-GDP series of Ghana is estimated by using OLS method. This is expressed in the
following testable function:
ln EXGt     ln REVGt  t (7)
Where variables are as defined. In this case, the hypothesis to be tested in support of the objective of this study is
stated as follows:
H 0 : Government revenue and expenditure are not cointegrated.
H 1 : Government revenue and expenditure are cointegrated.
93
This result is possible if the series of residuals ( t ) is stationary and displays no unit root. Hence, we apply the
estimated cointegration relationship to generate residual errors. The estimated OLS residual errors are then tested
for unit root. In respect of this, the following hypothesis is tested to find out the stationarity of the residuals:
H0 :  t is not stationary (no cointegration)
H 1 :  t is stationary (cointegration)
For the budget deficit of Ghana, the following hypothesis is tested to establish the statistical significance of the
cointegrating vector:
H 0 :  =1
H1 :  1
Also, the short-run relationship between government expenditure and revenue of Ghana is estimated by using
error correction model (VECM). The model uses the first difference of the variables. For Ghana, the testable
function is specified as follow:
 ln EXGt      ln EXGt ( 1)   ln REVGt ( 1)  ECT ( 1)   t (8)
Where ∆ is the first-order time difference, ∆ECT is the first difference of the error correction term and α is the
intercept.  t is the error term with zero mean whilst δ is the coefficient of the period change in government
revenue which tends to capture the short-term effects. Also, φ is the coefficient of the error correction term
which incorporates feedback in the relationship between government revenue and expenditure
For the budget deficit of Ghana, the study tests whether there is heteroscedasticity based on the following
hypothesis:
H 0 : There is no heteroscedasticity.
H 1 : There is heteroscedasticity.
In the case of Ghana, the hypothesis test is conducted by employing the white heteroscedasticity test.
For efficiency of the estimates, the study also tests for autocorrelation since the classical linear regression model
assumes the existence of no autocorrelation in the disturbance term. This is important because, in the absence of
the assumption, the OLS estimator may no longer have minimum variance among all linear unbiased estimators.
In this case, the estimates may not be as efficient as other linear unbiased estimators and that the t, F and 2
tests may give misleading conclusions [Gujarati, 2004: 442].
In testing for autocorrelation, the following hypothesis is stated:
H 0 : There is no autocorrelation in the disturbance term.
H1 : There is autocorrelation in the disturbance term.
4. Empirical Results
4.1 Overview
The most common practice, among the set of methods to evaluate budget deficit sustainability, is to investigate
past fiscal data to establish if there is cointegration between government revenue and expenditure. For the
economy of Ghana, we perform the econometric analysis by using the OLS estimator.
4.2 Stationary and Unit Root Test
Table 1 shows the results of the test statistics for both ADF and PP of government revenue-GDP and government
expenditure-GDP series of Ghana in respect of the study period. The variables have been tested, with and
without trend, in their log levels as per Panel A and first differences as per Panel B. Table 1 panel A indicates
that the results of ADF and PP tests do not allow for rejection of the null hypothesis of non-stationarity for the
log levels of government expenditure-GDP series (ln EXG) and government revenue-GDP series (ln REVG).
This occurred both for the test with trend and the constant without trend.
In this case, the test show that both government expenditure-GDP series and revenue-GDP series of Ghana
supports the null hypothesis when ADF and PP tests are used. This means that government expenditure and
revenue of Ghana are non-stationary in their log levels. However, many time series data need to be appropriately
94
differenced in order to achieve stationarity [Dolado, Jesús and Marmol, 1999: 3]. As a result, we proceed to test
and confirm the stationarity of these variables after their first difference. The unit root test on the variables in
their first difference produced estimates that show a reverse situation compared to the estimates in the log levels.
Table 1 panel B shows the test for first differences of government expenditure-GDP series (∆lnEXG) and
government revenue-GDP series (∆lnREVG) of Ghana. From the estimation, both tests reject the null
hypothesis of non-stationarity of the variables after first differencing. This is possible even at 1 percent level of
significance for both ADF and PP tests with trend and without trend.This leaves no controversy in terms of both
ADF and PP tests in achieving stationarity after first difference of government expenditure-GDP and
government revenue-GDP series.
Table 1. Results of the unit root tests

Panel A:Log Levels
Variable ADF PP
Constant Constant Constant Constant
No Trend Trend No Trend Trend
DataPeriod: 1960-2010
Ln EXG -1.831012 -2.852240 -1.824912 -2.554048
Ln REVG -1.668219 -2.914847 -1.754793 -2.523154
Panel B: FirstDifferences
Variable ADF PP
 lnEXG -5.620420*** -5.552860*** -6.633597*** -6.515187***
 lnREVG -6.382306*** -6.318479*** -6.380414*** -6.308134***
Note: *** indicatesrejection ofthe null hypothesisof non-stationaryat 1percentsignificancelevel basedon theMacKinnon critical values.
In this case, the estimation show that government expenditure-GDP and revenue-GDP series favour the existence
of unit root in the log levels, but indicate stationarity after first differencing. Hence, we conclude that both
government expenditure and revenue of Ghana are stationary and integrated of order one. Yet, before estimating
the cointegration relation between the variables, we proceed to find the causal link between these variables.
4.3 Granger Causality Test
The result of the Granger causality test is represented in table 2.
Table 2. Granger causality test

Null Hypothesis: Obs F-Statistic Probability
ln EXG does not Granger Causeln REVG 50 0.45684 0.50264
ln REVG does not Granger Causeln EXG 50 2.20932 0.14431
The estimates from table 2 indicate that we reject the null hypothesis of no causality between government
expenditure-GDP and revenue-GDP of Ghana in both cases. This means that the coefficients  a2i and   2k
are statistically not equal to zero. In this case, there is bi-directional causality such that both government
expenditure and revenue of Ghana have temporal precedence over each other. Yet, the existence of Granger
causality does not imply that the occurrence of revenue is the result of expenditure. It means that changes in
revenue precede changes in expenditure.
4.4 Cointegration Tests
4.4.1 Long-Run Relationship between Government Expenditure and Revenue
The results from the estimation are presented in tables 3 and 4.
95
Table 3. Long run cointegration results using OLS method

Dependent Variable: LNEXG
Variable Coefficient Std. Error t-Statistic Prob.

LNREVG 0.991195 0.071018 13.95699*** 0.0000
C 0.271404 0.203966 1.330633 0.1899
R-squared 0.808968
Adjusted R-squared 0.804815
F-statistic 194.7976
Prob(F-statistic) 0.000000
Durbin-Watson stat 1.895125
Note: *** denotes the rejection of the null hypothesis at 1 percent significance level.
Table 4. Long run cointegration results using OLS method

Dependent Variable: LNREVG

LNEXG 0.816155 0.058476 13.95699*** 0.0000
C 0.322608 0.182514 1.767581* 0.0838
R-squared 0.808968
Prob(F-statistic) 0.000000
Note: * and *** denote the rejection of the null hypothesis at 10 percent and 1 percent significance level respectively
The estimation of the long-run cointegration relationship between government expenditure-GDP and
revenue-GDP series of Ghana shows that both variables are statistically significant at 1 percent level. In this case,
both variables depend on each other. We therefore proceed to test the statistical significance of the cointegration
coefficients.
In a second step to the Engle-Granger procedure, the estimated cointegration relationship between government
revenue-GDP and expenditure-GDP series is used to generate residual errors. The generated residuals are then
subject to the test for unit roots. Within this procedure, the error term represents the deviation of government
expenditure and revenue from their long-run equilibrium relationship. Rejecting the null hypothesis of the
presence of unit root in the residuals implies rejection of the null hypothesis of no cointegration. The test result is
represented in table 5:
Table 5. Test for unit root in the residuals

Panel A:Level
Variable ADF PP
Data Period: 1960-2010
RESID -3.866473*** -3.884572** -3.866473*** -3.884572**
Notes: ** and *** indicate rejection of the null hypothesis of non-stationary at 5 percent and 1 percent significance level respectively based
on the MacKinnon critical values.
The estimations in table 5 show that the test for unit root in the residuals of Ghana’s budget deficit fails to accept
the null hypothesis at 5 percent significance level. Both the ADF and PP tests reject the null hypothesis at 5
percent significance level for stationarity of the residuals. It thus, assumes the presence of cointegration in the
residuals of government revenue-GDP and expenditure-GDP series of Ghana when there is trend in the variables.
On the other hand, when no trend in the variables is considered, ADF and PP tests reject the null hypothesis of
no stationarity at 1 percent significance level. Therefore, it can Ghana’s budget deficit is rejected at 5 percent
96
significance level. This result is achieved after testing the residuals in their levels. With this firm result of
stationarity and cointegration of the variables at their levels, it does not therefore call for further tests of the
variables at their first difference. By this conclusion, it means that the budget deficit and hence government
revenue and expenditure of Ghana are cointegrated.
Table 6. Results of the wald restriction test

Equation Coefficient Null Hypothesis F-Statistics
ln EXG  f (ln REVG) 0.991195 H0 :   1 0.015373[0.9019]
ln REVG  f (ln EXG) 0.816155 H0 :   1 9.884251[0.0029]
According to the results obtained, we fail to reject the null hypothesis at 10 percent significance level where
government expenditure-GDP series depends on government revenue-GDP series [ln EXG  f (ln REVG )] . It is
therefore concluded that the coefficient of lnREVG (β =0.991195) is statistically not different from one at 10
percent level of significance. For this reason, Ghana’s budget deficit is said to be sustainable in the strong sense.
In this case, it means that the series of government expenditure and revenue move together in the long run and
that the coefficient of government revenue is statistically equal to or not different from one.
However, when government revenue-GDP series is dependent on expenditure-GDP series [ln EXG  f (ln EXG)], it
rejects the null hypothesis and concludes that β is statistically different from one. This indicates that the budget
deficit of Ghana will explode over the long run. Hence, it is not possible for government to continue
experiencing stable debt-to-GDP ratio indefinitely in the long run [Xiomara and Greenidge, 2003: 2-4]. For this
reason, government cannot continue to service its debt which accumulates from budget deficit without large
future correction to the balance of income and expenditure. Thus, the revenue capacity of government will not be
able to support government expenditure in the long run. This may call for large fiscal adjustments.
From the preceding analysis, it shows that the study have satisfied both sufficient and necessary conditions and,
therefore concludes that government expenditure-GDP series of Ghana depends on government revenue-GDP
series for the period 1960 to 2010. It is therefore important to find out the extent of adjustment of disequilibrium
between government expenditure-GDP series and government revenue-GDP series of Ghana within the study
period.
4.4.2 Estimation of the Short-Run Dynamic Model
The vector error correction model (VECM) of Engle-Granger is then used to estimate the short-run dynamic
relationship between government expenditure and revenue of Ghana.
The results of the short-run model of the relationship between government revenue-GDP and government
expenditure-GDP is presented in Table 7.
Table 7. Estimation of short run dynamic model

Dependent Variable: ∆LNEXG

C 0.018320 0.031116 0.588761 0.5592
∆LNEXG(-1) 0.327530 0.211667 1.547384 0.0129
∆LNREVG(-1) -0.274931 0.237208 -1.159029 0.0253
ECT(-1) -0.526017 0.199678 -2.634327** 0.0118
R-squared 0.747201
Prob (F-statistic) 0.007976
Note: ** denotes rejection of the null hypothesis at 5 percent significance level.
In the case of Ghana, the coefficient of the error correction term achieved the conventional negative sign. It is
also statistically significant at 5 percent, which further confirms the long run cointegration relationship between
97
the variables. This means that there exists a long-run cointegration relationship between government expenditure
and revenue of Ghana. The implication is that in the long run when budget deficit, at any given time t, exceeds
that of the previous period ( t  1 ) for which EXGt 1 is greater than REVGt 1 because  is less than zero, the
error correction term works to push government expenditure back towards equilibrium. This long term stability
however, does not rule out the possibility of Ghana running short-term budget deficits as argued by Carneiro et
al [2005: 113-114].
Also, with an estimated coefficient of -0.53, it indicates that approximately 53 percent of disequilibrium
generated is restored every year following shocks to the system. In the case of Ghana therefore, the speed of
adjustment parameter is relatively large with the right negative sign, indicating that there is a greater rate of
convergence toward equilibrium. By this finding, it is concluded that any disequilibrium within the budget
deficit of Ghana in the short run is quickly adjusted and converged back to equilibrium in the long run. Hence,
the model combines flexibility in dynamic specification with desirable long-run properties.
5. Discussion
From the analysis, it shows that the OLS estimate is linear, unbiased and normally distributed. However, it is
important to test for heteroscedasticity and autocorrelation since the estimate may no longer be efficient. In
testing for heteroscedasticity, the estimate indicates that the number of observations multiplied by R-square (n.R²)
gives a value of 1.539232 at a probability of 0.463191.
For all practical purposes, the study does not reject the null hypothesis. It is therefore, concluded that there is no
heteroscedasticity in government expenditure-GDP and revenue-GDP series of Ghana over the study period.
This means that the assumption of equal variance of the disturbance term hold for government expenditure and
revenue of Ghana between 1960 and 2010.
The test for autocorrelation, in respect of Ghana, produces a Durbin-Watson statistic of 1.895125. This means
that the test for autocorrelation of Ghana’s budget deficit fails to reject the null hypothesis of no autocorrelation.
Thus, government expenditure and revenue exhibit no autocorrelation and hence the budget deficit of Ghana
exhibits minimum variance in the disturbance term. This makes it efficient in arriving at better conclusions.
Based on the estimation and analysis, the long-run cointegration relationship of Ghana shows that government
expenditure-GDP series is dependent on government revenue-GDP series for the period between 1960 and 2010.
The estimate in this case produces a long-run cointegration coefficient for government revenue-GDP series (β) of
0.991195. This indicates that the long-run coefficient of Ghana’s budget is statistically significant at 1 percent
level. The estimated value of β being less than one means that anytime government revenue changes by 1 unit
on the average, it leads to a corresponding change of 0.99 units in government expenditure of Ghana. By this
finding, it indicates that the relationship is characterized by persistently higher government spending relative to
revenues but spending and revenue share a long-run equilibrium relationship.
In addition, the estimate of the long-run cointegration relationship of Ghana’s budget deficit indicates an
R-squared value of 0.808968 and adjusted R-squared of 0.804815. This means that approximately 80 percent of
the variations in government expenditure of Ghana are explained by variations in government revenue. The
estimates have been obtained with F-statistic of 194.7976 at a probability of 0.0000. This rejects the null
hypothesis and confirms the long run cointegration relationship between government expenditure and revenue of
Ghana between 1960 and 2010. This fulfills the intertemporal budget constraint of government which imposes
restriction on the long run relationship between expenditure and revenue such that government runs some
surpluses in the future to offset prevailing deficits. The restriction allows government to rule out Ponzi scheme
where new issues of debt must necessarily be made to finance deficits. This gives government the ability to raise
the necessary funds either by balancing its budget in present value terms or borrowing temporarily to be able to
service its debt [Kustepeli and Onel, 2004: 2-3].
6. Conclusion
The study sought to evaluate the sustainability of budget deficit in Ghana between 1960 and 2010. The
estimation uses operational budget deficit due to the inclusion of real interest payment in primary deficit. This is
a good choice since interest payment ultimately limits the deficit finance through growth of accumulated debt.
Sustainability analysis requires government to be able to service its debts without large future correction to the
budget. This would avoid rolling over initial debts with the interest forever. The unit root tests favoured the
stationarity of the variables at 1 percent significance level after first differencing. This means government
expenditure and revenue are integrated of order one process. Also, a Granger causality test supported the
existence of bi-directional causality between the variables. Hence, past and present values of government
98
revenue provide important information to forecast future values of expenditure. The Engle-Granger cointegration
test achieved a cointegration vector of 0.991195 at 1 percent level of significance indicating long run
cointegration relationship between government expenditure and revenue. It shows that approximately 80 percent
of variations in government expenditure of Ghana are explained by variations in government revenue. The linear
restriction test showed that is statistically not different from β one at 10 percent Significance level. This indicates
sustainability of budget deficit in the strong sense. Also, the error correction model achieved the conventional
negative sign at 5 percent significance level. This indicates that approximately 53 percent of disequilibrium is
restored every year following shock to the system. This is relatively large indicating greater rate of convergence
toward equilibrium. The diagnostic tests showed that government expenditure and revenue of Ghana exhibits no
heteroscedasticity and autocorrelation. Also, a normality test based on Jarque-Bera supported the normality
assumption. It is recommended that efforts should be made to consistently increase government revenue as
revenue and expenditure must be stationary and integrated of the same order. Since reduction in government
expenditure is not plausible, the tax net of Ghana should be expanded to capture all “taxable” individuals and
firms. This would ensure that expenditure do not move too far away from revenue. Any policy to increase
expenditure in Ghana should consider past and present values of government revenue. This is because
expenditure and revenue take temporal precedence over each other.
References
Afonso, A., & Rault, C. (2007). Should we Care for Structural Breaks When Assessing Fiscal Sustainability?
Lisbon, Technical University of Lisbon. Retrieved from http://pascal.iseg.utl.pt/~depeco/wp/wp012008.pdf
Amir, K. (2005). Sustainability of the Fiscal Process in Developing Countries-Egypt, Iran and Turkey: A
Multicointegration Approach. Ottawa, Carleton University. Retrieved from
http://www.erf.org.eg/CMS/uploads/pdf/1184499958_Amir_Kia.pdf
Aráoz, M., Cerro, M., Meloni, O., & Genta, S. (2006). Fiscal Sustainability and Crises - The Case of Argentina.
UNSTA, University of Tucuman. Retrieved from
http://www.aaep.org.ar/anales/works/works2006/Araoz_Cerro_Meloni_SoriaGenta.pdf
Arestis, P., & Sawyer, M. (2006). The Intertemporal Budget Constraint and the Sustainability of Budget Deficits.
Leeds, University of Leeds. Retrieved from http://www.palgrave.com/products/title.aspx?pid=288001
Aristovnik, A., & Bostjan, B. (2007). Fiscal Sustainability in Selected Transition Countries. MPRA Paper No.
122, Slovenia, University of Ljubljana. Retrieved from
http://mpra.ub.uni-muenchen.de/122/1/MPRA_paper_122.pdf
Auerbach, J. A., & Laurence, J. K. (1995). Macroeconomics: An Integrated Approach. Ohio, South-Western
College Publishing.
Bajo-Rubio, O., Díaz-Roldán, C., & Esteve, V. (2005). US Deficit Sustainability Revisited: A Multiple
Structural Change Approach. University of Manchester, Instituto de Estudios Fiscales. Retrieved from
http://www.ief.es/documentos/recursos/publicaciones/papeles_trabajo/2005_19.pdf
Bank of Ghana. (2005). The HIPC Initiative and External Debt: An Empirical Assessment and Policy Challenges.
Accra, Bank of Ghana Research Department.
Bank of Ghana. (2007). Bank of Ghana Monetary Policy Report. Vol. 1, No. 6/2007. Accra, Bank of Ghana
Research Department.
Barua, S. (2005). An Examination of Revenue and Expenditure Causality in Bangladesh: 1974-2004. World
Bank Working Paper Series WP 0605 Dhaka, Bangladesh Bank Policy Analysis Unit (PAU). Retrieved
from http://siteresources.worldbank.org/PSGLP/Resources/wp0605.pdf
Bebi, H. (2000). Sustainability of Budget Deficits–A Case Study of Namibia (1991-1999). Nepru Working Paper
No. 73, Windhoek, Namibian Economic Policy Research Unit. Retrieved from
http://openlibrary.org/books/OL3985876M/Sustainability_of_budget_deficits
Blanchard, O. J., & Fischer, S. (1989). Lectures on Macroeconomics. Cambridge Massachusetts Institute of
Technology, MIT Press.
Carneiro, F. G., Joao, F., & Barry, B. (2005). Government Revenues and Expenditures in Guinea-Bissau:
Causality and Cointegration. Journal of Economic Development, 30(1). Texas, World Bank. Retrieved from
http://www.jed.or.kr/full-text/30-1/08_J680.PDF
99
De Castro, F., & Pablo Hernandez, D. C. (2002). On the Sustainability of the Spanish Public Budget
Performance, Madrid-Spain, Instituto De Estudios Fiscales. Retrieved from
http://www.ief.es/documentos/recursos/publicaciones/revistas/hac_pub/160_Castro.pdf
Delong, B. J. (2002). Macroeconomics (Rev. ed.). New York: McGraw-Hill Companies Inc.
Dholakia, R. H., & Navendu, K. (2005). Consistent Measurement of Fiscal Deficit and Debt of States in India.
Economic and Political Weekly, 38(10). India. Retrieved from
http://www.iimahd.ernet.in/publications/data/2004-07-05rdholakia.pdf
Díaz, A. C., Alejandro, I., & Panizza, U. (2004). Fiscal Sustainability in Emerging Market Countries with an
Application to Ecuador, Washington, D.C., 20577, Inter-American Development Bank.
http://www.iadb.org/res/publications/pubfiles/pubWP-511.pdf
Dinh, H. T. (1999). Fiscal Solvency and Sustainability in Economic Management. Southern Africa Region, the
World Bank. Retrieved from
http://www-wds.worldbank.org/servlet/WDSContentServer/WDSP/IB/1999/11/19/000094946_9911050543
2935/Rendered/PDF/multi_page.pdf%20Download%20Restriction:%20no
Dolado, J. J., Gonzalo, J., & Francesc, M. (1999). Cointegration. Madrid, University of Carlos.
Easterly, W., & Klaus, S. H. (1993). Fiscal Deficits and Macroeconomic Performance in Developing Countries.
The World Bank Research Observer, Washington DC, World Bank. Retrieved from
http://www-wds.worldbank.org/external/default/WDSContentServer/WDSP/IB/1999/09/23/000178830_98
101911471889/Rendered/INDEX/multi_page.txt
Fischer, S., & William, E. (1990). The Economics of the Government Budget. The World Bank Research
Observer, 2(5), Oxford, Oxford University Press. Retrieved from:
http://www1.worldbank.org/publicsector/pe/PEAMMarch2005/easterly.pdf
Gebhard, K., & Silika, P. (2006). Sustainability of Swiss Fiscal Policy. CESIFO Working Paper 1689(1),
Switzerland, University of St. Gallen. Retrieved from
http://www.cesifo-group.de/portal/page/portal/DocBase_Content/WP/WP-CESifo_Working_Papers/wp-ces
ifo-2006/wp-cesifo-2006-03/cesifo1_wp1689.pdf
Government of Ghana. (2005). Ghana-Enhanced Structural Adjustment Facility. Economic and Financial Policy
Framework Paper, 1998–2000. Accra, Ministry of Finance.
Gujarati, D. N. (2004). Basic Econometrics (4th ed.). New York: McGraw-Hill Companies.
Holmes, J. M., Otero, J., & Theodore, P. (2007). Are EU Budget Deficits Sustainable? New Zealand, Waikato
University. Retrieved from http://www.rcfea.org/RePEc/pdf/wp17_09.pdf
Institute of Economic Affairs. (2008). Annual 2007 Economic Review and Outlook. Accra, Institute of
Economic Affairs.
Institute of Statistical, Social and Economic Research. (1994). The State of the Ghanaian Economy in 1993.
Accra-Legon, Wilco Publicity Services Ltd.
Institute of Statistical, Social and Economic Research. (2007). The State of the Ghanaian Economy in 2006.
Accra-Legon, Wilco Publicity Services Ltd.
International Monetary Fund. (2001). Ghana - Enhanced Heavily Indebted Poor Countries (HIPC) Initiative –
Preliminary Document, Washington DC, International Monetary Fund, International Development
Association
International Monetary Fund. (2002). Assessing Sustainability, Washington DC, International Monetary Fund,
International Development Association.
Issahaku, A. N. (1999). The Political Economy of Economic Reform in Ghana: Implications for Sustainable
Development. Washington, World Employment Programme Research. Retrieved from
http://www.jsd-africa.com/Jsda/Spring%202000/articles/Reform%20in%20Ghana.pdf
Jacobs, D., Nicholaas, S., & Jan, H. (2002). Alternative Definitions of the Budget Deficit and its Impact on the
Sustainability of Fiscal Policy in South Africa, Pretoria, University of Pretoria.
Kustepeli, Y., & Gulcan, O. (2004). Fiscal Deficit Sustainability with a Structural Break: An Application to
Turkey. İzmir-Turkey, Dokuz Eylül University. Retrieved from:
100
http://www.docstoc.com/docs/68100677/FISCAL-DEFICIT-SUSTAINABILITY-WITH-A-STRUCTURA
L-BREAK-AN
Luporini, V. (1999). Sustainability of the Brazilian Fiscal Policy and Central Bank Independence. Cedeplar Belo
Horizonte, the Federal University of Minas Gerais. Retrieved from
http://www.cedeplar.ufmg.br/pesquisas/td/TD%20125.pdf
Mahendra, R. (2006). Assessing Sustainability of Fiji’s Public Debt: A Cointegration Analysis Approach. Fiji,
Reserve Bank of Fiji. Retrieved from
http://www.reservebank.gov.fj/docs/Spch%20by%20Dr%20Mahendra%20Reddy%20at%20Eco%20Assoc
%20Forum%20RBF_01mar06.pdf
Mankiw, G. N. (2003). Macroeconomics (5th ed.). New York: Worth Publishers Inc.
Moalusi, D. K. (2004). Causal Link between Government Spending and Revenue: A Case Study of Botswana.
Botswana, Fordham University. Retrieved from
http://www.fordham.edu/images/Undergraduate/economics/DKMFORD1.pdf
Mokoena, T. M. (2005). Testing for fiscal sustainability in South Africa. Pretoria: University of Pretoria.
Mutasa, C. (2008). The Challenges of Debt Sustainability in Africa: The Case of Ghana, Accra, S.K.B Asante
International Consultancy Services. Retrieved from
http://www.afrodad.org/site/Publications/Debt%20Sustainability/dsa%20ghana%20final.pdf
Nyamongo, M., Moses, S., & Niek, S. (2007). Government Revenue and Expenditure Nexus in South Afric.
South African Journal of Economic and Management, 10(2). Retrieved from
http://repository.up.ac.za/bitstream/handle/2263/3282/Nyamongo_Government%282007%29.pdf?sequence
=1
Organisation for Economic Co-Operation and Development. (2005). Sustainable Budget Policy – Concepts and
Approaches, Thailand, Allen Schick Brookings Institute.
Oshikoya, T., & Tarawalie, A. B. (2008). Sustainability of Fiscal Policy: The West African Monetary Zone
(WAMZ) Experience. Journal of Monetary and Economic Integration, 9(2). Lagos, ECOWAS. Retrieved
from http://www.wami-imao.org/ecomac/english/newreports/v11_no1/pdfdocs/v9n2_unit1.pdf
Pomeyie, P. (2001). Macroeconomics: An Introductory Textbook. Accra, Wade Laurel co.
Quintos, C. E. (1995). Sustainability of the Deficit Process with Structural Shifts. Journal of Business and
Economic Statistics, 13(4). Washington, American Statistical Association. Retrieved from
http://wenku.baidu.com/view/610c4f07cc175527072208f0.html
Raghbendra, J. (2001). Macroeconomics of Fiscal Policy in Developing Countries. Australia, Australian
National University. Retrieved from http://www.wider.unu.edu/stc/repec/pdfs/dp2001/dp2001-71.pdf
Straub, R., & Ivan, T. (2007). Assessing the Impact of a Change in the Composition of Public Spending. IMF
Working Paper, Washington DC, International Monetary Fund. Retrieved from
http://www.ecb.int/pub/pdf/scpwps/ecbwp795.
Wetzel, D. L., & Roumeen, I. (1991). The Macroeconomics of Public Sector Deficits: The Case of Ghana.
Washington DC, World Bank. Retrieved from
http://www-wds.worldbank.org/servlet/WDSContentServer/WDSP/IB/1991/05/01/000009265_3961001075
213/Rendered/PDF/multi_page.pdf
Xiomara, A., & Greenidge, K. (2003). Debt and Fiscal Sustainability in Barbados. Bridgetown, Central Bank of
Barbados. Retrieved from
http://centralbank.org.bb/webcbb.nsf/WorkingPapers/BA33B2A964946094042578090048D705/$FILE/WP
2002pp11_30.pdf
101
ISSN 1916-971X E-ISSN 1916-9728
The Optimal Licensing Strategy of an Outside Patentee in

Vertically-Related Markets
Ming-Chung Chang1, Jin-Li Hu2 & Chin-Hung Lin2
1
Department of Banking and Finance, Kainan University, Taiwan
2
Institute of Business and Management, National Chiao Tung University, Taiwan
Correspondence: Ming-Chung Chang, Department of Banking and Finance, Kainan University. No. 1 Kainan
Road, Luchu, Taoyuan County 33857, Taiwan. Tel: 886-3-341-2500 ext. 6212. E-mail:
[email protected]
Abstract
We extend the model in Kamien and Tauman (1986) by considering vertically-related markets where the outside
innovator transfers new technology by means of either a royalty or a fixed fee. Our conclusion is different with
Kamien and Tauman (1986) and announces that the optimal licensing strategy for an outside innovator is a
royalty contract with a non-exclusion licensing case. When the innovation size is small, the outside innovator’s
licensing behavior causes low social welfare.
Keywords: royalty, fixed fee, price discrimination, cournot competition
1. Introduction
Getting new technology through technology licensing involves low risk, yet increases a corporation’s profit.
Although research and development (R&D) is also a powerful way to stimulate profit growth, it needs a lot of
money for investment and those involved must spend a lot of time. However, the firm needs to shoulder
uncertain risk of whether the R&D will be successful or not. Many firms do not have enough capital to engage in
R&D activity instead of searching for new technology and acquiring technology licensing.
Technology licensing has recently been a very popular strategy for firms in almost all industries, because the
licensee can acquire the external knowledge to improve its production technology and the licensor can earn rent
so as to increase profit. Hence, technology licensing has become a powerful access to increase a firm’s revenue.
Licensing revenue is estimated at more than US$100 billion annually in the U.S. (Kline, 2003).
There is a vast amount of literature focusing on the decision of the patentee’s optimal licensing strategy. Formal
analysis on the patentee’s profit through licensing an innovation that reduces production costs can be traced back
to Arrow (1962), who concludes that a perfectly competitive industry provides a higher innovation incentive than
a monopoly industry. Kamien and Schwartz (1982) extend Arrow’s analysis to license an oligopolistic industry
by means of both a fixed fee and a royalty. Kamien and Tauman (1986) examine the optimal licensing strategy
by comparing a fixed-fee licensing contract and a royalty licensing contract. Katz and Shapiro (1985, 1986)
study the auction licensing method.
Much of the literature analyzes the innovator’s optimal licensing strategy by various model settings, such as the
innovator is either a product’s manufacturer in the market, i.e., the inside innovator, or a non-product’s
manufacturer that is outside of market, i.e., the outside innovator, or such as firms including an innovator
engaged in either Cournot competition or Bertrand competition. Along this line, we arrange the licensing
literature into the four types as follows.
i) An outside innovator and licensees engage a Cournot competition. Under this situation, Kamien and Tauman
(1986) present that the outside innovator has higher profit by the fixed-fee licensing method than that by the
royalty licensing method.
ii) An outside innovator and licensees engage in Bertrand competition. Under this situation, Muto (1993)
considers the optimal licensing strategy in a horizontal differentiated duopoly model by comparing three kinds of
licensing methods among fixed fee, royalty, and auction. He concludes that a royalty is the optimal licensing
102
strategy under a non-drastic innovation. Poddar and Sinha (2004) introduce a spatial framework with the
Hotelling linear city model and show that a royalty is always better than an auction and fixed fee for the outside
innovator in both drastic and non-drastic innovations.
iii) An inside innovator and firms engage in Cournot competition. Wang (1998), under a homogeneous Cournot
model, finds that the licensor prefers the royalty licensing method rather than the fixed-fee licensing method for
non-drastic innovation. Kamien and Tauman (2002) extend Wang (1998) by increasing the number of firms in
the market from 2 to n and conclude that the optimal licensing strategy among a fixed fee, a royalty, or auction
depends on the number of firms in the industry. Wang (2002), under a horizontal heterogeneous Cournot model,
discusses the licensing strategy and finds that the licensor is possible to license a drastic innovation by means of
royalty when the products are imperfect substitutes. Sen (2005) analyzes the licensor’s optimal licensing strategy
given the number of licensees is an integer. This assumption induces the innovator’s profit to become a step
function under the fixed fee licensing contract and the auction licensing contract. This reason induces the royalty
licensing to become a better licensing method.
iv) An inside innovator and firms engage in Bertrand competition. Wang and Yang (1999) demonstrate that the
royalty licensing method is better than a fixed-fee licensing method, no matter if the innovation size is drastic or
non-drastic as long as the products’ differentiations are not too much. Poddar and Sinha (2004) conclude that the
licensor prefers a royalty licensing method for non-drastic innovation. The licensor does not offer a technology
licensing when the innovation is drastic. The papers we note above, however, neglect that the real world exists
with many intermediate goods firms.
Arya and Mittendorf (2006) is one of the few studies to simultaneously discuss licensing and outsourcing. Their
model includes an upstream intermediate goods firm and two downstream final goods firms in a homogeneous
Cournot model. The licensor is an inside innovator with a cost-reducing technology in a final goods market, and
the upstream firm can take discriminatory pricing against the downstream firms. They conclude that the licensor
does not prefer a fixed-fee licensing method, but a fixed-fee licensing method is valuable for the inside innovator
to get a double marginalization gain when a fixed fee is used in conjunction with royalties. Sandonis and
Fauli-Oller (2006) consider that an upstream outside innovator, such as a laboratory that faces a downstream
heterogeneous duopoly market, decides either licensing by two-part tariff or merging with one of the firms to
become a vertically-integrated firm. Under the assumptions of linear demand and Cournot competition, they
conclude that when the innovation size is small, a vertical merger is profitable to the innovator. Moreover, when
the innovation size is large enough, a vertical merger can increase social welfare.
The study scope of recent literature has combined the four lines as mentioned above. Rey and Salant (2012)
establish the vertically-related market structure in which one or more upstream patent owners licensed to a
downstream industry to study the impact of the licensing policies. They conclude that when only one patent
owner in upstream and the monopoly increases the number of licenses, it intensifies downstream competition,
and thus dissipates profits. When the multiple patent owners in upstream, the royalty licensing method not only
increases aggregate licensing fees but thus reduces the downstream prices for consumers. Similarly,
Layne-Farrar and Schmidt (2010) discuss different types of licensing contracts in the vertically-related market.
The licensing scenario includes the cross-licensing agreements and non-linear licensing fees. In vertically-related
markets in which the labor union is the upstream and the downstream firm is a monopolistic final goods producer,
Mukherjee et al. (2008) conclude that the downstream producer can make the profit by licensing the technology
when the input market is imperfectly competitive. When the labor union is a centralized style, the licensing by a
monopolist is profitable under both uniform and discriminatory wage settings. They also find that licensing by
the monopolist is profitable no matter under Cournot or Bertrand competition. The same conclusion holds even
with decentralized unions. Kishimoto and Muto (2012) employ Nash bargaining process to examine a Cournot
duopoly market in which the patent owner negotiates with its rival firm about payments for licensing a
cost-reducing innovation. They take into account two licensing methods - either a fixed fee or a royalty, showing
that the royalty licensing is better than fixed fee licensing for both firms if the innovation is not drastic. The
royalty licensing is always superior to fixed fee licensing from the viewpoint of social welfare, but there exists a
case in which consumers prefer the fixed-fee licensing.
The model set-up in this paper uses the model set-up in Kamien and Tauman (1986), but involves the concept of
Arya and Mittendorf (2006). We contribute a vertically-related market in the model of Kamien and Tauman
(1986), in which the licensor is an outside innovator and it transfers a cost-reducing technology to both or one
downstream firm by means of either a royalty or a fixed fee. Two downstream firms engage in homogeneous
Cournot competition.
103
This is a three-stage game. At stage 1, the outside innovator decides the optimal licensing strategy as being either
a fixed fee or a royalty. At stage 2, the upstream intermediate goods firm sets the intermediate goods price to
maximize its profit. At stage 3, the two downstream firms compete in a homogeneous Cournot competition. We
use backward introduction to get the sub-game perfect Nash equilibrium (SPNE) of this game.
The remainder of this paper is organized as follows. Section 2 presents some licensing models. Section 3 offers
the optimal licensing strategy analysis. Section 4 provides the implications of consumer surplus and social
welfare. Section 5 reconsiders the optimal licensing strategy after involving the concept of price discrimination
on the intermediate goods. Section 6 concludes in this paper.
2. The Licensing Model
A model includes two downstream firms that conduct Cournot competition and one upstream firm to provide the
intermediate good. Both downstream firms produce a homogeneous product and the inverse demand function
that they face is p = a  qi, where p is the price of the product and qi is the supply quantity of the ith
downstream firm, i = 1, 2.
In the production process, we assume that (i) one unit of the final product needs one unit of the intermediate
good as an input factor; (ii) the unit production cost of the intermediate good is zero; (iii) the unit price of the
intermediate good is t; (iv) the production cost of final goods is ci for the ith downstream firm. The SPNE for the
fixed-fee licensing model is:
2a  7c1  5c2 2a  5c1  7c2 * 2a  c1  c 2
q1* = , q2* = ,t = ,
12 12 4
i* = (qi*)2, and u* = (2a  c1  c2 )

2
(1)
24
where 0 < ci < a, and subscript u stands for the case of the upstream firm. An outside patentee does not make
anything in the product market, but can license a cost-reducing technology to downstream firms.
The first stage of the game is an outside patentee to decide the optimal licensing contract either as a fixed fee or
a royalty rate. The downstream firm decides either to accept the contract or to reject the offer. In the second stage,
the upstream firm decides the price of the intermediate good. In the third stage, two firms conduct Cournot
competition.
2.1 Pre-Licensing
Assume that the initial unit production costs of two downstream firms are c1 = c2 = c. Substitute c1 = c2 = c into
Equation (1), and the profits for two downstream firms, the upstream firm, and the price for intermediate goods
are:
iP = ( a  c ) 2 , uP = (a  c) and tP = a  c

2
(2)
6 6 2
Here, the superscript “P” stands for the pre-licensing case.
2.2 After Licensing
We now consider two kinds of licensing methods: either a fixed fee or a royalty rate. The cost-reducing
innovation held by the outsider patentee makes the licensee’s unit production cost fall from c to c-, where
parameter  is defined as the innovation size and   0. According to the definition of Wang (1998), if the
innovation size is large enough, then a firm without a new technology will drop out of the market, i.e., the drastic
innovation case. On the contrary, non-drastic innovation means that no firm drops out of market when the
outside patentee only licenses to some of the firms.
2.2.1 Fixed-Fee Licensing Method
The outside patentee decides on whom to license the innovation by a contract in which the outside patentee
offers innovation  against a fixed fee to the downstream firm(s) by mode n, given the intermediate goods price t,
where n = E stands for the exclusive licensing case or n = N stands for the non-exclusive licensing case.
Under the exclusive licensing case, we assume that the licensee is firm 1. Firm 1’s unit production cost is c-,
and firm 2’s unit production cost is c. Substitute c1 = c   and c2 = c into Equation (1) and we obtain:
104
q1FE = 2a  2c  7 , q2FE = 2a  2c  5 , tFE = 2a  2c   ,

12 12 4
iFE = (qiFE)2, and uFE = (2a  2c   )

2
(3)
24
where superscript “FE” stands for the fixed-fee contract under the exclusive licensing case. Under the
non-drastic innovation, i.e., 0    (2/5)(a  c), q2FE > 0, the market structure is duopoly. Under the drastic
innovation condition, i.e.,  > (2/5)(a  c), q2FE = 0, the market structure is monopoly. The equilibrium solutions
under the drastic innovation are:
qmFE = a  c   , tmFE = a  c   , mFE = (qmFE)2, and umFE = (a  c   )

2
(4)
4 2 8
Subscript “m” stands for the drastic innovation case.
Under the non-exclusive licensing case, the outside patentee licenses to both firm 1 and firm 2. Firm 1 and firm
2’s unit production costs are c  . Substitute c1 = c2 = c   into Equation (1) and we obtain:
qiFN = a  c   , tFN =
a  c   ,  FN = (q FN)2, and  FN = (a  c   ) 2 (5)
i i u
6 2 6
where superscript “FN” stands for the fixed-fee contract under the non-exclusive licensing case.
At stage 1 of the game, the outside patentee decides the fixed licensing fee by comparing the change between
pre-licensing profit and after-licensing profit. Define the symbol  as the licensing fee charged by the licensor,
and we have:
FE = 28 (a  c)  49 if 0    2( a  c) , and

2
(6a)
144 5
= 5(a  c)  18 (a  c)  9 if  >

2 2
2(a  c) . (Drastic innovation) (6b)
144 5
FN =
2 (a  c)   2 for any   0. (6c)
36
Under the fixed-fee licensing contract, we take into account whether or not the licensor should license to one
firm or two firms. By comparing Equation (6a) and Equation (6c), we find FE  FN = [20(a  c) + 452]/144
 0 when   (0, 2(a  c)/5). By comparing Equations (6b) and (6c), we obtain that FE  FN = 5(a  c +
)2/144  0 when   (2(a  c)/5, ).
Lemma 1 Under the fixed-fee licensing method, the outside patentee prefers to license to only one firm instead of
two firms.
We can take an extreme example to illustrate the economic intuition of Lemma 1. When the innovation size is
large enough and under the fixed-fee licensing method, the outside innovator prefers to only license to one firm
and makes the licensee become a monopolist in the market. The outside patentee can then extract the licensee’s
monopoly profit by means of a fixed fee. On the contrary, if the outside patentee licenses to two firms, then it
induces the market structure to become a duopoly. Traditional wisdom tells us that the duopoly profit is always
smaller than the monopoly profit.
2.2.2 Royalty Licensing Method
The outside patentee provides the royalty licensing contract to the target firm. The outside patentee charges a
royalty licensing fee for transferring a new innovation  by mode n, given the price of the intermediate good t.
Under the exclusive licensing case, firm 1’s (licensee’s) unit production cost is c   + r, and firm 2’s unit
production cost is c, where symbol r stands for a royalty rate. Substitute c1 = c   + r and c2 = c into Equation (1)
and solve the optimal royalty rate for the licensor as rRE = (a  c)/7 + /2, where superscript “RE” stands for the
royalty contract under the exclusive licensing case. Since c1 = c   + rRE < c must hold, the condition that the
licensor uses the optimal royalty rate rRE = (a  c)/7 + /2 is  > (2/7)(a  c). We have equilibrium solutions as
follows:
105
q1RE = 2( a  c )  7 , q2RE = 38( a  c )  35 , tRE =

26(a  c)  7 ,
24 168 56
(46a  46c  35 )(2a  2c  7 ) RE (58a  58c  7 )(38a  38c  35 )
1RE = , 2 = , and
2016 14112
(26a  26c  7 ) 2
uRE = (7)
4704
When   (0, (2/7)(a  c)), the optimal royalty rate is a corner solution, i.e., rRE = , which satisfies c1 = c   +
rRE  c. The equilibrium solutions are as follows:
(a  c) RE (a  c) (a  c) 2 (a  c) 2
qiRE = , t = , i =
RE
, and u =
RE
(8)
6 2 9 6
Let q2RE in Equation (7) be zero to find the non-drastic innovation condition as   (38/35)( a  c), which induces
a duopoly market structure. On the contrary, when  > (38/35)( a  c), the market structure is a monopoly. The
equilibrium solutions under the monopoly market structure are:
ac ac (a  c   )2
qmRE = , tmRE = , mRE = (qmRE)2, and umRE = (9)
8 4 32
Under the non-exclusive licensing case, the outside patentee licenses to both firm 1 and firm 2. Firm 1 and firm
2’s unit production costs are c   + r. Substitute c1 = c2 = c   + r into Equation (1) and we obtain:
a  c   RN ac (a  c   ) 2
qiRN = , t = , iRN = (qiRN)2, and uRN = (10)
12 4 24
where superscript “RN” stands for the royalty contract under the non-exclusive licensing case. The optimal
royalty rate rRN = (a  c + )/2 must satisfy ci = c   + rRN < c, and hence we obtain  > a  c. On the contrary,
when   (0, a  c), the optimal royalty rate rRN =  satisfies ci = c   + rRN  c. Substitute c1 = c2 = c into
Equation (1) and we obtain:
a  c RN ac (a  c) 2
qiRN = , t = , iRN = (qiRN)2, and uRN = (11)
6 2 6
At stage 1, the equilibrium licensing revenue for a licensor calculated by the formulations of rq1 under the
licensing mode E and r(q1 + q2) under the licensing mode N is represented as follows:
 (a  c) 2( a  c )
RE = if 0    , and rRE =  (12a)
3 7
(2a  2c  7 ) 2 2( a  c ) 38(a  c) (a  c) 
= if <   , and rRE =  (12b)
336 7 35 7 2
=
(a  c   ) 2 if  > 38( a  c) , and rRE = ( a  c   ) . (Drastic innovation) (12c)
16 35 2
 (a  c)
RN = if 0    (a  c), and rRN =  (12d)
3
(a  c   ) 2 (a  c   )
= if   (a  c), and rRN = (12e)
12 2
Under the royalty licensing contract, we want to find the optimal licensing strategy. By comparing four intervals
of  in Equation (12), i.e., (0, (2/7)(a  c)), ((2/7)(a  c), (a  c)), ((a  c), (38/35)(a  c)), and ((38/35)(a  c), ),
we find the optimal licensing strategy for a licensor is to license to both firms. Hence, we have Lemma 2 as
follows.
Lemma 2 Under the royalty licensing method, the outside patentee prefers to license to both firms.
106
Since the royalty fee equals the multiplication of the royalty rate and the licensee’s output, we summarize two
main reasons to illustrate the economic intuition of Lemma 2 as follows: i) one licensee’s output is always less
than two licensees’ output, i.e., q1RE < q1RN + q2RN; ii) the licensor decides the royalty rate for maximizing the
licensing revenue at Stage 1, and hence both the optimal royalty rate in the excluding licensing case and the
optimal rate in the non-exclusion licensing case are appropriate, i.e., rRE  rRN. Based on the two reasons above,
we conclude that the result of Lemma 2 mainly depends on the final goods’ market output.
3. The Optimal Licensing Strategy for an Outside Patentee
3.1 The Performance of a Licensing Contract for an Outside Patentee
Based on the analysis in Section 2, we get an equilibrium result that the licensor would like to license to only one
firm by means of a fixed fee and license to both firms by means of a royalty. We next want to decide the optimal
licensing strategy for the licensor by comparing Equations (6a), (6b), (12d), and (12e). In three intervals, i.e., (0,
(2/5)(a  c)), ((2/5)(a  c), (a  c)), and ((a  c), ), we find the optimal licensing strategy for a licensor is to
license to both firms by the royalty licensing method. Equations (12d) and (12e) represent the optimal licensing
strategy for the licensor.
3.2 Decision-Making for a Licensee: Accept vs. Reject It
In this subsection we need to examine whether the licensee accepts or rejects the licensing contract provided by
the outside patentee. We assume that if the licensee does not become worse after accepting a licensing contract,
then it would like to accept the licensing contract.
According to Equations (12d) and (12e), when   (0, (a  c)) with rRN = , the licensee’s profit is iRN = (a 
c)2/36  iP = (a  c)2/36. The licensee accepts the outside patentee’s licensing offer. When   ((a  c), ) with
rRN = (a  c + )/2, the licensee’s profit is iRN = (a  c + )2/144  iP = (a  c)2/36. In this regime the licensee
accepts the outside patentee’s licensing offer. Hence, we conclude the optimal licensing strategy for the outside
patentee as follows.
Proposition 1. The royalty contract with non-exclusive licensing is an equilibrium licensing strategy for an
outside patentee under vertically-related markets.
The outside patentee’s best licensing strategy is to license to both firms by royalties, because it benefits the
licensor’s licensing revenue by increasing the market’s competitive degree. This finding is totally different from
that of Kamien and Tauman (1986) in which the best licensing contract is to license to only one firm by a fixed
fee when the licensor is an outside patentee. However, it is noteworthy that after we consider the intermediate
goods firm in the model, the result of Kamien and Tauman (1986) reverses. The main reason is that the
intermediate goods firm and the outside patentee commonly share the downstream firms’ profits, and hence the
outside patentee can induce the upstream firm to decrease the intermediate goods price by a royalty licensing
contract in order to extract the intermediate goods firm’s profit. We use a mathematical result to support our
above explanation such as that tRN < tFE with rRN =  for   (0, (2/5)(a  c)), tRN < tmFE with rRN =  for  
((2/5)(a  c), (a  c)) and tRN < tmFE with rRN = (a  c + )/2 for   ((a  c), ).
Our conclusion is the same with Arya and Mittendorf (2006) in which an innovator in vertically-related markets
does not prefer the fixed-fee licensing method, but the innovator in their model is an insider which is different
from ours. The inside innovator in the model set-up of Arya and Mittendorf (2006) can obtain double
marginalization gains when a fixed fee is adopted in conjunction with royalties, but the outside innovator in our
model set-up can decrease the intermediate goods pricing for extracting the profit from the intermediate goods’
supplier by a royalty licensing method. Our conclusion is also the same with Sandonis and Fauli-Oller (2006) in
which an outside innovator does not prefer the fixed-fee licensing method, but their model does not include the
intermediate goods’ firm and the product market structure is a heterogeneous duopoly market.
4. Intermediate Goods Price, Consumer Surplus, and Social Welfare
We examine the intermediate goods price, consumer surplus, and social welfare under the optimal licensing
strategy.
4.1 Intermediate Goods Price
Table 1 arranges the intermediate goods price and the royalty rate in equilibrium. Since the profits of the
intermediate goods firm and the outside licensor are respectively calculated by tRN(qiRN) and rRN(qiRN), both of
their profit sizes depend on the sizes of the intermediate goods price (tRN) decided in stage 2 and the royalty rate
(rRN) decided in stage 1.
107
Table 1. Intermediate goods price and royalty rate in equilibrium

Interval / Item Intermediate Goods Price (tRN) Royalty Rate (rRN)
  (0, (a  c)) (a  c)/2 
  ((a  c), ) (a  c + )/4 (a  c + )/2
When the innovation size is large enough, i.e.,   ((a  c), ), the outside licensor has a first-mover advantage
since rRN > tRN. However, when the innovation size is small, i.e.,   (0, (a  c)), the outside licensor may not
have a first-mover advantage since there is an uncertain relationship between rRN and tRN. We take an extreme
example to illustrate the economic intuition of the result. When the innovation size approaches to zero, the
licensing revenue of the licensor also approaches to zero. However, the producer still needs to purchase the
intermediate goods as an input factor. The price of the intermediate goods is charged by the upstream firm for
maximizing its profit. Hence, we have proposition 2 as follows.
Proposition 2. The licensor may not have a first-mover advantage when the innovation size is small.
4.2 Consumer Surplus Analysis
Since the product is homogeneous in our model, consumer surplus can be calculated by CSj = (q1j + q2j)2/2,
where superscript j represents the licensor’s licensing strategy. The optimal licensing strategy in our model is the
royalty licensing contract under the non-exclusive case, and many studies conclude that consumers like the
fixed-fee licensing method under the non-exclusive case (Wang, 2002). Hence, we here compare the consumer
surplus size between CSRN and CSFN as follows:
CSRN =
(a  c) 2 , where 0 <  < (a  c) (13a)
18
(a  c   ) 2
= , where (a  c) <  <  (13b)
72
(a  c   ) 2
CSFN = , where 0 <  <  (13c)
18
It is easy to calculate and obtain the result that CSRN < CSFN for all  > 0. Hence, we conclude that the fixed-fee
licensing contract induces a higher consumer surplus than the royalty licensing contract in the non-exclusive case.
The reason is that the fixed-fee contract does not affect the licensee’s marginal production cost. However, the
royalty contract can increase the licensee’s marginal production cost and then decrease the firm’s output. Hence,
the total quantity is greater under a fixed-fee contract with non-exclusive licensing than that under a royalty
contract with non-exclusive licensing. This result is the same as that in Wang (2002).
4.3 Social Welfare Analysis
Since there is an inconsistent preference between the licensor that would like the royalty licensing contract with
a non-exclusive case and the licensee that prefers the fixed-fee licensing contract with a non-exclusive case, we
next need to examine the optimal social choice. The social welfare function in our model includes the consumer
surplus, and the profits of the downstream firms, the upstream firm, and the outside patentee. We formulate the
+ uj + j, however, the social welfare
2
social welfare function under the licensing case as Sj = CSj +

i 1
i
j
+ u . Hence, we have:
2

P P P
function under the pre-licensing case is S = CS + P
i
i 1
( a  c )(5a  5c  12 )
SRN = , where 0 <  < (a  c) (14a)
36
11(a  c   ) 2
= , where (a  c) <  <  (14b)
72
11 2
5 ( a  c ) 2  11( a  c )  
S FN
= 2 , where  > 0 (14c)
18
108
SP = 5(a  c)
2
(14d)
18
By comparing the royalty licensing contract and the fixed-fee licensing contract under the non-exclusive case, it
is easy to calculate and obtain the result that SRN < SFN for all  > 0. It shows that social welfare is higher under
the fixed-fee licensing contract with the non-exclusive case than that under the royalty licensing contract with
the non-exclusive case. However, Proposition 1 concludes that the optimal licensing method for the outside
patentee is the royalty licensing contract with the non-exclusive case. Hence, we have the proposition as follows.
Proposition 3. The licensor and the social planner have an inconsistent licensing preference. However, the social
planner’s and the consumer’s licensing preferences are consistent.
The reason for the inconsistent licensing preference between the licensor and the social planner is that the total
output in the royalty licensing contract is lower than that in the fixed-fee licensing contract. This finding is
supported by the evidence that q1RN + q2RN = (a  c + )/6 < q1FN + q2FN = (a  c + )/3 for rRN = (a  c + )/2 and
q1RN + q2RN = (a  c)/3 < q1FN + q2FN = (a  c + )/3 for rRN = .
We next consider whether the technology licensing can increase social welfare by comparing SRN and SP, and we
obtain SRN < SP with   (0, (5/12)(a  c)), and SRN > SP with   ((5/12)(a  c), ). It shows that social welfare
under the no licensing case may be high when the innovation size is small. Hence, we have a proposition as
follows.
Proposition 4. Technology licensing induces low social welfare when the innovation size is small.
5. Price Discrimination on Intermediate Goods Pricing
The intermediate goods firm can adopt the price discrimination strategy on intermediate goods pricing when two
downstream firms have a cost differentiation caused by an excluded licensing case. Hence, we reconsider here
the licensor’s optimal licensing strategy when the intermediate goods firm has an ability to take a discriminatory
price.
5.1 Price Discrimination under a Fixed-Fee Contract with an Excluded Licensing Case
Because of an excluded licensing, the licensee’s, i.e., firm 1, per unit production cost is c1 = c   + t1, and the
non-licensee’s, i.e., firm 2, per unit production cost is c2 = c + t2, where t1(t2) is a discriminatory intermediate
goods price charged on firm 1(2). By backward induction, we have the SPNE as follows:
ac a  c   FE’ a  c
q1FE’ = , q2FE’ = a  c   , t1FE’ = , t2 = ,
6 6 2 2
iFE’ = (qiFE’)2, and uFE’ = (a  c )  (a  c   )

2
(15)
6
where   (0, (a  c)) and superscript “FE’” stands for the fixed-fee contract with the exclusive licensing case
under price discrimination. When  > (a  c), i.e., a drastic innovation, the SPNE is the same as that in Equation
(4).
The equilibrium licensing revenue for the licensor calculated by the formulation of 1FE’  1P is represented as
follows:
FE’ =  ( a  c   ) if 0    (a  c), and (16a)

9
=
5( a  c ) 2  18 ( a  c )  9 2 if  > (a  c). (Drastic innovation) (16b)
144
5.2 Price Discrimination under a Royalty Contract with an Excluded Licensing Case
Under the royalty licensing method, firm 1’s per unit production cost is c1 = c   + t1 + r, and firm 2’s per unit
production cost is c2 = c + t2. According to the calculated result, the optimal royalty rate for the licensor is rRE’ =
(a  c + 2)/4, and the SPNE for firm 2’s output in this regime is q2RE’ = (5a  5c  2)/24, where superscript
“RE’” stands for the royalty contract with the exclusive licensing case under price discrimination.
Let q2RE’  0, and the non-drastic innovation condition is   (5/2)(a  c). Here, rRE’ must satisfy the condition for
c1 = c   + t1 + r  c, and hence we have   (5/2)(a  c) which violates the non-drastic innovation condition. It
109
implies that under the duopoly market structure, the optimal royalty rate is not rRE’ = (a  c + 2)/4 instead of a
corner solution, i.e., rRE’ = . Hence, we have the SPNE under the duopoly market structure as follows:
qiRE’ = a  c , tiRE’ = a  c , 1RE’ = (a  c) , 2RE’ = 7(a  c) and uRE’ = (a  c )

2 2 2
(17)
6 2 36 36 6
Traditional wisdom tells us the intermediate goods firm will charge a high (low) price to a low (high) cost firm.
This conclusion is also confirmed in Equation (15) in our paper, i.e., t1FE’ > t2FE’.
If  > (5/2)(a  c), and the licensor only licenses to one firm, then the market structure will become a monopoly.
The SPNE for the drastic innovation in this regime is the same as that in Equation (9).
The equilibrium licensing revenue for the licensor is calculated by the formulation of rRE’q1RE’ and is represented
as follows:
RE’ =  ( a  c ) for   (0, (5/2)(a  c)), and rRE’ =  (18a)

6
(a  c   )2
= for   ((5/2)(a  c), ), and rRE’ = ( a  c   ) . (Drastic innovation) (18b)
16 2
5.3 The Optimal Licensing Strategy after Considering a Price Discrimination Regime
By comparing Equations (16) and (18), we obtain the optimal licensing strategy in a price discrimination regime
as follows:
RE’ =  ( a  c ) if 0    (a  c)/2, and rRE’ =  (19a)

6
FE’ =  ( a  c   ) if (a  c)/2 <   (a  c) (19b)

9
5(a  c) 2  18 (a  c)  9 2
= if (a  c) <   (5/2)(a  c), (Drastic innovation) (19c)
144
RE’ = (a  c   ) if  > (5/2)(a  c), and rRE’ =

2
( a  c   ) . (Drastic innovation) (19d)
16 2
We now consider the optimal licensing strategy after considering a price discrimination regime by comparing
Equation (19) and Equations (12d) and (12e). By a simple calculation, we find RN is always larger than RE’ for
every   (0, ). The result shows that the outside licensor prefers a royalty contract with a non-excluded
licensing case instead of an excluded licensing contract in which the intermediate goods firm has price
discrimination ability. The reason is that the licensor and the upstream firm commonly share the downstream
firm’s excess profit. If the upstream firm has price discrimination ability on the intermediate goods, then it will
crowd out the profit that the licensor can obtain from the licensee. Hence, we have a proposition as follows.
Proposition 5 In vertically-related markets, the outside licensor would like the intermediate goods firm to adopt
uniform pricing instead of a discriminatory price.
6. Concluding Remarks
Our model includes an upstream firm which provides the intermediate goods to two homogeneous downstream
firms, and an outside patentee with a new production technology that chooses the optimal licensing strategy by
means of either a royalty or a fixed fee for maximizing its licensing revenue. The paper most related to our study
is Kamien and Tauman (1986) in which their model does not consider the vertically-related market. The main
result in this paper is very different from Kamien and Tauman (1986) in which the outside innovator prefers a
fixed-fee licensing contract instead of a royalty licensing contract. After involving vertically-related markets in
the model of Kamien and Tauman (1986), we get an inverse result that the outside innovator’s optimal licensing
contract is by means of royalties instead of a fixed fee.
The implications of consumer surplus and social welfare are also important in the firm’s optimal licensing
behavior. We find there is an inconsistent licensing preference between the consumers and the licensor, and
between the social planner and the licensor. The consumers and the social planner prefer a fixed-fee licensing
110
contract, however, an outside patentee prefers a royalty licensing contract. Most importantly, given the firm’s
optimal licensing behavior, there will be low social welfare after licensing when the innovation size is small.
Under the optimal licensing strategy, the licensor first decides the optimal royalty rate, and then the upstream
firm chooses the optimal intermediate goods price. However, the optimal royalty rate is not necessarily higher
than the optimal intermediate goods price when the innovation size is small. Hence, we conclude that the
licensor may not have a first-mover advantage when the innovation size is small. Finally, we reconsider the
optimal licensing strategy after extending the assumption that the intermediate goods firm has price
discrimination ability. However, the equilibrium licensing strategy does not change. Since the licensor and the
intermediate goods firm commonly share the downstream firms’ excess profit, and the discriminatory price on
the intermediate goods can increase the upstream firm’s profit and then crowds out the licensor’s profit, the
licensor must not adopt a licensing strategy that lets the intermediate goods firm have an opportunity to realize
the discriminatory price.
Acknowledgements
The authors appreciate financial support from Taiwan’s National Science Council (NSC 101-2410-H-424 -012
and NSC99-2410-H-009-063).
References
Arrow, K. (1962). Economic welfare and the allocation of resources for inventions. In Neslon, R. (Ed.), The rate
and direction of inventive activity (pp. 609-625). Princeton: Princeton University Press.
Arya, A., & Mittendorf, B. (2006). Enhancing vertical efficiency through horizontal licensing. Journal of
Regulatory Economics, 29, 333-342. http://dx.doi.org/10.1007/s11149-006-7403-7
Kamien, M. I., & Schwartz, N. L. (1982). Market structure and innovation. Cambridge University Press.
Kamien, M. I., & Tauman, Y. (1986). Fee versus royalties and the private value off a patent. Quarterly Journal of
Economics, 101, 471-491. http://dx.doi.org/10.2307/1885693
Kamien, M., & Tauman, Y. (2002). Patent licensing: The inside story. Manchester School, 70, 7-15.
http://dx.doi.org/10.1111/1467-9957.00280
Katz, M., & Shapiro, C. (1985). On the licensing of innovations. RAND Journal of Economics, 16, 504-520.
Katz, M., & Shapiro, C. (1986). How to license intangible property. Quarterly Journal of Economics, 101,
Kishimoto, S., & Muto, S. (2012). Fee versus royalty policy in licensing through bargaining: An application of
the Nash bargaining solution. Bulletin of Economic Research, 64, 293-304.
Kline, D. (2003). Sharing the corporate crown jewels. MIT Sloan Management Review, 44, 89-93.
Layne-Farrar, A. S., & Schmidt, K. (2010). Licensing complementary patents: ‘Patent Trolls’, market structure,
and ‘Excessive’ royalties. Berkeley Technology Law Journal, 25, 1121-1143.
Mukherjee, A., Broll, U., & Mukherjee, S. (2008). Unionized labor market and licensing by a monopolist.
Journal of Economics, 93, 59-79. http://dx.doi.org/10.1007/s00712-007-0293-z
Muto, S. (1993). On licensing policies in Bertrand competition. Games and Economic Behavior, 5, 257-267.
http://dx.doi.org/10.1006/game.1993.1015
Poddar, S., & Sinha, U. B. (2004). On patent licensing and spatial competition. Economic Record, 80, 208-218.
Rey, P., & Salant, D. (2012). Abuse of dominance and licensing of intellectual property. International Journal of
Industrial Organization, 30, 518-527. http://dx.doi.org/10.1016/j.ijindorg.2012.05.003
Sandonis, J., & Fauli-Oller, R. (2006). On the competitive effects of vertical integration by a research laboratory.
International Journal of Industrial Organization, 24, 715-731.
http://dx.doi.org/10.1016/j.ijindorg.2005.08.014
Sen, D. (2005). Fee versus royalty reconsidered. Games and Economic Behavior, 53, 141-147.
http://dx.doi.org/10.1016/j.geb.2004.09.005
Wang, X. H. (1998). Fee versus royalty licensing in a Cournot duopoly model. Economics Letters, 60, 55-62.
111
Wang, X. H. (2002). Fee versus royalty licensing in a differentiated Cournot duopoly. Journal of Economics and
Business, 54, 253-266. http://dx.doi.org/10.1016/S0148-6195(01)00065-0
Wang, X., & Yang, B. (1999). On licensing under Bertrand competition. Australian Economic Papers, 38,
112
ISSN 1916-971X E-ISSN 1916-9728
Corporate Sustainability Reporting and Analysis of Sustainability

Reports in Turkey
Rabia Aktaş1, Koray Kayalidere1 & Mahmut Karğin1
1
Accounting and Finance Department, Celal Bayar University, Manisa, Turkey
Correspondence: Rabia Aktaş, Accounting and Finance Department, Celal Bayar University, Uncubozköy,
Manisa, Turkey. Tel: 90-236-233-0657. E-mail: [email protected]
Early version of this paper was presented at “First International Conference on Sustainable Business and
Transitions for Sustainable Development”, 11-13 October 2012, Konya-Turkey.
Abstract
Financial reports are the most important sources of information for financial information users. Investors, lenders,
and other creditors use financial information in their decision making process. Therefore, governmental bodies
of financial reporting have mandated disclosing of some financial reports (statement of financial position,
statement of comprehensive income, and statement of changes in equity) for public interest. It has been stated
that besides financial information, non-financial information is also important for the users. It is clear that
financial information contributes better decisions making when it is supported by non-financial information.
Non-financial reports inform stakeholders (e.g., investors, employees, customers, and non-governmental
organizations) and the general public about the firm’s activities involving environmental, social, and governance
(ESG) issues. In this study we have discussed elements of corporate sustainability reporting and detected
sustainability reports of selected public firms in Turkey. Sustainability reports of nine firms have been analyzed
based on the Global Reporting Initiative (GRI) indicators. The results show that selected firms’ sustainability
reports fulfill requirements related to Part I and Part II, “Profile Disclosures” and “Disclosures on Management
Approach”, which is mostly consistent with their reports’ application level. However, it is hard to say the same
thing for Part III. In fact, firms’ sustainability reports fail to consistently disclose “Performance Indicators”.
Keywords: corporate sustainability reporting, sustainability reports, GRI
1. Introduction
From the sustainability perspective corporations are being forced to redesign almost every part of their
operations; from their strategies to objectives, to technologies, to product design, to production process, to
business models etc. In other words, to maximize firm’s value and create sustainable value in long term
companies must achieve three elements of sustainable development, namely economical, social, and
environmental sustainability.
Sustainability and reporting sustainability have been considered and discussed by governments, organizations,
and academicians. The idea of sustainability is changing not only businesses’ cultures, but it is also changing
people’s habits, life style, and plans as well. Sustainability is being defined as follows: “Sustainable development
is development that meets the needs of current generations without compromising the ability of future
generations to meet their own needs”(Sustainability Framework, 2011, p.19). “The global challenge is to ensure
that organizations develop sustainably to reverse the previous erosion of natural resources, and to improve their
environmental, social, and financial performance. This requires radical changes in the way they do business and
the way we live our lives” (Sustainability Framework, 2011, p.5). Moradzadehfard and Moshashaei (2011, p.397)
state the World Business Council of the Sustainable Development’s (WBCSD) sustainability definition: “The
sustainable development is a simultaneous activity for the economic prosperity, environmental quality and the
social justice”.
Sustainability reporting has become an active area of research and been dominated by large multinational
corporations recently. The scope of reports, potential target audiences, and integration with financial reports have
113
increased as well. However it has been stated that there is relatively little information available on the process of
developing reports and how they are used. To be more reliable most of corporations have followed guidelines for
corporate sustainability reporting. The well known set of voluntary guideline is the Global Reporting Initiative
(GRI, 2006). The guideline basically focuses on context of corporate sustainability reports, corporations’
sustainability vision, their performances, and objectives towards sustainability (Roca & Searcy, 2012,
p.103-105).
Sustainability accounting and related terms are discussed by academics in conferences and practices frequently.
The relationship between accounting and sustainability is another topic that is also discussed (Schaltegger &
Burritt, 2010). Accountants affect sustainability and the quality of reporting sustainability. In fact, the
International Federation of Accountants (IFAC) believes that accountants can play key roles on sustainability.
IFAC published “Sustainability Framework 2.0 Professional Accountants as Integrators” and outlines statements
that accountants can support their organizations to achieve sustainable development by; “defining and clarifying
the terminology that an organization has decided to use (sustainability, corporate responsibility, or corporate
social responsibility) and what it means in relation to the organization; establishing leadership, vision, values,
and behaviors; ensuring appropriate governance structures are in place to strengthen implementation, monitoring,
and accountability; effective stakeholder engagement; setting goals and targets; establishing the business case;
integrating risk management and assessment; and engaging suppliers”.
Many Turkish firms have been paying attention to sustainability reporting lately. Most of these firms disclose
sustainability reports in their operations reports. “As a result of increasing importance of sustainable
development and the need for the companies to address and manage environmental, social and governance issues
in a competitive global market, the Istanbul Stock Exchange (ISE) and the Turkish Business Council for
Sustainable Development (TBCSD) have planned to work on a project titled “ISE Sustainability Index” (SKD,
2012). Some firms are the members of Sustainability Disclosure Database and disclose their repots in the
Database. Firms’ sustainability reports are prepared and disclosed in various levels as G3, G2, G1, and A, B, and
C level.
The GRI criteria or performance indicators have been used to evaluate firms’ sustainability reports by
researchers. In this study, selected nine Turkish firms’ sustainability reports are being analyzed based on the GRI
indicators.
2. Sustainability Reporting: Studies and Practices
Sustainability reporting has historically focused on private sector not on public sector (Williams et. al., 2011).
Farneti and Guthrie (2009) agreed this statement and suggested that public sector also should be analyzed in
perspective of sustainability and its reporting.
Sustainability reporting is mandatory in many countries. The laws and regulations outline how sustainability
reporting must be. However, Joseph (2012) argued that with the emphasis on voluntarism, sustainability
reporting is in a transition stage yet. In some countries sustainability reports reveal just some of its indicators
properly. Murguia and Böhling (2013) stated that in some Argentinean firms environmental and economic
indicators are the most contentious and least reported.
Number of firms who report sustainability has increased lately. “In the last decade, reporting of nonfinancial
information has become widespread. According to the Global Reporting Initiative (GRI), only 44 firms followed
GRI guidelines to report sustainability information in 2000 yet by 2010, the number of organizations releasing
sustainability reports, predominantly on a voluntary basis, grew to 1,973” (Ioannou & Serafeim, 2012, p.2).
Marimon et. al. (2012) indicate that even though many firms have participated in sustainability reporting, that is
not enough when it is compared to the numbers of total firms worldwide.
Ioannou and Serafeim (2012) have analyzed the consequences of mandatory corporate sustainability reporting by
using 58 countries’ data. They report that after the adoption of mandatory sustainability reporting laws and
regulations, the social responsibility of business leaders increases. They also conclude that mandatory corporate
sustainability improve and develop sustainability development, employee training, and corporate governance.
Corporate sustainability reporting applications are increasing in some countries. Gurvitsh and Sidorova (2012)
conducted a survey among 15 listed firms in Estonia and they concluded that corporate sustainability is not a
“stranger in the night” for those firms however it is a “friend, one would like to spend more time with”. Adams
and Frost (2008) found that companies are adapting environmental and social indicators of sustainability while
they take these indicators into account in strategic planning process, performance measurements, decision
making, and risk management. Chang et. al. (2011) analyzed corporate sustainability performance of 16
114
industries by Data Envelopment Analysis (DEA) and found that 7 of those industries improved their
sustainability performances.
Rowbottom and Lymer (2009) examined the website users of sustainability reports and their requires. They
found that vast majority of requests come from employees, private individuals, and consultants. They also
indicated that professional investors, creditors, and accounting firms mostly focused on the annual reports,
significantly less focused on the sustainability reports.
Searcy and Elkhawas (2012) analyzed 24 listed firms in terms of using Dow Jones Sustainability Index (DJSI) in
some aspects such as using its logo. They state that less than half of those firms use the logo of DJSI in their
sustainability reports and websites for increasing ability of stakeholders to understand sustainability reports
which would help them making different decisions than from their competitors.
Albu et. al. (2011, p.223) state that literature indicates four benefits of Corporate Social Responsibility (CSR):
CSR reduces direct costs (energy, materials, time loss, etc.); improves productivity of workers (increased
motivation, low absenteeism, reduced turnover); reduces management risk (easier access to credit, increased
value of the assets for investors, support by stakeholders, etc.); and improves the competitive image of the firm.
Some researchers investigated the relationship between corporate social responsibility and firm’s attributes and
specifics. The results show a positive association between proportion of Independent Directors (INDs) and
Corporate Social Responsibility Disclosure (CSRD). Some other results indicate that firm’s size has no effect on
CSR but Board Leadership Structure, Board Audit Committee, Return on Equity are positively related to CSR
(Rouf, 2011).
There are few studies related to sustainability reporting in Turkey. Başar and Başar (2006) analyzed 100 largest
listed Turkish firms’ sustainability profile. According to the researchers most of the firms’ sustainability reports
are disclosed in their operations reports, but not separately. The firms’ sustainability reports mostly reveal
information related to human resources, health and safety issues. In addition to that the study reveals that the
listed firms’ reports do not disclose enough details about sustainability profile of the firms. Senal and Aslantaş
Ateş (2012) state that sustainability development promotes using various accounting and costing methods such
as environmental accounting, environmental managerial accounting, and life cycle costing.
Based on the answers to the survey related sustainability, it can be seen that most firms have difficulties in
understanding and applying sustainability concepts. Especially environmental issues are not being priorities of
many firms. However many firms have interested mostly in social and economic elements of sustainability
(Sustainability Survey, 2011, p.9).
3. Corporate Sustainability Reporting and Dimensions of Sustainability
“Sustainability”, “environmental, social and governance” (ESG), “non-financial” or “corporate social
responsibility” (CSR) reporting have been used interchangeably in the past, to present environmental, social or
governance issues in the related reports (Ioannou & Serafeim, 2012, p.2).
Although “Sustainability” and “Corporate Social Responsibility” (CSR) terms have often been used
inter-changeably, they indicate different points. “Sustainability is more of an over-arching concept which seeks
to promote continuous long term growth in all the various forms of capital available to us—financial, natural and
social. By contrast many see CSR as a more limited concept, focused on shorter-term issues and activities such
as legal compliance, philanthropy and improvement in workforce conditions. In general it might be said all
organizations aspire to being responsible but few would claim to be truly sustainable” (IFAC, 2006, p.1).
In the published literature there is no commonly accepted definition of corporate reporting or a corporate report.
Roca and Searcy (2012, p.104) state that some researchers define a sustainability report as “a report which must
contain qualitative and quantitative information on the extent to which the company has managed to improve its
economic, environmental and social effectiveness and efficiency in the reporting period and integrate these
aspects in a sustainability management system”. Roca and Searcy (2012, p.105) also indicate that the World
Business Council for Sustainable Development’s definition is similar; WBCSD (2002) defines “sustainable
development reports as public reports by companies to provide internal and external stakeholders with a picture
of the corporate position and activities on economic, environmental and social dimensions”.
There are three dimensions of sustainability; economic viability, social responsibility, and environmental
responsibility. “While trade-offs can occur between these dimensions, they are interconnected in various ways.
For example, being socially and environmentally responsible (toward employees, communities, and other
stakeholders), leads to enhanced trust, and, therefore, makes good business sense. Social and environmental
responsibility cannot stand in isolation from economic viability. Organizations must continue to provide
115
products and services that people want in order to generate profits, growth, and new jobs. While pursuing a
commercial imperative, organizations must also take into account their social and environmental impact as part
of ensuring that they generate added value for an organization and its stakeholders” (Sustainability Framework,
2011, p.8).
4. Corporate Sustainability Reporting and Stakeholder Engagement
Stakeholder engagement is vital to firms in creating value and viable operations. Stakeholder engagement is
basically informing different stakeholder groups and considering their opinions in decision making process of a
firm. Successful firms are certainly keen to identify the means of engaging stakeholders in that process.
Somehow failing in stakeholder engagement would lead to several negative results: poor performance,
unsatisfied customers, frustrated employees, damaging supply chain, possible compromising a firm’s reputation
with the wider community. In fact the quality of sustainability reporting clearly depends on successful
stakeholder engagement. Reflecting stakeholder’s opinion in decision making process possibly could help better
dealing with anticipated issues and be more proactive compared to the firms that are not involved in this process
(Sustainable Framework, 2011, p.34-35).
Global Reporting Initiative (GRI) reporting principles that are stated in “G3 Sustainability Reporting Guidelines”
outline how stakeholders are included to the process of business: “The reporting organization should identify
its stakeholders and explain in the report how it has responded to their reasonable expectations and interests”
(Sustainable Framework, 2011, p.35). By doing that firms craft and implement their strategy, attain social,
environmental, and economical elements of sustainability goals. Therefore, stakeholder engagement helps to
build skills of being proactive instead of being reactive.
The 2008 KPMG International Survey of Corporate Responsibility Reporting presents how stakeholder
engagement is being considered by large companies in 22 countries (Global Fortune 250 (G250), 100 largest
companies by revenue (N100)). According to the survey more than 50% of the companies somehow structured
stakeholder engagement by formal or informal techniques. 60% of G250 companies disclose of information
about their stakeholders and how the engagement is being done. Most of these companies indicate that
stakeholder engagement help them to search a way of reducing risk and exploiting new creative business
opportunities with corporate responsibility (Sustainable Framework, 2011, p.36).
5. Global Reporting Initiative (GRI)
GRI is the most well known guideline for sustainability reporting. The objective of GRI is to mainstream
“disclosure on environmental, social and governance performance”. The GRI G3 was released in 2006 and
updated in 2011 as Version 3.1 (Roca & Searcy, 2012, p.105). Figure 1 presents how GRI guideline outlines the
reporting process. Inputs are the principles and guidance, outputs are standard disclosure.
116
Principles and Guidance
Guidance for Defining

Report Content
input
Principles for Defining

Report Content
input
Principles for Ensuring

Report Quality
input
Guidance for Report

Boundary Setting
input
Profile Management Performance

Standard Disclosures
Approach Indicators
output output output
Focused Sustainability Report
Figure 1. Overview of the GRI guidelines

Source: Sustainability Reporting Guideline, 2006, Version 3, p.4.
The GRI criteria or performance indicators have been used to analyze firms’ sustainability reports by researchers.
Roca and Searcy (2012) analyzed 94 corporations’ sustainability reports to identify the indicators disclosed in
the reports and also investigated the use of indicators of GRI. Their findings suggested that 31 of the 94 reports
included indicators explicitly identified as GRI indicators. The researchers stated that “the most reported GRI
indicators appeared in 28 of the reports, while the least reported indicators appeared in 5 of the reports”.
6. Analyzing Sustainability Reports in Turkey
6.1 Data and Methodology
In this study selected sustainability reports of 9 non-financial companies are analyzed. Numbers of the reports
based on application levels are: 1 level A report, 1 level B+ report, 5 level B reports, and 2 level C reports. The
application level of reports can be plus (+) if external assurance was utilized for the report. As detailed in Table 1,
117
application level C is not required to disclose Part II indicators. Level B is required to report fully on a minimum
of any 20 “Performance Indicators”, at least one from each of: economic, environment, human rights, labor,
society, and product responsibility (GRI Application Levels, 2010-2011, Version 3.1).
The content analysis method is used based on the indicators of GRI. 3 firms’ sustainability reports (Arçelik,
Aygaz, and Efes reports) analyses are done manually as abstracting from the content table of the reports.
However, 6 sustainability reports’ content analysis data (Bilim, Akçansa, Koç Holding, Coca-Cola, Opet, and
Zorlu Energy) is obtained from Sustainability Disclosure Database. All the selected sustainability reports were
prepared based on 2006 Guideline, G3 type.
Sustainability reports used in this study were scored based on indicators that guided by GRI. Fully reported
indicator is scored as (2), partially reported is scored as (1), and finally not reported indicator is scored as zero
(0). If a firm’s sustainability report meets all the principles and standards of GRI Guideline, the firm could reach
to the maximum score of (310).
Table 1. Minimum application level criteria and maximum score

Total Number of Max. Score
Indicators (indicators*score)
(Max.) Score:
Application Level C B A
Reported :2
Partially Reported :1
Not Reported :0
Part I number of indicators 24 42 42 42 84
each indicator or each indicator
Part II number of indicators Not required 34 68
category or category
Part III number of
Minimum 10 Minimum 20 55+ 79 158
indicators
Source: Generated from GRI Application Level, 3.1, 2010-2011 and (http://3blmedia.com/media/3bl.png).
Based on GRI to meet minimum requirements, Level C firm must disclose 24 indicators of Part I, minimum 10
indicators of Part III and none of Part II. Level B firm must disclose all indicators of Part I and Part II, and
minimum 20 indicators of Part III. Level A firm must disclose all indicators of Part I and Part II, and minimum
55 indicators of Part III. As provided in Table 1, if a firm discloses all 42 indicators of Part I, its score is
maximum 84; all 34 indicators of Part II, its score is maximum 68; and all 79 indicators of Part III, its score is
maximum 158. Hence the highest score a firm could reach is 310 as shown in Appendix.
6.2 Results
Results based on GRI indicators are summarized in Table 2. The Table actually is abstracted from Appendix.
The Table provides three parts of sustainability reports that are widely accepted and used by GRI Guide. First
part presents firms’ “Profile Disclosure” level that contains “Profile, Reporting Parameters, and Governance,
Commitments, and Engagement” elements. The second part presents “Disclosure on Management Approach
(DMA)”. The third part is related to “performance indicators” that are outlined as “Economic, Environmental,
Labor Practices and Decent Work, Human Rights, Society, and Product Responsibility”.
118
Table 2. Summary of firms’ content analysis

GRI Indicators / Firms/
Publication Year: Bilim Koç Zorlu Max.
2011-2012 P.Comp Akçansa Holding Coca-Cola Arçelik Aygaz Efes Opet Energy score
2010 2010 2011 2010 2009-2010 2010 2010

2007-2009 2010-2011
Type of Report/Application
G3-A G3-B G3-B G3-B G3-B+ G3-B G3-B G3-C G3-C
level
Part I: Profile Disclosures 84 84 84 84 84 84 84 58 84 84
1-2. Profile 24 24 24 24 24 24 24 22 24 24
3. Reporting Parameters 26 26 26 26 26 26 26 24 26 26
Report Profile 8 8 8 8 8 8 8 8 8 8
Report Scope and Boundary 14 14 14 14 14 14 14 12 14 14
GRI Content Index 2 2 2 2 2 2 2 2 2 2
Assurance 2 2 2 2 2 2 2 2 2 2
4. Governance,
Commitments, and 34 34 34 34 34 34 34 12 34 34
Engagement
Governance 20 20 20 20 20 20 20 8 20 20
Commitment to External
6 6 6 6 6 6 6 0 6 6
Initiatives
Stakeholder Engagement 8 8 8 8 8 8 8 4 8 8
Part II: Disclosures on
Management Approach 68 64 68 68 68 0 68 0 68 68
(DMA)
Part III: Performance
122 104 64 93 119 115 114 56 146 158
Indicators
Economic 14 12 8 14 13 15 17 9 18 18
Environmental 49 28 21 25 55 33 43 9 48 60
Labor Practices and Decent
22 24 17 21 26 25 19 15 28 28
Work
Human Rights 10 14 9 8 14 13 12 7 18 18
Society 10 10 6 7 5 13 10 7 16 16
Product Responsibility 17 16 3 18 6 16 13 9 18 18
TOTAL 274 252 216 245 271 199 266 114 298 310
Several results can be generated from Table 2. The first one is about Part I. As presented in Table 2, selected
firms sustainability reports Part I indicators are fully reported except in one firm’s report (Opet). Since the firms’
application level is C, those indicators are not disclosed. Those indicators could be reported voluntarily as other
C level firm did (Zorlu Energy). Selected firms’ sustainability reports clearly have disclosed “Profile
Disclosures”. It can be said that firms fulfill requirements related to Part I based on their reports application
level.
The second results are related to Part II, “Disclosures on Management Approach”. Almost seven selected reports
disclosed those indicators fully in their reports while two firms disclosed none of the indicators. This situation
can be explained by reports’ application levels. C level reports are not required to disclose those indicators
unless they voluntarily do it. Although level B reports are required to disclose Part II indicators, one B
application level firm (Aygaz) did not reported such indicators in its content table. As one of the elements of Part
II, firms’ stakeholder engagement scores are consistent among reports.
The third results state disclosure level of “Performance Indicators”, Part III. Most interesting result of the study
is in the third part. Based on the disclosing level, firms’ reports meet minimum requirements. However,
disclosing Part III indicators varies significantly among the same application level firms. For example Koç
Holding’s score is 64 while Arçelik’s score is 119 for the total of Part III indicators. There is inconsistency
between C level reports as well. One report has much higher score than the other. It can be concluded that firms’
sustainability reports fail to consistently disclose Performance Indicators.
In Part III category, only one firm’s report (Zorlu Energy) meets the maximum scores except one subcategory,
environmental indicators. Most of the firms’ environmental disclosure level is not close to the maximum score.
There is no consistency disclosing other performance indicators between firms’ reports as well. Except one
119
firm’s report, all the reports do not fully report “Economic, Labor Practices and Decent Work, Human Rights,
Society, and Product Responsibility” indicators.
7. Conclusions
Reporting sustainability is a key process to inform stakeholders whether the firm is achieving sustainable growth
and value for their interest. To be more reliable most corporations have followed guidelines for corporate
sustainability reporting. The well known set of voluntary guideline is the Global Reporting Initiative (GRI, 2006).
The guideline basically focuses on the context of corporate sustainability reports, corporations’ sustainability
vision, their performances, and objectives towards sustainability.
The GRI criteria or performance indicators have been used to analyze firms’ sustainability reports by researchers.
In this study selected nine Turkish firms’ sustainability reports have been analyzed based on the GRI indicators.
The results show that selected firms’ sustainability reports usually meet the minimum requirements of the GRI
standards based on their application level. However there are differences between firms’ scores even though they
are in the same application level. Some firms reveal more information than others, but in general firms do not
disclose many indicators voluntarily.
It can be said that firms fulfill requirements related to Part I and Part II, “Profile Disclosures” and “Disclosures
on Management Approach” based on their reports’ application level. However, it is hard to say the same thing
for the Part III. In fact firms’ sustainability reports fail to consistently disclose “Performance Indicators”.
The effort of preparing sustainability index in Istanbul Stock Exchange is the evidence that Turkey and Turkish
firms are paying more attention to reporting sustainability. However, most firms are still behind of this trend.
Regular reporting of sustainability and the disclosure level has several problems needed to be solved. Mandatory
regulations may provide better sustainability reporting environment. There are many possibilities for future
research in this field. For example, studies could explore relations between firms’ performance and level of
disclosure.
References
Adams, C. A., & Frost, G. R. (2008). Integrating Sustainability Reporting into Management Practices.
Accounting Forum, 32, 288-302. http://dx.doi.org/10.1016/j.accfor.2008.05.002
Albu, N., Albu, C. N., Gîrbină, M. M., & Sandu, M. I. (2011). The Implications of Corporate Social
Responsibility on the Accounting Profession: The Case of Romania. Corporate Social Responsibility,
XIII(29), 221-234.
Başar, A. B., & Başar, M. (2006). Social Responsibility Reporting: The Case of Turkey. Sosyal Bilimler Dergisi,
2, 213-230.
Chang, D. S., Kuo, L. R., & Chen, Y. (2011). Industrial Changes in Corporate Sustainability Performance - An
Empirical Overview Using Data Envelopment Analysis. Journal of Cleaner Production, xxx, 1-9.
Farneti, F., & Guthrie, J. (2009). Sustainability Reporting by Australian Public Sector Organisations: Why They
Report. Accounting Forum, 33, 89-98. http://dx.doi.org/10.1016/j.accfor.2009.04.002
GRI Application Levels. (2010-2011). 3.1.
Gurvitsh, N., & Sidorova, I. (2012). Survey of Sustainability Reporting Integrated into Annual Reports of
Estonian Companies for the Years 2007-2010: Based on Companies Listed on Tallinn Stock Exchange as of
October 2011. Procedia Economics and Finance, 2, 26-34.
IFAC, Information Paper. (2006). Why Sustainability Counts for Professional Accountants in Business.
Professional Accountants in Business Committee, 1-8.
Ioannou, I., & Serafeim, G. (2012). The Consequences of Mandatory Corporate Sustainability Reporting.
Working Paper, Harvard Business School.
Joseph, G. (2012). Ambiguous but Tethered: An Accounting Basis for Sustainability Reporting. Critical
Perspectives on Accounting, 23, 93-106. http://dx.doi.org/10.1016/j.cpa.2011.11.011
Marimon, F., Alonso-Almeida, M. M., Rodrigez, M. P., & Alejandro, K. A. C. (2012). The Worldwide Diffusion
of the Global Reporting Initiative: What is The Point?. Journal of Cleaner Production, 33, 132-144.
http://dx.doi.org/10.1016/j.jclepro.2012.04.017
120
Moradzadehfard, M., & Moshashaei, M. (2011). The Accountants’ Role in Organizations’ Sustainability
(Regarding Sustainability Framework of the International Federation of Accountants (IFAC)). European
Journal of Scientific Research, 59(3), 396-402.
Murguia, D. I., & Böhling, K. (2013). Sustainability Reporting on Large-Scale Mining Conflicts: The Case of
Bajo De La Alumbrera, Argentina. Journal of Cleaner Production, 41, 202-209.
Roca, L. C., & Searcy, C. (2012). An Analysis of Indicators Disclosed in Corporate Sustainability Reports.
Journal of Cleaner Production, 20, 103-118. http://dx.doi.org/10.1016/j.jclepro.2011.08.002
Rouf, A. (2011). The Corporate Social Responsibility Disclosure: A Study of Listed Companies in Bangladesh.
Business and Economics Research Journal, 2(3), 19-32.
Rowbottom, N., & Lymer, A. (2009). Exploring the Use of Online Corporate Sustainability Information.
Accounting Forum, 33, 176-186. http://dx.doi.org/10.1016/j.accfor.2009.01.003
Schaltegger, S., & Burritt, R. L. (2010). Sustainability Accounting for Companies: Catchphrase or Decision
Support for Business Leaders?. Journal of World Business, 45, 375-384.
http://dx.doi.org/10.1016/j.jwb.2009.08.002
Searcy, C., & Elkhawas, D. (2012). Corporate Sustainability Ratings: An Investigation into How Corporations
Use The Dow Jones Sustainability Index. Journal of Cleaner Production, 35, 79-92.
Senal, S., & Aslantaş Ateş, B. (2012). Kurumsal Sürdürülebilirlik için Muhasebe ve Raporlama. Muhasebe ve
Denetime Bakış. Nisan, 83-977.
SKD. (2012). http://www.tbcsd.org/?Lisan=EN, (August 2012).
Sustainability Disclosure Database. (August 2012). http://database.globalreporting.org/
Sustainability Framework 2.0. (2011). Professional Accountants as Integrators, p. 1-202.
Sustainability Reporting Guidelines (RG). (2000-2006). GRI, Version 3.0, Global Reporting Initiative, p. 1-44.
Sustainability Survey. (2011). Türk İş Dünyası’nda Sürdürülebilirlik Uygulamaları Değerlendirme Raporu. Eylül
2011, 1-50,.
Williams, B., Wilmshurst, T., & Clift, R. (2011). Sustainability Reporting by Local Government in Australia:
Current and Future Prospects. Accounting Forum, 35, 176-186.
http://dx.doi.org/10.1016/j.accfor.2011.06.004
Appendix
Koç
GRI Indicators / Firms/ Bilim Coca-Cola OPET Zorlu max
Akçansa Holding Arçelik Aygaz Efes
publication year 2011-12 P.Comp. İçecek Petrol. Energy score
A.Ş.
Year 2010 2007-2009 2010 2011 2010 2010-11 2009-10 2010 2010
Type of
G3-A G3-B G3-B G3-B G3-B+ G3-B G3-B G3-C G3-C
Report/Application level
Part I: Profile Disclosures
Score of profile 24 24 24 24 24 24 24 22 24 24
1.1. 2 2 2 2 2 2 2 2 2 2
1.2. 2 2 2 2 2 2 2 0 2 2
2.1. 2 2 2 2 2 2 2 2 2 2
2.2. 2 2 2 2 2 2 2 2 2 2
2.3. 2 2 2 2 2 2 2 2 2 2
2.4. 2 2 2 2 2 2 2 2 2 2
2.5. 2 2 2 2 2 2 2 2 2 2
2.6. 2 2 2 2 2 2 2 2 2 2
2.7. 2 2 2 2 2 2 2 2 2 2
2.8. 2 2 2 2 2 2 2 2 2 2
2.9. 2 2 2 2 2 2 2 2 2 2
2.10. 2 2 2 2 2 2 2 2 2 2
121
3. Reporting Parameters 26 26 26 26 26 26 26 24 26 26
Report Profile 8 8 8 8 8 8 8 8 8 8
3.1. 2 2 2 2 2 2 2 2 2 2
3.2. 2 2 2 2 2 2 2 2 2 2
3.3. 2 2 2 2 2 2 2 2 2 2
3.4. 2 2 2 2 2 2 2 2 2 2
Report Scope and
14 14 14 14 14 14 14 12 14 14
Boundary
3.5. 2 2 2 2 2 2 2 2 2 2
3.6. 2 2 2 2 2 2 2 2 2 2
3.7. 2 2 2 2 2 2 2 2 2 2
3.8. 2 2 2 2 2 2 2 2 2 2
3.9. 2 2 2 2 2 2 2 0 2 2
3.10. 2 2 2 2 2 2 2 2 2 2
3.11. 2 2 2 2 2 2 2 2 2 2
GRI Content Index 2 2 2 2 2 2 2 2 2 2
3.12. 2 2 2 2 2 2 2 2 2 2
Assurance 2 2 2 2 2 2 2 2 2 2
3.13. 2 2 2 2 2 2 2 2 2 2
4. Governance,
Commitments, and 34 34 34 34 34 34 34 12 34 34
Engagement
Governance 20 20 20 20 20 20 20 8 20 20
4.1. 2 2 2 2 2 2 2 2 2 2
4.2. 2 2 2 2 2 2 2 2 2 2
4.3. 2 2 2 2 2 2 2 2 2 2
4.4. 2 2 2 2 2 2 2 2 2 2
4.5. 2 2 2 2 2 2 2 0 2 2
4.6. 2 2 2 2 2 2 2 0 2 2
4.7. 2 2 2 2 2 2 2 0 2 2
4.8. 2 2 2 2 2 2 2 0 2 2
4.9. 2 2 2 2 2 2 2 0 2 2
4.10. 2 2 2 2 2 2 2 0 2 2
Commitment to External
6 6 6 6 6 6 6 0 6 6
Initiatives
4.11. 2 2 2 2 2 2 2 0 2 2
4.12. 2 2 2 2 2 2 2 0 2 2
4.13. 2 2 2 2 2 2 2 0 2 2
Stakeholder Engagement 8 8 8 8 8 8 8 4 8 8
4.14. 2 2 2 2 2 2 2 2 2 2
4.15. 2 2 2 2 2 2 2 2 2 2
4.16. 2 2 2 2 2 2 2 0 2 2
4.17. 2 2 2 2 2 2 2 0 2 2
Part I - Total 84 84 84 84 84 84 84 58 84 84
Part II: Disclosures on
Management Approach
(DMA)
-DMA EC
-Economic performance 2 2 2 2 2 0 2 0 2 2
-Market presence 2 2 2 2 2 0 2 0 2 2
-Indirect economic impacts 2 2 2 2 2 0 2 0 2 2
-DMA EN
-Materials 2 2 2 2 2 0 2 0 2 2
-Energy 2 2 2 2 2 0 2 0 2 2
-Water 2 1 2 2 2 0 2 0 2 2
-Biodiversity 2 2 2 2 2 0 2 0 2 2
122
-Emissions, effluents and

2 2 2 2 2 0 2 0 2 2
waste
-Products and services 2 1 2 2 2 0 2 0 2 2
-Compliance-EN 2 2 2 2 2 0 2 0 2 2
-Transport 2 0 2 2 2 0 2 0 2 2
-Overall 2 2 2 2 2 0 2 0 2 2
-DMA LA
-Employment 2 2 2 2 2 0 2 0 2 2
-Labor/management
2 2 2 2 2 0 2 0 2 2
relations
-Occupational health and
2 2 2 2 2 0 2 0 2 2
safety
-Training and education 2 2 2 2 2 0 2 0 2 2
-Diversity and equal
2 2 2 2 2 0 2 0 2 2
opportunity
-DMA HR
-Investment and
2 2 2 2 2 0 2 0 2 2
procurement practices
-Non-discrimination 2 2 2 2 2 0 2 0 2 2
-Freedom of association and
2 2 2 2 2 0 2 0 2 2
collective bargaining
-Child labor 2 2 2 2 2 0 2 0 2 2
-Forced and compulsory
2 2 2 2 2 0 2 0 2 2
labor
-Security practices 2 2 2 2 2 0 2 0 2 2
-Indigenous rights 2 2 2 2 2 0 2 0 2 2
-DMA SO
-Community 2 2 2 2 2 0 2 0 2 2
-Corruption 2 2 2 2 2 0 2 0 2 2
-Public policy 2 2 2 2 2 0 2 0 2 2
-Anti-competitive behavior 2 2 2 2 2 0 2 0 2 2
-Compliance-SO 2 2 2 2 2 0 2 0 2 2
-DMA PR
-Customer health and safety 2 2 2 2 2 0 2 0 2 2
-Product and service
2 2 2 2 2 0 2 0 2 2
labelling
-Marketing communications 2 2 2 2 2 0 2 0 2 2
-Customer privacy 2 2 2 2 2 0 2 0 2 2
-Compliance-PR 2 2 2 2 2 0 2 0 2 2
Part II - Total 68 64 68 68 68 0 68 0 68 68
Part III: Performance
Indicators
Economic 14 12 8 14 13 15 17 9 18 18
EC1. 2 2 2 2 2 2 2 1 2 2
EC2 1 0 1 2 1 1 2 1 2 2
EC3 2 2 1 2 1 2 1 1 2 2
EC4 2 2 0 2 1 2 2 1 2 2
EC5 2 2 1 0 2 2 2 0 2 2
EC6 1 2 0 1 1 2 2 1 2 2
EC7 1 1 1 1 2 2 2 0 2 2
EC8 2 1 1 2 2 1 2 2 2 2
EC9 1 0 1 2 1 1 2 2 2 2
Environmental 49 28 21 25 55 33 43 9 48 60
EN1 2 2 0 0 2 0 2 0 2 2
EN2 2 2 0 0 1 1 2 0 2 2
EN3 2 2 0 2 2 0 2 0 2 2
123
EN4 2 2 0 0 1 0 2 0 2 2
EN5 2 2 2 2 2 2 2 1 2 2
EN6 2 0 2 1 2 2 1 0 2 2
EN7 2 0 1 1 2 1 1 0 2 2
EN8 2 2 2 1 2 2 2 1 2 2
EN9 0 0 0 1 2 2 2 0 2 2
EN10 0 0 2 1 2 2 0 1 2 2
EN11 2 2 2 2 2 2 2 0 2 2
EN12 2 2 0 0 2 0 0 0 2 2
EN13 2 0 2 0 2 1 0 0 2 2
EN14 2 2 0 0 2 0 2 0 1 2
EN15 2 0 0 0 2 0 9 0 1 2
EN16 2 2 0 1 2 0 2 0 1 2
EN17 2 0 0 0 2 0 1 0 1 2
EN18 2 2 2 1 2 2 2 1 1 2
EN19 0 0 0 0 2 2 0 0 1 2
EN20 1 2 1 1 2 0 0 1 1 2
EN21 2 0 1 2 2 1 2 2 1 2
EN22 1 2 1 1 2 2 0 1 1 2
EN23 2 0 0 2 2 2 0 0 1 2
EN24 0 0 0 0 2 1 0 0 1 2
EN25 2 0 0 0 2 2 2 0 1 2
EN26 2 0 2 2 2 2 2 1 2 2
EN27 1 0 0 1 1 1 1 0 2 2
EN28 2 0 0 2 1 2 0 0 2 2
EN29 2 0 0 1 1 0 1 0 2 2
EN30 2 2 1 0 2 1 1 0 2 2
Labor Practices and
22 24 17 21 26 25 19 15 28 28
Decent Work
LA1 2 2 2 2 2 2 1 1 2 2
LA2 0 1 0 2 2 1 1 1 2 2
LA3 2 2 0 2 2 2 2 2 2 2
LA4 2 2 2 2 2 2 2 0 2 2
LA5 2 1 2 2 2 2 2 2 2 2
LA6 1 2 2 1 2 2 2 0 2 2
LA7 2 2 0 2 1 2 1 0 2 2
LA8 1 1 1 1 2 1 0 2 2 2
LA9 0 2 1 2 2 2 0 0 2 2
LA10 2 1 2 2 2 2 1 1 2 2
LA11 2 2 1 1 2 1 2 1 2 2
LA12 2 2 2 1 2 2 1 2 2 2
LA13 2 2 0 1 1 2 2 1 2 2
LA14 2 2 2 0 2 2 2 2 2 2
Human Rights 10 14 9 8 14 13 12 7 18 18
HR1 1 1 0 0 1 1 0 0 2 2
HR2 1 1 1 0 1 1 0 0 2 2
HR3 1 0 0 0 2 0 2 0 2 2
HR4 1 2 2 2 2 2 2 2 2 2
HR5 2 2 2 2 2 2 2 1 2 2
HR6 2 2 2 2 2 1 2 2 2 2
HR7 2 2 2 2 2 2 2 2 2 2
HR8 0 2 0 0 2 2 2 0 2 2
HR9 0 2 0 0 0 2 0 0 2 2
Society 10 10 6 7 5 13 10 7 16 16
SO1 1 0 1 0 2 1 1 0 2 2
SO2 1 0 1 1 1 2 1 2 2 2
124
SO3 2 2 1 0 0 1 0 1 2 2
SO4 2 0 1 0 0 1 2 2 2 2
SO5 0 2 0 2 0 2 2 2 2 2
SO6 0 2 2 2 2 2 2 0 2 2
SO7 2 2 0 2 0 2 0 0 2 2
SO8 2 2 0 0 0 2 2 0 2 2
Product Responsibility 17 16 3 18 6 16 13 9 18 18
PR1 1 0 0 2 1 1 0 0 2 2
PR2 2 2 0 2 0 2 2 0 2 2
PR3 2 2 1 2 2 2 2 2 2 2
PR4 2 2 0 2 0 2 2 2 2 2
PR5 2 2 2 2 1 2 1 2 2 2
PR6 2 2 0 2 2 1 2 1 2 2
PR7 2 2 0 2 0 2 2 2 2 2
PR8 2 2 0 2 0 2 2 0 2 2
PR9 2 2 0 2 0 2 0 0 2 2
Part III -Total 122 104 64 93 119 115 114 56 146 158
TOTAL 274 252 216 245 271 199 266 114 298 310
125
ISSN 1916-971X E-ISSN 1916-9728
Top Companies Ranking Based on Financial Ratio with AHP-TOPSIS

Combined Approach and Indices of Tehran Stock Exchange
A Comparative Study
Seyed-Hasan Hosseini1, Mohammad Esmaeil Ezazi1, Mohamad Rasoul Heshmati2 & Seyed-Mohamad Reza
Hosseini Moghadam2
1
Faculty of Management and economics, School of Business management, University of Sistan and Baluchestan,
Zahedan, Iran
2
Departments of Management, Faculty of Economic and Administrative Sciences, Ferdowsi University of
Mashhad, Mashhad, Iran
Correspondence: Seyed-Hasan Hosseini, Faculty of accounting and management, School of financial management,
University of Sistan and Baluchestan, Zahedan, P.O. Box: 98155-987, Iran. Tel: 98-91-4841-1580. E-mail:
[email protected]
Abstract
This paper aims to explore the relationship between ranking of the top-50 listed companies on Tehran Stock
Exchange (TSE) for the years 2009- 2011 in terms of their liquidity, operation, leverage and profitability ratios
using combined AHP-TOPSIS approach and the ranking made by the stock exchange. Ranking of the companies
on the stock exchange is done based on their state in terms of the above ratios and it serves as a criterion for
decision making on investment. Using a questionnaire, views of experts, scholars and the capital market
authorities on the effect of financial ratios were gathered and then using AHP- TOPSIS technique the companies
were ranked based on these ratios. The obtained results from the Spearman Test show a weak correlation
between the rankings based on AHP-TOPSIS approach and the ranking of the stock exchange. Finally, our
results indicate that financial ratios of the top stock exchange selected companies are crucial factors in
investment and ranking. This paper contributes to a signaled need for investigation on how and why financial
statements of top listed companies cannot be regarded as a critical factor in ranking and decision making on
investment.
Keywords: AHP-TOPSIS, financial ratios, ranking, Tehran Stock Exchange (TSE)
1. Introducation
1.1 Introduce the Problem
Companies ranking is a practical tool for distinguishing companies in distress from the healthy ones. Ranking
also guarantees survival of companies (Li & Sun, 2008). It is necessary to make a great number of business
decisions, e.g. in case of investment decisions, loan granting, and generally in case of establishment of any
business relationship. In recent years, many developed countries by foundation of capital market (i.e. foundation
of stock exchanges and over the counter markets (OTC) have secured their economic development
(Babic&Plazibat, 1998). Development of stock exchange and OTC has resulted in generation and expansion of
financial services which at different levels provide investors with consultation services and rank companies and
stock exchanges (O'Hara & Vetere,1993).Investors, managers, creditors and in general, stakeholders each using
some criteria measuring business unit and its performance. In this regard, analysis of financial statements can be
a right solution for evaluation and ranking of companies. Ranking can reveal weaknesses and strengths of the
companies as well as opportunities and threats for them. In fact, it is a full-length mirror reflecting their state of
affairs. This tool has very crucial role in decisions regarding companies' trading, investments and financing
(Conner, 1973).
Sometimes, decision maker uses knowledge and experience based models simultaneously and often by accurate
evaluations and application of various methods (Babic&Plazibat, 1998). Different criteria may be applied to
126
company ranking using particular indicators and different ranking methods. In the Iranian capital market, due to
absence of financial services institutes, the current reliable ranking is the ranking made by the stock exchange.
One of the published reports concerns the report on the top listed companies on the stock exchange which is
prepared based on share liquidity, amount of company's investment in the market and company's situation in
terms of financial excellence using harmonic mean method. Ranking is a complex and multi-criteria concept
which, given its calculation simplicity and optimality, serves as a suitable tool to examine and analyze the
enterprise excellence and its situation.
1.2 Explore Importance of the Problem
In this research, for prioritization and ranking of the top-50 listed companies based on their financial ratios and
then comparing it with the ranking of the stock exchange, Analytic Hierarchy Process (AHP) technique in the
hand of each key indicator (liquidity, leverage, operation and profitability ratios) was employed, while a relative
weight was assigned to each sub-indicator. Next, using Technique for Order Preferences by Similarity Ideal
Solution (TOPSIS) the under study firms were ranked based on each one of these indicators. Finally, the
obtained ranking was compared with the ranking provided by the stock exchange and based on the stock
exchange's indices and then their correlation was examined. In fact, we will use the TOPSIS method for final
ranking and analytic hierarchy process to determine the importance of the criteria in the top-50 listed companies
in terms of liquidity, operational, leverage and profitability ratio.
1.3 An Overview of Ranking Based on Different Criteria
Studies on indices of successful companies indicate a significant relationship between firms ranking based on
performance measurement criteria and financial criteria (Johnson &Soenen, 2003).Another study in S&P index
revealed that such factors as share price, sales, and profit margin suggested higher rating of favorable factors for
high ranking firms (Polonchek&Krehbiel, 1994). Following the prior research, Omran and Ragab (2004)
examined presence of a linear relationship between share return and financial ratios and then investigated
presence of a linear relationship between share return and financial ratios (Omran&Ragab, 2004) .They found
that the ratio return on equity (ROE) was significant for all models. Hassanzadeh et al (2010) in their study
found that there is a significant association between firms' financial ratios and bank managers' decision on
granting them credit(Hassanzadeh& et all,2010).Lev and Thiagrajan (1993) following Penman's studies and
using financial ratios concluded that fundamental signals are associated to share return(Lev &
Thiagrajan,1993).In addition, recent findings suggest that although financial information plays a crucial role in
prediction of return on investment (ROI), yet the effect of each financial factor depends on market condition and
under this condition this effect is not stable(Knif& Miranda,2000).Besides, other studies indicate that financial
statements and financial ratios are used for distinction of successful companies from unsuccessful ones
(Piotroski,2000;Mohanram,2005;Michou,2007). Their results indicate that companies with a higher book-to-
market value have a higher return on average.
2. Method
This research is of descriptive-correlation type in which top-fifty listed companies for the period 2009-2011 were
examined by census. The research statistical data was gathered by referring to the Securities and Exchange
Organization (SEO) and using Rahavard-e-Novin software. The top-50 listed companies on TSE are presented in
the table in each one of the under study years according to their industry.
Table 1. Top-50 listed companies on tse in each one of the under study years per industry
Industry type 2009 2010 2011
Financial brokers, investment & holding 19 22 19
Cement, lime, chalk & metal ore extraction 6 4 6
Automobile manufacturing & base metals 10 11 9
Food and chemical products 6 4 4
Pharmaceutical materials& products 3 5 3
Miscellaneous industries 6 4 9
Total (Number) 50 50 50
In present research, for the purpose of ranking based on financial ratios, a questionnaire was composed by means
of which weight of financial ratio in the ranking is obtained. Once the questionnaire's validity has been
confirmed, its reliability using inconsistency rate was found to be 0.06 which was smaller than 0.1 and hence
127
approved. Next, the questionnaire was handed out to 20 experts and respective officials of the capital market as
well as professors of finance and accounting from various universities and they were asked to give their opinion
on the question as to what extent each one of the financial ratios should be considered as a critical indicator in
firms ranking. Given the type, purpose, hypotheses and questionnaire of the research, a 1-to-9-hour scale was
employed to form the matrix of paired comparisons in order to evaluate weight of the indicators and to rank
companies using AHP and TOPSIS techniques. In the next step, using AHP technique and Expert Choice
software, the indicators were assigned weight and then profiting from TOPSIS technique under TOPSIS (2005)
software the firms were ranked. Finally, utilizing SPSS software, Spearman Correlation Coefficient (rs) and
relation (1) were used to examine and analyze the correlation between the rankings and its significance.
∑
1 (1)
In this relation, n is the number of data, ∑(di2) is the sum of square difference between rankings of two variables
x and y. Variable x represents firms ranking based on the stock exchange indices and variable y represents firms
ranking based on financial ratios in the combined AHP-TOPSIS approach. To study the correlation significance,
relation 2 is utilized:
1 (2)
Where the statistical hypotheses are:

H0: ρ = 0, there is no significant relationship.
H1: ρ ≠ 0, there is significant relationship
In addition, given the obtained correlation coefficient value, and intensity and direction of the relationship,
method of weight assignment to ratios and firms ranking is determined. The indicators have been selected
conform to opinion of a group of experts and managers. Another expert group might have selected a somewhat
different set of indicators, yet this would not affect the main idea in this research. As was stated, this research
aims to form a methodological framework and to propose a model which can substitute the composed indicator
for companies ranking. To assign weight and generalize experts' view on the four understudy financial criteria
(financial ratios), AHP technique was employed. In this technique, after construction of the model and
development of the decision tree (hierarchy structure), the matrix of paired comparisons for indicators which
represents significance and priority of the index i relative to index j was formed and subsequently de-scaled.
Next, matrix inconsistency rate was calculated and finally, for prioritization and ranking of the top-50 companies,
relative weight of the indicators was obtained, as is presented in Table 2 and Figure 1.
Table 2. Matrix of paired comparisons of research's chief criteria in expert choice

Chief criteria Liquidity ratio operational ratio Leverage ratio Profitability ratio
Liquidity ratio 1 0.33 0.5 0.2
operational ratio 1 3.5 0.5
Leverage ratio 1 0.5
Profitability ratio 1
Notes: Liquidity: 0.0.86, operation: 0.322, Leverage: 0.153, Profitability: 0.438.
As is seen, profitability ratio with a relative weight of 0.438 has the highest significance and the index liquidity
ratio with a relative weight of 0.086 has the least significance. A paired comparisons inconsistency rate of 0.06
was obtained which is smaller than 0.10. Hence, we infer that these comparisons have a reasonable consistency.
Sub-criteria's relative weight compared to chief criteria (four financial ratios) and their prioritization in Expert
Choice software are presented in figures 2 and 5 according to which relative weight of the four chief criteria was
obtained.
128
Fiigure1. Prioritiization of four chief criteria ((financial ratioo) according too Expert Choicce software
(A: Liquidiity; B: Operatiion; C: Leveraage; D: Profitabbility)
Figgure2. Prioritizzation of liquiddity ratio usingg Expert Choicce software

(A: Currrent; B: Instannt)
Figgure 3. Prioritizzation of operaation ratio usinng Expert Choiice software

( Period of receiving claim
(A: ms; B: Inventorry flow; C: Asssets flow)
Figgure 4. Prioritizzation of Leveerage ratio usinng Expert Choiice software

(A
A: Debt to equuity ratio; B: D
Debt ratio)
Figuure 5. Prioritizaation of profitaability ratio usiing Expert Chooice software

(A: P
Profit margin; B: ROE; C: R
ROA; D: ROI)
Following this process, to t apply TOPS mpanies based on each one of

SIS technique to ranking of tthe top-50 com o the
financial rratios and choiice of the best option from aamong the avaailable options, first, decisionn making matrrix is
formed annd using entroopy weight asssignment metthod the indiccators are assiigned a weighht. To de-scale e the
indicators’’ measurementt units and to allow algebraiic operation onn them, each eelement of deccision matrix using
u
relation 3 iis standardizedd and finally thhe de-scaled m
matrix is built.
(3)
∑
Next, the weighted de-sscaled matrix and positive ((A+) and negaative (A-) focuuses were calcuulated. To find the
129
answer of positive and negative focuses for the indices, relations 4 and 5 are used.
, ,….., | , | | 1,2, … . . , (4)
, ,….., | , | | 1,2, … . . , (5)
In these relations, J+ and J- represent the indicator sets of positive and negative indicators in decision making,
respectively. Answer of the focuses for each indicator in each year is presented in Table 3.
Table 3. Positive and negative focuses for each index in each year
Year A+ and A- Profitability Leverage Operation Liquidity
2009 , ,….., 0.0622 0.0003 0.0092 0.0068

, ,….., -0.0090 0.0084 -0.0013 0.0004
2010 , ,….., 0.0209 0.0000 0.0580 0.0172
, ,….., -0.0160 0.0071 0.0000 0.0004
2011 , ,….., 0.0398 0.0000 0.0292 0.0122
, ,….., 0.0005 0.0067 0.0000 0.0000
In the next stage, using relations 6 and 7 which respectively represent indicators difference from positive focus
and negative focus for option i, Euclidean difference of each option was calculated:
→ 1,2, … . . , 6
∑ → 1,2, … . . , (7)
Having found all the positive and negative spaces, each decision option was determined by linking it to a number
(8). It is noteworthy that the result should always be between 0 and 1.
→ 1,2, … . . , , 0 1 8
Finally, the firms were ranked per year and indicators (financial ratios), so as the greater the relative distance
(CLi) of decision making options is and the closer they are to (1), the higher their priority and rank will be.
3. Results
The hypothesis of this study suggests that most companies with appropriate financial ratios in terms of liquidity,
operation, leverage and profitability ratios have a higher rank. So such companies enjoy a higher priority. Now to
prove this claim, we need to find a significant relationship between two types of rankings based on proportion of
each ratio. Given the large number of the group financial brokerage firms, investment, and holding among the
top listed companies, the correlation between the two rankings has been analyzed with and without inclusion of
this group of companies. Table 4 presents the results obtained from test of the hypotheses and correlation
coefficient between these two types of ranking. In this research, given the proposed hypotheses, five correlations
between the rankings were specified and calculated and analyzed based on all or each one of the financial ratios
(Table 4). In analysis of the main hypothesis, there an inverse relationship was found between two types of
ranking with and without presence of financial brokerage, investment and holding companies in 2009 but this
relationship in 2010-2011 has turned into a direct relationship, indicating a doubled interest in ranking and
investors' investment relative to the past. Given the stronger and greater correlation, if financial brokerage,
investment and holding companies are excluded from the group of the superior firms, it can be said that
relationship between ranking of other companies by the stock exchange and their ranking based on financial ratio
using AHP-TOPSIS approach becomes more aligned. With regard to the sub-hypothesis liquidity, correlation
coefficient with and without presence of financial brokerage, investment and holding companies was positive in
130
all three years, indicating a direct relationship. Hence, it can be stated that in each year relative to previous year,
attention to liquidity ratio has increased in the ranking and investment. In addition, there is no difference
between financial brokerage, investment and holding companies and other companies regarding convergence of
the ranking by the stock exchange and that based on the financial ratios using AHP-TOPSIS technique.
In the second sub-hypothesis, correlation coefficient was negative in all the three years of 2009-2011 and its
intensity has increased year by year. This indicates a greater inverse attitude to the operation ratio in the ranking
by the stock exchange during the time period in question and investors have had greater negative attitude to this
ratio as well as to share trading. However, as is seen, with exclusion of financial brokerage, investment and
holding companies, the inverse correlation in 2009 has turned into a direct correlation in 2010 and again into
inverse correlation in 2011. This suggests a greater interest in operation ratio in the ranking of other companies
during 2010 and the relationship between ranking of these companies by the stock exchange and their ranking
based on operation ratios using AHP-TOPSIS approach has had more convergence in this year. But this attitude
and convergence has been weakened in 2011. The obtained results regarding the third sub-hypothesis, confirms
that the negative and very weak correlation in 2009 has turned into a positive correlation in 2010-2011 and
intensity of this relationship has enhanced. This change in the relationship direction indicates the positive and
greater interest in leverage ratio in ranking or investors' positive attention to leverage ratio in their share trading
and investment. This finding is more remarkable in case of exclusion of financial brokerage, investment and
holding companies in 2011. Finally, the obtained results in Table 4 indicate that intensity and direction of
relationship of the ratio profitability from negative in 2009 at increasing rate has turned into positive relationship
in the next years, in the sense that during the under study years, in each year relative to previous year, greater and
more positive attitude towards profitability ratio has been observed in the ranking and in fact, investors have
shown increasingly greater and more positive attitude to share reading and their investment in each year
compared to previous year and this trend has been intensified by exclusion of financial brokerage, investment
and holding companies.
4. Discussion
Firms ranking, while promoting competition and market efficiency, is a useful guide for investors and market
operators. The results obtained from study of correlation between the ranks made by the financial ratios using
AHP-TOPSIS combined approach and those by indicators of the stock exchange regarding top-50 listed
companies in the years 2009- 2011 indicate that in 2009, there is an inverse relationship (with and without
presence of financial brokerage, investment and holding companies) between these two types of ranking, in the
sense that contrary to our expectation, the higher ranked companies on the stock exchange, were not ranked as
much higher based on the financial ratios of 2009. This relationship in the years 2009 and 2010 with an
increasing correlation coefficient has turned into a direct correlation. It means that a doubled attention has been
paid to financial ratios in ranking and investment decisions. In addition, the companies other than financial
brokerage; investment and holding have shown more convergence in these two types of ranking. In general, it
can be said that there is no significant association between ranking of the top-50 TSE listed companies based on
financial ratios in AHP-TOPSIS combined approach and their ranking based on the indicators of the stock
exchange and the presented items in firms' financial statements are not regarded an adequate approximation for
their estimation of their excellence.
131
Table 4. Correlations between the rankings of TSE and AHP-TOPSIS approach based on financial ratios
2009 2010 2011
Relationship Relationship Relationship
Hypothesis Correlation Correlation Correlation
intensity & intensity & intensity &
coefficient coefficient coefficient
direction direction direction
Negative & Positive & Positive &
A -0.306 0.116 0.224
financial relatively weak very weak weak
ratio Negative & Positive & Positive &
B -0.171 0.145 0.358
weak very weak moderate
Positive & very Positive & Positive &
A 0.036 0.199 0.199
liquidity weak weak weak
ratio Positive & very Positive & Positive &
B 0.027 0.262 0.149
weak moderate weak
Negative & Negative & Negative &
A -0.376 -0.250 -0.140
operation relatively weak weak very weak
ratio Negative & Positive & Positive &
B -0.256 0.073 -0.145
weak very weak very weak
Negative & Positive & Positive&
A -0.054 0.018 0.177
leverage very weak very weak weak
ratio Negative & Negative & Positive &
B -0.101 -0.010 0.400
very weak very weak moderate
Negative & Positive & Positive &
A -0.114 0.203 0.255
profitability very weak weak moderate
ratio Negative & Positive & Positive
B -0.113 0.090 0.458
very weak very weak &stronge
A: With presence of financial brokers, investment and holding.
B: Without financial brokers, investment & holding.
References
Babic, Z., & Plazibat, N. (1998).Ranking of enterprises based on multicriterialanalysis. International journal of
production economics, 56-57, 29-35. http://dx.doi.org/10.1016/S0925-5273(97)00133-3
Conner, M. C. (1973). On the Usefulness of Financial Ratios to Investors. Journal of Accounting Review, 48(2),
551-556. http://dx.doi.org/10.1111/j.1468-5957.1987.tb00108.x
Hassanzadeh, R., Nezhadirani, F., & Lootfelahihagi, M. (2010). The comparative investigation of financial
effects on decision-Making of bank manager. Journal of Beyond Management, 3(11), 185-211.
Johnson, R., & Soenen, L. (2003). Indicators of Successful Companies. European Management Journal, 21(3),
364-369. http://dx.doi.org/10.1016/S0263-2373(03)00050-1
Knif, J., Hogholm, K., & Miranda, F. G. (2000). Ranked market information as a stock return indicator. The
Finnish Journal of Business Economics, 2, 233-244.
Lev, B., & Thiagrajan, S. (1993). Fundamental Information Analysis. Journal of Accounting Research, 31(2),
Li, H., & Sun, J. (2008). Ranking-order case-based reasoning for financial distress prediction. Knowledge-Based
Systems,21, 868–878. http://dx.doi.org/10.1016/j.knosys.2008.03.047
Michou, M. (2007). Investing: The Use of Financial Statements to Separate Winners from Losers. Business
School of Edinburgh, University of Edinburgh, United Kingdom.
Mohanram, P. (2005). Separating Winners from Losers among Low Book-to-Market Stocks using Financial
Statement Analysis. Review of Accounting Studies, 10, 133-170.
O'Hara, T. H., & Vetere, J. (1993). An analysis of the Standard and Poor's Stock Appreciation Ranking System.
Journal of Economics and Business, 45(2), 179-182. http://dx.doi.org/10.1016/0148-6195(93)90031-I
Omran, M., & Ragab, A. (2004). Linear versus non-Linear Relationship between Financial Ratios and Stock
132
Return: An Empirical Evidence from Egyptian Firms. Review of Accounting and finance, 3(2), 84-102.
http://dx.doi.org/10.1108/eb043404
Piotroski, J. I. (2000). Value Investing: the Use of Historical Financial Statement Information to Separate
Winners from Losers. Journal of Accounting Research, 38, 1-41.
Polonchek, J., & Krehbiel, T. (1994). Price and Volume Effects Associated with Changes in the Dow Jones
Averages. The Quarterly Review of Economics and Finance, 34(4), 305-316.
http://dx.doi.org/10.1016/1062-9769(94)90016-7
133
ISSN 1916-971X E-ISSN 1916-9728
Migrant Workers' Remittances and External Trade Balance in

Sub-Sahara African Countries
Henry Okodua1 & Wumi Kolawole Olayiwola1
1
Department of Economics, Covenant University, Ota, Ogun State, Nigeria
Correspondence: Henry Okodua, Department of Economics, Covenant University, Ota, Ogun State, Nigeria.
E-mail: [email protected]
Received: November 28, 2012 Accepted: December 31, 2012 Online Published: February 22, 2013
Abstract
Workers’ remittances represent a major source of private external finance for many developing countries.
Moreover, these flows which have been on the upward trend in recent times are now widely regarded as
important financial flows to many developing countries that receive them in large quantity. One major concern
about remittances is that in countries receiving significant flows, the local currencies could appreciate artificially
due to over-valuation; this might in turn be harmful to the overall trade balance and long-run economic growth
of the receiving economies.This study investigates the possibility of this phenomenon in some selected
Sub-Saharan African countries. The hypothesized link between workers’ remittances and external trade balance
was specified in a linear dynamic panel data model and estimated using the system Generalized Method of
Moments.A major finding shows that remittance inflows have a contemporaneous negative but statistically
insignificant impact on external trade balance across the sampled countries. This result suggests that remittance
flows may not be helpful in promoting the goal of maintaining a sustainable external trade balance in the
selected SSA economies.
Keywords: remittances, trade balance, sub-saharanafrica, dynamic panel data model, system generalized method
of moments
1. Introduction
Workers’ remittances flow unarguably has been on the upward trend in recent times and these flows represent a
major source of private external finance for many developing countries. Lartey, Mandelman and Acosta (2010)
believe that remittances are becoming increasingly important as a source of foreign income in terms of both
magnitude and growth rate, exceeding the inflow of foreign aid and private capital in many countries.
Remittances currently represent about one-third of total financial flows to the developing world. Remittances
also double as the main transmitter of migration’s development benefits to migrants sending economies. The
literature on remittances is largely conclusive on the stable nature of these flows and their potential to
significantly contribute to the poverty reduction process by enhancing the living standards of the beneficiaries.
However, there is still a running debate on the possibility of the unpleasant experiences of a real exchange rate
appreciation and a loss of competitiveness in the tradable sector of many small developing economies that
receive significant inflows.
Official data on remittances inflow to Sub-Saharan Africa (SSA) reveal that the flow of remittances to the region
has been far more stable than official aid flows and foreign direct investment (FDI). Besides, remittances flow is
so resilient that they do not significantly decline even in conditions of instability and poor governance. Hence,
remittance flows represent one of the least volatile sources of foreign exchange earnings. Workers’ remittances
flow has steadily increased since the mid 1980s. Officially recorded remittances were an estimated US$206
billion in 2006, compared to US$19.6 billion in 1985 (World Development Indicators, 2006). Remittances have
been the second most important source of external finance for developing countries, being twice the size of
Official Development Aid (ODA) and almost as large as FDI. World Bank (2009) reports that recorded
remittances to developing countries in 2008 were estimated to have reached US$305 billion. This is equivalent to
nearly two percent of aggregate developing country gross domestic product (GDP) and well over half of
estimated FDI inflows (US$490 billion). The 2008 estimated remittances to developing countries are over twice
as large as ODA of US$119 billion received by developing countries.
134
The importance of remittances for some countries in the SSA region can be best illustrated by expressing them
as a ratio to GDP, while in others, the absolute total of per-capita value of remittances flows are more revealing.
When considered as a share of GDP, workers’ remittances can in fact be conveniently regarded as a vital source
of finance for many developing countries. These flows contribute to the poverty reduction process by enhancing
the living standards of the beneficiaries.Workers’ remittances can also contribute to the poverty reduction
process through the multiplier effects of flows which create additional demand, employment and income. Page
and Adams (2003) found that a 10% increase of remittances per capita leads to a decline of the poverty head
count by 3.5%, due to multiplier effects on GDP growth.
Two major forces are expected to ensure the growth and sustenance of remittance flows: Globalization and the
aging populations (Olayiwola et al., 2008). Globalization and the aging of developed economy populations will
ensure that demand for migrant workers remains robust for years to come. Consequently, the volume of
remittances will most likely continue to grow, since migrants will continue to support the elderly and other
dependants in their countries of origin. However, challenges remain in determining how best to channel the flow
of remittances through formal financial institutions to promote economic growth and development in sending
countries (Chami et al. 2008).
The impact of remittances on the real exchange rate and export competitiveness, and the Dutch disease effect, is
one important area of the debate. In countries receiving significant remittances flow, the local currencies could
appreciate, which might be harmful to their long-run economic growth. Remittances may exhibit the Dutch
disease effects on the competitiveness of the tradable sector of receiving economies if these flows cause an
over-valuation of the local currencies. Chami et al. (2003) noted that remittances inflow may be large in some
countries when compared to the size of the economy and supply constraints may also pose a significant
hindrance to the expansion of the non-tradable sector. In addition, a significant portion of remittances receipt
may actually be spent on domestic goods; in such cases, the Dutch disease phenomenon arises and policymakers
will need to be mindful of this possibility. Moreover, significant remittances receipts may reduce the labour
supply or labour market participation of recipients in which case remittances could lead adverse economic
developmental outcomes in the SSA region.
As a consequence, remittances flow via the over-valuation of local currencies of receiving economies may cause a
deterioration of the external trade. Since remittances increase purchasing power in general within the receiving
economy, they also promote domestic demand and preferences here may actually be in favour of imported goods
and services. This fact often results in the deterioration of the external trade balance of remittance receiving
economies and could bring about a worsening of the overall balance of payments (BOP) position of these
economies. The big question at this point is; to what extent does workers’ remittance inflows impact on foreign
trade balance in SSA countries? This study empirically examines this question by investigating and characterizing
the relationship between trade balance and remittance flows in some selected remittance receiving SSAcountries.
The remainder of the paper is divided into sections as follows: section two deals with literature review while
section three is concerned with data and methodology and it comprises the empirical model, definition of
variables, data sources, and model estimation technique. Section four deals with empirical results and
discussions and policy implications of findings. Section five provides some general concluding remarks.
The impact of remittances on the real exchange rate and export competitiveness, and the Dutch disease effect,
represent an area with inconclusive debate within the remittance literature. In the case of countries that receive
significant remittances, the supply surge in foreign currencies could constitute pressure on the local currencies,
which as a natural response could appreciate. And this might be very harmful to the long-run economic growth
objective of the receiving economy (a Dutch disease effect). Remittances inflow may be so significant in volume
as to result in an artificial appreciation of the real exchange rate of the receiving economy. In this case,
remittances inflow may reduce the foreign trade competitiveness which in turn, weakens the real external
balance and by implication, the current account balance position of the recipient economy. Consequently,
remittances in this context halt the receiving country’s trade balance via a reduction in exports of traded goods.
Overall, the occurrence of remittance receipts in volumes that reduce the foreign trade competitiveness of the
receiving economy, will adversely impact on the trade balance as well as the economic development of the
receiving economy.
The above concerns that massive inflow of remittances to small developing economies could generate a resource
allocation from the tradable to the non-tradable sector have been raised in a number of studies (see: McCormick
and Wahba, 2000; Amuedo-Dorantes and Pozo, 2004; Lopez et al., 2007; Acosta et al., 2009; etc.). Rodrik (2008)
135
provides evidence that real exchange rate overvaluation undermines long-term economic growth, particularly for
developing countries, andin those countries tradable goods production suffers disproportionately from weak
institutions and market failures. This underscores the importance of the implications of remittances for real
exchange rate movements.Acosta et al. (2009) find that in addition to the usual nominal exchange rate channel,
remittances result in a shrinkage of, and resource re-allocations away from, the tradable sector through (i)
increasing prices in the non-tradable sector, and (ii) reducing the labour supply to, and thereby increasing the
production costs of, the otherwise labour-intensive non-tradable sector. Studies in Latin America
(Amuedo-Dorantes and Pozo, 2004) and Cape Verde (Bourdet and Falck, 2006) provide additional evidences in
support of a Dutch disease effect of remittances on the competitiveness of the tradable sector of the receiving
economies.
Chami, et al. (2003), argue that in countries where remittances inflows are large compared to the size of the
economy, where supply constraints are a significant hindrance to the expansion of the non-tradables sector, and
where a significant portion of remittances are spent on domestic goods, policymakers will need to be alert to the
possibility of a Dutch disease phenomenon. Moreover, remittances may reduce the labour supply or labour
market participation of recipients. If these negative factors dominate, remittances could be detrimental to
economic development in SSA.
López et al. (2008) provide a very useful discussion on the channel through which remittances, real exchange
rate, and Dutch disease may be linked. According to them, workers’ remittances can be viewed as a capital
inflow, and a surge in inflows resulting from a positive remittances income shocks produce some extra spending
on both tradables and non-tradables. However, there will be a relative price change between tradables and
non-tradables in favour of non-tradables and this makes production of the latter more profitable. Consequently,
the price shift and resulting resource re-allocation erodes the competitiveness of export-oriented sectors and at
the same time, hurt import-competing sectors. The final result of this real exchange rate appreciation is normally
increased import flows and lower export sales.
On the positive side however, some studies have evidences that sizeable remittances inflows do not always
depress the external trade balance of the receiving economy. Bouhga-Hagbe (2004) found that in the case of
Morocco, remittances receipts almost covered the trade deficit and actually explains the observed surpluses of the
external current account, and the overall BOP.Meanwhile, the BOP surpluses have in turn contributed to the
improvement of Morocco’s external position as evidenced in the accumulation of foreign exchange reserves,
which now cover the external public debt.Rajan and Subramanian (2005) believe that remittances may in fact be
self-correcting as an overvalued currency deters remittances, and hence Dutch disease effects are not sustained.
This study contributes to the literature by extending the debate to the effect of remittance flows on external trade
via real exchange rate appreciation within the SSA context. The study also contributes by accounting for the
possibility of dynamic endogeneity in the hypothesized relationship and this has the overall effect of ensuring
that the quality of results obtained in the study are much more reliable.
3. Data and Methodology
The impact of remittance flows on external trade balance using a linear dynamic panel data model is empirically
analysed in this study.Specifically, this study evaluates whether remittance flows significantly impact on
external trade balance in recipient economies. The underlying assumption here is that remittances may be
positively correlated with real exchange rate appreciation in the recipient economy and thereby hindering the
external foreign trade competitiveness of the recipient economy. Significant inflows of remittances may result in
the artificial appreciation of the real exchange rate of the receiving economy and consequently penalize the
traded goods sector (since its exportables now become more expensive and less competitive) in the other
economies.
3.1 The Empirical Model
In modeling the impact of workers' remittances on external trade balance of the receiving economies of SSA, the
work of Obstfeld and Rogoff (1996) is considered a major motivation because they argue that a positive transfer
of resources to a country hurts its competitiveness in world markets by reducing the range of goods it exports.
The reduction in competitiveness takes place because the transfer reduces the country's real exchange rate. The
specified empirical model is made up of a dynamic panel model. This equation is an attempt to verify
empirically whether remittances inflow brings about a significant variation in the external trade balance of the
receiving economies.
, , , , , 1
136
The explicit form of the equation (1) is provided below:

2
1, 2, 3, … ,30 1, 2, 3, … ,10
These variables are expected on a priori grounds to be signed as follows:
, , , , , 0 , 0
Without any loss of generality, equations (2) may now be re-written in matrix notations to obtain the following
linear dynamic panel data model:
. β β (3)
is a vector of strictly exogenous covariates (ones dependent on neither current nor past ); and may be
written as:
, , ′
is a vector of predetermined covariates (which may include the lag of ETB) and endogenous covariates, all of
which may be correlated with the . These include:
, , ′
are vectors of parameters to be estimated.
is the usual error component decomposition of the error term;
are unobserved individual-specific effects;
are the observation-specific (idiosyncratic) errors;
comprises variables that are defined as follows:
on the other hand comprises variables that are defined as follows:

one period lag of the dependent variable,
REER = real effective exchange rate
WR = worker’s remittances receipt
Equation (3) may be rewritten in first differenced notations in order to get a consistent estimate of as N →∞
with T fixed as follows:
. . . β . β D. . 4
The unobserved individual-level effects, has been eliminated from the differenced equation (4) because it does
not vary over time. The Ds are the first difference operators.
3.2 Definition of Variables and Data Sources
Data for all variables in this study were sourced from the World Bank, Africa development indicators and the
International Monetary Fund (IMF), World Economic Outlook (2005, 2012). The study employs data on the
variables and covering a period of ten years (2002-2011). The choice of this period is explained by the
availability of data across the selected countries as well as the fact of a dramatic rise in recorded remittance
flows to the region over this period. The definition of variables employed in the study and data sources are
presented in the Table 1.
137
Table 1. Definition of variables

Variable Definition
Workers’ remittances (WR) Workers’ remittances received comprise of current transfers by migrant workers. It is
measured as a ratio of GDP
Inflation rate (INF) This is the annual percentage change in consumer price index (CPI)
Growth rate of Gross Domestic Product (GDPgr) This is the annual percentage change in the value of the real GDP
Trade Openness(OPEN) Degree of trade openness measured by the volume of trade (export plus imports)
divided by the GDP for a given year.
External Trade Balance (ETB) ETB is the difference between the monetary value of exports and imports of output in
an economy over a given year
Real Effective Exchange Rate (REER) This is the nominal exchange rate that has accounted for the inflation differentials in
each country over time.
The study is limited to the thirty SSA countries (Note 2) that reported inward remittances receipts for the period-
2002 and 2011. Remittance flows isrestricted to inter-household unilateral and unrequited transfer of cash
earnings, meaning that such transfer is void of any form of quid pro quo terms, across national boundaries only.
The implication is that remittances in forms of material transfers by migrant workers to their home countries,
compensation of employees, or unrequited inter-household cash transfers within each economy under
investigation, are not covered in this study. It is important to clarify here that the study is restricted to the
macroeconomic impact of remittances on the receiving economies and not on their microeconomic impact.
3.3 Model Estimation Technique
There are two major and important complications arising from efforts to estimate the dynamic panel data
regression model referred to in equation (3) using macroeconomic panel data: first, the presence of endogenous
and/or predetermined covariates, and second, the small time-series and cross-sectional dimensions of the typical
panel data set. The dynamic panel data regression model is in fact further characterized by some sources of
persistence over time. There is the problem of autocorrelation which is due to the presence of a lagged dependent
variable among the regressors and the other is the problem of heteroskedasticity.
Working within the context of remittance flows, current country remittance realizations will affect future trade
performance via its effects on real exchange rate and this may, in turn, affect future country remittance
realizations. Thus, giving rise to what may be termed as “dynamic endogeneity”. The argument here centers on
the fact that cross-sectional variation in observed country economic structures is driven by both unobservable
heterogeneity and the country’s history. As such, any attempt to explain the role of remittance flows or its effect
on trade performance of selected countries that does not recognize these sources of endogeneity may be biased.
The problem of endogeneity that is often associated with the use of panel data analysis are thus resolved in this
study by the choice of the System Generalized Method of Moments (GMM) Estimator to estimate the relation
between remittance flows and external trade balance. This methodology not only eliminates any bias that may
arise from ignoring dynamic endogeneity, but also provides theoretically based and powerful instruments that
accounts for simultaneity while eliminating any unobservable heterogeneity. Dynamic panel estimation is most
useful in situations where some unobservable factor affects both the dependent variable and the explanatory
variables, and some explanatory variables are strongly related to past values of the dependent variable. This is
likely to be the case in regressions of remittance flows on external trade balance. This is because remittance
flows tend to exert a strong, immediate and persistent effect on trade performance.
In the presence of heteroskedasticity and serial correlation, the two-step System–GMM uses a consistent
estimate of the weighting matrix, taking the residuals from the one-step estimate (Davidson and MacKinnon,
2004). Bun and Windmeijer (2009) emphasized that the good performance of the system GMM estimator
relative to the difference GMM estimator in terms of finite sample bias and root mean square error, has made it
the estimator of choice in many applied panel data settings. In multivariate dynamic panel models, the System–
GMM estimator is also shown to perform better than the Differenced–GMM when series are persistent and there
is a dramatic reduction in the finite sample bias due to the exploitation of additional moment conditions
(Blundell et al. 2000).
Bond et al. (2001) provide a useful insight in the GMM estimation of dynamic panel data models, arguing that
the pooled ordinary least square (OLS) and the least squares dummy variable (LSDV) estimators should be
considered respectively as the upper and lower bound. Variations outside these boundaries may therefore be
138
considered as a sign that the estimates are biased probably owing to a weak instrument problem. Thus, if this is
the case, the use of System–GMM is highly recommended and its estimates should lie between OLS and LSDV.
In view of the obvious strengths of the Blundell and Bond (1998), extended version of the GMM estimator, (also
known as system GMM estimator) in overcoming complications that may arise from efforts to estimate the usual
linear dynamic panel data models, this estimator was considered appropriate and applied to estimate the
specified model for this study.
4. Empirical Results and Discussions
Summary of the dynamic panel data model estimation results for equation (3) are presented in Table 2. The
system GMM estimator is categorized into the one-step and two-step options, these are reported in columns 2
and 3 respectively. The OLS and the LSDV results are reported in columns 1 and 4 respectively. Apart from
providing some additional robustness check, the results in columns 1 and 4 will also provide a guide based on
the position of Bond et al. (2001) that suggest the pooled OLS and the LSDV estimators should be considered
respectively as the upper and lower bound for the system GMM coefficients. With this guide in place, it will be
easy to tell when each coefficient estimate is either downward or upward biased.
All estimates in columns 2 and 3 are robust to heteroskedasticity or autocorrelation. The related endogenous and
predetermined variables on the right hand side of this specification include the REER, WR and lagged ETB
respectively. To control for endogeneity of these variables that appear as regressors, internal instruments are
utilized; and these include the lagged levels of the standard differenced equation (equation 4) and lagged
differences of the levels equation (equation iii). Correlation coefficients (Note 1) between residuals from the
base regression and independent variables were computed as an additional check of potential endogeneity
problems. An investigation of these coefficients of correlations suggests that none of the independent variables is
highly correlated with predicted residuals.
Table 2. Estimated empirical results

Dependent Variable: ETB
Pooled OLS SYSTEM-GMM LSDV
One-Step Two-Step
Instrument Weight collapsed collapsed
Regressors (1) (2) (3) (4)
ETB(-1) 0.978* 0.860* 0.879* 0.455*
(47.38) (11.27) (14.55) (5.97)
WR -1.42e-10 -9.81e-11 -1.88e-10 -4.88e-10
(-0.75) (-0.35) (-0.72) (-1.47)
REER 0.0174 0.0578*** 0.0626*** 0.0198
(1.27) (1.82) (1.99) (0.40)
GDPgr 0.200*** 0.271*** 0.267* 0.108
(1.76) (1.95) (3.48) (0.86)
INF -0.0607 -0.0580 -0.0366 -0.0658
(-1.21) (-1.09) (-0.85) (-1.22)
OPEN 0.857 5.813 4.219*** 3.940
(0.43) (1.24) (1.77) (0.65)
Constant -3.073*** -10.19* -10.12** 20.61**
(-1.70) (-3.03) (-2.64) (2.15)
Time Dummy No Yes Yes Yes
Group Dummy No No No Yes
Observations 270 270 270 270
No. of countries 30 30 30 30
Instrument count - 27 27 -
F-stat (Wald χ2 ) - 42.79 65.55 -
F-stat (p-value) [0.0000] [0.0000] [0.0000] -
AR(2) - [0.360] [0.437] -
AR(3) - [0.477] [0.466] -
Sargan Test (OIR) - [0.504] [0.368] -
Hansen Test (OIR) - [0.729] [0.810] -
Notes: Robust standard errors are with Windmeijer (2005) finite-sample correction for the two-step covariance matrix.
t statistics are reported in curly brackets below each coefficient estimate. P-values are reported in square brackets. * indicates significant at 1
percent level.** indicates significant at 5 percent level.*** indicates significant at 10 percent level
139
Some specification tests are examined as a starting point to determine the reliability of coefficient estimates
reported in Table 2. The working assumption here is that the idiosyncratic errors in the system GMM
estimators are serially uncorrelated for consistent estimations. The assumption that the full disturbance is
autocorrelated because it contains fixed effects is also made hence; the system GMM estimator is the most
appropriate tool to eliminate this potential source of trouble. The Arellano-Bond tests for autocorrelation are
reported as AR(2) and AR(3) in the lower portion of table 2.The p-values are greater than 0.05 in the one-step
and two-step system GMM estimates their values indicate that there is no evidence of serial correlation at the five
percent level of significance. Given these results, the estimates can be regarded as consistent and the instruments
are valid.
The test of over-identifying restrictions of whether the instruments, as a group, appear exogenous is implemented
by the Sargan and Hansen J tests. Here the Hansen J statistic, which is the minimized value of the two-step
system GMM criterion function, and is robust to autocorrelation, is of tremendous importance.Only the
respective p-values are reported for this test results in the lower part of Table 2. Of course, the null hypothesis that
the population moment condition is valid is not rejected if 0.05.In columns 2 and 3, the summary statistics
indicate that the system dynamic panel model of the selected 30 SSA countries has 27 instruments and 16
parameters in both the one-step and two-step system GMM options. This represents a total of 11 over-identifying
restriction in each case. Thus, the Hansen–J statistic does not reject the Over-Identifying Restrictions (OIR)
hence, the instrument set can be considered valid. The F-statistic which measures the overall significance of all
regressors in the estimated model is satisfactorily significant at the one percent level. This is indicative of the
fact that all the exogenous variables, in each estimated result, jointly explained significantly, the systematic
variations in external trade balance across the sampled SSA countries over the study period.
A look at the control variables reveals that many of the coefficient estimates are satisfactorily consistent with
theoretical expectations. The Blundell–Bond robust estimates of lagged external trade balance are positively
signed. As can be seen in columns 2 and 3 of Table 2, past realizations of external trade balance positively
impact on its contemporaneous levels. These external trade balance dynamics are significant at the 1 percent
level in these specifications. In specific terms, a 100 percent increase in external trade balance dynamics will
explain about 87.9 percent of the increase in the contemporaneous realizations of external trade balance within
the sampled SSA countries. This of course is when the two-step collapsed instruments option is considered.
External trade balance dynamics suggest here that external trade balance itself has a way of feeding on its past
realizations within the study group.
Surprisingly, the workers’ remittance variable is negatively signed and statistically insignificant in both the
one-step and two-step system GMM robust estimates. This clearly suggests that although workers’ remittance
flows to SSA exhibit some potential to depress external trade balance across countries in the sample group for
this study, this tendency does not constitute a source of worry for now as it is statistically insignificant.A
possible explanation for this observed relationship may be found in the trade patterns of most SSA countries.
Many of these economies are characterized by high export concentration index owing to exports of some few
primary products. This fact often severely limits SSA capacity to diversify and expand exports in response to
changing demand conditions in the foreign trade sector. Imports by SSA economies on the other hand, comprise
mainly processed or manufactured goods and high-tech services. A sustained significant increase in remittances
receipts by the SSA economies will most likely depress the trade balance of the regional economies on account
of consumption pattern that frequently reveals preferences for imported goods and services.
Real effective exchange rate (REER) variable is positively signed and is statistically significant at the 10 percent
level when the one-step and two-step system GMM are considered. In more definitive terms, a 100 percent
increase in REER produces about 6.26 percent contemporaneous increase in external trade balance across the
selected economies. Given this finding, it can be remarked here that real exchange rate depreciation in the
selected SSA economies helps improve external trade balance across these economies. This is both expected and
healthy for these SSA economies.
Interestingly, growth rate of GDP (GDPgr) is also found to be highly statistically significant at the one percent
level and positively related to external trade balance. Contemporaneously, a 100 percent increase in GDPgr leads
to about 26.7 percent improvement in external trade balance of the selected economies. This agrees with
theoretical expectation and is indicative of export oriented growth economies across the selected SSA economies.
Trade openness (OPEN) variable is also positively signed and statistically significant at the ten percent level.
This result suggests that the greater the degree of trade openness the greater the external trade balance for these
economies.
140
5. Policy Implications of Findings

Two major policy issues naturally arise from the empirical findings in this study. First, a steady increase in
workers’ remittances inflow to SSA economies is observed to depress external trade balance over time and
across the selected countries within the sub-region. In terms of quantum, this relationship is statistically
insignificant; however, appropriate policies are needed to ensure that these financial flows are properly
channeled into economically productive uses so as to begin to contribute positively to the external sector of the
economies. Second, the positive sign and significant nature of the REER variable in explaining variations in the
external trade balance also require some further comments. Increase in this variable indicates a depreciation of
the local currency so the positive sign of REER is in line with theoretical expectation. This means that
appropriate foreign exchange policies (policies in support of weaker national currencies in order to promote
external trade) may be adopted across the SSA countries to maintain a healthy trade balance for the region.
6. Conclusion
Given the major findings in this study which clearly highlight the negative but insignificant role of workers’
remittances on external trade balance in some selected SSA economies; it is instructive to remark here in
conclusion, that policy incentives designed to attract more workers’ remittances flow to SSA countries may not
be able to help in promoting the goal of maintaining a healthy external trade balance in the selected SSA
economies for now. There is therefore the need for adoption of policy options that encourages the economically
productive use of remittance inflows into SSA countries. Since remittances are regarded as desirable financial
flows in general, the sustenance of these flows and their productive use in the SSA region, demand a central role
for governments and monetary authorities in terms of the provision of relevant policy direction.
Every related policy measure therefore should be targeted towards the reorientation of senders and recipients of
remittances so as to ensure that these flows are regularly engaged productively. Moreover, it will not be out of
place if policy incentives are given a sectoral focus such that remittances are used productively in sectors that are
external trade oriented. This will hopefully allow for a sustainably stable external trade balance in the SSA
region.
Acknowledgements
This paper is a part of the first author’s PhD Dissertation now revised in article form for journal publication.
Professor Donald N. Ike and the second author supervised the PhD Thesis and their contributions in this regard
are hereby acknowledged. The management of Covenant University, Nigeria is also acknowledged here for
funding the entire PhD programme under the staff development scholarship scheme.
References
Acosta, P. A., Lartey, K. K., & Mandelman, F. S. (2009). Remittances and Dutch Disease. Journal of
International Economics, 79, 102-116. http://dx.doi.org/10.1016/j.jinteco.2009.06.007
Adams, R., & Page, J. (2003). International Migration, Remittances, and Poverty in Developing Countries.
World Bank Policy Research Working Paper 3179, Washington D. C. World Bank.
Amuedo-Dorantes, C., & Pozo, S. (2004). Workers' Remittances and the Real Exchange Rate: A Paradox of
Gifts. World Development, 32(8), 1407-1417. http://dx.doi.org/10.1016/j.worlddev.2004.02.004
Blundell, R. Bond, S., & Windmeijer, F. (2000). Estimation in Dynamic Panel Data Models: Improving on the
Performance of the Standard GMM Estimator. Institute for Fiscal Studies. IFS Working Papers W00/12
Blundell, R., & Bond, S. (1998). Initial Conditions and Moment Restrictions in Dynamic Panel Data Models.
Journal of Econometrics, 87, 11-143. http://dx.doi.org/10.1016/S0304-4076(98)00009-8
Bond, S., Hoeffler, A., & Temple, J. (2001). GMM Estimation of Empirical Growth Models. Economics Group,
Nuffield College, University of Oxford. Economics Papers, 2001-W21.
Bouhga-Hagbe, J. (2004). A Theory of Workers' Remittances with an Application to Morocco. IMF Working
Paper 04/194, Washington D. C., International Monetary Fund.
Bourdet, Y., & Falck, H. (2006). Emigrants' Remittances and Dutch Disease in Cape Verde. International
Economic Journal, forthcoming. http://dx.doi.org/10.1080/10168730600879323
Brown, P., & Richard, C. (1994). Migrants' Remittances, Savings and Investment in the South Pacific.
International Labor Review, 133(3), 357-67.
Bun, M., & Windmeijer, F. (2009). The Weak Instrument Problem of the System GMM Estimator in Dynamic
Panel Data Models. Tinbergen Institute Discussion Paper, TI 2009-086/4.
141
Chami, R. A., Fullenkamp, C., & Jahjah, S. (2003). Are Immigrant Remittances Flows a Source of Capital for
Development? IMF Working Paper 03/189. Washington D. C., International Monetary Fund.
Chami, R. A., Barajas, T. C., Fullenkamp, C., Gapen, M., & Montiel, P. (2008). Macroeconomic Consequences
of Remittances. Washington DC, International Monetary Fund, IMF Occasional Paper 259.OCCA.
Davidson, R., & Mackinnon, G. J. (2004). Econometric Theory and Methods. New York: Oxford University
Press.
Lartey, E. K. K., Mandelman, F. S., & Acosta, P. A. (2010). Remittances, Exchange Rate Regimes and the Dutch
Disease: A Panel Data Analysis. Washington, D. C., World Bank.
Lopez, H., Molina, L., & Bussolo, M. (2007). Remittances and the real exchange rate. World Bank Policy
Research Working Paper 4213. Washington, D.C. World Bank
López, J. H., Molina, L., & Bussolo, M. (2008). Remittances, the Real Exchange Rate, and the Dutch Disease
Phenomenon. In Pablo F. & J. H. López (Eds.), Remittances and Development Lessons from Latin America.
Washington D. C. World Bank.
McCormick, B., & Wahba, J. (2000). Overseas Employment and Remittances to a Dual Econ-omy. Economic
Journal, 110, 509-534. http://dx.doi.org/10.1111/1468-0297.00535
Obstfeld, M., & Rogoff, K. (1996). Foundations of International Macroeconomics. Cambridge, M A: MIT Press.
Olayiwola, W. K., Oyinloye, O., & Akinrinola, L. (2008). An Empirical Assessment of Old-Age Support in
SSA- Evidence from Ghana. Final Report Submitted to African Economic Research Consortium (AERC),
Nairobi Kenya.
Rajan, R., & Subramanian, A. (2005). What Undermines Aid's Impact on Growth? IMF Working Paper 05/126,
Washington D. C., International Monetary Fund.
Rodrik, D. (2008). The Real Exchange Rate and Economic Growth. Brookings Papers on Eco-nomic Activity, 2,
365-412.
Roodman, D. (2009). How to do xtabond2: An introduction to difference and system GMM in Stata. Stata
Journal, 9(1), 86-136.
Windmeijer, A. F. (2005). Finite Sample Correction for the variance of Linear Efficient Two-Step GMM
Estimators. Journal of Econometrics, 126(1), 25-51. http://dx.doi.org/10.1016/j.jeconom.2004.02.005
World Bank. (2006). World Development Indicators. Washington, D. C. World Bank. Retrieved from
http://siteresources.worldbank.org/INTSTATINAFR/Resources/ADI_2006_text.pdf (accessed April 17,
2010)
Notes
Note 1. Details of these results are not reported here but can be made available on request.
Note 2. These countries are: Benin, Botswana,Burkina Faso, Cameroon, Cape Verde, Congo, Côte d'Ivoire,
Ethiopia, Gabon, Gambia, Ghana, Guinea, Guinea-Bissau, Kenya, Lesotho, Madagascar, Malawi,
Mali,Mozambique, Namibia, Niger, Nigeria, Rwanda, Senegal, Seychelles, Sierra Leone, South Africa, Swaziland,
Tanzania, andTogo.
142
ISSN 1916-971X E-ISSN 1916-9728
The Profitability of Banking Sector in Republic of Macedonia

Meri Boshkoska1
1
Faculty of administration and management of information systems in Bitola, Republic of Macedonia
Correspondence: Meri Boshkoska, Faculty of administration and management of information systems in Bitola,
Republic of Macedonia. E-mail: [email protected]
Abstract
The banking sector has a dominant role in the process of mediating and dispersing the finances of the economical
subject with suffice towards the ones with a deficit. Although it is well known fact that the banking sector plays the
major role in the enhancing and developing of a country’s economy, it is evident that this sector is under big
influence of the changes within the financial environment-regarding the situation of the capital market on a global
scale. If the banking sector functions efficiently, give loans to support the enterprises, it leads to bigger
consumption-meaning that for the population of the country the banking sector is highly valued and trusted.
The paper examines the structure and the profitability of the banking sector in Republic of Macedonia comparing
with the banking sectors of some countries in the region.
Keywords: bank’s profitability, ROA, ROE, net-interest margin, non-performing loans, Basel III Accord
1. Introduction
It is essential for a country to have well structured and profitable banking sector in order to have more competitive
and successful banking sector. For the existence of this kind of sector of particular importance is the range of profit
that is realized by the banks. The stability of the financial system depends on the profitability of the banking sector.
General problem of the banking sector in the Republic of Macedonia is the downward trend of its profitability in
the past few years.
The main objective of this paper is to answer the following questions:
 Who are the most important reasons for the negative trend in the Macedonian banking sector's profitability?
 What needs to be changed in this sector to improve this situation?
Further down in this paper, through comparative analysis (with the banking sectors of other countries in the region)
and research of its profitability, we will show the real situation in which at this point is the Macedonian banking
sector.
The basic hypothesis that is tested in the paper is: If the structure of the banking sector in Republic of Macedonia is
consolidates, then the trend of banks' profitability will be positive.To prove the hypothesis we used methods that
will be based on:
 Content analysis of various domestic and foreign banking system reports, banking system indicators,
publications and Financial statements of the National Bank of the Republic of Macedonia (NBRM).
 Comparative analysis.
The research paper is structured as follows. At the beginning, before we give the analysis of its profitability, it is
crucial to present the structure and competition of the Macedonian banking sector in order to draw a conclusion
whether are needed changes in the bank’s structure.
The scope of the research is covering explanation of the mentioned objective and will depend on the knowledge
that already exists in the literature and is provided by different authors that made past research in this field as well
as on the data and analysis of the financial statements that are provided by home and foreign financial institutions.
Popovska (2008) provide a complex analysis of the performance of the commercial banking sector. Her findings
showed that the basic problems in the Macedonian banking sector are: inadequate competition, high concentration
of bank capital in a small number of big banks and undeveloped structure of services for the clients.
143
Accordingg to CEA (2005) the structurre of the Macedonian bankinng sector is higghly concentraated and resultss in a
low compeetition in this sector.
s
Findings oof Davcev and Hourvouliadees (2009) show wed that the moost important ddeterminants oof bank profitab bility
are: Returnn on total Asseets, total loss onn loans and opperative expensses are significcant only in thee ROE profitab
bility.
Croatian BBanking Assocciation (2009) eexamined the reasons for thee decline in thee profitability of banks in Crroatia
compared to other counntries in the reegion. Accordiing to this anaalysis, there arre two reasonss for this declined:
short-term
m (profit and sttability of the banking systeem) and long-tterm (ability tto attract addittional capital).. The
general coonclusion is thhat low bank pprofit means thhat the econom
my remains wiithout one of tthe most impo ortant
factors forr its growth andd developmentt.
Madzova ((2011) examinned the opportuunities and chaallenges of intrroducing of Baasel III. She emmphasized that their
introductioon will result in
i an increasinng in the bankk costs. These indicate that w we can expectt a reduction in the
profitabilitty of the Maceedonian bankinng sector, but nnot a problem iin their fulfillm
ment.
The mentiioned researchh and findings represent thee basis of our research, whicch is a small contribution to the
modest ressearch that wee have on thiss topic. The paaper contributees to the existting literature by researching the
performannce of the bannking sector oof which depends the econoomic growth aand developm ment of a counntry’s
economy.
2. The Strructure of the Banking Secttor in Republlic of Macedon
nia
In this parrt, we will disccuss the data aabout the num
mber of banks functioning onn the territory of the Republlic of
Macedoniaa, and analysis is made wheether the numbber of the bannks means effiicient functionning of the ban nking
sector.
The bankinng sector of thee Republic of M Macedonia connsists of 17 privately owned bbanks and 8 deepositories and d loan
houses (saavings houses).. (Note 1) Althhough there aree fully functionnal depositoriess and loan houuses (saving houses),
still the maajor role in thee financial marrket is played bby the banks. IIn 13 of these bbanks, major sshareholders arre the
foreign ow wners.
Fiigure 1. Share of each bank iin the total asseets of the bankking system
Source: N
NBRM (2012). Report on Bannking System of the Republic of Macedonnia in the seconnd quarter of 2012.
2
Skopje, p. 9.
Most of thhe total assetss of the bankiing sector are concentrated in the three llargest banks which indicate the
existence oof a large conncentration in tthe banking syystem. (Figuree 1). This meaans that the M Macedonian ban
nking
system is iin a group of monopolistic
m coompetitive maarkets, with a rrelatively low llevel of compeetition.
The high cconcentration inn the banking sector in Maceedonia also meeans that there aare banks, poppularly called by
b the
IMF and thhe WB “pockeet banks”- meaaning that thesee are smaller bbanks that are uused as a cheaap source of fin nance
for certain groups and/orr individuals inn the country. S
Some of these ssmaller banks (banks with asssets less than 6.200
6
million deenars) (Note 2)) realized smalll-scale activities, lower commpetition, and passive operattion on the ma arket.
These bankks should seekk for a solutionn in finding strrategic investorrs or in emergiing with other banks.
144
Having into consideration the critical point mentioned above about the banking sector in Republic of Macedonia
we can offer a few suggestions that follow. Firstly, in order to stop the opening of new and operating of the existing
small noncompetitive banks, that more than often are the reason why the total profit of the banking sector is getting
lower, it is necessary to take several precaution measures. The most exceptional attention should be paid to
establishing the conditions under which a permit for opening a bank is granted-meaning, increasing of the
minimum requested ownership capital (which according to the latest Banking Law is increased from 3, 5 to 5
million Euros) (NBRM, 2006) and the authority and the reputation of the person who intends to open a bank. Next,
attention should be paid to the qualities and qualifications of the person who is meant to be the general
manager-weather that person has the capacity for proper managing of the bank - to be closely controlled especially
when it comes to giving loans with a high-risk factor-whether the loans are given to persons closely related to the
managerial staff of the bank (so called connected crediting). Furthermore, it is necessary to have liable information
on the legality of the capital that the person who wants to open a bank is giving as a deposit. Then, it is also
important to establish a strict supervision according to the international standards, to publish on time the eventual
measures for correction and enhancing of the functioning taken by the National Bank of Republic of Macedonia
(NBRM) about the banks that encounter functional problems.
3. Comparative Analysis of Macedonian Banking Sector Profitability
3.1 Macedonian Banking Sector – It’s Profitability
In order to compare the profitability of the Macedonian banking system, further in this text, first we will present
information about the profits that the Macedonian banks makes (ranked according to how big they are) and we will
make an analysis of the profitability with the banking sectors in some countries in the region.
The profitability of the banking system of a country is measured by the rate of return of assets (ROA), the rate of
return of equity (ROE) and Net interest margin in order to better measurement the bank performances.
ROA indicates how well bank management used its total amount of assets (loans and investments) to earn
income, or how much profit is created of each dollar in assets. (Center for Financial Training, 2010)
Although ROA provides useful information about bank profitability that is not what the bank owners care about
most. They are more concerned about how much bank is earning on their equity investment that is measured by the
return on equity -ROE. (Mishkin, 2012)
ROE shows how much the bank earned in comparison with the share holders’ capital (Petkovski, 2009).
Investors can compare a bank's ROA and ROE to those of other banks to see how it performed relative to the
other banks. A bank's ratios for several years can be reviewed to determine whether they have remained the same,
increased, or decreased. A bank's ratios that decrease or stay the same are cause for concern. (Center for
Financial Training, 2010)
A bank’s net interest margin is the difference between the interest it receives on its securities and loans and the
interest it pays on deposits and debt, divided by the total value of its earning assets. (Hubbard and O’Brien,
2012)
In table 1 are given standards for ROA and ROE indicators from lowest to highest. These standards will help us
to determine on what level of this scale is the Macedonian banking sector.
Table 1. Standards for ROA and ROE

RATING ROA RATING ROE
Low - Bellow 0.5% 15% - after taxation
Satisfying - 0.5% - 1% 25% - before taxation
Good - 1% - 2%
High - Above 2%
Source: Popovska, K (2008). Commercial Banking – Success and profitability. Skopje,”Institute of economics” – Skopje, p. 167.
Qin.X and Pastory.D. (2012). Commercial Banks Profitability Position: The Case of Tanzania, p.137.
In the following text will be analyzed the trend of profitability observed by groups of banks and the entire banking
sector in the country. The analysis is made by the following key indicators of the bank’s profitability: ROAA,
ROAE and Net interest rate/operational costs.
145
Table 2. R
Rate of return of
o average asseets (ROAA)
Group 9/30/2010 9/30/2011 09/30/2012*
Large bank
ks 1.0% 0.9% 0.8%
Medium-siize banks -0.5% -1.7% 0.2%
Small-size banks -2.4% -2.8% -4.8%
Banking syystem 0.5% 0.1% 0.3%
Table 3. R
Rate of return of
o average equiity (ROAE)
Group 9/30//2010 9/30/2011 30.09.20122*
Large banks 11.1%
% 9.3% 7.7%
Medium-size banks -4.2%
% -14.5% 1.5%
Small-size baanks -6.1%
% -8.7% -23.9%
Banking system 4.0%
% 1.0% 2.3%
* Quarterly inndicators are reduuced on an annual basis.
Source: NBR
RM - according daata from individuaal banks.
As shown in Tables 2 annd 3, the large bbanks have thee largest share in the financiaal results of thee complete ban
nking
wer compared to the previouus period. But, in Septemberr 2012, it is no
sector, thoough their proffitability is low otable
that the mmedium-size banks raised thheir profits coompared to thhe same periodd the previous considered years y
(Septembeer 2010, 2011)) and the small-size banks hhave losses andd their profitabbility and it is at the lowest level
compared to the other tw wo groups of banks (NBRM, 2011).
Reduced leending activityy, reduced operational efficieency, low markket share, increease in non-peerforming loans, the
higher reseervations amouunt of bad loanns are among tthe reasons forr the lowered pprofitability off small-sized banks.
This tendss to support ourr previous sugggestion for meergers or underrtaking of smaaller banks.
Table 4. N
Net interest ratee/operational ccosts
Group 9/30/2011 30.09.2012
Large banks 103.9 131.1
Medium-size b
banks 72.6 85.4
Small-size bank
ks 35.8 62.8
Banking system
m 88.4 93.7
- opperative costs - interesst income - otheer regular incom

me
- innterest expensees - impairm
ment - nett commission iincome
Figure 2. A
Absolute increease/decrease oof the major revvenue and exppenditure comppared to the saame period lastt year
Source: NB
BRM (2012): Report
R on bankking system off the Republic of Macedoniaa in the third quuarter of 2012, p.60
In Table 4 are presented interest

i incom
me and operative costs observeed by groups oof banks and for the entire ban
nking
o Macedonia. From the data it can be conclluded that therre is a trend of increased cove
system of tthe Republic of erage
146
of the operative costs with interest income. However, despite this trend, the amount of the operative costs is
relatively high. On 09.30.2012, this amount was about 200 million denars higher than the same period last year
(Figure 2).
An important component of bank operating expenses is the interest payments that it most make on its liabilities,
particular on its deposit. (Mishkin, 2007) The banks explain increased operative costs as a result of 11.3% annual
increase of deposit insurance premiums that in fact corresponds to the annual growth of bank deposits. Increased
operative costs are one of the factors for decreased profitability of the total banking system.
3.2 International Comparative Analysis of the Macedonian Banking Sector
In order to get a better perspective on the level of profitability of the Macedonian banking sector, it is necessary to
make a comparison to the profitability of other banking sectors in some of the countries in the region. (Note 3) The
analysis included more countries in the region: Bosnia and Herzegovina, Serbia, Bulgaria, Romania, Czech
Republic, Hungary, Poland, Slovenia and Croatia.
We compare three key profitability indicators: ROA, ROE and non-performing loans.
14.00%
12.00%
10.00%
8.00%
6.00%
4.00%
ROA
2.00%
ROE
0.00%
-2.00%
-4.00%
Figure 3. ROA and ROE in certain countries

Sources: UniCredit CEE Strategic Analysis (2012): CEE Banking Outlook.
NBRM (2012): Report on Banking System of Republic of Macedonia in 2011.
The comparison with the mentioned countries showed that ROA and ROE of the Macedonian banking system are
in the middle of the list of analyzed countries (behind Serbia, Bulgaria, Czech Republic, Poland and Croatia), but
still they are below the determined standards (Figure 3).
Along with the previously mentioned indicators, in order to evaluate the profitability of the banking sector, the
percentage of the non-performing loans that the banks have in their portfolios can serve, too. A negative change in
this indicator can significantly affect the profitability of the banks. Increased bad loans means that banks allocate
higher provisions for risky loans - higher costs which, particularly, has a negative impact on the financial results of
smaller banks.
147
Figuree 4. Rate of noonperforming loans, by sector

Source: N
NBRM (2012). Report on Baanking System
m of Republic of Macedoniaa in the seconnd quarter of 2012.
2
Skopje, p.19.
The percenntage of the non-performing

n g loans is conntinually goingg up (except ffor March 20112) and the big ggest
contributioon to the grow
wth rate actuallly has the nonn-performing loans given too the companiies (Figure 4).. The
companiess cannot repayy the loans duue to the recession and the long period oof illiquidity oof the Macedo onian
economy. Reduced liquuidity suggestts that firm’s investments are less proffitable. The ppercentages off the
non-perforrming loans inn Macedonia annd in several oother countries in the region aare presented iin Table 5.
Table 5. N
Nonperformingg loans (in %) bby countries
Countries Noonperforming Loaans (in %) Date
Albania 14..4% 2011
Macedonia 9.1 2011
B&H 11..7% 2011
Croatia 11..5% 2011
Slovenia 3.77 2010
Montenegro 25%
% 2011
Bulgaria 13..5% 2011
Romania 13..4% 2011
Serbia 18..6% 2011
Czech Repubblic 5.66% 2011
Hungary 10..4% 2011
Poland 8.44% 2011
Source: htttp://data.worlddbank.org/indicator/FB.AST
T.NPER.ZS
As shownn in Table 5, the t percentagee of the nonpeerforming loanns in Macedonnia, which is 99.1%, is no su uch a
problem iff we compare the percentagges of Albaniaa (14.4%), of C Croatia (11.5% %), of Montennegro (25%) an nd of
Bulgaria (13.5%). On thhe other hand, Slovenia has only 3.7% andd Czech Repubblic has 5.6% of non-performing
loans and iit is a good inddicator that mucch more can bee done to lowerr the Macedonnian percentagee of these loanss-that
will eventuually lead to ennhancement off the profitabiliity of the bankks. There are a lot of reasons for the existen
nce of
these non--performing loans, and some of them are ass follows:
- Banks doo not have reaal informationn about the com mpanies’ posssibilities of deevelopment annd how capable the
certain bussiness is to gennerate profit. The banks do noot have a solid basis to evaluaate how capable they are to return
the loans tthey take;
148
- Absence of adequate credit risk management policy;

- Lame capacity of the human resources for a proper revision of the credit risks and absence of adequate training in
the field
Popovska, suggests that in order to lower the percentage of the non-performing loans “it is necessary to establish
rating agencies, more efficient exchange of credit information within the banking system, to have efficient policy
and procedures concerned with the risk management and to have a prudent supervision”. (Popovska, 2008)
4. Basel III Accord and Bank's Profitability
The appearance of a large number of financial crises, and particular the development and transmission of the
global financial crisis of 2007/08, imposes the need for taking appropriate corrective actions in the way of
functioning of the international financial system.
In that direction, the Basel Committee on Banking Supervision continuously monitor the changes that occur on
international financial market, such as: global financial crisis, new banking products and services, necessity for
finding new ways for identifying and managing the banking risks. Therefore, Committee in 2010 introduced new
Basel 3 standards in order to strengthen regulatory framework and to increase stability and resilience of the
financial sector. (BIS, 2013) The most significant changes that should be made concerning:
 Increasing the required level of capital for covering the risks undertaken by banks, with own capital (share
capital - ordinary shares, reserves and retained earnings). Capital adequacy ratio is a measure of the amount of a
bank's capital expressed as a percentage of its risk weighted credit exposures. This indicator measures how
resistant is the banking system on losses. (RBNZ, 2007)
 Total capital required rose to 8% — 10.5%.
 New Liquidity Coverage Ratio (LCR) which increases the level of high quality (cash or assets) assets that can
be converted into cash at little or no loss of value in private markets to meet its liquidity needs in 30-day period.
(BIS, 2012). This standard will reduce the bank run, improve the financial sector stability. Reliance on more
liquid assets will have a negative impact on bank’s profitability.
Also, there is one more liquidity standard - Net Stable Funding Ratio (NSFR) which should provide better
management of medium and long-term financial performance of the banks. This ratio should encourage banks to
rely on more stable source of funding (ex.deposits) and increase the competitiveness among banks.
After the Group of Governors and Heads of Supervision of the Basel Committee reached agreement on new
Basel capital and liquidity reform package, NBRM prepare information for introduction of Basel 3 standards. The
information contains an analysis of the increasing the level of capital required to cover the risks, the introduction of
international liquidity standard, and timeframe of their implementation and application. Information also contains
an analysis for evaluating the required capital level for coverage the risks in our banking system. (NBRM, 2010)
The analysis confirms that the Macedonian banks fulfill the new capital requirement rates. The capital adequacy
ratio is well above the new prescribed rate of 10.5% and accounted - 17%.
In order to ensure that the Macedonian banks operate according with generally accepted international standards,
NBRM adopted a new decision on the methodology for determining the capital adequacy. With this decision
change are made in the way the bank defined its own capital and are part of the activities for the application of
capital standards defined by the changes in the new Basel Capital Accord.
149
Table 6. Capital adequacy ratio (CAR) in 2011

Country CAR
Macedonia 17%
Czech Republic 16%
Hungary 12%
Poland 13.60%
Romania 13%
Slovenia 11.9%
Bosnia and Herzegovina 16
Bulgaria 17.80%
Croatia 19%
Source: UniCredit CEE Strategic Analysis (2012): CEE Banking Outlook.
Research has shown that the ratio of capital adequacy in the Macedonian banking system of 17% is among the
highest compared to the considered countries. Although, the high CAR promotes stability of the banking system, it
may serve to explain the low level of ROE in the banking system of the Republic of Macedonia (Table 6).
Since the main goal of this research is to show the level of profitability of the Macedonian banking system, in this
context, we should mention that there will be an increase in bank costs as a result of the introduction of the New
Basel Accord.
Thomas W. Killian (2010), says that higher common equity requirements of Basel III (from 2% to 4.5%) could
be expected to reduce bank profitability as expressed by return on common equity, put pressure on earnings per
share and lower growth potential.
In particular, the introduction of new international standards will increase the cost of funding, need for
reorganization and will reduce the profitability of the Macedonian banking system.
In this context it is important to mention that the necessary increase in the required capital may be an issue for
small-sized banks. These banks could raise the capital through their enlargement. We had a last example of such
consolidation of the banking capital in Macedonia in 2012, when Ziraat Bank AD merged to HalkBank. The
process of bank consolidation is supported by NBRM and is expected to continue in the future in order to improve
profitability, competitiveness in the banking system and improve the quality of the banking services.
According to bank experts the Macedonian banks continually cooperate with the NBRM in the field of
implementation of the new Basel Accord. To what measure the introduction of these standards will affect the
profitability of the Macedonian banking sector no one can be sure. It is well known that Basel III standards will be
phased in gradually by 2019 that goes in favor to banks.
5. Conclusion
As a result of the research done for this thesis we can accept the basic hypothesis and we can make a list of several
points as a conclusion, and they are as follows:
1) The research on the profitability of the Macedonian banking sector showed that the small-sized banks have a
major role in decreased profitability of the entire sector.
2) It is necessary to lower the number of banks existing in Macedonia through the processes of merging,
overtaking of the smaller banks by the big ones, in order to strengthen the competition in the banking sector.
3) Bank consolidation will have the following benefits for smaller banks:
- Increased market share
- Expansion of business activities
- Increased type and quality of service for their clients
- Increased competitiveness
- Increased profitability
- Rationalized expenses
- Strengthen the bank capital.
150
4) It should be pointed that the introduction of the new Basel standards will influence the level of profitability of
banks in Macedonia. Furthermore, this process is expected to be implemented without difficulty in the expected
timeframe.
In order to improve the banking sector as a whole, it is also important to undertake some structural reforms of the
banking sector, which will be a topic of further discussion in the future.
References
Center for Economic Analysis. (2005). The efficiency of the Macedonian Banking Sector. CEA Journal of
Economics. Skopje.
Center for Financial Training. (2010). Banking Systems, 2nd Edition 2010, South-Western, Cengage Learning.
Croatian Banking Association. (2009). Profitability of banks in Croatia: How to evaluate downfall? Retrieved
from http://www.hub.hr/Default.aspx?art=1935&sec=566
Davcev, L., & Hourvouliades, N. (2009). Profitability Parameters in the Banking System of Macedonia.
Proceedings of International Conference on Applied Economics 2009. pp. 133-141.
Hubbard, G., & O’Brien, A. (2012). Money, Banking and Financial System. USA: Pearson Education, Inc.
Killian, T. W. (2010). Basel III and Its Implications: A Closer Look. Sandler O’Neill & Partners.
KPMG. (2011). Basel III: Issues and Implications White Paper, USA. Retrieved from
http://www.kpmg.com/mk/en/issuesandinsights/articlespublications/brochures/pages/baseliiiissuesandimpli
cationswhitepaper.aspx
Madzova, V. (2011). Introduction of Basel III: Opportunities and Challenges. Yearbook, 3. University “Goce
Delcev” – Stip. R.Macedonia. P.73-84.
Mishkin, F. (2007a). The economy of money, banking and financial markets (8th ed.) Pearson Education, Inc.,
USA.
Mishkin, F. (2012b). The economy of money, banking and financial markets (10th ed.). Pearson Education, Inc.,
USA.
NBRM. (2006). Law on National Bank of Macedonia. Official Journal of the Republic of Macedonia, 3/02, 51/03,
85/03, 40/04, 61/05, 129/06. Skopje.
NBRM. (2010). Changes of Basel Capital Accord (Introduction of Basel 3) and preliminary estimates of their
impact on the capital adequacy of banks in Republic of Macedonia. Skopje.
NBRM. (2012). Report on Banking System of Republic of Macedonia in 2011. Skopje.
NBRM. (2012). Report on Banking System of the Republic of Macedonia in the second quarter of 2012. Skopje.
Petkovski, M. (2009). Financial markets and Institutions. Skopje, Faculty of Economics.
Popovska, K. (2008). Commercial Banking – Success and profitability. Skopje, “Institute of economics” – Skopje.
Qin, X., & Pastory, D. (2012). Commercial Banks Profitability Position: The Case of Tanzania. International
Journal of Business and Management, 7(13). http://dx.doi.org/10.5539/ijbm.v7n13p136.
RBNZ. (2007). Capital adequacy ratios for banks - simplified explanation and example of calculation. New
Zealand.
UniCredit CEE Strategic Analysis. (2012). CEE Banking Outlook.
Notes
Note 1. This is the situation in the Macedonian banking sector in 30.09.2011.
Note 2. 1 CAD = 49 MKD.
Note 3. 2011 is last year for which exists comparable data.
Note 4. http://data.worldbank.org/indicator/FB.AST.NPER.ZS
Note 5. http://www.bis.org/press/p130106a.pdf
Note 6. www.bis.org. Bank of international settlement.
151
ISSN 1916-971X E-ISSN 1916-9728
Can Financial Ratios Reliably Measure the Performance of

Banks in Bahrain?
Naser J. Najjar1
1
College of Business Administration, Kingdom University, Manama, Bahrain
Correspondence: Naser J. Najjar, College of Business Administration, Kingdom University, Manama, Bahrain.
Tel: 97-33-683-3091. E-mail: [email protected]
Received: January 7, 2013 Accepted: January 21, 2013 Online Published: February 22, 2013
Abstract
The aim of this study is to analyze the financial performance of major banks in Bahrain. This study covers the
calculation of important financial ratios of major financial institutions in Bahrain as well as comparing their
performance in the context of the global financial crisis. It also compares ratios of conventional banks with
Islamic financial institutions in Bahrain. These ratios define profitability, financial performance, size and type of
banks.
The analysis of ratios shows the differences in financial management practices of banks in the respective areas.
The study reveals that there are wide differences in the ratios used by different banks, especially before and after
the financial crisis.
This study helps identify best practice in the areas of profitability management, liquidity management, and
interest rate risk management.
The result of the analysis of ratios for measuring financial performance shows that there is corporate excellence
in asset management and value equity shares. This analysis can be used as a basis for preventative actions for
future bankruptcy and market risk. The components in financial statements for Islamic banks differ from
conventional banks.
The study recommends that banking institutions in Bahrain should use this ratio analysis to prevent unpredicted
financial problems and take corrective measures or provisions to avoid such events for financial institutions.
Keywords: Central Bank of Bahrain (CBB), conventional banks, islamic banks, financial ratios, liquidity
1. Introduction
This study compares the performance of a group of banks in the Kingdom of Bahrain, depending mostly on their
published financial statements and applying certain financial ratios as an indicator of financial management and
evaluating financial performance.
1.1 Issue of the Study
The purpose of this work is to provide a new interpretation for financial ratios and how they can help analysts
evaluate the overall financial condition of banks and organizations.
It will calculate and compare the important financial ratios used by major commercial banks as part of the
financial management practices in the Kingdom of Bahrain.
This study highlights the performance of Islamic banks in comparison to conventional banks- as a result the
findings are more applicable to the market in Bahrain.
1.2 Objectives and Scope of the Study
The objective of this study is to understand and analyze the different financial ratios which quantify many
aspects of a banking business in Bahrain, by making comparisons between different types and size of banking
institutions and comparing different time periods for a bank particular bank.
This analysis is primarily used to compare financial figures of banks over a period of time, a method sometimes
called trend analysis. Through trend analysis, it is easy to identify positive and negative trends, and to adjust
business practices accordingly.
152
There are several limitations in this study when comparing ratios from one financial period to another or when
comparing the financial ratios of two or more banking institutions
• If we are making a comparative analysis of a company's financial statements over a certain period of time, there
should be a study of any changes in accounting policies that occurred during the same time span.
• When comparing business with the industry, there should also be an analysis of material differences in
accounting policies between one company and industry norms.
• When comparing ratios from various fiscal periods or between different accounting methods, results may vary
widely in reported figures.
• Determination of the nature of adjustments is another limitation. In many cases, these adjustments can
significantly affect the ratios, whether ratios are calculated before or after adjustments are made to the balance
sheet or income statement, such as non-recurring items and inventory or pro forma adjustments.
• Carefully examine any departures from industry norms.
1.3 Significance of the Study
This study will support both Islamic and conventional banks to understand the nature and operation of financial
performance of banks in general, as well as enabling them to benchmark themselves with the sample taken for
the purpose of this study.
1.4 Research Questions
• What are the ratios which can indicate the factors associated with high financial performance?
• Does the size of the bank determine the ratios used?
• What is the impact of financial ratios in indicating future performance management?
• Is there a significant difference between ratios applied for Islamic banks and those used for conventional
banks?
• Is there a difference in performance measured by ratios before and after the financial crisis?
2. Background
2.1 Bahrain's Economy
Bahrain is the first country where oil was discovered and the first that announced that it is draining. Oil
Revenues and natural gas currently account for approximately 10% of its Gross Domestic Product (GDP).
Thus, Bahrain has worked to become a financial center in the GCC.
The Kingdom of Bahrain has become a major financial center where many international financial institutions
operate both onshore and offshore, without impediments. Currently the financial sector is the largest contributor
to GDP at 30%.
Around 370 offshore representative offices and units are located in Bahrain, and 65 American firms. The
Kingdom of Bahrain has made a concerted effort to become one of the leading Islamic finance centers in the
Arab world, and has standardized regulation for the Islamic banking industry. Currently with its 32 Islamic
commercial banks, a range of investment and leasing banks, and Islamic insurance companies (Takaful), Bahrain
has the largest concentration of Islamic financial institutions in the Middle East.
Bahrain has worked to develop other industries of service such as information technology industry, healthcare
and education.
The government used its oil income to build a top advanced infrastructure in telecommunications and
transportation.
The government moved toward privatizing the production of electricity and water by licensing for constructing
an independent power plant at a cost of $500 million. The company started operations in May 2006.
[http://www.traveldocs.com/bh/economy.htm]
Kingdom of Bahrain Financial Sector:
The financial sector in the Kingdom of Bahrain is diversified and well-developed. The sector consists of a
variety of both Islamic and conventional financial institutions and markets; the financial sector is well-positioned,
offering a wide range of financial products and services which makes Bahrain the leading financial center in the
153
GCC. The financial sector represents the largest single employer in Bahrain, with Bahrainis representing over 80%
of the work force.
One of the key drivers of growth in the country is this sector; its contribution is 27% of Bahrain's GDP.
The Central Bank of Bahrain (CBB) is the main regulator and supervisor of banks in Bahrain, controlling the
on-going regulation, licensing and supervision. To qualify for a license the requirements are capital adequacy,
business conduct, risk management, disclosure requirements and reporting actions.
[http://www.cbb.gov.bh/cmsrule/default.js]
2.2 Types of Banks in Bahrain
Bahrain has attracted many corporations and banks to expand their business and open their branches and
subsidiaries in Bahrain.
There are many banks operating in Bahrain under the regulation and supervision of CBB, investment and
commercial banks. The Commercial banks are divided into conventional and Islamic banks.
There are also offshore commercial banks which do not offer regular banking services within Bahrain but offer
financing for large business and projects.
[http://www.marcopolis.net]
The topic of financial ratios analysis has always been important for both researchers and those involved in
finance. Susan W. (2003) stated that lenders often rely on ratio analysis because it allows them to see how one’s
business is doing and how compare it compares to other businesses which they have loaned money to. The
potential loan creditor also requires information about security that the Ratio Analysis can be provided
(Hingorani N.L. and A.R. Ramanathan, 1986).
Ratio analysis is a useful tool for business owners as well. It measures the health of the business whether it is a
bank or a multinational corporation. It helps to measure the performance of the business, to diagnose potential
problems and to see how well it is doing over time (Spathis, K., and Doumpos M., 2002).
The current ratios, total debt ratios, and profit margin ratios are the three important ratios that indicate the health
of your business (Susan W., 2003).
Current ratio is an excellent diagnostic tool, because it measures whether or not your business has enough
resources to pay its bills over the next 12 months. Total debt ratio shows how much your business is in debt,
making it an excellent way to check your business’s long-term solvency (Lynch, R. and Willamson, 1983). The
profit margin shows how much net profit your business’s sales are producing.
Howard Finch (2005) stated that financial ratios are one of the most common tools of managerial decision
making. A ratio is to compare one number to another—mathematically, a simple division problem. Financial
ratios involve the comparison of various figures from the financial statements in order to gain information about
a company's performance. It is the interpretation, rather than the calculation, that makes financial ratios a useful
tool for business managers. The future decisions can be evaluated in terms of measuring the behavior of the firm
value, given a decision to be analyzed (Chien, T., Danw, S. Z. 2004). However, examining historical financial
statements analysis has the limitations in reaching the optimal financial decisions (Ignacio Velez-Pareja, 2007).
Ratios may serve as indicators, clues, or red flags regarding not worthy relationships between variables used to
measure the firm's performance in terms of profitability, asset utilization, liquidity, leverage, or market valuation
(Sree Rama Murthy, Y., 2004).
Nenide et al. (2003) used ratio calculations with multivariate analysis relying on a large database for predicting
the performance of business firms.
In their paper, adjustment techniques were recommended for researchers using multivariate statistical analysis on
large databases for ratio calculation, so that results of the analysis would be meaningful and that inferences could
be drawn from the data. In order to illustrate and explain techniques for data error identification, they used
sample data from a Balance Sheet and Income Statement of 250 firms, sampled from the Financial Statement
Database of the Kauffman Center for Entrepreneurial Leadership. They were also able to demonstrate how to
handle the problem of denominators being negative or approaching zero when calculating ratios, and effective
techniques for transforming the data to achieve approximation of normal distributions.
Salmi and Teppo Martikainen (1994) provided a critical review of the theoretical and empirical basis of four
central areas of financial ratio analysis.
154
The research areas reviewed were the functional form of the financial ratios, distributional characteristics of
financial ratios, classification of financial ratios, and the estimation of the internal rate of return from financial
statements. A common feature of all these areas of financial ratio analysis seems to be that while significant
regularities can be observed, they are not necessarily stable across the different ratios, industries, and time
periods. This leaves much space for the development of a more robust theoretical basis and for further empirical
research.
Jerome Osteryoung et al. (1992) stated that financial ratios were broadly used by academic researchers, financial
analysts, lenders, and small business managers. Most commonly, researchers use ratios as predictor variables in
models that forecast business distress and failure. Other common applications include trend analysis studies of
individual firm performance and cross-sectional studies that compare individual firm ratios against average
industry ratios. These average industry ratios are usually reported for different size firms in each industry, under
the assumption that ratios are different for firms of different sizes. (Robert Morris Associates' Annual Statement
Studies and the Financial Research Associates' Financial Studies of the Small Business both use size categories
in the reporting of financial ratios). They have found evidence that financial ratios differ across different size
firms. In their study they examined the differences between the financial ratios of large public and small private
firms across a large number of narrowly defined industry groups.
If certain ratios are found to be constant across the different size groups, it would indicate that those ratios could
be used for ratio comparison purposes, regardless of whether the ratios are from large or small firms. If the ratios
are found to be different, the results will emphasize the importance of identifying the appropriate source of
industry and its average ratio for comparison purposes when examining a particular ratio.
A J Singh (2002) in his study identified commonly used ratios in the lodging industry and discovered their
importance for lodging financial executives. In his point of view, financial ratios have always been a valuable
tool for lodging industry managers. Ratios allow the user to summarize and analyze related data to provide
meaningful information for making decisions.
The Operating and Profitability ratios clearly stand out as the most important ratios for lodging managers. This
study contributes to educating managers about the range of ratios, their relative importance, and opportunities for
using ratios, which were considered less useful in the past.
3. Methodology
3.1 Data Collection
This study will utilize data collected from secondary sources, which are the annual reports of commercial banks
in the Kingdom of Bahrain for the period of 2005-2009.
For purposes of international comparison; data was drawn from various internet based sources local banks and
international foreign banks.
3.2 Population and Sample
The population in this study includes all 109 Banks in The Kingdom of Bahrain (82 Conventional banks retail
and wholesale and 27 Islamic banks).
The sample was drawn from the different types of banks in Bahrain. It was selected to include all types of banks
with different categories and the availability of data for the covered period of the study.
Table 1. Sample banks

Name Type
National Bank of Bahrain Conventional – Retail (Local)
Bank of Bahrain and Kuwait Conventional – Retail (GCC)
Gulf Finance House Islamic – Wholesale
Kuwait Finance House Islamic – Retail
BNP Paribas Conventional – Wholesale (International)
3.3 Analysis Plan

A set of financial ratios will be used to measure the performance of the banks by analyzing the relationship
155
between independent and dependent variables. By comparing these ratios, the results will be used to measure
the effect of the financial crisis on the performance of the banks.
The independent variables are: the size of the bank, the type of bank, and the policies and regulations that the
CBB sets which cannot be controlled or changed by the bank.
Dependent variables are: the performance of the bank, the policies and regulations, and internal controls and
processes of the banks.
3.3.1 Definition of Variables
Profitability Management Ratios
Profitability reflects the final result of business operations. There are two types of profitability ratios: profit
margin ratios and rate of return ratios.
The profit margin ratios show the relationship between profit and revenue. The rate of return ratios reflects the
relationship between profit and investment. Some rate of return ratios are: return on assets, earning power,
return on capital employed, and return on equity.
Return on Equity (ROE) is calculated by dividing the total profit after tax by total shareholders equity. From the
investor’s point of view, ROE on a Post – Tax basis is a better measure of profitability. While ROE (post tax)
itself is not an indicator of investors return on investment (investors return on investment would rather depend on
dividend declared plus capital appreciation if any) we can argue that a higher ROE leads to better return for the
shareholders.
Return on Assets (ROA) is another good measure of performance and profitability. However, ROA does not
reflect the impact of capital structure decisions (this financial leverage is also called gearing on the firms
earnings).
The DuPont model expresses profitability as a percentage of total assets and total shareholders equity. This
model is very useful for evaluating how a bank is doing over time.
Dupont Model Profitability Ratios
ROE = Profit Margin × Asset Yield × Leverage
or
ROE = Return on Assets × Financial Leverage
Profit Margin = Profit after Tax / total income
Return on Assets can be expressed as
ROA = Profit Margin × Asset Yield
Liquidity Management Ratios
Liquidity refers to the ability of a firm to meet its obligations in the short run, usually one year. Liquidity ratios
are generally based on the relationship between current assets and current liabilities.
Liquidity risk management refers to the ability of any banking institution to maintain the right cash balance. The
term cash can be explained as currency held with the banks’ plus balances with Central Banks of Bahrain.
Profit sub – optimization is possible if a bank is maintaining high level of cash balances. Banks may lose profits
if they have an opportunity to invest these funds. Excess cash earns a zero interest. If such cash is invested in the
form of commercial loans or securities then the bank would earn a return which directly contributes to the profit.
Cash accumulated through the receipt of deposits requires banks to pay interest to the depositors.
Liquidity Management Ratios
Liquid Assets Ratio = Total Cash Resources / Assets * 100
Total Cash Resources = Cash + Treasury Bills + Placements with Banks
Loans to Deposit Ratio = Loans / Deposits * 100
Interest Rate Risk Management Ratios
Interest rate risk can be defined as the risk, which arises when interest rates are changed. This has a direct impact
on the interest earned on loans and investments and the interest paid on deposits (Prasanna Chandra, 2007).
Interest rate risk management is required in order to manage the net interest margin, and to control the risk posed
by changing interest rates while trying to take advantage of changing interest rates.
156
The risk of interest rates changing can be controlled by matching the re-pricing maturities of assets and liabilities.
If both the asset and the liability are reprised at the same time, a bank will be able to maintain the net interest
margin as interest cost and interest earnings either by going up or down simultaneously.
Interest Rate Risk Management Ratios:
Break Even Yield = Interest Expense as % Assets*
Interest Yield = Interest Revenue as % of Assets*
Net Interest Margin = Net Interest Income as % Assets
(*assets = average asset of current year and previous year)
4. Results and Interpretation
Analysis (Profitability ratios)
Table 2. DUPONT model profitability ratios

2005 2006 2007 2008 2009
BNP Return on Equity 13.67% 14.24% 13.17% 5.85% 8.06%
Profit Margin 28.76% 27.94% 25.2% 12.60% 16.10%
Asset yield 1.73% 1.94% 1.83% 1.32% 1.95%
Leverage 27.35 26.27 28.52 35.20 25.61
Return on assets 0.50% 0.54% 0.49% 0.16% 0.31%
GFH Return on Equity 12.11% 31.68% 39.03% 30.19% (168.09%)
Profit margin 62.35% 61.20% 58.54% 47.26% (171.49%)
Asset Yield 9.26% 23.03% 26.12% 17.72% 3.78%
Leverage 2.09 2.25 2.55 3.6 3.79
Return on assets 5.77% 14.09% 15.29% 8.3% (44.35%)
KFH Return on Equity 22.78% 23.19% 19.89% 11.02% 0.90%
Profit margin 48.39% 54.16% 56.96% 46.35% 5.66%
Asset Yield 12.10% 9.51% 7.8% 6.45% 4.06%
Leverage 3.89 4.5 4.47 3.68 3.91
Return on assets 5.84% 5.15% 4.44% 2.99% 0.23%
NBB Return on Equity 14.00% 16.59% 17.05% 15.98% 17.74%
Profit margin 62.18% 64.05% 63.90% 55.44% 59.24%
Asset Yield 3.28% 3.43% 3.42% 3.08% 3.41%
Leverage 6.87 7.55 7.80 9.36 8.77
Return on assets 2.04% 2.19% 2.18% 1.70% 2.02%
BBK Return on Equity 16.89% 17.42% 12.61% 12.90% 15.15%
Profit margin 52.39% 50.84% 34.19% 23.55% 37.83%
Asset Yield 3.73% 3.80% 4.18% 5.3% 4.06%
Leverage 8.6 9.02 8.81 10.35 9.87
Return on assets 1.95% 1.93% 1.4% 1.25% 1.53%
Profit margins of Bahrain commercial banks have tended to vary from one financial institution to another, for
example in the year 2005, GFH reported a profit margin as high as 62.35% while in the same year BNP reported
a PM of only 28.76%.
An analysis of the profit margin ratios of various banks from 2005 to 2009 shows that NBB was successful in
managing profit margins even though there was a slight decline trend in 2008 and 2009- this also indicated that
NBB was able to control operating costs well.
157
Table 3. Profit margin (%)

2005 2006 2007 2008 2009
BBK 52.39 50.84 34.19 23.55 37.83
BNP 28.76 27.94 25.20 12.60 16.10
GFH 62.35 61.20 58.54 17.26 (171.49)
KFH 48.39 54.16 56.96 46.35 5.66

NBB 62.18 64.05 63.90 55.44 59.24
GFH reported lower PM in 2008 at 17.26% and it was negative in 2009 at 117.49%. In 2009 KFH reported its
lowest PM in 5 years at 5.66%. GFH revenue decreased by 89% in 2009 compared to 2008. The huge provision
for impairment affected the income of GFH in 2009.
All five banks reported a decline trend in PM in 2008 and 2009 which shows the global financial crisis largely
hit the performance of financial institutions of Bahrain.
Asset yields of Bahraini banks during the period of 2005 to 2009, show stability in the case of NBB and BBK,
whereas BNP, GFH and KFH reported a decline trend in 2008 and 2009.
Table 4. Asset yield (%)

2005 2006 2007 2008 2009
BBK 3.73 3.80 4.18 5.3 4.06

BNP 1.73 1.94 1.83 1.32 1.95
GFH 9.26 23.03 26.12 17.72 3.78
KFH 12.10 9.51 7.8 6.45 4.06
NBB 3.28 3.43 3.42 3.08 3.41
GFH reported its lowest results in 2009 at 3.78% compared to the previous four years. KFH also reported its
lowest results in 2009 at 4.06% compared to previous years. Only BNP was able to report its highest result in
2009 at 1.95% compared to previous years. Stability in asset yields can be achieved by adopting good interest
rate risk management practices and better asset mix decisions.
Higher leverage improves profitability and cost of equity is high compared to cost of deposits. Therefore, a
bank’s management would prefer to be high on deposits and low on equity. However, a low equity implies a
high leverage ratio, which is risky. A bank with a high leverage ratio (that is low capital to assets ratio) faces the
problem of solvency risk) during periods of economic crisis.
Financial Leverage in the five Bahraini Financial Institutions during the period of 2005 to 2009 were in the range
of 35.20 times (BNP in year 2008) to 2.09 times (GFH in year 2005). BNP was able to practice stable and high
level leverage in 5 years. Only GFH and KFH reported low leverage approximately between 2 to 4 times from
2005 to 2009.
Table 5. Leverage
2005 2006 2007 2008 2009
BBK 8.06 9.02 8.8 10.35 9.87
BNP 27.35 26.27 28.52 35.20 25.61
GFH 2.09 2.25 2.55 3.6 3.79
KFH 3.89 4.5 4.47 3.68 3.91
NBB 6.87 7.55 7.80 9.36 8.77
158
Here, BBK and BNP are showing the leverage ratio more than 8 times, which indicates that the business size of
these banks is larger than the other three banks (GFH, KFH and NBB). Income of the Islamic Financial
Institutions such as GFH and KFH mostly rely on Murabaha contract fees and profit. These banks mostly
depend on the profit sharing rather than interest income.
NBB and BBK are able to generate higher ROE compared to the same level of operational profitability – ROA.
Table 6. Return on assets (%)

2005 2006 2007 2008 2009
BBK 1.95 1.93 1.4 1.25 1.53
BNP 0.5 0.54 0.49 0.16 0.31
GFH 5.77 14.09 15.29 8.3 (44.3)
KFH 5.84 5.15 4.44 2.99 0.23

NBB 2.04 2.19 2.18 1.70 2.02
Table 7. Return on equity (%)

2005 2006 2007 2008 2009
BBK 16.89 17.42 12.61 12.90 15.15
BNP 13.67 14.24 13.17 5.85 8.06

GFH 12.11 31.68 39.03 30.19 -168.0
KFH 22.78 23.19 19.89 11.02 0.90
NBB 14.00 16.59 17.05 15.98 17.74
In 2006, NBB, which had an ROA of 2.19%, reported a ROE of 16.59%. If we agree that shareholders and
investors are ultimately interested in return on equity (ROE) then the performance of NBB is better, compared to
that of GFH and KFH.
This result can be possible because the average leverage ratio of NBB is 8 times for 5 years in that year as
compared to average leverage ratio of GFH and KFH which is 5 and 2.8 for 5 years. BNP and BBK also
reported good result in 5 years. However in 2008 and 2009, the figures show a decline trend in return on equity.
GFH and KFH reported worst results in 2009 at 168.09% in the case of GFH and 0.90% in the case of KFH.
Earnings Per Share is the indicator for the performance of the organization.
Table 8. Earnings Per share (%)

2005 2006 2007 2008 2009
BBK 0.052 0.047 0.043 0.034 0.042
BNP 7.02 8.49 8.03 2.99 5.20
GFH 136 29.92 47.55 29.08 - 68.41
KFH NA NA NA NA NA
NBB 56.6 56.9 64.10 44.7 55.10
BBK and BNP are showing stability in their performance in profitability of Earning per Share but GFH suffered
losses in 2009 - large provisions against the net profit showing negative earnings per share.
Analysis of ROE trends in the five banking institutions in Bahrain over the period of 2005 to 2009 indicates that
years 2005, 2006, and 2007 were good years for these banking institutions in general.
Furthermore, the ratios reported above also show that the good performance in 2005, 2006 and 2007 was mainly
due to good profit margins generated by these institutions in those years. Similarly ratios from 2008 and 2009 r
159
show that poor profit margins had a significant impact on return in equity in these years.
Asset Interest Yield and leverage variations from year to year have been less important. The conclusion therefore
is that, due to economic crisis the performance of Bahrain banks affected severely which forced them to focus on
factors influencing profit margins like cost management and credit risk management practices.
Bank BNP prefers a high leverage multiplier at 26% average for 5 years from 2005 to 2009 while GFH shows
lowest leverage at 2.8% as average for 5 years. In 2008 and 2009 there are no significance changes in leverage
and capital to asset ratios compared to previous period.
There were no significant changes in the leverage ratios or capital to assets ratios of banks during this period, as
compared to previous years. BNP and KFH experienced a slight reduction of financial leverage ratio, indicating
that they are not able to deploy as much assets for every unit of capital in this period. Better leveraging would
help banks to improve profitability and reduce spreads.
Analysis (Liquidity Management Ratios)
The period from 2005 to 2009, demonstrated that liquidity management ratios for the GFH showed a high level
of liquidity. It was reported that this was the worst performance in 2009. Although NBB also reported an average
of 27% liquidity ratio in 5 years, their profit margin was still stable.
Regarding the Loan to Deposit ratio, KFH showed good results in 2008 and 2009 where all other institutions
showed below 100% result. GFH had the lowest result in 2009 at 28.32%.
KFH experienced good results in 2009 at 175%. BNP had a low liquid asset ratio in 2005 at 0.56% which moved
to a higher liquid asset ratio in 2009 to 2.73%. At the same time the bank’s loan to deposit ratio declined from
105.87% in 2006, to 92.98% in 2009. According to the annual reports, companies having a lower level policy of
lending, and a higher liquid assets ratio, may be as a result of company liquidity policy. We can conclude that
apart from KFH, all other banks showed a decline trend in Loan to Deposit ratio because the financial crisis had
depressed the market causing these banks to change their lending policies.
Analysis (Interest Rate Risk Management Ratios)
Comparisons of the interest rate management ratios from 2005 to 2009 indicate that there was a considerable
decline in Asset Interest Yield in 2009. If there had been a decline in Break Even Yield more than the decline of
Asset Interest Yield the impact would be an improvement in Net Interest Margins.
The change in Asset Yield, Break Even Yield and Net Interest Margin has to be viewed in the light of a general
decline or increase in interest rates in the Bahrain economy. These ratios indicate that banks have been able to
manage and overcome the interest rate risk created by declining interest rates. Apart from the effect of the
financial crisis, the decline in Break Even Yield shows there is change in the interest risk gap management
strategies of Bahraini banks. However, GFH was the only bank showing a negative result in Net Interest Margin
in 2009 at a rate of -0.91% (Due to huge provisions for Impairment)
KFH recorded the lowest Break Even Yield at 0.69% in 2009. This Islamic bank mostly depended on the
contract fees, investment activities, and shares of income from associates rather than interest income.
All five banks demonstrated a decline in Asset Interest Yield in 2008 and 2009 because of the decline in interest
income or profit sharing income (Islamic banks) compared to the amount of assets.
Major Results
1. Profitability ratios showed an increasing trend in the first three years and a decline trend in 2008 and 2009.
GFH suffered huge losses in 2009. (Impact of financial crisis)
2. BBK and BNP achieved a Leverage ratio of more than 8% and GFH and KFH less than 8%. NBB achieved a
ratio higher than 8% in 2008 and 2009. This indicates that BBK and BNP had the highest share of the market,
and highest income from Interest and Fee income. There is an implication that major income depends on profit
sharing income, which probably indicates why the Islamic banks show low leverage. (Type of the banks)
3. Asset Interest Yield shows good performance in 2005, 2006 and 2007 but declined in 2008 and 2009. It
happened even in the case of Islamic banks GFH and KFH. This shows the performance of assets decreased in
2008 and 2009
Findings as per Research Questions
1). Difference in the performance measured by Ratios before and after crisis.
The Global financial crisis largely hit the banking activities in the Gulf region. As per the analysis of financial
160
performance of five major financial institutions, the financial ratios indicate stable and good performance in
2005, 2006 and 2007, but in most of the ratios showing declining trend in 2008 and 2009.
For example:
Average Profit margin of BBK in 2005, 2006 and 2007 was 52.39%, 50.84% and 34.19% but it declined in 2008
and 2009 to 23.55% and 37.83%.
BNP recorded 28.76%, 27.94%, and 25.20% in 2005 2006 and 2007 but declined in 2008 and 2009 to 12.60%
and 16.10% respectively.
GFH recorded 62.35%, 61.20%, 58.54 % in 2005, 2006 and 2007 but decreased to 17.26% in 2008 and showed
negative ratio (- 117.49) in 2009.
KFH showed 48.39%, 54.16%, and 56.96% in 2005, 2006 and 2007. However, in 2008 and 2009 it declined to
46.35% and 5.66% respectively.
NBB showed 62.18%, 64.05%, and 63.90% in 2005, 2006, and 2007 and declined to 55.44% and 59.24% in
2008 and 2009.
If profit earning is the main motive for the business of banking, then the above decline trend shows that the
recession in market affected the investing and lending transactions of the banks in Bahrain.
It also indicates that the earning per share showed a declining trend in 2008 and 2009 and these are negative in
the case of GFH in 2009.
2). Ratios indicate the factors associated with high financial performance Profitability Ratios could be, net profit
to total assets, a profit margin, ROI, net profit after taxes to Net worth, Return on Equity, or Return on Assets.
The variety of ratios is mainly used to measure the profitability of the business and management’s overall
effectiveness. The profit of the business is the indicator of the improved financial performance of the
organization. If there is decline in profit margin, this also affects the return on equity and return on total assets.
GFH showed a profit margin of (-1171.49%) which affected the return on assets at (-44.35%) and return on
equity at (-168.09%).
These ratios are related to the factors such as cost management, management of interest risk, leverage and
liquidity conditions.
3). Relationship between the size of the bank and the ratios applied.
According to the study, ratios, excluding Leverage ratios, are not directly linked to the size of a bank. The profit
of a bank cannot be measured according to its size. Even a small bank can achieve high profit and can show
highly positive ratios, whilst a bank with a large asset may not yield good income in particular year.
For example, ratios derived from the components of financial statements, showed that GFH’s asset size in 2006
was $ 1,500,884,000, and its net profit was $211,586,000. But in 2009 the asset size was $1,642,336,000 and its
net profit decreased to net loss ($728.379, 000).
Financial performance of institutions can be affected by the complexity of managing large sized financial
transactions. The leverage may be a good position but profitability may show a decline trend due to poor
management or due to external factors.
We can conclude that the size of the bank can influence its efficiency in the way it is run, ultimately impacting
on the profitability of the firm.
4). The effect of the financial ratio indicators on the future performance management.
These financial ratios are prepared on the basis of historical financial statements and they are useful indicators of
a firm's current financial performance and current financial situation. These financial ratios can be used to
analyze current trends and to compare the firm's financial position to those of others. They can predict future
bankruptcy.
5). Significant difference between ratios applied to Islamic banks compared to the ones used for conventional
banks.
In this matter the financial statements of Islamic banks GFH and KFH are do not show interest income, therefore
interest rate risk analysis can’t be applied to these banks. Thus the ratios that include interest income in the
numerator or the denominator should be replaced with income from Murabaha and fees and commissions.
Interest expense should be replaced by Profits distributed to depositors’ because Islamic banks do not pay
interest. Instead, they share their profits with depositors. Some ratios will still be zero or undefined as the
161
equivalents for numerators and denominators do not exist for Islamic banks. For instance, a risk ratio, ‘Provisions
to Earning Assets’ will by definition be zero because provisions for loan losses are zero because no loans are
made.
5. Conclusion
The tools of analysis are helpful in making business decisions, evaluating performance and forecasting future
developments (Bhatia, Manohar L. 1986). Control of business activity is crucial for efficiency. Managerial
action follows meaningful information flows. These ratios provide a relevant basis, but all ratios may not serve
the objective of the control. A profit performance measure, which is widely prevalent, is the Return on
Investment, which is considered a primary yardstick for the measurement of operational efficiency. The DuPont
chart underlines areas which need managerial control for achieving the basic goal of maximizing the return on
capital employed in financial institutions.
This study examined the impact of financial ratios on the financial performance of Bahraini commercial and
Islamic banks. The results of the analysis showed that, the financial performance of banks was strongly and
positively influenced by their operational efficiency, asset management, and, their size. However, ratios
themselves cannot indicate the size of the banks, which means that banks can only be measured by their overall
financial performance.
6. Recommendation
The Return of Equity (ROE) of the five Bahraini banks has varied much during the period 2005 to 2009.
This study has used the DuPont model to explain the variations in ROE (profitability) through Profit Margin,
Asset Yield and Leverage ratios. Analysis of ROE trends in the five Bahraini banks over the period of 2005 to
2009 indicates that years 2005, 2006 and 2007 were on average, good years for the banks. Furthermore, the
ratios reported above also show that the good performance in 2005, 2006 and 2007 was mainly due to good
profit margins generated by banks in those years. On the other hand in 2008 and 2009, ratios showed that poor
profit margins had a significant impact on Return on Equity in these years. Asset Interest Yield and Leverage
variations from year to year have been less important.
Banks in the Kingdom of Bahrain should focus on factors such as costs and provisions, which influence profit
margins. The decrease in Return on Equity ultimately affects the interests of shareholders. The risk of
fluctuations in interest rates penetrates the interest incomes, which affect the overall income of the banks. If the
banks increase their income by other means such as free income, profit sharing, investment activities, which
have the lowest income expenses, it would increase the overall profitability. KFH shows low rating of breakeven
yield ratio compared to other banks.
While most of the banks show a continuous increase in the Liquidity Rate, GFH is still operating in the same
environment and within the same regulations, which maintain liquidity levels as high as 77%. Liquid asset ratios
of banks have in general increased over the period. High level of liquidity affects the interest revenue therefore
banks should have effective liquidity management policies. These policies should maintain adequate cash
balances and prevent excess cash balances or deficit cash balances.
Net Interest Margins increased in 2005, 2006 and 2007, even though 2008 and 2009 were not good years in
terms of profitability due to the economic crisis. Despite this, we can recommend that the banks should have
effective interest rate risk management polices to deal with the effects of fluctuation in interest rates, which can
result in lower margins.
A low Leverage ratio indicates that the bank prefers to follow a safe path as it grows, while a high Leverage ratio
indicates that the bank is taking more risk. However, a higher Leverage ratio means that the bank chooses to
improve its profitability. At the same time, finance experts say that cost of equity is high, compared to the cost of
deposits and therefore a bank’s management may prefer to finance its asset growth with more deposits and less
of capital (equity) resulting in a high Leverage ratio.
In conclusion, the study provides banking authorities with understanding of activities that would enable them to
review the financial performance of their bank. The results of this study should provide support in making the
right decisions to enhance the financial positions of the bank.
Finally, even though this study has focused on certain ratios, there are many other financial ratios that might be
used such Economic Value Added (EVA), as well as other measures of performance that are non-financial such
as productivity, customer satisfaction, social responsibility, and the use of technology.
162
References
Annual Report. (2005, 2006, 2007, 2008, 2009). Bank of Bahrain and Kuwait, BNP Paribas, Gulf Finance House,
Kuwait Finance House, National Bank of Bahrain.
Bhatia, M. L. (1986). Profit Centers: Concepts, Practices and Perspectives. Somaiya Publications Pvt. Ltd,
Mumbai.
Chandra, P. (2007). Financial Management: Theory and Practice. New York: McGraw-Hill.
http://highered.mcgraw-hill.com/sites/0070656657
Chien, T., & Danw, S. Z. (2004). Performance Measurement of Taiwan Commercial Banks. International
Journal of Productivity and Performance Management, 53(5), 425-434.
Hingorani, N. L., & Ramanathan, A. R. (1986). Management Accounting. Sultan Chand: Delhi.
Howard Finch. (2005). Financial Ratios. Retrieved from
http://www.referenceforbusiness.com/management/Ex-Gov/Financial-Ratios.html#b
Lynch, R. M., & Willamson, R. W. (1983). Accounting for Management. New York: McGraw-Hill.
Nenide, B., Pricer, R. W., & Camp, S. M. (2003). The use of financial ratios for research: problems associated
with and recommendations for using large databases. Unpublished manuscript retrieved August 3, 2010
from http://www.fintel.us/download/ProblemswithandRecommendationsforLargeDatabases.pdf.
Osteryoung, J., Constand, R. L., & Nast, D. (1992). Financial ratios in large public and small private firms.
Journal of Small Business Management, 30(3). Available for authorized users in the PDF and HTML
formats from Internet: EBSCOhost at FinELib [Cited 5-Oct-2002]. ABI Inform: ProQuest Direct database
search keywords used: financial ratio proportionality.
Salmi, & Martikainen, T. (1994). A Review of the Theoretical and Empirical Basis of Financial Ratio Analysis.
The Finnish Journal of Business Economics, 4(94), 426-448. Retrieved from
http://lipas.uwasa.fi/~ts/ejre/ejre.html
Singh, A. J., & Schmidgall, R. S. (2002). Analysis of financial ratios commonly used by US lodging financial
executives. Journal of Retail & Leisure Property, 2, 201-213.
http://dx.doi.org/10.1057/palgrave.rlp.5090210
Spathis, K., & Doumpos, M. (2002). Assessing profitability factors in the Greek banking system: A multi criteria
methodology. International transaction in Operational Research, 9, 517.
Sree Rama Murthy, Y. (2004). Financial Ratios of Major Commercial Banks.
http://dx.doi.org/10.2139/ssrn.1015238
Ward, S. (2003). Is Your Business Sick? Give Your Business a Health Checkup with These Three Ratios.
Retrieved from http://sbinfocanada.about.com/od/management/a/3ratios.htm
163
ISSN 1916-971X E-ISSN 1916-9728
Testing the CAPM for the Brazilian Stock Market: A Study of

Dynamic Beta Using Multivariate GARCH
Lucas Lúcio Godeiro1
1
Department of Agro technology and Social Science, Federal Rural University of Semi-Arid, Brazil
Correspondence: Lucas Lúcio Godeiro, Department of Agro technology and Social Science, Federal Rural
University of Semi-Arid, 59625-900, Francisco Mota Avenue 572, Mossoró-RN, Brazil. Tel: 55-84-3315-8555.
Received: January 11, 2013 Accepted: February 7, 2013 Online Published: February 22, 2013
Abstract
The paper tests the CAPM for the Brazilian stock market using dynamic betas. The sample involves 28 stocks
included in the Ibovespa portfolio as of March 21, 2012 and that were traded during the period from Jan. 01,
1995 to March 20, 2012. Dynamic betas were estimated and conditional betas contributed with larger
explanatory power of excess cross section returns. The main contribution of the paper is the estimation of
dynamic betas for Ibovespa shares, which can be useful for investors using Long-Short strategies.
Keywords: CAPM, multivariate GARCH, dynamic betas
1. Introduction
The Brazilian stock market has passed for changes in recent years. On scenario a macroeconomic stability after
the Real Plan, the Ibovespa, principal index of Bovespa passed to approximately 4300 points in 1995 for about
60000 points in 2011. All this evolution occurred for several reasons, how investment foreign flow, increase of
individual investors participation, inter alia. Despite this evolution, the stock market is a very volatile market. To
get an idea, the 2008 crisis did Ibovespa fall of 70000 points to 29000 points in a few weeks. This requires an
efficient risk management by managers and investors, so they are not surprised by sudden movements of the
market.
One of the most widely used models in risk management by financial market is the Capital Asset Pricing Model
(CAPM) developed by Sharpe (1964). The model use the betas how principal measure of non-diversifiable risk
and emerged how alternative to efficient frontier Model Markowitz’s (1952). The previous model to CAPM was
required many algorithms, even for a small number of assets or portfolios. However, there is a discussion in the
literature about the validity or otherwise of the CAPM. Several tests indicate validity, and various tests indicate
that the model fault on several occasions. Another discussion in the literature is about the behavior of betas.
Many studies admit that betas are constant over time and many reject this hypothesis.
Tambosi Filho et al. (2007) tested the traditional capm and the conditional CAPM in the Brazilian, American and
Argentine market. Further according to the above authors, the CAPM conditional differs from the Conditional
CAPM by incorporating the variation of the betas over time, allowing identifying variances and covariances that
change in time.
Tambosi Filho et al. (2007) have used stock portfolios in their tests and selected the stocks that composed the
portfolio via the criterion of liquidity and survivors stocks. As proxy of human capital the authors used the
growth rate of labor income explained by GDP. The results obtained were as follows: the traditional capm
without human capital explains the expected returns, the conditional CAPM without human capital also explains
the expected returns in all markets more efficiently than static, as it showed a better fit, the conditional CAPM
with human capital not explains the expected returns in any market, and the traditional capm with human capital
also did not get a good power explanation.
Flister et al. (2011) tested whether the conditional CAPM is able to explain anomalies of size, moment and
book-to-market. They concluded that the conditional CAPM, using a series of regressions of short term, showed
gain negligible compared to alphas calculated with unconditional CAPM. They also found that the betas
calculated by conditional CAPM vary with time, but not enough to the unconditional model alphas’ were
164
explained by the conditional model.

Paiva (2005) tests the CAPM according to the methodology of Fama and MacBeth (1973) and finds evidence
that the betas of companies that remaining in Ibovespa for a long period of time tend to be smaller than one, and
with this characteristic the actions are defensive to investors.
Alencastro (2009) tests the traditional capm according to the methodology of Fama and Macbeth (1973) using
stock portfolios. He systematizes research by dividing the period after the implementation of the Real Plan in
three parts: 1994-1999, 1999-2004, 2004-2009. The author analyzes the microeconomic scenario of the stock
market and also the macroeconomic situation of the period. The results were: validity of the CAPM in the first
period, with low explanatory power, in the second period, the model is not validated, in the third period the
model was validated after correction of heteroscedasticity and elimination of two outliers.
However, which is verified is a lack of empirical tests of the CAPM with dynamic betas in the Brazilian stock
market, estimated with Multivariate GARCH DCC (Dynamic Conditional Correlation). Therefore, the research
aims to test the CAPM for the Brazilian stock market using dynamic betas.
2. Tests of the CAPM with Dynamic Approach
The literature of tests of CAPM with dynamic approach born with the need of analysis of non-diversifiable risk
in time domain, and also by the fact that the model coefficients aren’t stable.
Harvey (1989b) proposes a test of the capital asset pricing model that allows the conditional covariance vary in
time. Also according to the author, conditional covariances time-varying have been modeled with the ARCH-M
models (autoregressive conditional heteroskedasticity in the-mean model) of Engle et al. (1987). However, the
disadvantage of the ARCH-M model, according to Harvey (1989a), is that it is not aggregated, ie, what works for
a particular asset is not true for portfolios. So Harvey (1989a) chooses to use a new approach in which the
conditional covariances are approximated by the product of the innovations of linear forecasts for the returns
given the factors of the information set.
Harvey (1989b) uses monthly returns of NYSE stocks between 1941 and 1987 and the returns of U.S. Treasury
bonds as risk-free rate. As a proxy for the risk premium he uses the difference between the yields of BAA and
AAA treasuries, according to Moody’s. Harvey (1989b) divided into ten portfolios the assets surveyed. The risk
premiums, the return of dividends, among others, are used as instrumental variables. A dummy variable with a
value of 1 for the month of January was used as a control variable.
In ordinary least squares regression by Harvey (1989b), introduction of instrumental variables improves for
13.4% the model adjusting as compared to the result obtained by Campbell (1987) on the same sample. Then, the
author tests the hypothesis that the conditional covariances are constants. The previous regression residuals are
multiplied by the excesses of return of the assets and by the market excess return and regressed on the
instrumental variables. If the null hypothesis is true, only the intercept of the regression, which is the risk free
rate, should be significant and different from zero. For all portfolios, the null hypothesis is rejected, indicating
that the covariances vary over time.
The next step of the study is to verify the relationship between the market return and volatility for each stock
researched. The test is done using the following system:
  rmt  Z t 1 m ' 
 t  umt emt     (1)
rmt   * ( Z t 1 m ) 2  
'
 
The goal is to test whether the parameter  * that measures the “cost of risk”, is constant, and Z t 1 represents
the set of instrumental variables. This parameter indicates the expected return in excess of the market divided by
the standard deviation of the market. The parameter can be estimated by the equation:
 t  rmt   * mt2 (2)
Where  mt 2
is the volatility of the market, estimated by the procedure of Davidian and Carroll (1987). All the
tests carried out by the author also reject the null hypothesis, indicating that the parameter varies over time.
The next step of this research was to estimate the coefficient which measures the relative risk aversion, as well as
conditional covariance by the system:
165
  rjt  Z t 1 j  
'
 
 t  u jt umt e jt     rmt  Zt 1 m   (3)
'
 
  r  r Z  r Z 

 jt t 1 j   mt t 1 m  
'
 mt 
The results obtained by the researcher indicated that smaller firms had higher coefficient of relative risk
aversion. Following was made a comparison between models with risk aversion parameter constant and varying
in time. The author concludes that the model with constant parameter has better explanatory power of the data.
Ng (1991) tests the CAPM with multivariate GARCH approach to verify if the proxy of the market portfolio lies
on the border of mean-variance conditional, if the relationship in cross-section between risk premium of an asset
and their covariances are linear or proportional, and if the relationship between risk premium market and its
conditional variance is constant over time.
The study of the author cited above uses a multivariate GARCH process which allows conditional returns excess,
conditional variances and conditional covariances change over time. The study finds evidence that the price of
risk varies with the market volatility and the market risk premium is linearly related to the conditional variance
of the market with negative intercept, which is inconsistent with the proposition of Merton (1980) of CRRA type
preferences. However, the negative intercept is consistent with Bollerslev et al. (1988) and Harvey (1989b). The
work of Ng (1991), however, uses monthly data and a model with constant conditional correlation, which is not a
good assumption when working with financial data.
Wang (2003) presents a new test of the conditional CAPM, an extension of the work of Jagannathan and Wang
(1996) and the three-factor model of Fama and French (1993). Also according to Wang (2003), a dynamic
version of the CAPM has generally better performance than static models.
According to this author, dynamic models are attractive and challenging to be tested, mainly because they do not
have a guide showing how the betas vary with variables that represent conditioning information. The authors use
a nonparametric representation of the stochastic discount factor implicit in a conditional linear model of pricing
with factors. This methodology allows to perform tests that are free functional specifications bad of the dynamic
of conditional betas, of the risk premium and the discount stochastic factor. As results, the authors found that the
shape of nonparametric CAPM performs better than the unconditional CAPM and the conditional CAPM is
rejected statistically. Other implications of this research are that pricing errors has a strong pattern in volatility,
but not on average.
Also according to Wang (2003), a critical issue in the discussion of empirical tests of the CAPM is how to
measure expected returns that vary over time. Different ways of modeling the systematic risk with conditional
and non-conditional models can produce various opinions if the moment’s strategy and gains are consistent with
the time-varying expected returns. Also according to the author above, nonparametric tests can avoid the effects
of misspecification of the model, but usually the underlying nonparametric estimators converge more slowly
than parametric estimators for equilibrium.
The sample used by Wang (2003) consists of stocks traded on the NYSE between January 1947 and December
1995. The author divides the sample into four panels according to company size and the conditioning variables.
Then, the betas are estimated by means of a multivariate function to verify whether the betas are nonlinear. A
visual inspection of the graph of betas against time suggests that betas are not linear. The next step was to test the
nonlinearity of betas through the LM test of Andrews (1993) and following the model of explicit beta of Ghysels
(1998). In fact, the author notes that betas are not linear in relationship to returns, which gives evidence that the
conditional model is not valid.
The work of Kumar et.al (2008) investigates the role of information on the cross-section of asset returns when
investors are faced with uncertain information. For both, the authors use variables such as oil prices, market
volatility and exchange rates as variables to be tested.
Kumar et al. (2008) reject the hypothesis of CAPM that information is perfect and the prices converge to the
equilibrium even with the new information because, in their opinion, many market professionals explain market
movements in terms of “resolution of uncertainty “. Yet according to the authors, the assumption that investors
have complete information about the generating process of asset returns is clearly extreme. In reality, investors
are uncertain as to parameters that governing these processes, besides facing the risk inherent in the production
and investment. Another research problem concerns the quality of investor information and how to incorporate
them into their forecasting models.
166
To facilitate comparability with the standard CAPM, Kumar et. al (2008) constructed a model in which returns
are multivariate normal, but investors are uncertain as to the first and second moment of the joint distribution of
returns and information. A crucial implication of learning from information of uncertain quality is that both the
first and the second moment of the conditional expected returns are dependent on information, and therefore
stochastic. In equilibrium, assets are priced according to their intrinsic systematic risk. Furthermore, the market
risk premium and volatility are dependent on information. The hypothesis put forward by the authors is that the
portion of the change in aggregate risk of the market depends on macroeconomic factors such as the price of oil,
for example. A fluctuation in oil prices causes changes in cash flows of the companies and thus influences the
prices of their assets in the stock exchange.
To investigate the relationship between innovations in volatility and cross section of returns, Kumar et. al. (2008)
used the following specification:
Rat  R ft   a   a ( mt2 )   a age    ( mt2 )* age   a ( Rmt  R ft ) (4)
The coefficient a had a negative sign, indicating an inverse relationship between innovations in volatility and
excess return. In the following section, is tested if innovations in oil prices affect the returns via the equation:
R at  R ft   a   a O IL   ( R m t  R ft ) (5)
The results of Kumar et. al. (2008) show that increases in oil prices negatively affects the cash flows of the
companies that consume oil and not positively affects the cash flows of firms producing. The results of author
may suffer some sort of questioning, given that high oil prices could positively affect the cash flows of oil
companies. If the oil companies have a large weight in the composition of the stock index, which is the case of
Petrobras in Brazil, the aggregate result may be an improvement in market expectations.
Then, the authors simulate the effects of an innovation in the exchange rate may cause in the stock return of
firms exporters and importers. For importer firms there is an inverse relation and to the exporter firms the
parameters were also negative, but not significant.
The next step of the research of Kumar et. al. (2008) was to test whether the dispersion of opinions of market
analysts explain the cross section of excess returns. The authors found an inverse relationship between the
increase in the dispersion of analysts’ opinions and excess returns. Finally, we analyze the effect of the
announcement of share repurchases and dividend payments on the estimated betas of firms. There is a fall in the
average estimated betas when announced dividends and share repurchases, indicating fall in non-diversifiable
risk.
Mamaysky, Spiegel and Zhang (2008) use the Kalman Filter to estimate the trajectory of the betas of investment
funds. The method is justified by the inability of the static model of capture the dynamics of the market, an
important component of analysis by managers. The results indicate that funds follow strategies quite dynamic.
The betas estimated by Kalman filter are smoother and less prone to sudden changes when compared to betas
estimated by ordinary least squares.
Hueng and Huang (2008) investigated the asymmetrical relationship between risk and return using the CAPM
with time-varying betas. The authors specify with the following equation the time-varying betas:
R a t  R ft   a t   a t ( R m t  R ft )  v t
 a t   a t  1   t (6)
 a t   a t  1   t
The research of Hueng and Huang (2008) examined the daily closing prices of 358 assets comprising the S & P
500 from 1987 to 2003. The authors compose their portfolios according to market sectors which companies
belong. As a result, the authors find that there is indeed an asymmetrical relationship between risk and return, as
there is a considerable difference between static betas and average dynamic betas.
This research estimates the dynamic betas by two methods: DCC-Multivariate GARCH and Kalman Filter. The
estimation by two different methods aims to identify which types of betas have greater explanatory power of the
cross section of returns.
3. Data
Were researched the daily closing prices of the shares of Ibovespa traded throughout the period of Jan. 01, 1995
to March 20, 2012. The period of Ibovespa composition taken as a reference was the last day of the sample. All
167
prices were adjusted for dividends, splits and any other earnings. Also were researched the daily closing
Ibovespa as a proxy for the market return and the Selic rate as a proxy of the risk-free return. A total of 28 shares
were surveyed with 4493 observations each. The total sample, including Ibovespa and Selic, has the total of
134,790 observations. Then were calculated the compound return of the shares and the index, for thus, be
obtained the excess returns of the market and of each stock. Table 1 shows the surveyed companies and their
ticker on the BOVESPA.
Table 1. Companies and ticker

Company Type Ticker
Ambev PN ambv4
Bradesco PN bbdc4
Braskem PN brkm5
Brasil Telecom PN brto4
Banco do Brasil ON bbas3
Cemig PN cmig4
Souza Cruz ON cruz3
Siderúrgica Nacional ON csna3
Eletrobras ON elet3
Eletrobras PN elet6
Embraer ON embr3
Fibria ON fibr3
Gerdau PN ggbr4
Gerdau Metalúrgica PN goau4
Itaú Investimentos PN itsa4
Itaú Unibanco PN itub4
Klabin PN klbn4
Lojas Americanas PN lame4
Light ON ligt3
Lojas Renner ON lren3
Pão de Açúcar PN pcar4
Petrobrás ON petr3
Petrobrás PN petr4
Usiminas ON usim3
Usiminas PN usim5
Vale ON vale3
Vale PN vale5
Vivo PN vivt4
Source: Authors
4. Econometric Model
This session shows the Econometric Models used by research. The Models are the Multivariate Generalized
Autoregressive Conditional Heteroskedasticity (MGARCH) and Kalman Filter.
The betas estimated with MGARCH are called conditional betas. The model is known DCC (Dynamic
Conditional Correlation)-MGARCH because the correlations are time-varying. For each asset is estimated a
univariate model, but also for the Ibovespa and then are estimated the multivariate part of the model and the
unconditional correlation. With the outputs of MGARCH are calculated conditional variances of asset and
market, the conditional covariance between the asset and the market and unconditional correlation between the
two. To calculate the beta conditional splits the conditional covariance between the asset and the market by the
conditional variance of market. The Kalman filter estimates from the variance-covariance matrix and state space
representation of what is called structural betas. To obtain the structural betas will be estimated level local model
with the level of the variance fixed. It is expected that the structural betas have a behavior softer than the
conditional betas, given the characteristics of the cited methods which will be detailed below.
168
4.1 Estimation of the CAPM with Multivariate GARCH

The research uses the model MGARCH following the approach of Tse and Tsui (2002). This method allows the
estimation of dynamic betas from the conditional covariance and conditional variance obtained. The choice of
model is justified because the authors incorporate correlations varying in time, while satisfying the condition that
the conditional variance matrix is positive definite. The MGARCH model of Tse and Tsui (2002) is an
innovation model of Bollerslev (1988) and the model of Engle and Kroner (1995).
Considering rt , t  1,..., T the set of multivariate observations of excess returns of stocks and the market, each
with K elements, where rt  (r1t ,..., rKt ) . The conditional variance, assuming that rt time-varying, is defined
by:
Var (rt | t 1 )  t (7)
Where t represents the set of information at time t. The variance of the elements of t is represented by
 it , for i  1,..., K and the covariances are represented by  ijt , where 1  i  j  K . Defining Dt the
diagonal matrix K  K in which the ith diagonal element is  ijt , you can define  t  Dt1rt .  t represents
the standardized residual and it is assumed that it is IID with zero mean and variance matrix  t    ijt  .
Therefore, the correlation matrix for rt is denoted by t  Dt t Dt . The conditional variance follows
formulation vech-diagonal developed by Bollerslev (1988). Then, each term in the conditional variance follows a
univariate GARCH (p, q) given by equation:
p q
 it2  i   bih i2,t  h   aih ri 2,t  h , i  1,..., K (8)
h 1 h 1
p q
Where i , aih , bih are nonnegative and a  b
h 1
ih
h 1
ih  1 for all i  1,..., K . The conditional correlation matrix
varying in time is defined by the equation:

t  (1  1  2 )  1t 1   2 t 1 (9)
Where   ijt  is a positive definite matrix parameters, of size K  K , time variant with unitary diagonal
elements and  t 1 is a matrix which the elements are functions of the lagged observations of rt . The
parameters 1 and  2 are nonnegative and it is assumed that the restriction of which 1   2  1 .
It is observed that  t 1 is analogous to ri ,t 1 of GARCH (1,1). However, with  t is, according to Tse and
2
Tsui (2002), a standardized measure,  t 1 need depend on standardized residuals lagged t . Defining
 t   ijt  ,  t 1 follows the specification:
M
 i ,t  h  j ,t  h
 ij ,t 1  h 1
, 1 i  j  K (10)
 M M

   i ,t  h   j ,t  h 
 h 1 h 1 
 t 1 is the correlation matrix of  t 1 ,...,  t  M  . Defining t 1 a matrix K  M given by t 1   t 1 ,...,  t  M  .
1/ 2
If Bt 1 , is a diagonal matrix K  K in where the ith diagonal element is    i2,t  h 
M
for i  1,..., K , we
 
 h 1 
have:
 t 1  Bt11 t 1 t' 1 Bt11 (11)
The conditional log-likelihood  t of the observation rt is given by:
169
1 1
 t   ln | Dt t Dt |  rt ' Dt1t1 Dt1rt
2 2 (12)
1 1 K 1
 t   ln |  t |   ln  it2  rt ' Dt1t1 Dt1rt
2 2 i 1 2
Define   (1 , a11 ,..., a1q , b11 ,..., b1 p , 2 ,..., aKq , 12 ,...,  K 1, K ,1 , 2 ) as the vector of parameters and
maximizing  in relation to  , we have ˆ , where    t .

4.2 Estimation of the CAPM with Kalman Filter
The CAPM with beta time-varying by making the hypothesis that the risk-free rate is zero is defined by:
rt  t r 'mt   t (13)
The beta time-varying, according to Hamilton (1994), is given by the equation:
(t 1   )  F (t   )  vt 1 (14)
If the eigenvalues of the matrix F ( k  k ) are inside the unit circle, so  is the vector of mean values of
coefficients. Thus, the state vector will be t  t   and (13) can be rewritten as follows:
rt   r 'mt  t r 'mt   t (15)
Define t1 as a set of conditional information to rmt in t-1 and Pt|t 1 as the variance-covariance matrix of the
process, the maximum likelihood function is represented by the equation:
T
T  1 T

t 1
log f (rt | rmt , t 1 )    
 
2
log(2 )     log(r 'mt Pt|t 1rmt   )
  t 1
2
2
(16)
 1  (r   r '   r ' )
T 2
    t ' mt t \t 1 2 mt
 2  t 1 (rmt Pt|t 1rmt   )
5. Results
Table I.1 in appendix I presents descriptive statistics for the excess returns of assets surveyed, for the excess
market return and the Selic rate.
It is observed that thirteen of assets and the Ibovespa don’t have excess return positive. The analysis table
identifies some “stylized facts” of financial time series described by Daníelsson (2011). The serial
autocorrelation found in series through the test Q indicates that there are clusters of volatility, in other words,
periods of high volatility which are preceded by periods of low volatility. The values of the autocorrelations of
squared excess returns show that there is nonlinear dependence of the series, because the square of the excess
return has a larger structure than the excess return. Finally, the value of kurtosis greater than three found in all
the asset is signal that distributions have fat tails and, therefore don’t follow a normal distribution.
The first procedure to test the CAPM with dynamic betas is the estimation of own betas. The estimation is
performed with DCC-MGARCH model, developed by Tse and Tsui (2002). In the mean equation were estimated
vector autoregressive to remove the serial correlation and used the residuals of this estimation. The model
provides the conditional covariances and conditional variances. To retrieve the conditional betas were divided
the conditional covariance between the excess returns of the assets and excess return of the market by the
conditional variance of excess return of the market. In the financial market, it is known that the market rises with
low volatility and falls with high volatility. When the series exhibit this asymmetry is used TGARCH model
developed by Glosten et al. (1993). This model uses a dummy variable equal to 1 if the return of innovation in
variance equation is negative. Thus, greater weight is given to negative returns. For each asset models were
estimated with and without asymmetry and chosen the best model according to the Schwarz information
criterion. Table 2 shows the results.
170
Table 2. DCC-MGARCH model

Asset   10 4 a1 a2 b1 b2 1 2 1 2  i ,ibov
ambv4 0.08* 0.21 -0.15 0.92 0.94 0.02 0.51
ibov 0.13 0.10 0.02 0.84
bbdc4 0.13 0.08 -0.04 0.89 0.14 -0.04* 0.96 0.02 0.86
ibov 0.15 -0.03 0.07 0.85 0.22 -0.07
brkm5 0.10* 0.15 -0.10 0.92 -0.02* 0.07* 0.97 0.01* 0.67
ibov 0.14 -0.03 0.05 0.86 0.23 -0.07
brto4 0.04* 0.08 -0.07 1.33 -0.36* 0.09* -0.07* 0.96 0.02 0.75
ibov 0.01 -0.03 0.04 1.63 -0.65 0.28 -0.26
bbas3 0.02 0.16 -0.15 1.41 -0.43 0.97 0.01 0.76
ibov 0.17 0.09 0.07 0.42 0.36
cmig4 0.11 0.04 0.90 0.06 0.98 0.01 0.82
ibov 0.14 0.02* 0.86 0.14
cruz3 0.06* 0.19 -0.13 0.93 0.97 0.01** 0.51
ibov 0.12 0.10 0.01 0.85
csna3 0.10 0.07 -0.05 1.12 -0.18 0.14 -0.09 0.97 0.01 0.86
ibov 0.12 -0.03 0.06 1.02 -0.14 0.22 -0.09
elet3 0.08* 0.18 -0.10 0.91 0.98 0.01 0.89
ibov 0.12 0.09 0.02 0.85
elet6 0.04* 0.07 -0.05 1.30 -0.33 0.14 -0.12 0.98 0.01 0.89
ibov 0.01 -0.03 0.04 1.62 -0.65 0.27 -0.26
embr3 0.06 0.14 0.87 0.98 0.01 0.43
ibov 0.12 0.11 0.85
fibr3 0.07 0.15 -0.14 1.44 -0.46 0.98 0.01 0.68
ibov 0.18 0.10 0.07 0.30 0.47
ggbr4 0.58 0.06 0.81 0.09 0.97 0.02 0.89
ibov 0.14 0.02 0.87 0.14
goau4 0.18 0.10 -0.07 0.90 0.12 -0.03 0.97 0.02 0.26
ibov 0.12 -0.03 0.05 0.87 0.24 -0.11
itsa4 0.09 0.02 0.91 0.08 0.97 0.02 0.84
ibov 0.14 0.02 0.87 0.14
itub4 0.10* 0.06 -0.04 0.92 0.12 -0.03* 0.96 0.02 0.85
ibov 0.12 -0.03 0.05 0.87 0.25 -0.13
klbn4 0.42 0.09 0.82 0.08 0.98 0.01* 0.66
ibov 0.14 0.02 0.86 0.15
lame4 0.02 0.15 -0.14 1.47 -1.49 0.05 -0.03 0.98 0.01 0.65
ibov 0.03 -0.03 0.04 1.46 -0.50 0.26 -0.22
ligt3 0.04* 0.19 -0.14 0.94 0.96 0.02 0.64
ibov 0.12 0.10 0.01 0.85
lren3 0.03* 0.08* -0.07 1.47 -0.49 0.19 -0.17 0.76 0.03* 0.55
ibov 0.00 0.07 -0.06 1.72 -0.74 0.28 -0.26
pcar4 0.01* 0.05* 0.93 0.04 0.97 0.01 0.51
ibov 0.08 0.01 0.89 0.15
petr3 0.13* 0.04 0.90 0.08 0.97 0.01 0.81
ibov 0.08 0.01 0.88 0.15
petr4 0.20 0.03 0.86 0.12 0.97 0.01 0.88
ibov 0.08 0.01 0.88 0.15
usim3 0.10* 0.63 -0.47 1.14 -0.25 0.97 0.02 0.59
ibov 0.18 0.10 0.08 0.35** 0.41*
usim5 0.18* 0.08* -0.05* 0.90 0.09* -0.03* 0.97 0.02 0.84
ibov 0.13 -0.03 0.06 0.86 0.22 -0.09
vale3 0.14 0.05 0.89 0.06 0.98 0.01 0.39
ibov 0.14 0.02* 0.86 0.14
vale5 0.00 0.16 -0.15 1.69 -0.71 0.97 0.01 0.93
ibov 0.00 0.13 -0.12 1.76 -0.77
vivt4 0.17* 0.06 0.86 0.08 0.97 0.01 0.80
ibov 0.14 0.01* 0.87 0.14
Source: Authors.
Notes: *, ** and *** indicate no significance at levels of 1%, 5% and 10%, respectively.
171
The estimated parameters were significant at the 5% level and all the conditional correlations were significant at
1%. In only nine of twenty-eight assets researched the best estimated model showed no asymmetry. The
multivariate estimation obtained better performance with errors following the t-student distribution for all pairs of
assets. The asset of greatest unconditional correlation with the Ibovespa was the preferred share of Vale, which is
justified by the same have a great weight in the composition of the index. The stock that had the lowest
unconditional correlation was preferred share of Gerdau Metalúrgica. Shares of banks surveyed have
unconditional correlation high with the Ibovespa.
The parameters of asymmetry estimated are consistent with the theory, indicating that there is volatility increase if
the return is negative. To control residuals did not need more than two lags of DCC-MGARCH. To test whether
residuals aren´t autocorrelated was used the Q test of Ljung Box, as shown the Table 3.
Table 3. Diagnostic checking for DCC-MGARCH
Asset Q1 (5) Q1 (10) Q1 (15) Q1 (20) Q2 (5) Q2 (10) Q2 (15) Q2 (20)

ambv4 5.14 8.85 13.77 16.48 3.67 6.97 9.13 12.51
ibov 14.17 21.68 30.49 41.46 9.66 24.55 31.65 39.39
bbdc4 5.70 12.26 18.37 31.92 7.00 14.87 17.43 20.39
ibov 7.98 15.57 18.71 24.69 4.20 12.35 14.77 15.23
brkm5 2.39 10.05 19.73 21.69 6.68 11.03 19.32 26.44
ibov 7.91 9.45 12.76 16.22 10.97 25.28 29.18 30.26
brto4 1.11 4.95 8.55 16.81 1.72 3.42 4.17 8.69
ibov 18.95 28.64 37.19 42.77 9.80 19.93 26.05 30.11
bbas3 4.08 8.82 15.66 19.67 1.96 5.99 16.29 24.29
ibov 12.36 20.35 26.18 32.27 6.61 22.94 25.76 30.94
cmig4 0.63 3.35 10.30 12.85 3.27 4.19 8.66 17.23
ibov 8.04 14.10 24.66 30.27 15.20 24.68 27.44 28.19
cruz3 0.57 3.21 5.73 9.99 1.46 12.30 18.47 27.98
ibov 9.63 16.90 22.99 31.89 19.91 36.50 38.89 42.32
csna3 14.01 23.57 28.68 33.07 5.65 12.24 25.95 29.59
ibov 11.65 16.91 23.03 26.50 22.90 31.75 34.80 35.30
elet3 4.85 7.48 13.53 18.62 7.74 14.97 24.02 29.88
ibov 13.38 20.35 38.64 30.33 8.53 21.30 34.85 30.53
elet6 6.53 9.83 12.58 16.61 8.85 10.07 13.44 14.36
ibov 16.27 25.24 36.28 40.90 2.80 8.35 11.14 12,65
embr3 11.19 31.86 39.65 64.59 0.25 0.68 0.99 1.70
ibov 9.49 16.46 23.81 31.39 15.99 31.40 34.70 40.02
fibr3 1.67 2.80 7.18 18.98 4.87 17.28 21.11 24.64
ibov 13.68 22.49 26.79 33.57 5.25 14.18 18.56 21.67
ggbr4 7.26 8.49 10.98 16.94 9.54 11.68 13.15 20.89
ibov 21.89 28.47 39.41 50.21 21.91 30.59 44.86 47.74
goau4 8.28 9.44 13.59 25.44 8.03 11.53 17.39 20.91
ibov 14.07 24.69 27.34 28.94 15.45 25.61 36.52 38.72
itsa4 1.13 6.35 11.62 16.09 3.12 6.66 13.33 14.81
ibov 13.28 22.73 32.87 37.12 6.40 23.73 28.88 30.20
itub4 2.13 4.55 8.24 25.35 6.90 20.02 24.86 29.90
ibov 14.20 23.36 25.75 32.05 11.96 20.34 21.56 26.70
klbn4 4.54 6.37 11.29 15.30 10.17 12.73 14.68 15.79
ibov 8.96 17.98 23.43 26.74 8.73 22.92 26.16 26.79
lame4 3.80 10.43 12.60 17.54 2.46 13.29 16.01 21.56
ibov 12.42 17.84 24.79 26.34 21.58 33.43 41.03 43.87
ligt3 2.66 3.85 6.13 8.13 2.74 7.52 12.19 18.01
ibov 8.25 13.98 18.94 23.01 9.88 25.00 26.96 30.78
lren3 26.67 28.53 30.80 35.06 10.44 14.95 19.21 26.63
ibov 2.78 7.95 11.00 14.42 1.35 5.81 7.15 10.82
172
pcar4 2.50 10.33 13.44 15.41 10.14 15.81 17.48 19.67

ibov 1.74 7.95 8.49 18.68 2.82 8.68 9.96 15.14
petr3 2.08 6.28 10.38 13.18 5.41 10.40 12.91 14.64
ibov 2.70 10.08 11.77 21.05 2.14 14.00 16.63 20.95
petr4 9.88 12.41 14.42 16.08 5.08 7.48 15.68 17.48
ibov 2.20 9.45 13.59 20.96 3.10 11.87 17.76 23.77
usim3 26.85 33.92 36.82 39.51 1.78 3.73 4.84 9.78
ibov 7.91 14.07 21.23 28.37 20.54 32.90 38.69 43.52
usim5 1.15 3.97 8.33 10.24 5.86 9.90 11.66 20.27
ibov 10.29 22.09 29.49 34.75 18.88 27.77 32.05 33.70
vale3 7.56 10.41 15.43 16.61 7.66 19.68 22.09 23.95
ibov 14.20 25.93 36.46 44.04 13.98 23.34 31.53 33.20
vale5 1.30 4.57 9.06 15.97 12.46 15.11 17.73 18.44
ibov 12.76 22.96 37.74 42.33 5.71 22.36 33.30 35.22
vivt4 5.21 7.42 9.68 14.27 9.81 22.02 24.06 28.37
ibov 9.69 14.89 17.48 19.58 6.91 12.90 16.91 23.47
Source: Authors.
The next step of the study was to estimate the structural betas through a level local model with variance of the
level fixed and time-varying betas. Estimates of residuals and diagnosis of the structural model can be found in
Appendix I. The structural model estimates the betas smoothed, as shown in Figure 1 (see Appendix II). The
betas estimated by two methods are tested to see if they have explanatory power on average excess returns in
cross section.
The betas estimated (see table 4), except ELET6, ELET3 and USIM5 are smaller than 1, characterizing that the
stocks are considered defensive for investors. This result was expected and is in line with the results found by
Paiva (2005), given that the shares that remain long in Ibovespa are considered safer by agents.
The structural beta follows the movements of the conditional betas. The conditional betas have more “nervous”
behavior that the structural betas. Besides the structural and conditional betas, the conditional variances were
estimated for each asset searched. The analysis of the conditional variance is important in the financial market,
because it is from the same which is calculated VaR (Value at Risk), another widely used measure of risk. Then
we calculated the conditional correlations between each asset and the Ibovespa, as shown Figure 3. The study of
correlation time-varying allows the investor to know what stock is more impacted event of a change in market
index and, thus, make their hedging strategies. There are times wherein the stocks of Renner (LREN3), Vale
(VALE3) and Embraer (EMBR3) have a negative correlation with the Ibovespa, indicating a fall in the index
increases the price of these stocks. The stock of Vivo (vivt4) has a decreasing correlation with the Ibovespa over
time, from an average of 0.75 at the beginning of the series to 0.50 in recent times.
Table 4 shows the average betas estimated by the two models. Note that, on average the betas estimated by the
two models have similar values, although their trajectories are often distinct.
Table 4. Average dynamic betas

Asset Average Conditional Betas Average Structural Betas
ambv4 0.49 0.50
bbdc4 0.85 0.85
brkm5 0.77 0.77
brto4 0.93 0.89
bbas3 0.82 0.43
cmig4 0.89 0.89
cruz3 0.52 0.52
csna3 0.89 0.89
elet3 1.05 1.03
elet6 1.08 1.04
embr3 0.48 0.50
fibr3 0.65 0.66
ggbr4 0.91 0.90
173
goau4 0.68 0.81

itsa4 0.79 0.80
itub4 0.82 0.82
klbn4 0.73 0.74
lame4 0.69 0.68
ligt3 0.81 0.79
lren3 0.95 0.94
pcar4 0.49 0.50
petr3 0.91 0.89
petr4 0.92 0.91
usim3 0.58 0.58
usim5 1.04 1.05
vale3 0.60 0.73
vale5 0.82 0.82
vivt4 0.71 0.70
Source: Authors.
Then it was verified whether the average conditional betas and average structural betas explain the cross section
of excess returns. The results are shown in Table 5.
Table 5. Estimation of the equation of cross section of excess returns with conditional betas
Ra   0  1 a   2  a2   3 Sea   a
Model
0 1 2 3 Q1 (5) Q2 (5) JB Test (p-value)
1 -0.0066** 0.0210* -0.0135** -0.0003*** 8.86 2.35 0.98

2 -0.0059 0.0163 -0.01058 1.43 2.67 0.15
3 0.0014* -0.0018* 8.43 5.84 0.43
4 0.0005*** 11.96 1.31 0.30
Source: Authors.
Notes: *, ** and *** indicate no significance at levels of 1%, 5% and 10%, respectively.
The survey found that the conditional beta has better explanatory power of returns than the betas estimated by
the structural model. However, when tested without the intercept, the conditional beta wasn’t significant,
showing that the CAPM fails on many occasions and not always the asset at higher non-diversifiable risk offers
the highest expected return. Therefore, the research concludes that, for the Brazilian stock market in the period
surveyed there is no evidence that the action with greater non-diversifiable risk represented by beta offered
higher return to investors.
The structural betas don’t explain the cross section of returns, showing that the CAPM, tested using the
methodology of Fama and McBeth (1973) doesn’t fit well to the stock market in Brazil.
One of the research findings is that the constant beta serves only for reference if action follows or not the
market’s performance in a given period, but not enough to serve as a good measure of risk. As demonstrated, the
beta isn’t constant. The dynamic betas are a better approximation of market risk as they increase in times of
crisis and increased risk aversion.
The main contribution of this research is the estimation of conditional betas. It is important the use of alternative
risk measure that not only VaR, which is the most used on the market. The conditional beta, unlike VaR, uses the
conditional variance and conditional covariance between the market and active. The disadvantage of the beta to
the VaR is that the beta is not expressed in monetary value, but as a sensitivity coefficient of share in relation to
the market.
Another contribution of estimation of dynamic betas for the Brazilian market is the use of different weights for
different stocks in Long-Short. Assuming that betas vary, the investor can change each time the position of each
asset in your portfolio, not just applying equal values in each of the assets.
Unlike Hueng and Huang (2008), this work did not find an asymmetric relation between risk and return, because
the dynamic betas estimated have values close to the static betas estimated for Brazilian market by Godeiro
174
(2012).
The results are in line with Flister et al. (2011), in that time-varying betas provide a better fit for the CAPM. The
work Tambosi Filho et al. (2007) also points to the validity of the conditional CAPM.
Given the failure of beta in provide a reliable measure of risk, the research measured the relationship between the
market conditional variances, estimated by DCC-MGARCH and excess return of assets. The goal is to estimate
the “cost of risk”, in other words is to check in as increase the risk when it increases the expected return.
The parameters of premium and cost of risk were estimated following the methodology developed by Harvey
(1989b). The author assesses the relationship between the innovations of the excess return and the variance or
standard deviation of market. Also according to Harvey (1989b) the parameters of risk premium represents the
risk aversion of the agents. To retrieve the parameters were used the conditional variance and standard deviation
conditional estimated by MGARCH, as well as innovations in excess returns of assets and market. It is expected
that an inverse relationship occurs between excess return and volatility, indicating that an increase in the
volatility decreases excess return expected.
Table 6. Parameters of relative risk aversion

 it  rmt   mt2
 it  rmt   * mt
Asset  Q1 (10) Q2 (10) * Q1 (10) Q2 (10)
ambv4 0.1087*** 8.25 15.35 0.2174*** 8.25 15.35
bbdc4 0.6176* 10.29 14.14 1.3793* 10.10 12.42
brkm5 1.3341 2.18 10.11 2.8095 15.48 9.51
brto4 0.5079*** 3.56 12.37 0.9786*** 7.60 8.40
bbas3 1.1581 16.11 8.26 2.3162 16.11 15.35
cmig4 -2.8183 13.63 17.07 -4.7073 13.38 19.58
cruz3 0.4732*** 2.82 6.29 0.8864*** 2.81 6.39
csna3 0.8249** 16.88 5.55 1.6226** 16.89 5.54
elet3 1.5975* 22.09 10.41 3.1424** 22.12 10.36
elet6 -1.9075 30.21 6.08 -3.9365 30.53 6.26
embr3 1.4833 17.73 8.82 3.3816 17.11 8.59
fibr3 1.8182 4.84 9.95 3.4090 5.24 10.27
ggbr4 -2.2992 18.41 33.64 -4.0567 18.59 30.85
goau4 -0.5911** 8.61 4.77 1.1823** 8.61 4.77
itsa4 -0.7033* 5.58 11.97 -1.6294* 5.29 7.05
itub4 -5.0057 11.33 14.51 9.4190 11.03 14.14
klbn4 -3.4198 15.92 10.40 -4.2357 15.44 8.79
lame4 -4.3396 13.17 19.77 -9.2651 13.25 19.75
ligt3 1.3684 5.57 13.75 2.3793* 5.61 13.18
lren3 -6.9712 16.54 10.74 -15.8659 15.77 10.06
pcar4 1.2336 12.16 14.42 1.9296* 12.48 14.31
petr3 -2.7105 10.23 16.79 -5.4485 11.12 18.51
petr4 -2.9347 17.47 11.61 -5.6543 19.05 11.20
usim3 2.2612 4.33 5.73 4.7372 4.10 5.74
usim5 1.1078 4.38 6.09 2.5264 4.48 6.99
vale3 -2.7735 16.31 25.75 -4.3373 14.66 29.53
vale5 -1.6266 18.03 15.55 -3.2533* 18.03 15.55
vivt4 -2.4876 12.68 18.03 -5.5686 12.54 17.30
Source: Authors.
Notes: *, ** and *** indicate no significance at levels of 1%, 5% and 10% respectively.
Fourteen of the twenty-eight Ibovespa stocks analyzed presented results in accordance with the results of Kumar
et. al (2008), since these actions had an inverse relationship between return and volatility. This type of behavior
is not typical of stocks classified as defensive by market agents.
The analysis of Table 6 provides the identification of some facts. Assets with the coefficients of relative risk
175
aversion negatives are a good alternative in times of increased in volatility, because they have a direct
relationship with the conditional variance. These actions, as shown in the table are shares of the electricity sector,
as CMIG4 and shares that have a large weight in the index, as Petrobras and Vale. This is because in times of
crisis and increased risk aversion there is a change in the preference basket of asset of investors. They tend to
decrease positions in stocks Small Caps and increasing position in treasuries or shares considered safer by the
market. The ordinary shares of Usiminas had the highest inverse relation with volatility, indicating that they are
significantly impaired when there is increase in risk aversion. Therefore, one arrives at the conclusion that the
conditional variance has a good explanatory power on the excess return of assets and the methodology developed
by Harvey (1989b) has a good performance in the Brazilian stock market, unlike the methodology of Fama and
Mcbeth (1973) who did not achieve a satisfactory result.
6. Conclusions
The research proposed a test of the capital asset pricing model for the Brazilian stock market using the
MGARCH model developed by Tse and Tsui (2002). For comparison were also obtained estimates of betas
smoothed in time. It was found that the estimated betas have a significant increase in times of crisis, indicating
increase in non-diversifiable risk these periods. The increased risk occurs because of the loss of attractiveness of
equities papers and hence the fall in their prices.
The volatility of twenty-eight asset surveyed was estimated from the DCC-MGARCH model, verifying that the
crisis of 2008 is the period when risk aversion reaches its peak, because in this period occur the highest level of
conditional variance. Stocks of companies like Bradesco, Petrobras and Vale had a lower volatility than other
companies in the sub-prime crisis, which strengthens the argument that these companies are defensive because
are most demanded in times of crisis. The parameters of relative risk aversion estimated also suggest the same
conclusion.
The test methodology of Fama and Mcbeth (1973) isn’t valid for Brazil in the period surveyed, since there was a
no significant relationship between excess returns and betas, ie, stocks with higher non-diversifiable risk weren’t
the stocks with greatest excess return. The parameters that measure relative risk aversion according to Harvey
(1989b) were significant and are valid as a measure of risk for Brazil during the interval studied.
The estimate of conditional correlations shows relevant information. Despite the Vale (VALE5) preference stock
have the highest unconditional correlation with the Ibovespa; analyzing the conditional correlation appears that
there were periods in which the correlation between the variables mentioned was less than 0.20. Likewise, the
share of Gerdau Metalúrgica (GOAU4) had the lowest unconditional correlation with the market index, when
dynamic analysis is done there are times when the conditional correlation comes close to one, indicating that the
share follows the movements of market in these intervals.
It is important when working with financial data, the modeling of stylized facts. This research attempted to do
this by not only estimating a model for the conditional mean, but also for the conditional variances. The stylized
fact of fat tails was modeled using t-distribution. Regarding the asymmetry was used GJR model, aiming to
replicate the fact that the risk averse agent gives greater weight to negative than positive results in its utility
function.
For future research we suggest that betas are estimated by various methods offered by MGARCH models, such
as CCC, BEKK and DCC of Engle himself with checking if they have the power to explain the cross section of
excess returns, and also can be analyzed assets contained in other indexes besides the Ibovespa.
References
Alencastro, D. (2009). Empirical analysis of the basic CAPM to Brazil after the implementation of the Real Plan.
Dissertation (p. 74). Porto Alegre: PUCRS.
Andrews, D. W. K. (1993). Tests for parameter instability and structural change with unknown change point.
Econometrica, 61, 821-856. http://dx.doi.org/10.2307/2951764
Bollerslev, T., Engle, R. F., & Wooldridge, J. A. (1988). capital asset pricing model with time varying
covariances. Journal of Political Economy, 96, 116-131. http://dx.doi.org/10.1086/261527
Campbell, J. Y. (1987). Stock returns and term structure. Journal of Financial Economics, 18, 373-400.
http://dx.doi.org/10.1016/0304-405X(87)90045-6
Campbell, J. Y., Lo, A. W., & Mackinlay, A. C. (1997). The Econometrics of Financial Markets (p. 651).
Princeton, NJ: Princeton University Press.
Danielsson, J. (2011). Financial Risk Forecasting (p. 325). New York: Wiley Finance.
176
Davidian, M., & Carrol, R. J. (1987). Variance function estimation. Journal of the American Statistical
Association, 82, 179-1091. http://dx.doi.org/10.1080/01621459.1987.10478543
Engle, R. F., & Kroner, K. F. (1995). Multivariate Simultaneous Generalized ARCH. Econometric Theory, 11,
122-150. http://dx.doi.org/10.1017/S0266466600009063
Fama, E. F., & Macbeth, J. D. (1973). Risk, return and equilibrium: empirical tests. Journal of Political
Economy, 81, 607-636. http://dx.doi.org/10.1086/260061
Flister, F. V., Bressan, A. A., & Amaral, H. F. (2011). Conditional CAPM in the Brazilian Market: A Study of the
Effects Momentum, Size and Book-to-Market between 1995 and 2008. Brazilian Review of Finance, 9(1),
105-129.
Ghysels, E. (1998). On stable factor structures in the pricing of risk: do time varying betas help or hurt. Journal
of Finance, 53, 549-574. http://dx.doi.org/10.1111/0022-1082.224803
Glosten, L. R., Jaganathan, R., & Runkle, D. (1993). On the relation between the expected value and the
volatility of the normal excess return on stocks. Journal of Finance, 48, 1779-1801.
Godeiro, L. L. (2012). An empirical application of the CAPM to the Brazilian capital market (p. 60).
(Dissertation). São Paulo: PUCSP.
Hamilton, J. D. (1994). Time Series Analysis (p. 783). Princeton, NJ: Princeton University Press.
Harvey, C. (1989a). Time-Varying Conditional Covariances in Tests of Asset Pricing Models. Journal of
Financial Economics, 24, 289-317. http://dx.doi.org/10.1016/0304-405X(89)90049-4
Harvey, C. (1989b). Is expected compensation for market volatility constant through time? Working paper, Duke
University.
Huang, P., & Hueng, C. J. (2008). Conditional Risk-Return Relationship in a Time-Varying Beta Model.
Quantitative Finance, 8, 381-390. http://dx.doi.org/10.1080/14697680701191361
Jagannathan, R., & Wang, Z. (1996). The Conditional CAPM and the Cross-Section of Expected Returns.
Journal of Finance, 51(1), 3-53. http://dx.doi.org/10.1111/j.1540-6261.1996.tb05201.x
Kumar, P., Sorescu, S. M., Boehme, R. D., & Danielsen, B. R. (2008). Estimation Risk, Information, and the
Conditional CAPM: Theory and Evidence. Review of Financial Studies, 21(3), 1037-1075.
http://dx.doi.org/10.1093/rfs/hhn016
Mamaysky, R., Spiegel, M. E., & Zhang, H. (2008). Estimating the Dynamics of Mutual Fund Alphas and Betas.
Review of Financial Studies, 21(1), 233-264. http://dx.doi.org/10.1093/rfs/hhm049
Markowitz, H. (1952). Portfolio Selection. Journal of Finance, 7(1), 77-91.
Mas-Colell, A., Whinston, M. D., & Green, J. R. (1995). Microeconomic Theory (p. 985). Oxford Oxford
University Press.
Merton, R. C. (1980). On estimating the expected return on the market: an exploratory investigation. Journal of
Financial Economics, 8, 323-361. http://dx.doi.org/10.1016/0304-405X(80)90007-0
Ng, L. (1991). Tests of the CAPM with Time-Varying Covariances: A Multivariate GARCH Approach. The
Journal of Finance, 46, 1507-1521. http://dx.doi.org/10.1111/j.1540-6261.1991.tb04628.x
Paiva, F. D. (2005). Asset Pricing Models of Financial Single Factor: An Empirical Test of the D-CAPM and
CAPM. Research and Administration Books, 12, 49-65.
Sharpe, W. F. (1964). Capital Asset Prices: A Theory of Market Equilibrium Under Conditions of Risk. Journal
of Finance, 19, 425-442.
Tambosi Filho, E., Garcia, F. G., & Bertucci, L. A. (2007). Empirically Testing the CAPM conditional of
expected returns of portfolios of Brazilian, Argentine and American market. Journal of Management USP,
14(4), 63-75.
Tse, Y. K., & Tsui, A. K. C. (2002). A Multivariate Generalized Autoregressive Conditional Heteroscedasticity
Model with Time-Varying Correlations. Journal of Business & Economic Statistics, 20(3), 351-36.
Varian, H. R. (1992). Microeconomic Analysis (p. 559). New York: W. W. Norton & Company Inc.
Wang, K. Q. (2003). Asset Pricing with Conditioning Information: A New Test. Journal of Finance, 58,
177
Appendix I
Table I.1. Descriptive statistics of daily data of excess return of assets, the excess return of the market and the
selic rate
Asset Mean Standard Min Max Skewness Kurtosis AC (One lag) AC (One lag) JB LB LB squared of returns
Deviation returns of squared returns (p-value) 20 lags (p-value) 20 lags (p-value)
selic 0.0694% 0.03% 0.03% 0.21% 1.80 6.57 0.99 0.99 0.00 0.00 0.00
ibov -0.0083% 2.22% -17.34% 28.71% 0.45 17.05 0.03 0.21 0.00 0.00 0.00
ambv4 0.0419% 2.18% -17.81% 16.02% -0.16 10.67 0.08 0.22 0.00 0.00 0.00
bbdc4 0.0173% 2.52% -21.79% 28.58% 0.27 11.95 0.08 0.26 0.00 0.00 0.00
brkm5 -0.0343% 2.75% -15.47% 19.21% 0.13 6.32 0.10 0.18 0.00 0.00 0.00
brto4 -0.0283% 2.93% -18.06% 20.90% 0.06 7.41 0.05 0.18 0.00 0.00 0.00
bbas3 -0.0105% 2.83% -16.73% 18.77% 0.15 6.49 0.00 0.20 0.00 0.01 0.00
cmig4 0.0062% 2.78% -27.91% 26.29% 0.26 12.20 0.08 0.21 0.00 0.00 0.00
cruz3 0.0438% 2.29% -17.12% 16,30% 0.07 6.74 -0.01 0.18 0.00 0.00 0.00
csna3 0.0363% 2.82% -18.82% 19.58% 0.09 7.35 0.07 0.22 0.00 0.00 0.00
elet3 -0.0473% 3.15% -17.92% 28.30% 0.48 9.79 0.10 0.26 0.00 0.00 0.00
elet6 -0.0319% 3.12% -19.14% 32.44% 0.64 11.81 0.06 0.26 0.00 0.00 0.00
embr3 -0.0370% 2.92% -38.46% 29.95% 0.48 24.09 0.15 0.20 0.00 0.00 0.00
fibr3 -0.0456% 2.73% -19.10% 17.59% 0.35 7.67 0.09 0.23 0.00 0.00 0.00
ggbr4 0.0277% 2.84% -16.19% 20.83% 0.20 6.91 0.11 0.22 0.00 0.00 0.00
goau4 0.0322% 2.68% -27.63% 21.62% -0.03 9.86 0.07 0.24 0.00 0.00 0.00
itsa4 0.0284% 2.43% -21.37% 22.38% 0.25 9.56 0.05 0.22 0.00 0.00 0.00
itub4 0.0284% 2.47% -15.79% 20.95% 0.32 7.65 0.11 0.19 0.00 0.00 0.00
klbn4 -0.0086% 2.97% -15.31% 19.93% 0.48 7.26 -0.05 0.19 0.00 0.03 0.00
lame4 0.0201% 3.01% -17.45% 24.67% 0.48 8.34 0.04 0.18 0.00 0.00 0.00
ligt3 -0.0822% 3.15% -26.35% 24.02% 0.19 9.86 0.06 0.21 0.00 0.00 0.00
lren3 0.0396% 2.13% -26.90% 40.39% 1.89 52.64 0.08 0.37 0.00 0.00 0.00
pcar4 -0.0023% 2.47% -24.62% 30.90% 0.41 20.32 0.05 0.25 0.00 0.00 0.00
petr3 0.0210% 2.78% -22.43% 20.37% -0.11 9.78 0.08 0.24 0.00 0.00 0.00
petr4 0.0114% 2.64% -21.26% 20.95% -0.11 10.84 0.09 0.33 0.00 0.00 0.00
usim3 -0.0050% 2.81% -19.31% 37.39% 1.13 19.66 0.11 0.18 0.00 0.00 0.00
usim5 -0.0084% 3.06% -18.28% 16.58% 0.03 6.06 0.10 0.20 0.00 0.00 0.00
vale3 0.0083% 2.56% -20.60% 29.75% 0.30 12.82 0.03 0.08 0.00 0.00 0.00
vale5 0.0195% 2.56% -16.49% 38.40% 0.83 18.24 0.03 0.10 0.00 0.00 0.00
vivt4 -0.0033% 2.63% -20.68% 40.77% 0.90 25.03 0.03 0.25 0.00 0.00 0.00
Source: Authors.
Table I.2. Estimation of equation of excess returns in cross section with structural betas
Ra   0  1 a   2  a2   3 Sea   a
Model 0 1 2 3 Q1 (5) Q2 (5) JB test (p-value)
1 -0.0736*** 0.2951*** -0.2034*** -0.0094*** 8.93 1.56 0.21

2 -0.0690*** 0.2115*** -0.1488*** 11.41 1.47 0.20
3 0.0079*** -0.0086*** 12.14 1.35 0.30
4 0.0012*** 11.93 1.26 0.30
Source: Authors.
Notes: *,** and *** indicate no significance at levels of 1%, 5% and 10%.
178
Table I.3. Estimation of variances according to the structural model

Asset Level Irregular Beta
ambv4 0.00 0.00 0.0004
bbdc4 0.00 0.00 0.0019
brkm5 0.00 0.00 0.0010
brto4 0.00 0.00 0.0021
bbas3 0.00 0.00 0.0027
cmig4 0.00 0.00 0.0002
cruz3 0.00 0.00 0.0009
csna3 0.00 0.00 0.0008
elet3 0.00 0.00 0.0002
elet6 0.00 0.00 0.0001
embr3 0.00 0.00 0.0010
fibr3 0.00 0.00 0.0003
ggbr4 0.00 0.00 0.0002
goau4 0.00 0.00 0.0002
itsa4 0.00 0.00 0.0007
itub4 0.00 0.00 0.0003
klbn4 0.00 0.00 0.0007
lame4 0.00 0.00 0.0002
ligt3 0.00 0.00 0.0015
lren3 0.00 0.00 0.0020
pcar4 0.00 0.00 0.0028
petr3 0.00 0.00 0.0011
petr4 0.00 0.00 0.0003
usim3 0.00 0.00 0.0007
usim5 0.00 0.00 0.0005
vale3 0.00 0.00 0.0018
vale5 0.00 0.00 0.0005
vivt4 0.00 0.00 0.0014
Source: Authors.
Notes: *,** and *** indicate no significance at levels of 1%, 5% and 10%.
Table I.4. Diagnostic checking for structural model

Asset Q1 (5) Q1 (10) Q1 (15) Q1 (20)
ambv4 0.28 1.21 7.15 19.88
bbdc4 3.03 9.64 13.56 24.12
brkm5 1.91 14.38 19.99 26.53
brto4 4.69 5.81 8.86 10.51
bbas3 7.54 40.41 57.63 62.86
cmig4 1.61 3.00 6.69 9.89
cruz3 1.04 1.20 2.76 7.20
csna3 10.92 14.68 23.70 27.97
elet3 0.26 2.09 6.72 10.04
elet6 1.39 6.69 9.65 11.20
embr3 0.67 6.53 12.17 30.20
fibr3 1.72 3.73 5.94 10.65
ggbr4 4.48 5.03 7.96 14.54
goau4 3.19 9.73 10.15 10.82
itsa4 3.55 5.56 9.30 10.97
itub4 1.57 4.63 9.41 21.48
klbn4 0.91 1.62 5.50 13.16
lame4 23.65 27.32 29.55 53.09
ligt3 0.89 2.86 5.42 7.99
lren3 20.65 21.41 25.01 28.04
pcar4 8.94 16.21 26.32 29.27
petr3 3.50 16.98 22.29 32.37
petr4 6.58 12.67 17.62 20.46
usim3 2.12 3.09 5.90 14.57
usim5 1.36 3.81 8.53 13.82
vale3 3.01 4.86 11.79 15.31
vale5 2.71 3.51 7.95 14.83
vivt4 6.55 8.13 9.42 17.36
Source: Authors.
179
Appendixx II
Figure 11. Dynamic bettas

Source: Authhors.
180
CondVar_Vambv4er
0.015 CondVar_Vbbdc4er CondVar_Vbrkm5er
0.0075 CondVar_Vbbas3er CondVar_Vcmig4er
0.0075 0.0075 0.010
Source: Authors.
0.010
0.0025 0.005 0.0025 0.0025 0.005
www.ccsenet.org/ijef
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
0.0075 CondVar_Vcruz3er CondVar_Vcsna3er CondVar_Velet3er
0.015 CondVar_Velet6er CondVar_Vembr3er
0.010 0.015 0.02
0.0050 0.010
0.0025 0.005 0.005 0.005 0.01
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
CondVar_Vitub4er
0.0075 CondVar_Vfibr3er 0.0075 CondVar_Vggbr4er CondVar_Vgoau4er CondVar_Vitsa4er 0.0075
0.015 0.0075
0.0050 0.0050 0.0050
0.0025 0.0025 0.005 0.0025 0.0025
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
181
CondVar_Vklbn4er 0.015 CondVar_Vligt3er
0.0075 CondVar_Vlame4er1 CondVar_Vlren3er1 0.015 CondVar_Vpcar4er
0.010 0.010
0.010 0.010
0.0025 0.005 0.005 0.005 0.005
Figure 2. Volatility of assets

1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 2006 2008 2010 2012 1995 2000 2005 2010
International Journal of Economics and Finance
CondVar_Vpetr3er 0.010 0.10 CondVar_Vusim3er CondVar_Vusim5er CondVar_Vvale3er

CondVar_Vpetr4er
0.0075 0.0075 0.0075
0.005 0.05
0.0025 0.0025 0.0025
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
CondVar_Vvale5er CondVar_Vvivt4er CondVar_Vbrto4er
0.02 0.015 0.0075

0.01 0.005 0.0025
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
Vol. 5, No. 3; 2013
CORR_VIboverbbdc_Vbbdc4er CORR_VIboverBBAS3_Vbbas3er CORR_VIbovercmig4_Vcmig4er
CORR_VIbover_Vambv4er CORR_VIboverbrto4_Vbrto4er
0.6 0.75 0.75 0.75 0.8
Source: Authors.
0.4 0.50 0.50
0.25 0.4
www.ccsenet.org/ijef
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
CORR_VIbovercruz3_Vcruz3er CORR_VIboverCSNA3_Vcsna3er CORR_VIboverelet3_Velet3er CORR_VIboverelet6_Velet6er CORR_VIboverEMBR3_Vembr3er
1.0 1.0
0.5 0.75 0.5
0.8
0.6 0.5
0.3 0.25 0.0
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
CORR_VIboverGGBR4_Vggbr4er CORR_VIbovergoau4_Vgoau4er CORR_VIboverITSA4_Vitsa4er CORR_VIboverITUB4_Vitub4er
0.75 CORR_VIboverFIBR3_Vfibr3er 1 1.0
0.75 0.75
0.25 0.50 0.5
0.25 0
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
182
CORR_VIboverKLBN4_Vklbn4er CORR_VIboverLREN1_Vlren3er1
CORR_VIbover1_Vlame4er1 CORR_VIboverLIGT3_Vligt3er
0.8 0.8 CORR_VIboverPCAR4_Vpcar4er
0.6 0.5 0.6

0.4 0.4 0.0
0.2 0.2
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
CORR_VIboverPETR3_Vpetr3er CORR_VIboverpetr4_Vpetr4er CORR_VIboverUSIM5_Vusim5er
CORR_VIboverUSIM3_Vusim3er CORR_VIboverVALE3_Vvale3er
International Journal of Economics and Finance
1
0.8 0.8 0.8 0.75
0.25
0.4 0.6 0 0.4
Figure 3. Conditional correlation between assets and Ibovespa

1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
CORR_VIboverVALE5_Vvale5er CORR_VIboverVIVT4_Vvivt4er CORR_VIboverbrkm5_Vbrkm5er
0.8
0.75 0.75
0.50
0.25 0.4
1995 2000 2005 2010 1995 2000 2005 2010 1995 2000 2005 2010
Vol. 5, No. 3; 2013
ISSN 1916-971X E-ISSN 1916-9728
The Regional Integration Agreements: A New Face of Protectionism

Montej Abida1
1
School of Business, University of Sfax, Tunisia
Correspondence: Montej Abida, School of Business, BP 1081 Sfax 3018, Tunisia. Tel: 216-98-487-913. E-mail:
[email protected]
Abstract
The regional integration agreements can be a strategy of trade diversion and thus we can say that there is a
violation of the rules of free trade. By creating preferential rules which are inconsistent with the principles of the
WTO, the strategy of regional integration can increase the risk of trade disputes with third party countries and
can therefore generate a commercial environment full of threats and reprisals. Third countries, especially
developing countries have small markets, may find themselves marginalized further when the members of the
regional group adopt the principle of discrimination.
The philosophy of the WTO paves the way for a transition from regional integration towards a multilateral
integration. The question that arises is whether regional integration agreements meet this conception of the WTO
or they represent a new form of protectionism hindering free trade. We know that the regional groups hold
private information about the actions and decisions they adopt in the intra-group. The uncertainty and hidden
information can cause conflicts internationally.
The WTO, we can assume as the Principal, should create incentives that can guide regional groups that can be
assumed as Peripheral Agents, to comply with Article XXIV and the elimination of regional protectionism.
Control actions and trade policies of regional groups should be done on a regular basis to make the game of
international trade more fair.
Keywords: regional integration, protectionism, free trade, trade policy, W.T.O, Principal-Agent model
1. Introduction
An increasingly common strategy tends to want substituting a regionalization of the world economy for
multilateralism. Indeed, after the end of the Cold War, the countries were divided into regional entities. It is
obvious that despite the economic crisis and the inability of a single national strategy, the policies of regional
integration become desirable. This is, especially, due to the success of European integration that has encouraged
the countries to regroup. Many countries launch new regional integration projects or revive old ones.
The strategy of regional integration helped to divide the world into three major economic regions, each of them
is centered around a pole: Europe around the European Union, America around the United States, Asia around
Japan. Each of these three major economic powers is trying to attract as many countries as possible by signing
the preferential agreements. It is mainly the developing countries cling to these three areas in order to find the
breath required to start their own economies.
Even if the regional integration does not fit into the philosophy of multilateralism, the WTO is the exception to
the principle of non-discrimination and authorizes the establishment of a regional group. Indeed, the WTO was
believed that if one starts with regional trade cooperation, we will eventually trade cooperation at the
international level. But in reality, the most of the agreements of the regional integration are not interpreted and
applied as necessary as the rules set forth in Article XXIV of the WTO. The question is whether the regional
integration may undermine multilateralism or not. In other words, we must know what the implications of the
strategy of regional integration are in the multilateral trading system.
The most of countries have understood that it is better to come together to achieve the economic development
(Shadikhodjaev 2009). Going it alone is not profitable. Therefore, the regional integration strategy encourages
the member countries to cooperation and economic development. It is creating customs unions and free trade
183
areas. This strategy aims to remove the barriers for the competition and to allow a more efficient allocation of
resources to increase the production and to improve the well-being of each country.
It was after the Second World War, and more exactly after the creation of the European Economic Community
(EEC) in 1957, the strategy of regional integration has been reappeared. Indeed, the regional integration
agreements which are signed in 1547-1548 between England and Scotland show that these agreements are not
new. In general, it is from the 16th century several regional integration agreements have been signed in order to
eliminate or reduce, on a mutual basis, all the barriers to the free movement of goods.
It is very appropriate to report back a wave of the regional integration agreements, as well as deepening and
widening of existing of the agreements in recent decades. Indeed, in 31 July 2010, 474 RTAs had been notified
to the GATT / WTO, the notifications for goods and those for services have been recognized separately.
Concerning these RTAs, 351 were notified under the Article XXIV of the GATT 1947 or GATT 1994; 31 under
the Enabling Clause and 92 under the Article V of the GATS. On the same date, 283 agreements were entered by
force.
2. Regionalism and Multilateralism: A Review of the Literature
It is legitimate to ask if the strategy of regional integration will have an influence on the multilateral trading
system. The new international political and economic order created by the end of the Cold War seems to be
involved in the abandonment of the enlargement strategy and regional integration. The question is why countries
are moving towards regional integration strategy while there is a multilateral trading system. Is that because the
regional integration agreements are more effective than multilateral trade agreements in progress towards free
trade? It is also important to know if the regional integration agreements help or hinder multilateralism (Siroën
2000a, Trotignon 2009).
Several theoretical and empirical studies (Hauser and Zimmermann 2001, Khavand 2001, Crawford and Laird
2001, OECD 2003, Mashayekhi and Taisuke 2005, Baldwin 2009), examined the interplay between the
multilateral trading system and Preferential Trade Agreements (PTAs ). Most of these studies have asked
whether PTAs promote or hinder multilateral tariff reductions (Hauser and Zimmermann 2001, Khavand 2001,
Baldwin and Jaimovich 2007 and 2009). The objective is to analyze the effect of preferential tariff reductions on
the Law of the Most Favoured Nation (MFN).
The examples studied in this work focus on the case of free trade agreements North American "NAFTA"
(Baldwin 2009, Goyette 2006, Deblock and Dorval 1993, Tshiyembe 2012) and the case of the European Union
(Baldwin 2010, Kearney 2008, Clergerie, Gruber and Ramband 2010, European Commission 2008). In recent
decades, the case of the formation of a regional group in Asia has attracted several researchers (Figuière and
Guilhot 2009, 2007). Indeed, Baldwin (2010) suggests that we must learn from the example of regional
integration in Europe. It assumes that the spontaneous cooperation that created the factory in Asia has not been
codified. He suggested to institutionalize the spontaneous cooperation that already exists on trade, services and
investment.
By using the domino theory, Baldwin and Venables (1995) presented a model of enlargement of the European
Union. Indeed, this theory shows that the potential loss of market share encourages non-members to adhere to
existing PTAs, which generates a process of action and reaction or infection. The model of the domino effect
was used again by Baldwin and Jaimovich (2009) in order to see whether the bilateral agreements are contagious
or not. These two authors have shown that exporters of non-member countries push their politicians to join the
PTAs existing or create new ones to avoid damage that could result in preferential trade liberalization.
Baldwin (2009) examined the systemic effects, that is to say, the static and dynamic effects, PTAs on
multilateralism. Static effects are, for example, the existence of conflicting rules, including on trade remedies.
Dynamic effects are, among others, the impact of PTAs on the probability that countries undertake new
multilateral negotiations.
From their sides, Baldwin and Seghezza (2007) found that regionalism does not impede MFN tariff reduction.
Indeed, they showed that the MFN tariffs (Most Favored Nation) and PTAs (Preferential Trade Agreements) are
complementary and do not substitute since margins of preferences tend to be low or zero for products where
countries apply high MFN tariffs.
However, Crawford and Laird (2001) showed that the rapid growth of trade agreements is a threat to the
multilateral trading system. The orientation towards regional integration strategy seems to be explained by the
fact that free trade can take much longer delay in the multilateral trading system in the regional integration
agreements. If the multilateral trading system is credible, countries no longer consider the strategy of regional
184
integration. Market access, which is the main concern of countries is better guaranteed in regional integration
agreements in the multilateral trade agreements.
The first risk that the strategy of regional integration can cause is the negative impact on the multilateral trading
system. But until now, no research has proved the existence of a full scale trade and investment. The effects of
various regional integration agreements on trade in third countries can be summarized in a high probability of
trade diversion towards countries of this agreement. It is thanks to the early work of Viner (1950) we observed
that a regional integration agreement may result in a diversion of trade for the goods listed. This diversion
depends on the form of the agreement of regional integration and trade policies of member countries.
The second risk of a regional integration strategy can be summarized in a division of the commercial geography
in regional blocs. The creation of dominant economic powers and protectionist may present a danger to the
multilateral trading system. Indeed, the use of aggressive strategies by the major economic powers will have a
negative impact on international trade relations: the use of a protectionist trade policy by one of the regional
groups necessarily leads to retaliation and anti-reprisals therefore a trade war is likely to be triggered.
The signing of several regional integration agreements concerned third countries about the future of their trade.
The diversion of trade and investment to a regional grouping, which is not consistent with fundamental
principles of the WTO can leave multilateralism in crisis. Indeed, whether the measures adopted by supporters of
a trade agreement is not extended to third countries and regional preferences are maintained, there is a risk that
third countries do not enjoy the same access to the large market of this regional group. The regional integration
agreements attach great importance to the commercial interests of all parts of the regional group. Third countries
that do not grant reciprocity, can not take advantage of market access of a regional grouping.
The solution that countries have adopted to avoid the risk of trade diversion is to join a regional group
(Chaponnière and Vérez 2009). Thus, we saw the signing of several preferential trade agreements. While many
agreements are not limited to countries of the same region, we speak of a PTA that can be a free-trade area or a
customs union with a common external tariff.
The desire to join a regional group is explained by the fact that multilateral trade agreements are not credible and
effective. If the approach of multilateralism was successful, the country will no longer consider integrating into
regional groups. We can conclude that the multilateral trading system is in crisis because many countries wanted
to try the regional integration strategy. If countries have changed the tactics of the game by integrating into
regional groups is that the multilateral trading system is not working as it should. The regional integration
strategy therefore prevails on the strategy of multilateralism. Countries thought that the regional integration
agreements have better results than those of multilateral trade agreements.
The failure of the multilateral trading system to provide many benefits encourages countries to use regional
integration strategy to take at least one market access of a regional group. Most countries have understood that
the multilateral trading system has not yielded good results, especially with regard to market access: tariff will be
more easily removed as part of a regional group in the framework of a multilateral trading system. Thus, the
failure of the strategy of multilateralism has led to the success of the regional integration strategy.
Thus, faced with discrimination exercised by the commercial strategy of regional integration, the countries are
still trying to fit into a regional group or through a membership, either by signing a preferential agreement such
the fourth ACP-EEC Convention This tactic allows them to enter the market of a regional group while
maintaining good relations with the member countries to avoid protectionist practices.
3. The Forms of Regional Integration Agreements
The regional integration can be either natural or strategic. Indeed, according to Jacquemin and Sapir (1991),
"The natural integration" involves trading partners 'natural' which are geographically close, they pursue a liberal
trade policy towards third countries. With respect to the strategic integration and according to the same authors
(Jacquemin and Sapir). It means that members of a regional group pursue a common commercial policy at the
expense of third countries.
The regional integration agreements are customs unions, such as the European Economic Community, to free
trade zones which are based on reciprocal agreements such as EFTA and ELENA, by passing through
non-reciprocal preferential agreements as the fourth ACP-EEC Convention of Lomé. These non-reciprocal
preferential aims to facilitate the expansion of trade in some developing countries with no demand for reciprocity
on the part of developed countries. The most of regional agreements involve free trade zones, whereas the
agreements number of the creation of the customs unions is limited to a few examples such as the European
Union, CARICOM and MERCOSUR.
185
4. The Characteristics of Regional Integration Agreements

Since the inception of GATT creation until 1994, 108 RTAs have been notified to the GATT which 98
agreements have been notified under Article XXIV and the rest has been notified under the Enabling Clause (For
more details, see page 98-117 Study of the WTO. "The regionalism and the world trading system", Geneva,
April 1995). According to all of these agreements, 38 were notified between 1970 and 1979 and 37 between
1990 and 1994. It is the European Union that has signed the largest number of regional agreements: 41 of the 98
agreements which were notified to the GATT under the title of the Article XXIV took EEC as a partner. But,
more generally, it is the Western Europe that adopted the most of the strategies of the regional agreements. In
fact, the creation of the EEC (1957) and EFTA (1960) facilitated the signing of 76 regional agreements. In
contrast, the regional agreements are almost absent among the countries in the developing world. This is
explained in one part by the adoption of self-reliant development strategy, in the other part, by the existence of
political unrest of these countries.
In addition, the majority of regional agreements are essentially the creation of free trade zones and more
specifically those of a reciprocal type. The integration agreements that are creating the customs unions are very
limited and mainly concern the EEC, CARICOM and MERCOSUR. But it should be noted that several regional
agreements have either been replaced by other agreements, or ceased to be practiced. Therefore, the considerable
number of regional agreements must not give us a false idea of the importance of regional integration strategy.
Moreover, all the agreements began with the partial reduction of tariffs; the others went to the gradual removal
of tariff barriers.
5. The Polarization of International Economic Relations
The regional integration strategy has resulted in the organization of major geographical areas in delimited and
protected trading blocs. Several economists have seen in regional integration agreements the sign of the advent
of three regional groupings are the United States, the European Union and the Asia-Pacific region. The risk that
these three great powers can cause to themselves is likely a deterioration in the international economic relations.
Therefore, we are witnessing a polarization of international economic relations around these three major
economic powers that are trying to attract the maximum number of countries to win more markets.
The poles of the trade can only be economic superpowers that dominate all markets of goods and services. They
are often made up of several countries to create a dynamic business inside the pole and try to attract the
maximum number of other countries through a complete bond or a signature of a preferential agreement. Each
pole tries to exert an influence on the international trading environment by encouraging countries to multiply
trade relations with him. The pole strategy necessarily entails a regional concentration of trade in some areas that
are economically powerful.
Regional integration facilitates the construction of hegemonic and powerful regions. The creation of the
European Union is a good example to demonstrate the existence of a hegemonic power capable of exerting
influence on other countries. The enlargement process of the European Union helped the formation of an
economic power that prompted other countries to join the group, or at least to sign preferential agreements in
order to receive certain benefits such as market access to the European Union.
The existence of hegemonic powers has much influence on the international trade relations that countries should
set the trading bloc with whom they want to work. Each regional group aims to attract the maximum number of
countries to form an economic or commercial power. The main purpose of an economic or commercial bloc is
the domination of maximum market via a policy of enlargement or a signature of preferential agreements with
third countries.
The regional integration agreements are therefore creating commercial powers. One of the objectives of regional
agreements is to address the major economic powers that dominate the commercial world. The race for regional
agreements or the enlargement strategy of a regional group shows that the followed objectives have a hegemonic
order. The more trading block is stronger. The more the third countries and other regional groups are wary and
take more precautions in the use of their trade policies because they know it is difficult to address such market
power.
The strategy of regional enlargement is mainly adopted by the three most powerful trading blocs in order to
dominate the global markets: the United States is seeking to create a free trade area by partnering with Canada
and Mexico. The European Union is following its enlargement policy to other countries such as Eastern
countries. As for Japan, it continues to adopt the opening policies, especially in Asia.
186
For economic and political reasons, a regional group would largely dominant over the rest of the world. This
domination allows it to control the game of international trade by attracting the maximum of countries that are
playing solitary. It is in the sharing of markets that the confrontations are taking place between the most
powerful regional groups. The developing countries, which are looking for the economic start, are more attracted
by the regional groups who seek to be dominant. Thus, regional integration can be considered as a regional
hegemonic strategy as it creates a competition of dominance between the major regional groups.
6. The WTO and the Three Most Powerful Regional Blocs
The WTO gives great importance to the three major economic powers in Europe, North America and the
Far-East. This importance is based on the issue of competition and the type of relationship between these three
trading blocs. "The only valid and viable perspective is that Europe, North America and the Far-East
arrangements strengthen their regional arrangements of integration and cooperation in a multilateral stable - a
common house - that provides clear and predictable rules and mechanisms to resolve the trade disputes
coherently and objectively. If we want to strengthen and not to lose the multilateral system, we have no choice
but to execute the Uruguay Round "(Declaration of the former General Director of GATT Peter Sutherland at a
symposium on "Global competition: Europe against North America against Far-East," which was held in
Cernobbio, Italy, September 4, 1993. News of the Uruguay Round.'' NUR 064. 6 September 1993. PP6).
The WTO insists on the fact that the regional groups should be open to the rest of the world. The regional
integration strategy should be directed towards the construction of an effective multilateral trading system and
without disloyal practices that violate the international competition. When commercial relations between these
three blocks are dominated by a non-cooperative business environment, we begin to exchange blows by
following commercial practices that are against the law of free trade. The protectionist policies, applied in the
30s, showed us the disastrous consequences of international trade relations.
The WTO is trying to develop a cooperative environment between the three major economic powers by
encouraging them to improve their commercial relations by trying to find an arrangement whenever there is a
problem. "It would be morally and economically condemnable if the countries of North America, Western
Europe and the Far-East, which have mostly benefited from the multilateral trading system, refuse to assume any
responsibility to reinforce it now whereas it showed its universal value "(Declaration of the former General
Director of GATT Peter Sutherland at a conference on" Global competition: Europe against North America
against the Far-East ", which took place in Cernobbio, Italy, September 4, 1993." News of the Uruguay Round."
NUR 064 6 September 1993. Pp 4). During the Uruguay Round, there has been an increase in the number of
trade issues between the regional blocs. The agricultural conflict between the EU and North America has
impeded the conduct of international commercial negotiations and has nearly destroyed the Euro-American
economic relations.
7. Application to Regional Integration Model Principal-Agent
The Principal-Agent model has been applied in several industries such as insurance service (Borche 1962),
public procurement, industrial economy, the relationship between shareholder and manager, the relationship
between the State and the private concessionaire ... We will use the Principal-Agent model to analyze the
interaction between Preferential Trade Agreements and the Multilateral Trading System. In this paper, we
develop a model of Principal-Agent type assuming that the WTO as the Principal and regional groups as Agents
devices. The WTO should create mechanisms to encourage regional groups to respect the rules of international
trade and in particular Article XXIV. The goal is to create mechanisms that should be optimal for the Principal
and that solve the problem of information asymmetry (see Kadan, Reny and Swinkels 2001).
The Principal-Agent model is developed by the agency theory. Indeed, the general form of the agency
relationship is defined as "a contract by which one or more persons (the mondant or the Principal) engage
another person (the Agent or mondataire) to perform an action on their behalf, which involves the delegation to
the Agent of a certain decision-making powers "(Jensen and Meckling 1976, p. 313). In general, "most agency
problems involve a combination of adverse selection, moral hazard and risk-sharing" (Laffont 1987 p.18).
The existence of a conflict of interest between the Principal and the Agents has led many economists to look at
this problem. Indeed, the presence of information asymmetry is the basis of the existence of conflicting interests
between the Principal and the Agents. Problems of information asymmetry are present in two forms which
depend on the nature of the information: the first concerns the anti-selection or adverse selection and the second
form illustrates the problem of moral hazard or risk.
187
The problem of adverse selection is based on the uncertainty about the agent type: the signing of the contract, the
Principal does not know the hidden information by the Agent. So the Agent has private information about an
exogenous variable. In contrast, the problem of moral hazard is a situation of hidden action after the contract
signature: it is the behavior of the Agent that is unknown. Therefore, the Principal does not observe the action of
the Agent which is an endogenous variable. Principal-Agent models can raise the case of adverse selection and
the moral hazard: the Principal has no idea about the exogenous and endogenous variables.
The work of Grossman and Hart (1983), Rogerson (1985) and Jewit (1988) have developed Principal-Agent
models in order to justify the approach of first order moral hazard. It should be noted that the agency relationship
necessarily entails risk sharing. Theorists have argued that the type of contract that governs the relationship of
agency may limit the opportunism of the Agent by encouraging them to exercise. Grossman and Hart (1983)
point out that "in general, the action that is optimal for the agent depends on the extent of risk sharing between
the Principal and the Agent." The objective is to determine the "optimal degree of risk sharing, in view of this
dependence." These two authors have analyzed the problem of socialization of risk and struggle against
opportunism (moral hazard stowaway).
Moral hazard model implicitly assumes that the Principal has the sole purpose of designing incentive contracts.
For example, in models Holsmtrom (1979) and Grossman and Hart (1983), the Principal usually have interests to
offer a contract that provides a variable remuneration for the Agent based on the realized profit.
The role of the WTO is to find a mechanism to reveal the exact information that holds the regional group. We
call the principle of revelation in the game of the regional grouping, determination or detailed specification of a
set of strategies for regional groups and an allocation rule for the WTO. As information is asymmetric and
regional groups have the opportunity to influence the outcome in their favor, revealing a mechanism must be put
in place by the WTO to eliminate any obstacle that hinders multilateralism. That is to say that the WTO should
create a revealing mechanism that encourages regional groups to prevent damaging the trade of third countries.
Looking for a way to encourage regional groups to reveal their private information is the first concern that the
WTO should do when it wants to eliminate all risks affecting the final result. Indeed, revealing mechanisms
should be created to cope with the exploitation of hidden information. Thus, each revealing mechanism tends to
encourage regional groups to be more honest by deploying more efforts in proclaiming their true characteristics.
8. Modeling the Regional Agreements as a Principal-Agent Model
The regional integration agreements can be modeled as a principal-agent model. The WTO can play the role of
Principal since it is this organization that protects the commercial interests of all the Contracting Parties. In turn,
the regional groups will form the peripheral agents that have a private information about the decisions and
principles taken about trade with third countries. The WTO is going to defend the interests of third countries by
trying to obtain the hidden information held by the regional groups. It must find a capable mechanism to
encourage the regional groups to respect the rules of free trade and in particular Article XXIV.
8.1 A Model of Hazard Moral Type
The existence of asymmetric information can make regional integration agreements dangerous for the
multilateral trading system. In fact, the members of a regional group can damage the trade of third countries by
adopting rules that promote only the intra-group trade. Each regional group has an interest to avoid disclose an
accurate information about the decisions and the actions that will be taken within the group.
This asymmetry of information is made of selfishness of regional groups: the objectives of the WTO and
regional groups may not concord. The WTO is supposed to defend the commercial interests of all contracting
parties, while the regional groups defend their own interests. The existence of a divergence of interests between
the WTO, which seeks to find the general interest and the regional groups that seeks to find particular interest,
obliges WTO to create incentives mechanisms. These mechanisms should make the game of trade between the
regional groups and the third countries more transparent and cooperative.
Each regional group is the only one who knows the secrets of his actions. It can easily announce that it does not
violate the rules of international trade and it does not affect the trade of third countries. But in reality there is
often a preference to intra-group. The members of a regional group star with exporting and importing their
products with the group before moving to the trade of third countries. These members can take advantage of
certain discriminatory rules that the WTO does not know.
The asymmetric information and strategic behavior of the regional groups will have a negative impact on the
multilateral trading system. The private information held by the groups of pressure gives them an advantage,
especially in trade. Each member tries to be careful using unobservable methods that guarantee a good protection
188
of the interests of all members of the regional group. Most of the regional groups give only the minimum
information about the trade policy and the actions they will take.
Therefore, the presence of asymmetric information between the WTO and the regional groups may increase the
unfair practices and disrupt the trade relations between the regional groups and third countries. The WTO should
increase the efforts to control the policies and the actions of the regional groups to collect the maximum
information.
Thus the moral hazard is present for all regional groups as soon as the WTO can not perfectly control the future
actions of the regional groups. When the WTO agreements are given for the creation of a regional group, he does
not know if the members do not divert trade to this group. There is always a risk that the members of a regional
group do not comply with Article XXIV.
The WTO can not observe a variable correlated with the action of the regional group: the trade with the third
countries. Then, we say that we are in second rank situation. So we will talk about the efforts to designate the
trade made by the countries that are not directly observable. The role of the WTO is to encourage the regional
groups not to divert trade to the members of these groups. The WTO must offer for the regional group a contract
that consists in canceling the agreement of the regional integration if there is a violation of Article XXIV. So the
informational asymmetry appears so after the signing of the contract: the action of the regional group is not
directly observable.
8.2 The Model Assumption
The most important assumption of this model is to recognize that the WTO has the power to control the actions
and commercial policies of the regional groups. It may also adopt a system of penalties against any regional
group that does not comply with the Article XXIV. This system of penalties can even lead to the cancellation of
the agreements of regional integration if it presents a danger to the game of international trade.
A regional group should choose an action   from a set of possible actions   :    .
The implementation of the contract involves a set of states of the world e with e   .
The regional group has a trade policy that allows you to export a product y  y  , e that allows to generate a
profit   , e .
The WTO represents the interests of all the Contracting Parties, in particular the interest of third countries, if
there is regional integration agreement. We assume that the overall profit is shared between the WTO, which
represents the benefit of third countries and the regional groups.
ry  : a part of profit taken by the regional group.
  r y  : a part of profit taken by the WTO (part of the benefit of third countries).
The trade of the member countries of a regional group is a signal that provides information about the action
chosen by the regional group.
The preferences of the WTO and the regional group are described as the utility function of Von
Neumann-Morgenstern.
U G  u G   r  y  : Utility function of the WTO (1)
With: uG
/
0 and uG
//
0
U I  u I r  y   C  : Utility function of the regional group. (2)
With: uI
/
0 , uI
//
0 and C/0 , C //  0
Choosing an action (  ) from the part of the regional group leads to a law defined on the domain of profit.
Suppose that ( G  ,   ) the function of the distribution of the random ( ~ ) (probability that the benefit is greater
than a given level) and ( g  ,   ) its density of probability. The Initial richness is normalized to 0.
Cδ : a function of the commercial policy that depends on the level of effort of the regional group (a level of
effort that complies with Article XXIV). That is what we will name the disutility.
189
The WTO maximizes the expected utility of the profit in the third countries under the incentive constraints and
participation.
8.3 The Incentive Constraint
To the satisfaction of the incentive constraint, it is necessary that the regional group provides a level of effort
which seems as advantageous as possible in the context of the WTO. The incentive constraint touches upon the
effort to comply the Article XXIV. This constraint can be written as follows:
 u r y  C  g,  dy   u r y  C ˆg ,  dy

I I
(IC)
(3)
8.4 The Constraint of Participation
The constraint of participation consists in the advantage of the regional group participation in the contract. This
means that the regional group would agree to participate in the contract unless it gives a value superior or equal
to ( U I ). ( U I )is the utility that the regional group can get if it has no relationship with the WTO. This is what we
call the next-best opportunity). The participation constraint is written as follows:
 u ry   C δgΠ, δdy  U

I I
(IR) (4)
8.5 The Resolution of the WTO Agenda

  
The WTO must choose the contract r y that maximizes its expected utility, taking into account the impact
of this contract on the decisions of the regional group:
Max  uG   r  y  g  ,   dy
(5)
Under the constraints of the incentive and the participation:
(IC):
 u ry   C δ g π, δ dy  
I
u ry   C δ̂ g π, δdy
I
(IR):
 u I r  y   C   g  ,   dy  U
I
With G, at least, three times continuously differentiable on its domain of definition of density ( g π, δ ). In the
level ( δ̂ ) of an insignificant effort different from ( δ ): the most advantageous effort from the standpoint of the
WTO.
The resolution of the WTO agenda is difficult to solve because of the existence of inequality in the incentive
constraint. For this reason, we assume that the regional group put up only two possible actions ( δ ) and ( δ )
with the probability of occurrence ( g ) and ( g ).
( g ) stochastically dominates the first order ( g : δ δ ). The disutility function (the function of trade policy
depends on the level of effort) of regional group is the largest one: ( C  C ).
We will assume that the WTO is neutrally in risk and that the regional group presents an aversion for the risk.
Maximizing the expected utility of the WTO with respect to r gives the following optimality condition:
1  g y   (6)
 β  α 1  
u I  g y  
Where β and α are appropriate multipliers.

β : Multiplier associated with the participation constraint.
α : Multiplier associated with the incentive constraint.
In case of α = 0 (incentive dimension), we find the condition of sharing the optimal of Borch risk where r (y) is
constant. When α is positive, r (y) must vary with y. This reflects an arbitration between risk sharing, on the
one hand, and the incitement to the effort, on the other hand. If the arbitration is not taken into account, the
country would provide the minimum effort compatible with the maintenance of the contract.
190
g
r (y) varies with the magnitude : ratio of two densities of probability in the general case, that is to say,
g
likelihood ratio. According to Laffont, the ratio measures the informative power of y designed as a signal.
Suppose that the action of the regional group is an unidimensional continuous variable of effort. In this case, the
general writing of the incentive constraint can not be maintained. The only way of proceeding is to adopt an
approach for the first order: we replace with the condition arg max indicating that, at the optimum, the expected
utility of the regional group is stationary in its effort.
The condition of stationarity:
 u ry  g y, δ dy  C δ   0

I
(7)
At the optimum, the expected marginal utility of effort must be equal to the marginal disutility. It is contained of
a necessary first order condition but not sufficient. It does not always correspond to a correct approach: it offers
a contract which sometimes does not verify the global incentive constraints of the original problem. Thus, we are
led to study the conditions under which the first order approach is valid (Grossman and Hart 1983, Jewitt 1988,
Rogerson 1985).
In order to obtain these conditions, by considering firstly a simple extension of the dichotomous case, we assume
that the regional group controls a family of probability laws as the following :
g y, δ   δ g y, δ   1  δ  g y, δ  (8)
This means that the regional group chooses his effort as a convex linear combination of two fixed probability
laws ( g , g ). This relationship reflects a condition of linearity of the distribution function of the product of effort
(spanning condition: condition of collection). If this condition is valid, the first order approach is valid, as well.
But in the general case, this recovery condition is not necessary. We can take the linear approximation of the
*
family law to the vicinity of the optimal effort ( δ ) that the WTO wishes to embody. With the approach of the
first order, we treat the problem as if the regional group chose its effort from a hypothetical form of laws.
   
g  y, δ   g y, δ*  χ g δ y, δ* (9)
χ is very small, it is used when a cost function is :
 
C    C  *   C    * 
  (10)
( g  ) is linear as in the case when the recovery condition is fulfilled. A problem remains:the WTO may want a
law other than ( g  ) in the family. If this is the case, it means that the WTO wants a sudden change in the
appeared effort. The linear approximation is not sufficient.
The Lagrangian of the optimization problem is: (11)
  I g y, δ 
  G  I

I
u y  ry   β  u ry   Cδ  U   β  u ry 

δ  Cδ g y, δ dy

  g 
Max  (Static optimization) compared to r. and δ .

uG y   r* y  g * y, δ 
βα δ first optimization condition (12)
I
u r y 
gy, δ 
By using (7) (link optimization), we obtain the condition of the optimization effort.
 I   *
 u y  r y  g δ* y, δ dy  α  u r y  g * y, δ  dy  C δ 
G
 δ  (13)
191
(12): ratio likelihood written as a continuous form. If the condition of Milgrom and Shannon (1994:
Monotonicity theorem) is satisfied, this ratio is increasing. This means that the Article XXIV thinks in the same
time when its effort level thinks. The equation (12) indicates how the need to encourage the regional group led
the WTO to deviate from a pure risk-sharing problem.
• If β  0 , this means that the incentive dimension is absent, there is an optimal solution similar to that of
Borch (1962), that is to say a solution of Pareto optimal risk sharing. This is called a solution of the first rank.
• If β is positive, this solution diverges from the first rank g δ    Log g  .
gδ  δ
Comparing with the solution of the first rank, the deviation greater than the likelihood Log g  . The difference is
even more important than the informational content of the signal collected by WTO. The WTO itself is
important when β is positive.
The WTO wants effort of the regional group to be greater than in the case of the first rank of risk sharing. To
encourage him to provide greater effort, the WTO must periodically control the trade policies of the regional
group and should verify if there is a violation in the Article XXIV. Since the WTO has all the power, the
regional group does not have to cheat if it wants its integration agreement not to be canceled.
8.6 The Conclusion of the Model
In this model, we have seen that there is a problem of ignorance of the actions and decisions of the regional
groups. The hidden information held by regional groups can lead to unfair practices that may impede the third
countries trade, and thus cause a destructive trade war.
The WTO should be able to observe perfectly all the actions of the regional groups that affect the usefulness of
the third countries. In other words, the WTO must ensure that the commercial interests of third countries are not
damaged. This means that the WTO should make more extensive control efforts on the trade policies of the
regional groups.
If the WTO can observe the action of the regional groups, it is in the case of the first rank. In the case of trade
diversion, the WTO must cancel the regional integration agreement. In contrast, in the case of compliance with
the Article XXIV, the WTO should encourage the regional groups to continue in this direction.
The trade policies of regional groupings are more mostly controlled by the WTO, especially those who may
present a threat to the multilateralism. It is the existence of certain rules that benefit the members of a regional
group that worries the WTO and leaves him more attentive towards the unfair practices employed by these
regional groupings. What is essential for The WTO is not to allow the regional groups adopt trade policies that
do not comply with the rules and principles of free trade.
The Negotiations of the Uruguay Round enabled surveillance increased by the GATT trade policies of all the
Contracting Parties. Indeed, a periodic review mechanism is set up to control all the unfair practices that go
against the rules of the WTO. The Trade Policy Review of the contracting parties, which is one of the main
concerns of the WTO, is carried out according the importance of countries or the regional groups.
The establishment of review mechanism of trade policies of contracting parties intended to ensure greater
efficiency in the evaluation of the effect of the Contracting Parties business strategies to the multilateral trading
system. This mechanism can help us build a more transparent business environment, which, as we know, is one
of the fundamental principles of the WTO. The trade policy review affects especially the main business entities
and in the place, the most powerful regional group. For example, the importance of the European Union
exchange requires the review mechanism to be replaced every two years.
By controlling, every two years, the trade policies of major economic powers clearly explain that these trade
blocs pose a threat to the multilateral trading system. The major trading powers are primarily responsible for the
success or failure of the multilateralism.
Over 80 regional agreements reviewed by the WTO, only six were considered consistent with WTO principles
(the European Economic Union is not included). This explains that the most regional groups use rules that are
incompatible with the multilateralism strategy. The adoption of trade policies, oriented only to the regional
preference, is the most abundant of all forms of the regional integration strategy. This preference concerned the
WTO and forces him to reinforce the control of the regional integration agreements whatever their forms are.
The 69 working groups, who were appointed to examine the 98 agreements were notified in January 1995 under
the Article XXIV, concluded that there are two regional integration agreements that are consistent with Article
192
XXIV. These two agreements are those that were signed between the Community and the Caribbean Common
Market (CARICOM), in the one hand, and those of the Customs Union that were signed between the States of
the former Czech and Slovak Federal Republic, in the other hand. For other agreements, the working groups
were unable to reach a consensus as most member countries of an integration agreement announced that this
latter is consistent with Article XXIV of the WTO while the third countries insist that such agreements are not
compatible with this article.
However, among the examples of non-compliance integration agreements with Article XXIV of the WTO, we
may mention the tariff treatment of EU imports of citrus industry products from certain countries of the
Mediterranean region. This tariff treatment, was concluded in 1982, was considered by the U.S. incompatible
with the principle of the most favored nation (MFN). This judgment has been confirmed by a special group that
was appointed to resolve this case. Moreover, two other complaints filed by third countries on non-compliance
integration agreements with WTO rules, were against the European Community.
9. Conclusion
The interplay between the regional groups remains uncooperative because everyone is trying to dominate the
trade by attracting the maximum number of countries. This non-cooperative environment requires the presence
of the WTO that must both set the rules of the international trade and control the trade policies of countries
especially those regional groups that dominate the commercial exchanges.
We have shown that the regional agreements can be modeled as a Principal-Agent model. The resolution of this
model is based on a very important assumption that is to give all the power to the WTO to cancel the regional
agreements that constitute a new form of protectionism and pose a threat to the multilateralism. It has become
necessary that the WTO should establish a legal authority that obliges the regional groups to respect the rules of
free trade and in particular Article XXIV.
The regional groupings aim to strengthen the political economic and social relations between different countries
in a group regardless of its level of development. The regional integration aims to accelerate the economic
growth in member countries. The financial aid should be give to the least developed countries so that they can
face the most serious economic problems. These aids also aim to promote convergence among the member
countries of the regional group.
Most countries, especially the least developed, are seeking to ally with major trading blocs. These countries
consider being a member of a powerful regional group as it enjoys lot of benefits such as free movement of
goods, services and factors of production. In addition, the solidarity principle, adopted by most regional groups
and served to assist countries with the greatest difficulties, has prompted several countries to move towards the
regional integration strategy.
According to the WTO, the regional integration strategy must be consistent (complementary or compatible) with
a globalization of trade and investment. The WTO considers that the regional and multilateral approaches should
be complementary. The regional integration strategy should lead to a significant multilateral trade liberalization.
With a liberal trade policy, this strategy should have a positive impact on the global trade.
But, the current economic and financial crises are not in favor of the strategy of multilateralism. The existence of
several factors (high unemployment, sluggish economic growth, rising food prices, ...) has encouraged countries
to use protectionist trade practices especially in harsh forms (such as measures so-called "gray zone "and the
agreements of voluntary restraint) or soft (such as sanitary and phytosanitary barriers and technical barriers to
trade)
The WTO must find a mechanism that requires countries to comply with Article XXIV, or more precisely a
mechanism which removes all unfair practices employed by regional groups. Without this mechanism, there will
always exist trade problems between countries. Consequently, the game environment of international trade is
often tense. The Revealing mechanism created by the WTO must steer countries towards legal free trade and
without free protectionist practices.
References
Baldwin, R. (2010). Sequencing Regionalism: Theory, European Practice, and Lessons for Asia. CEPR
Discussion Papers 7852, C.E.P.R. Discussion Papers.
Baldwin, R. (2009). Integration of the North American Economy and New-Paradigm Globalization. CEPR
Baldwin, R., & Jaimovich, D. (2009). Are Free Trade Agreements Contagious? CCES Discussion Paper Series
12, Center for Research on Contemporary Economic Systems, Graduate School of Economics, Hitotsubashi
193
University.
Baldwin, R., & Seghezza, E. (2007). Are Trade Blocs Building or Stumbling Blocks? New Evidence, CEPR
Baldwin, R., & Venables, A. (1995). Regional Economic Integration. In Dans Grossman G., Rogoff K. (Eds),
Handbook of international economics, III (pp. 1597-1644). Amsterdam, North-Holland.
Borch, K. (1962). Equilibrium in a Reinsurance Market. Econometrica, 30, 424-444.
Chaponnière, J. R., & Vérez, J. C. (2009). L'évolution des échanges commerciaux entre l'Union européenne et la
Turquie depuis l'union douanière de 1995. Economie Appliquée, 62(4), 99-130.
Clergerie, J. L., Gruber, A., & Ramband, P. et. (2010). L'Union Européenne. Dalloz, 928.
Commission Européenne. (2008). Une stratégie pour l'avenir de l'union douanière, Communication de la
Commission au Conseil, au Parlement européen et Comité économique et social européen. COM/2008/0169
final.
Crawford, J. A., & Laird, S. (2001). Regional Trade Agreements and the WTO. The North American Journal of
Economics and Finance, 12(2).
Deblock, C., & Dorval, B. (1993). Une intégration régionale stratégique: le cas nord-américain. Etudes
Internationales, 2(3), 595-629. http://dx.doi.org/10.7202/703211ar
Draper, P. (2010). Rethinking the (European) Foundations of African Economic Integration: A Political Economy
Essay, Document de travail du centre pour le développement de l'OCDE no 293. Voir également l'article du
même auteur dans Éclairage, 10(4).
Figuière, C., & Guilhot, L. (2009). L'intégration économique régionale : parangon de la complémentarité entre
l'économie internationale et l'EPI, Colloque international ''Économie politique internationale et nouvelles
régulations de la mondialisation.'' Centre de Recherche sur l'Intégration Économique et Financière,
Université de Poitiers, Poitiers, 14-15 mai 2009, 17 p.
Figuière, C., & Guilhot, L. (2007). Vers une typologie des "processus" régionaux. Le cas de l'Asie Orientale.
Revue Tiers-Monde, 192, 895-917. http://dx.doi.org/10.3917/rtm.192.0895
GATT. (1993). La concurrence mondiale : Europe contre Amérique du Nord contre Extrême-Orient, Colloque
qui a eu lieu à Cernobbio, ITALIE, le 4 septembre 1993. Nouvelles de l'Uruguay Round''.NUR 064. 6
septembre 1993. p. 6.
Goyette, G. (2006). Protectionnisme et réciprocité commerciale : répertoire des obstacles au commerce
maintenus par les Etats-Unis, Cahier de Recherche. Centre Etudes Internationales et mondialisation. Institut
d'études internationales de Montréal. Université du Québec à Montréal.
Grossman, S. J., & Hart, O. D. (1983). An Analysis of the Principal-Agent Problem. Econometrica, 51(1), 7-45.
Holmstrom, B. (1979). Moral Hazard and Observability. Bell Journal of Economics, 10, 74-91.
Hauser, H., & Zimmermann, T. A. (2001). Régionalisme ou multilatéralisme, La vie économique; Revue de
politique économique.
Jensen, M. C., & Meckling W. H. (1976). Theory of the Firm: Managerial Behavior, Agency Costs and
Ownership Structure. Journal of Financial Economics, 3(4), 305-360.
Jewitt, I. (1988). Justifying the First-Order Approach to Principal-Agent Problems. Econometrica, 56(5),
Kadan, O., Reny, P. J., & Swinkels, J. M. (2001). Existence of Optimal Mechanisms in Principal-Agent Problems,
1 Janvier 2011 IMF Working Paper n° 2011-002.
Kearney, P. (2008). 40 ans d'Union douanière, Revue du Marché commun et de l'Union européenne, n. 522,
octobre-novembre, pp. 581-586.
Khavand, F. (2001). Commerce international: le Régionalisme menace t-il l'universalisme?, Institut International
d'Etudes Stratégiques - Paris, France.
Khavand, F. (1995). Le nouvel ordre commercial mondial, Nathan. Paris. 192p. (Collection Circa; 34).
194
Laffont, J. J. (1987). Le risque moral dans la relation de mondat. Revue Economique, 38(1), Janvier.
Mashayekhi, M., & Taisuke, I. (2005). Multilatéralisme et régionalisme: la nouvelle interface, CNUCED;
Nations Unies. New York et Genève.
McMillan, J. (2010). La politique réglementaire communautaire pour la libre circulation des marchandises et
l'approfondissement/ achèvement du marché intérieur, Revue du droit de l'Union européenne, n. 2, pp.
229-285.
Milgrom, P., & Shannon, C. (1994). Montone Comparative Statics. Econometrica, 64, 157-181.
OCDE. (2001a). Politiques de la concurrence et des échanges: options pour une plus grande cohérence, OCDE.
Paris.
OCDE. (2001b). Intégration régionale : effets commerciaux et autres effets économiques observés, TD/TC/WP
2001, 19/ FINAL Paris.
OCDE. (2003). Le régionalisme et le système commercial multilatéral, OCDE. Paris.
OMC. (1997). Dossier spécial : le commerce et la politique de la concurrence, Rapport Annuel de l'OMC 1997.
Volume 1. Organisation Mondiale du Commerce. Genève.
OMC. (1995). Le régionalisme et le système commercial mondial, Rapport Annuel de l'OMC 1997. Genève,
Avril 1995.
Paugam, J. M. (2007). Commerce mondial : avec ou sans l'OMC? Problèmes économiques, 2(915), 17 janvier.
Piergiuseppe, F., & Valensisi, G. (2011). Intégration commerciale régionale et opportunités de développement:
l'expérience africaine, International Centre for Trade and Sustainable Development. Eclairage, 10(6) .
Rogerson, W. (1985). The First-Order Approach to Principal-Agent Problems. Econometrica, 53, 1357-1368.
Shadikhodjaev, S. (2009). Trade Integration in the CIS Region: a Thorny Path Towards a Customs Union.
Journal of International Economic Law, 12(3), 555-578. http://dx.doi.org/10.1093/jiel/jgp023
Siroën, J. M. (2000a). Régionalisme et multilatéralisme, Les cahiers français 11/2000 n° 299. pp. 34-40.
Siroën, J. M. (2000b). La régionalisation de l'économie mondiale, La découverte. Paris. 121 p. (Collection
Repères; 288).
Trotignon, J. (2009). L'intégration régionale favorise-t-elle la multilatéralisation des échanges? Revue française
d'économie, 23(3), 213-246.
Tshiyembe, M. (2012). Régionalisme et problèmes d'intégration économique: ALENA, MERCOSUR, Union
Européenne, Union Africaine, L'Harmattan. Paris.
Venables, A. J. (2000). Winners and Losers From Regional Integration Agreements. Discussion Paper, Centre for
Economic Policy Research. London.
Viner, J. (1950). The Customs Union Issue. New York, Carnegie Endowment for International Peace.
195
ISSN 1916-971X E-ISSN 1916-9728
Monetary Policy and Its Implications for Balance of Payments

Stability in Nigeria: 1980-2010
Anthony Ilegbinosa Imoisi1, Lekan Moses Olatunji2 & Bosco Itoro Ekpeyong2
1
College of Social & Management Sciences, McPherson University, Ogun State, Nigeria
2
School of Graduate Studies, Faculty of Social Sciences, Department of Economics, University of Port Harcourt,
Rivers State, Nigeria
Correspondence: Anthony Ilegbinosa Imoisi, College of Social & Management Sciences, McPherson University,
Seriki Sotayo, Km 96, Lagos-Ibadan Expressway, P.M.B. 2094, Abeokuta, Ogun State, Nigeria. Tel:
80-34-525-743. E-mail: [email protected]
Abstract
This study examines the efficacy of monetary policy in achieving Balance of Payments stability in Nigeria. One
of the stabilization policies with which the government of Nigeria manages the economy is that of Monetary
Policy. Monetary Policy formulation in Nigeria is usually targeted at achieving some macro-economic objectives
amongst which is equilibrium in the country’s Balance of Payments (BOP). The general objective of this
research was to examine the relationship between the Balance of Payments position in Nigeria and monetary
policy adopted in the country. The research was conducted using an Ordinary Least Squares (OLS) technique of
multiple regression models using statistical time series data from 1980-2010. The estimated result shows a
positive relationship between the dependent variable (Balance of Payments) and the Independent variables
(Money Supply, Exchange Rate and Interest Rate). Specifically, Money Supply and Interest Rate had significant
relationship with Balance of Payments, whereas Exchange Rate was not statistically significant. Based on the
results, it was therefore recommended that the government should promote the exportation of Nigerian products
especially the Non oil products, as this will bring in more foreign exchange into the country, boost productive
activities and improve the balance of payments position of the country. Also, the Central Bank of Nigeria should
ensure that the monetary policies adopted in the country are complemented with effective fiscal policies to foster
economic growth and development in the Nigerian economy.
Keywords: balance of payments, monetary policy, fiscal policy, exchange rate, interest rate, money supply
1. Introduction
Monetary Policy is a key component of any pro-growth economic system and much so in developing economies
such as the Nigerian Economy (Taylor: 2004). In general terms, monetary policy refers to a combination of
measures designed to regulate the value, supply and cost of money in an economy in consonance with the
expected level of economic activity (Nnanna; 2001). For most economies, Nigerian economy inclusive, the
objectives of monetary policy includes price stability, maintenance of Balance of Payments equilibrium,
promotion of employment and output growth. Gbosi (2002), posits that monetary policy aims at controlling
money supply so as to counteract all undesirable trends in the economy, these undesirable trends may include;
unemployment, inflation, sluggish economic growth or disequilibrium in the Balance of Payments. Monetary
policy may either be expansionary or restrictive. An expansionary monetary policy is designed to stimulate the
growth of aggregate demand through increase in the rate of money supply thereby making credit more available
and interest rates lower. An expansionary monetary policy is more appropriate when aggregate demand is low in
relation to the capacity of the economy to produce goods and services. On the contrary, if the quantity of money
is reduced or restricted, money income will rise slowly so that consumers spend less and funds for investment
are difficult to acquire thereby decreasing aggregate investment (restrictive monetary policy).
Thus, to regulate monetary policy in the Nigerian economy, the Central Bank of Nigeria (CBN) employs various
instruments which include; Open Market Operation (OMO), Reserve Requirement (RR) and Discount Rate (DR),
CBN (1994). The success of monetary policy depends on the operating economic environment, institutional
framework adopted, and the choice and mix of the instrument used. However, the current monetary policy
196
framework focuses on the maintenance of price stability and Balance of Payments equilibrium, while the
promotion of economic growth and employment generation are secondary goals of the policy. Owing to the
nature of Nigeria’s export and import, there existed a persistent Balance of Payments deficit in the economy.
Invariably, Nigeria has paid more to foreign countries than she receives. Thus, the attendant result affects the
economy, leading to gross depletion of Nigeria’s Foreign Reserves. It has also attracted reduction in the
country’s productive capacities and persistent inflationary pressures.
Sequel to the above, articulated efforts have been made by monetary authorities especially, Central Bank of
Nigeria (CBN), on how to drastically reduce the Balance of Payment s deficits in the economy. This is usually
done through the formulation and implementation of appropriate monetary policy measures. The objective of this
study is to; identify the extent to which monetary policy has achieved Balance of Payment stability in Nigeria.
Identify the causes and effects of Balance of Payments deficits in the Nigeria economy from 1980-2010.
However, our basic focus here is to identify and define the extent to which monetary policy measures have
achieved Balance of Payments stability in Nigeria? Providing answers to justify this question is the basic thrust
and focus of the study.
A thought-provoking issue, which has occupied the mind of economists and monetary authorities for decade is
the effectiveness of monetary policy in achieving macro-economic objectives. Notwithstanding however, there is
the lack of consensus among economists on how it actually works and/or the magnitude of its effect on the
economy. Nkoro (2003) observes that there exists a remarkable and strong agreement that monetary policy has
some measure of effects on the economy. The Nigerian economy, as in other economies has an apex bank;
Central bank of Nigeria (CBN), which has the authority and mandate of manipulating or regulatory monetary
policy, using monetary instruments with the aim of achieving desired macro-economic objectives. In Nigeria,
these broad objectives include the mandate to conduct and regulate monetary and financial policies with a view o
promoting economic growth and development in Nigeria, (Nkoro, 2003).
However, the primary objective of monetary policy in any modern economy is the maintenance of price stability
which is fundamental to the attainment of sustainable growth. Nnanna (2001) observed that the pursuit of price
stability invariably implies the indirect pursuit of objectives such as Balance of Payments (BOP) equilibrium.
Anyanwu (1993) posits that an excess supply of money in the economy will result to excess demand for goods
and services and in turn causes rise in prices and also, affect the Balance of Payments position. With the
achievement of price stability, the uncertainties of general price level will not materially affect consumption and
investment decisions. Rather, economic agents will take long-term decision without much reservation about
price change in the macro-economy. The condition in the financial markets and institutions would create a high
degree of confidence, such that the financial infrastructure of the economy is able to meet the requirements of
market participants (Nkoro 2003). In other words, an unstable and crisis-ridden financial system will render the
transmission mechanism of monetary policy less effective, making the achievement and maintenance of strong
macroeconomic fundamentals difficult.
2.1 Instruments of Monetary Policy
In its effort to achieve a stable macroeconomic policy, the Central Bank of Nigeria adopts and employs monetary
policy instruments. As identified by Iyoha (2002), the instrument of discretional monetary policy includes:
Open Market Operation: - This involves the sales or purchase of government securities (Treasury bills) to and
from deposit money banks and non-banking institutions with the view to regulate the costs and availability of
credit.
Discount Rate: - This is the rate at which the Central Bank of Nigeria lends to deposit money banks. The interest
rate charged by the Central Bank of Nigeria is known as discount rate. By varying discount rate, the CBN can
influence credit availability as lender of last resort to the deposit money banks. Its direct impact is on credit cost
that has direct impacts on banks.
Reserve Requirement: - The reserve requirement sets minimum balance on the liquidity of deposit money banks,
viz-a-viz their balances. It has two uses, to ensure the solvency of the banking system and control the expansion
of credit creation as an objective of monetary policy.
Moral Suasion - This is a process by which the Central Bank of Nigeria makes known to the deposit money
banks officials through informal (oral or written) discussions on the direction in which they wish monetary
policy to proceed and the contribution which is expected of the deposit money banks.
197
Direct Control of Banking System: - It involves the imposition of quantitative ceiling on the overall and /or
selective distribution of credit by the Central Bank of Nigeria. The tools are selective, not general, it is also
direct.
Direct Regulation of Interest Rate: - This is generally used in Less Developing Countries (LDCs) and not in
More Developed Countries (MDCs). In MDCs, interest rate is determined by market forces to a large extent.
However, in LDCs, interest rate is regulated or administered. In particular, interest rate is forced within, which
both the deposits and the lending rates are expected to be maintained by the deposit money banks.
2.2 Conceptual Framework on Balance of Payments
The balance of payments is defined as a systematic record of economic and financial transactions for a given
period of time, say one year, between residents of an economy and non residents - rest of the world. These
transactions involves the provision and receipts of real resources – goods, services and income – and changes in
claims on and liabilities to the rest of the world. Specifically, the balance of payments records transaction in
goods, services and income, changes in ownership and other changes in an economy’s holdings of monetary gold,
Special Drawing Rights (SDRs) and claims on and liabilities to the rest of the world. It also records unrequited
or unilateral transfers – the provision or receipts of an economic value without the acceptance or relinquishing of
something of equal value. Generally, transactions involving payments to a country by non-resident are classified
as credit entries. Those involving payments by country to non-residents are debt entries.
Basically, the balance of payments is divided into the current and capital account. The capital account is made up
of portfolio and direct investment, either long or short term capital and capital transfers. While the current
account records all current transactions, which are transactions that include either the export or import of goods
and services. They include merchandise and services. The capital account also refers to charges in financial
assets and liabilities, portfolio investment, external loan drawings and amortization and charges in short-term
capital movements. However, it should be noted that development in the other sectors – real, monetary and
public – has implications for the balance of payments. As a result, current account deficit may not necessarily be
an inappropriate policy to pursue especially in a country that is for example, importing to increase domestic
investment. However, in a short-term, import bills may remain unpaid or external reserves could be drawn down.
A long-term and more viable solution lies in ensuring balance of payments viability. A viable balance of
payments position may be defined as a current account position, which can be financed on a sustainable basis by
net capital movements on terms that are compatible with reasonable development, growth prospects and debt
servicing capacity as well as macro-economic stability. It can be seen that the balance of payments is linked with
the other accounts in a general equilibrium framework. This implies that disequilibrium in one sector; say
external sector is transmitted to the other sectors and vice versa. Thus, there is need to achieve both internal and
external balance.
According to Marsha (1994), two types of policy measures are used in dealing with balance of payments
problems. These are expenditure switching measures and expenditure reducing policies. Expenditure reducing
policies refer to fiscal policy (conducted by changing government expenditure and /or taxes) and monetary
policy which refers to changes in money supply, which in turn affect interest rate. Expenditure switching policies
refers to devaluation (depreciation) and revaluation (appreciation) of the country’s currency. The aim of
expenditure reducing policies is to reduce domestic expenditure on consumption and increase expenditure on
investment, thus, releasing goods and services for exports while leaving aggregate output unchanged. The aim of
expenditure switching policies is to switch domestic demand from imported goods to home made goods.
However, the extent to which expenditure switching policies is achieved depends on elasticity of supply and
demand for tradable goods. If the depreciation of the nominal exchange rate is matched by increase in wages,
absorption and inflation, the real exchange rate would not depreciate and so the balance of payments would not
improve. However, expenditure reducing policies have costs in terms of loss of output, investment and
employment. The loss will be minimized if resources can be easily moved to the tradable goods sector.
Alternatively bridging external loans may be contributed to sustain investment and output.
2.3 Trends in Nigeria’s Balance of Payments
From available data gathered, it showed that in the first half of 1994, there persisted a pressure on the balance of
payments position in the country. An overall deficit amounting to N42, 623.3 million was recorded in the
balance of payments compared with the surplus of N13, 615.9 million in the corresponding period of 1993. This
deficit was noticeable due to the huge current account deficit which substantially outweighed the surplus
recorded in the capital account. The deficit was financed through a reschedule of debt service responsibility
estimated at N24, 906.4 million ($1, 138.0 million) during this period. As a result of this development, the
198
external debt reduced from N29, 093.0 million ($1, 329.3 million) at the end of 1993 to N26, 722.8 million ($1,
222.0 million) at the end of 1994, (CBN, 2010). Nigeria’s overall balance of payments however, recorded a
surplus of N19, 531.3 million during the first half of 1995.
The balance of payments position of the country however plunged back into a deficit in 1996, 1998 and 1999.
This continued decline in the country’s balance of payments was assigned to the engorged deficits in the current
account which offset the surplus recorded in the capital account. As in the previous years, the financing of the
deficit was largely through further accrual of external debt responsibility which fell and amounted to N89, 813
million. Due to this deficit, the main external sector policy adopted in the year 2000 was to build the country’s
external reserves. The intended purpose was to restore confidence in the Nigerian Naira (N) and in the entire
economy. This effect of this external sector policy was reflected in the balance of payments position of the
country. As a result of this policy, the country recorded a surplus in the balance of payments position i.e. Nigeria
recorded a surplus of N314, 139.2 million in 2000, and N24, 738.7 million in 2001.
The weak position in the country’s current account was due to the deterioration in the services and income
account which outweighed the surplus recorded in the merchandise trade and involved net transfer account,
(Gbosi, 2001).In recent years, there have persistent deficit in the country’s balance of payments (See table 1).
Nigeria’s balance of payments recorded remarkable improvement during the period 2004-2005. However, the
situation worsened in 2008 as a result of the global financial and econosmic meltdown coupled with the falling
prices of crude oil in the international oil market (Gbosi, 2009).
2.4 Balance of Payments Theories
There are two basic theories that have been propounded to addressing balance of payments imbalance, these
include:
Inflationary theory: Inflation is a state of persistent rise in the general price level and hence falling value of
money, Dullo (1974). It is a malign condition that eats accumulated wealth and diverts the energies of the
economy. Countries report by the IMF, shows that the cause of Nigeria’s inflation are; increase in money supply
despite decrease in foreign exchange reserves (a decrease in foreign exchange reserve has the effect of
decreasing money supply). Budget deficit is also stated to be a contributory factor. Faced with increasing
population and the need to improve the standard of living, the Nigerian government has embarked on various
programmes to accelerate the rate of economic growth and provide government services, thereby increasing
expenditure within a limited scope of public borrowing leading to fiscal deficits.
Structural Theory: This theory argues that balance of payments disequilibrium abates due to an inherently
inefficient or imbalanced economy Gbosi (2001). Two specifications of structural problems that affect the
Nigerian economy are:
Weakness in fiscal system: This leads to budget deficit, expenditure increases due to population increase and the
need for development, while the revenue system and tax rate of the Nigerian economy are inadequate to obtain
the needed growth in revenue. What is needed is restructuring and improvement of the country’s revenue system
and increase in taxes. The revenue system of the economy should be elastic relative to economic growth, that is,
revenue should grow proportionally with higher GNP.
High External Debt Burden: Debt sustainability analysis of Nigeria by the IMF indicates that the country’s debt
has been increasing since1960. Over a period of 30 years, the external debt has risen by 2,899 percent.
Determining whether or not the level of debt is sustainable in the country is one of the most fundamental issues.
There is no conclusive level of measure amongst economists to determine when an external debt is sustainable or
not. However, for debt to be sustainable over the long term, a country’s rate of economic growth should be
higher than the rate of interest on foreign loans.
Structural inadequacies of Nigeria arose mainly from the flowing sources: Dependence on one primary
commodity (especially petroleum) as a major source of foreign exchange earner. This commodity is open to
world price fluctuations which affects the current account of the balance of payments; Excessive debt service
payment due to high non-concessional interest rates; and Weak industrial base by the manufacturing sector of the
country.
3. Method of the Study
Here, the research methodology used for the study will be discussed. Emphases were laid on the sources of data,
model specification, apriori expectation of variables, etc.
199
3.1 Sources of Data

The data used for this study is mainly secondary data. Some of these data sources include the publications of
Central Bank of Nigeria (CBN), National Bureau of Statistics (NBS), published and unpublished sources,
magazines, journals, text books and News papers.
3.2 Model Specification
The model was analyzed using an econometric model of multiple regression analysis to test the relationship
between the dependent variable (BOP) and independent variables (Money Supply, Interest Rate and Exchange
Rate). The model is specified as thus:
BOP = f (MS, IR, EXR) (1)
We can also specify the above equation in an econometric form
BOPt = a0 + a1MSt + a2IRt + a3EXRt + Ut (2)
While the log-linear function of the model is specified as thus:
Log BOPt = Log a0 + a1Log MSt + a2Log IRt + a3 Log EXRt + Ut (3)
Where:
BOPt = Balance of Payments (BOP) at time‘t’
MSt = Money Supply (MS) at time‘t’
IRt= Interest Rate (IR) at time‘t’
EXRt= Exchange Rate (EXR) at time‘t’
Ut = Error term at time‘t’
3.3 Apriori Expectation
Based on economic theory, an increase in money supply will bring about an increase in the total money in
circulation in the country. This will therefore increase aggregate demand and lead to a rise in productive
activities and investment opportunities in the economy. This rise in the productive activities in the economy will
lead to a rise in the export of goods and services, thus, leading to a rise in the balance of payments position of the
country. Hence, the coefficient of MS will be positive; i.e. a1 > 0. Also, from economic theory, an increase in the
rate of interest will discourage investors from borrowing funds from the financial sector, thus, reducing the level
of investment and productive activities in the economy. Such reduction in the productive activities of the country
will lead to fall in the balance of payments position in the country. Hence, the coefficient of IR will be negative;
a2 < 0. Finally, economic theory tells us that an increase in exchange rate (price of a country’s currency with
respect to other currencies of the world) will make the domestic currency cheaper in the foreign exchange market.
This depreciation in the domestic currency will make our exports cheaper and imports expensive in the
international market for goods and services. As a result of this, our goods and services would be more in demand
thus, leading to a rise in the balance of payments position of the country. Hence, a3 > 0
200
4. Data Presentation and Analysis
Table 1. Data on money supply, interest rate, balance of payments and exchange rate in Nigeria from 1980-2010
YEAR Money Supply (MS) M1 Interest Rate (IR) Balance of Payments (BOP) Exchange Rate (EXR)
1980 9650.7 9.5 2402.2 0.5464
1981 9915.3 10 -3020.8 0.6100
1982 10219.8 11.8 -1398.3 0.6729
1983 11517.8 11.5 -301.3 0.7241
1984 12497.1 13 354.9 0.7649
1985 13878.0 11.8 349.1 0.8938
1986 13560.4 12 -784.3 2.0206
1987 15195.7 19.2 159.2 4.0179
1988 22232.1 17.6 -2294.1 4.5367
1989 26268.8 24.6 8727.8 7.3916
1990 39156.2 27.7 18498.2 8.0378
1991 50071.7 20.8 5959.6 9.9095
1992 75970.3 31.2 -65271.8 17.2984
1993 118715.4 18.3 13615.9 22.0511
1994 169391.5 21 -42623.3 21.8861
1995 201414.5 20.8 19531.3 21.8861
1996 227464.4 20.9 -53152 21.8861
1997 268622.9 23.3 1076.3 21.8861
1998 318576.0 21.3 -220675.1 21.8860
1999 393078.8 27.2 -326634.3 92.5284
2000 637731.1 21.6 314139.2 109.5500
2001 816707.6 21.3 24738.7 112.4864
2002 946253.4 29.7 -567353.3 126.4000
2003 1225559.3 22.5 -162839.7 135.4067
2004 1330657.8 20.7 1128379.4 132.8417
2005 1725395.8 19.5 1364845.5 130.5550
2006 2280648.9 18.70 68348.84 128.2700
2007 3116272.2 18.36 62187.02 117.9860
2008 4857312.2 18.70 63648.72 130.7500
2009 5003866.6 22.90 80826.64 147.6000
2010 4930589.4 19.99 68887.46 139.1750
Source: CBN, Nigeria Major Economic, Financial and Banking Indicators, 2009.
CBN Statistical Bulletin, 2010.
National Bureau of Statistics 2009.
The table above presents the relationship between money supply, interest rate, exchange rate (Independent
Variables) and balance of payments (Dependent Variable) in Nigeria for the period under review i.e. 1980-2010.
From Table I, we can see that during the period of 1986-1987, when the country introduced the Structural
Adjustment Programme; whose objectives included diversifying the productive base of the economy, adoption of
realistic exchange rate through the establishment of Foreign Exchange Market, minimal inflationary growth rate
etc, there was an improvement in the position of the country’s balance of payments, i.e. the country recorded a
surplus in her balance of payments position. As depicted from the table, during the 1986-87, as the exchange rate
of the Nigerian naira to the United States Dollars depreciated from N2.02 - $1 to N4.02 - $1, the balance of
payments position moved from a deficit of 748.3 to a surplus of 159.2. This depreciation of the Nigerian naira
will now make the exports of Nigerian products cheaper and imports of foreign goods into the economy more
expensive. Thus, it will improve the nation’s balance of payments position. Also, from the table, from 1994-1995,
due to financial sector reforms by the Central Bank of Nigeria we noticed an increase in the money supply in the
economy i.e. from N169,391.5 million to N201,414.5 million. This increase in money supply moved the
201
country’s balance of payments position from a deficit of 42623.3 to a surplus of 19531.3. This shows that
monetary policy such as changes in the money supply can be used to influence the balance of payments position
of the country, (Gbosi, 2001).
4.1 Data Analysis and Interpretation of Result
BOP = 167714.4 + 0.9326MS – 1.1425IR + 9.6246EXR
t-stat (0.89) (4.21) (-2.81) (-1.13)
2 2
R = 0.76, Adjusted R = 0.72, F =10.10, Durbin Watson = 1.1364
Source: IBM SPSS version 19
The analysis was estimated both in linear form and log-linear form, the linear estimation gave us a better result
and was adopted for our analysis based on the goodness of fit of the regression model. From our result, the
coefficient of the Independent variables (Money Supply, Interest Rate and Exchange Rate) appeared with the
correct sign, thus, conforming to economic theory or apriori expectation.
The coefficient of autonomous balance of payments is 167714.4, meaning that if all the independent variables
are held constant; there will be a positive variation in the balance of payments position in the country to the tune
of 167714.4.
Money Supply appeared with the right sign, which is a positive sign and thus, conform to economic theory. The
indication of this is that there is a positive relationship between money supply and balance of payments in
Nigeria for the period under review. From our result, we observed that the coefficient of money supply is 0.9326
indicating that a unit increase in money supply, would lead to a 0.9326 increase in balance of payments. The
estimated coefficient of interest rate appeared with the right sign which is negative and thus, conforms to
economic theory, which states that there is an inverse relationship between balance of payments and interest rate.
Our result shows that the coefficient of interest rate is -1.1425, meaning that a unit increase in interest rate would
lead to a 1.1425 decrease in balance of payments.
Similarly, the coefficient of exchange rate appeared with the right sign, which is a positive sign, thus,
conforming to apriori expectation. This means that there is a positive relationship between exchange rate and
balance of payments in Nigeria for the period under review. Our result showed that the coefficient of exchange
rate is 9.6246, meaning that an increase in exchange rate would lead to a 9.6246 increase in the balance of
payments position in the country.
Our result also showed that the coefficient of multiple regression (R2) is 0.76, meaning that 76% of the
dependent variable (balance of payments) is explained by the independent variables (money supply, interest rate
and exchange rate), while the other 24% is explained by factors not included in the model, but are captured by
the error term for the period under review (1980-2010). This also indicates that the goodness of fit of the
regression result is strong and implies that 76% variation in balance of payments is explained by money supply,
interest rate and exchange rate.
The test of significance from our result showed that only money supply and interest rate were statistically
significant for the period under review at 5% level of significance. This is due to the fact that their calculated t
value (tcal) in absolute terms is greater than their theoretical value of t (ttab). For instance, the tcal of money supply
in absolute terms (4.21) is greater than the ttab (1.70), implying that there is a significant relationship between
money supply and balance of payments. Also, the tcal of interest rate in absolute terms (2.81) is greater than the
ttab (1.70), indicating that the relationship between balance of payments and interest rate is statistically significant.
However, exchange rate was not statistically significant for the period under review because its calculated t value
in absolute term (tcal = 1.13) is less than the theoretical value of t (ttab = 1.70).
The F test, which shows the significance of the entire regression model from our result, was significant. This is
due to the fact that the observed F-cal ratio (10.10) is greater than the theoretical value of F (1.88) and this
further confirms the value of the R2.
4.2 Discussion of Findings
For this study, quantitative tools were employed to verify the impact of monetary policies on balance of
payments stability in Nigeria for the period under review (1980-2010). From our result, we discovered the
following:
1). Money Supply which is the total money in circulation in the economy contributed significantly to the balance
of payments position in the country. This is revealed by the positive value of the coefficient of money supply.
Hence, the Central Bank of Nigeria has been relying on policies regarding the manipulation of money supply as
202
one of its monetary policies in ensuring stability in the country’s balance of payments position. This means that
if the total money in circulation (money supply) is increasing in Nigeria, there would be an increase in the
balance of payments position in the country for the period under review. When there is an increase in money
supply in the country, it would lead to a fall in interest rates; this fall in interest rates will encourage borrowing
from the financial institutions for investment purposes. An increase in investment opportunities would lead to an
increase the level of productive activities in the country. This rise in productive activities will lead to an increase
in exports of goods and services to other countries of the world, thus, increasing the level of foreign exchange
earnings into the country and leading to a favourable position in the country’s balance of payments.
2). Interest rate which is the cost of borrowing funds from deposit money banks contributed positively to the
balance of payments position in the country for the period under review. Regulation of interest rate is also one of
the monetary policies used by the Central Bank on Nigeria in achieving balance of payments stability. If there is
a rise in the rate of interest in the Nigerian economy, this will discourage borrowing from deposit money banks
for investment purposes in the country. However, reverse would be the case if there is a fall in interest rate. This
fall in interest rate in the country will increase the level of borrowing from the deposit money banks and will
ultimately lead to an increase in the level of investment opportunities in the country. Due to this, there will be an
increase in the production of goods and services in the country. As a result of the increase in production of goods
and services, the prices of these goods and services will reduce, thus making our exports cheaper and improving
the country’s balance of payments position.
3). Exchange rate which is the price of the domestic currency (in this case, the naira) in terms of the other
currencies of the world had a positive impact on balance of payments but was not significant for the period under
review (1980-2010). This means that in order to ensure balance of payments stability in the country, the
monetary authorities have been relying more on expenditure reducing polices such as monetary policy than on
expenditure switching policies such as devaluation of the country’s currency. If the Exchange Rate of the
Nigerian Currency (Naira) is increasing in relation to other currencies of the world, the balance of payments
position for the Nigerian economy will also be increasing and as such will lead to a favourable position in
Nigeria’s Balance of Payments. This is possible because as the Exchange Rate of the Nigerian Naira is
increasing in relation to other currencies of the world, it would make the country’s exports cheaper and imports
more expensive, thus switching expenditure from foreign goods to domestic goods.
5. Conclusion and Recommendations
This work examines monetary policy and its impact on balance of payments (BOP) stability in Nigeria from
1980 - 2010. Balance of payments position of any country is one of the indicators of economic growth and as
such countries try to make sure that they realize favourable balance for payments position. The paper presents
monetary policy as a key component of any pro - growth economic strategy. In Nigeria, the design and
implementation of monetary policy is the responsibility of the Central Bank of Nigeria (CBN). In line with the
above, monetary policy as an intervention strategy by the CBN is aimed primarily at maintenance of price
stability and balance of payments (BOP) equilibrium whereas the promotion of economic growth and full
employment are secondary. Therefore, it is the responsibility of the government to initiate policies that will
guide against political tussles and instability. Based on this, the following recommendations were made: The
Central Bank of Nigeria (CBN) should intensify the process of regular monitoring of the operation of deposit
money banks to ensure compliance with prudent guidelines and promote transparency in the banking operations;
The government should protect infant industries through the following ways: by raising high tariff on those
goods that are produced outside the country; encourage import substitution; support local industries and
manufacturing sector by giving them incentives such as; Tax holiday etc; the government should embark on
efficient and effective expenditure switching policy or devaluation of Nigeria Currency (Naira), as devaluation
of the country’s currency will make exports cheaper and imports more expensive, thus, leading to a favourable
balance of payments position in the country. It should be noted that despite the efforts of the Central Bank of
Nigeria in trying to use monetary policy in ensuring balance of payments stability in Nigeria, it has not made the
desired impact with regards to the stabilization of the balance of payments position in the country. Also, for
monetary policy to be effective in ensuring stability in the balance of payments position of the economy, it
should be complemented with an effective fiscal policy.
References
Ajayi, S. L. (1974). An Econometric Case study of Relative importance of Monetary and Fiscal Policy in Nigeria.
Bangladesh Economic Review, 11(2).
Anderson, L. C., & Jordan, J. L. (1968). Monetary and Fiscal Action A test of their Relative Importance in
Economic Stabilization. Federal Reserve Bank of St. Louis Review reprint series no. 34.
203
Anyanwu, J. C. (1993). Monetary Economics: Theory, Policy and Institution. Joane Educational Publishing Ltd,
Onitsha.
Central Bank of Nigeria. (2010). Statistical Bulletin, 20. Abuja.
Central Bank of Nigeria (2009). Nigeria Major Economic, Financial and Banking Indicators. Abuja.
Fieleke, N. S. (1996). What is Balance of Payments? Federal Reserve Bank of Boston Economic and Financial
Review.
Fischer, S. (1994). Modern Central Banking: The Future of Central Bank. The Tercentenary Symposium of the
Bank of England: Cambridge University Press.
Folawewo, A. O., & Osinubi, T. S. (2006). Monetary Policy and Macroeconomic Instability in Nigeria: A
Rational Expectation Approach. Kenila Raj.
Gbanador, C. A. (2005). Devaluation and Balance of Payments Stability in an Oil Producing Economy: lessons
from Nigeria Experience. The Nigerian Journal of Monetary Economics (NJOME), 5, 118-131
Gbosi, A. N. (2002). Financial Sector Instability and Challenges to Nigeria’s Monetary Authorities. African
Heritage Publishers, Port Harcourt
Ikhide, S. I., & Alawode, A. A. (1994). Financial Sector Reforms, Macroeconomics Instability and the order of
Economic Research Liberalization; the Evidence from Nigeria. African Economic Research Consortium
(AERC) Final report.
Ikhide, S. I. (1996). Financial Sector Reforms and Monetary Policy in Nigeria. IDSS working paper 68.
Iyoha, M. A. (2002). Macroeconomics: Theory and Policy. March Publishers, Benin City.
Iyoha, M. A. (2003). An Overview of leading Issues in the Structure and Development of Nigerian Economy
since 1960. In Iyoha, M. A. & Itsede, C. O. (Eds.), Nigerian Economy: Structure Growth and Development.
Mindex Publishing Benin City.
Jhingan, M. L. (1993). Macroeconomic Theory. Delhi: Konar Publishers.
Nkoro, E. (2003). Analysis of the Impact of Monetary policy on Economic Development in Nigeria (1980-2003).
University of Benin City.
Nnanna, O. J. (2001). Monetary Management Objectives, Tools and the Role of Central Banks in the Region.
Regional Forum on Economic and Financial Managements for Parliamentarian. Nigeria: WAIFEM.
Odozi, V. A. (1995). The Conduct of Monetary and Banking Policies by the Central Bank of Nigeria. Economic
and Financial review, 33(1).
Oduma, J. S. (1980). How effective have Fiscal and Monetary Policies been in Nigeria? CSER Print paper no. 7
Abu Zaria
Okaha, G. O. (1986). Theoretical Basis of Monetary Policy in Africa. Economic and Financial review, 10.
Ojo, M. O. (1987). Monetary Policy Instrument in Nigeria: Their Changing Nature and Implication. The
Nigerian Bank.
Sanuse, J. O. (2002). The Evolution of Monetary Management in Nigeria and its Impacts on Economic
Development. CBN bullion, 26(1).
Soludo, C. (2001). Debt Poverty and Inequalities: Towards an Exit Strategy for Nigeria and Africa. Proceedings
from International Conference on Sustainable Debt Strategy. Abuja Nigeria Processed.
Tailor, J. B. (2004). Improvements in Monetary Policy and Implications for Nigeria. Key Note Address, Money
Market Association of Nigeria, Abuja, Nigeria.
Uchendu, O. A. (1996). The Transmission of Monetary Policy in Nigeria. Central Bank of Nigeria Economic
and Financial Review, 34(2), 608.
Udegbunam, R. I. (2003). Monetary and Financial Policy. In Iyoha, M. A. & C. O. Itsede (Eds.), Nigerian
Economy: Structure Growth and Development. Mindex Publishing Benin City.
204
ISSN 1916-971X E-ISSN 1916-9728
Does Fiscal Policy Matters for Growth?

Empirical Evidence from Pakistan
Rabia Nazir1,2, Mumtaz Anwar2, Mamoona Irshad2 & Ayza Shoukat2
1
Department of Economics, The Islamia University of Bahawalpur, Pakistan
2
University of Punjab Lahore, Pakistan
Correspondence: Rabia Nazir, Lecturer, The Islamia University of Bahawalpur, Bahawalnagar Campus, District
Bahawalnagar Province, Punjab, Pakistan. Tel: 923-36-635-0651. E-mail: [email protected]
Received: December 18, 2012 Accepted: February 9, 2013 Online Published: February 26, 2013
Abstract
The present study is designed to investigate the short and long run impact of fiscal strategy variables on GDP
growth of Pakistan by employing Johanson co-integration technique and ECM. Data on GDP per capita, per
capita real public revenues, government final consumption expenditures, discount rate, trade openness, and gross
fixed capital formation has been gained from various sources like world development indicators, FBS Pakistan
and the economic survey of Pakistan (various sources). In long run government consumption expenditures and
public revenues both are affecting GDP significantly with negative and positive coefficients respectively.
Moreover ECM indicates that approximately 37% of the disequilibrium error is corrected in each period which is
a good speed of convergence. The Reduction of government consumption expenditures and enhancing revenue
generation efficiency is recommended for better outcomes of fiscal policy in Pakistan.
Keywords: fiscal policy, short and long run, ADF, cointigration, GDP growth
1. Introduction
Concept of economic growth is getting tremendous importance from researchers and policy makers because of
the fact that economic progress plays a vibrant role in the development of social, economic and political welfare
of nations. Fiscal trade and monetary policies play an important role to boost the level of economic growth
(Jawaid et al., 2011).
Among above stated macro level policies fiscal policy is key government policy to lead the economy towards
faster economic growth. Fiscal policy of the current period reflects the pattern of the economy in the coming
time period. Fiscal policy uses government spending and taxes to control and stimulate economic growth.
Theoretically fiscal policy can be employed to affect inflation, collective demand, and level of economic activity,
allocation and distribution of resources and to avoid economic depression. Fiscal policy actions consist of two
set of actions: first is discretionary action in which government sets tax rate, tax base and the size of the
government. On the other hand are the automatic stabilizers in which the variables change due to change in the
economic environment like during economic boom tax revenues will increase automatically and spending on
social benefits decline and vice versa in times of economic downfall (Akram et al., 2011). But practically there
are many limitations of using fiscal policy.
Economic survey of Pakistan 2005-06 segregated objectives of fiscal policy in micro and macro objectives. An
equitable distribution of resources, income and social services, increased investment in public goods, fulfilling
the basic needs of the poor and enhancing efficiency of public and private enterprises are micro economic
objectives and evolution of the economy as a whole is the macroeconomic perspective. Casey McCracken (2006)
claimed that it is the responsibility of the government to manage its fiscal policy in an efficient way. The
Government should decide about the nature and amount of taxes and expenditures in well-defined manner
because taxes and expenditures can affect the incomes and working of the millions of the targeted population.
Government can use expansionary or deflationary fiscal stance to meet the above stated objectives.
What will be the shape of the fiscal policy of any country is mostly determined by the political ideology of the
political leaders as is claimed by the political economists. According to them distributional conflict on fiscal
policy is classified in three categories, first group says that governments try to tie hands of the successors mostly
through use of the budget deficit, second group blames policy makers for delaying fiscal decisions and third
205
claims overspending is made to create trouble for the upcoming leadership which depend on the personality of
the policy maker (Aghion & Bolton, 1990; Persson & Svensson, 1989; Alesina & Tabellini, 1990; Tabellini &
Alesina, 1990; Rogoff & Sibert, 1988; Rogoff, 1990).
1.1 Research Question
Starting from the 1960s, East Asian countries achieved sustained long run GDP growth with the help of
persistent and good economic administration. According to Kakar (2011) decline in economic growth is mainly
attributed to the same macroeconomic imbalance which has put the nation to the abyss of poverty as well. Kakar
(2011) has also suggested that the government should be able to demarcate between the short run and long run
objectives of public policy. So the research problem at hand is very much important for the development of a
struggling country like Pakistan. Many studies has determined the influence of fiscal policy on the economic
development of Pakistan .This paper is also an attempt to identify whether Pakistan has been able to use its fiscal
policy to promote its economic growth both in the long or short run?
The paper is arranged as follows: section 1.3 evaluates the theoretical and empirical literature being conducted
on the influence of fiscal policy. In section 1.4 we have developed the hypothesis of the study. Section 2 is
comprised of data and methodology and section 3 and 4 explains the results and conclude the study respectively.
Wide range of literature is available on the important role of fiscal policy in fastening economic growth. There
are two schools of thought who explain the role of public policy in two distinct ways; they are neo classical
school of thought and the Keynesians. New classical school of thought blames fiscal policy for creating
crowding out of private investment and causing inflation. Increase in the public debt cause an increase in interest
rate which put downward pressure on output and inflation. Moreover due to increase in public debt public will
anticipate higher taxes in future which further raise labor supply, lower consumption, real wages, current
economic activity and inflation (Shaheen and Turner, 2008). All of the economists believe that the classical
model holds for the time frame of a decade or more. Afonso et al., (2005) claimed that in addition to economic
factors many social, cultural, environmental and geographic factors are involved in the long run so no single
explanation can be given to fully predict the behavior of fiscal policy that’s why now days the concept of the
quality of public finances is becoming popular to improve the role of fiscal policy for raising long run growth
potential.
New Keynesian School explains crowding in or multiplier affects in a way that public spending increase demand
and economic activity which ultimately leads to make fiscal policy successful (Shaheen and Turner, 2008). In
societies with developed and efficient money and capital markets LM curve is perfectly elastic or the IS curve is
insensitive to changes in interest rate so in such a situation monetary policy would fail to raise the income and
fiscal policy will work here. But economists agree that Keynesians framework only woks in a time span of a few
months or less (Afonso et al, 2005).
Literature on fiscal policy can be divided into theoretical and empirical parts; empirical literature on fiscal policy
is further divided into following categories: Studies which identify networks through which fiscal policy moves
economic growth and empirical studies which explain the progressive or adverse impact of fiscal policy
variables on economic progress in long and short time framework.
In empirical studies we first focus on the studies which have identified different channels through which fiscal
policy variables exert their impact on economic growth. (Gerson, 1998) identified three types of fiscal policies
which can affect economic growth of an economy: Firstly, are the policies which through focusing on education
and health level of the workers increase the skill of labor. Fiscal policy through focusing on the development of
the human capital and improving the productivity of workers exert positive impact on the economy. Second
channel works through increasing the productivity of the stock of physical capital both by increasing the efficacy
of the existing capital stock through investing in new projects to build up new physical capital stock. Investment
in new technology, infrastructure and research and development via fiscal policy can bring up good outcomes for
growth. The fiscal policy which encourages open trade (for facilitating import of modern technology) also
proves successful (Barro, 1990). Third are the policies which can increase or decrease the supply of labor and
quantity of capital stock through tax incentives or disincentives.
Baldacci et al., (2003) in their paper “Transmission Channels in Low-Income Countries” proved that factor
output is the most important channel in developing or low income countries through which public policy
enhances economic growth while for developed countries private investment is major transmission route. Their
empirical report states that factor product was four times more efficient as a channel to promote economic
growth than investment in poor income countries.
206
Afonso et al., (2005) have also identified certain channels through which economic progress of a country is
moved on by the use of public policy. First identified channel is size of the public sector which can either retard
or accelerate economic growth. If the government sector is very large with efficient institutions then it will fasten
the pace of economic growth but if large public sector is accompanied with higher tax burdens and inefficient
public administration then fiscal policy may fail to increase economic growth. Second is the composition and
efficiency of public expenditures which can change the behavior of fiscal policy. If the expenditures are more
development oriented i-e development expenditures on infrastructure, education, research and development; and
ther are being done in an efficient manner than fiscal policy can be successful. Third is Structure and efficiency
of the revenue system like tax system including tax reforms and introduction of less distortionary taxes; is yet
another key to success. Fourth one is good fiscal governance as another channel to promote economic growth via
fiscal policy. Fifth channel is appropriate functioning of markets and business environment which can be
improved via efficient and less distortionary revenue and expenditures systems and through solid fiscal
institutions. Jawaid et al, (2011) proved that economic development has a positive relationship with monetary
and fiscal policies both in short and long run in Pakistan. In their study monetary policy was found more
effective than fiscal policy to enhance growth. Trade policy was also observed as having a very significant affect
over the growth rate of Pakistan. Study recommended focusing more on monetary policy to generate increasing
pace of economic growth.
Turnovsky (2002) analyzed the working of fiscal policy in an economy with non-scale growth level by taking
into account both the public and the private sector and hence found a substantial impact of fiscal variables on
growth during transition periods. He suggested that for regulated economy, government investment may bring
better results than government consumption.
Dong Fu et al., (2003) analyzed the relationship between fiscal policy and the U.S economy’s growth. They used
pairwise combination of several fiscal indicators and developed VAR technique to evaluate simultaneous shocks
to more than one variables. The technique was used to check the impulse response for coincident, unexpected
and corresponding shocks to pairwise mixtures of fiscal indicators. Size of the federal government and economic
growth were negatively related for their sample. Deficit came out to be the most inconsistent indicator while tax
revenues were the steadiest indicators of fiscal policy.
Gemmell et al., (2006) used panel data of OECD countries and found that in long run; distortionary taxes and
productive expenses have an adverse and positive impact on growth of OECD countries respectively. The long
run effects were homogeneous in all of the OECD countries while the short run effects were quite different in all
such countries. These long run effects were mostly achieved within the time frame of 1-3 years. The Short run
effects were significant and persevering provided that the fiscal policy changes were not retreated. This result
was also supported by Knellera, (1999) who explored that economic growth is adversely influenced by
distortionary type of taxes and positively influenced by productive government expenditures.
McCracken (2006) also analyzed fiscal policy growth via geographical analysis and by using two models one of
which disaggregated the fiscal variables. He identified negative relation between government size and growth
with transfer payments and income taxes having the most harmful impact. (area of study and the sample size).
Ogbole et al., (2007) conducted a comparative study to analyze the difference in growth behavior during
regulation and deregulation period of Nigerian economy. They used the time series data taken from central bank
of Nigeria. Impact of fiscal policy to enhance growth was different during regulation and deregulation periods.
GDP was 14% higher in deregulation period as compared to regulation period. An appropriate policy mix,
prudent public spending, setting the rational fiscal policy targets and broadening the country’s economic base
was recommended by the authors to enhance growth of the economy.
Romer & Romer (2007) by using narrative records assessed impact of taxation on economic movements to
explore magnitude, timing and principal incentive for all chief postwar tax policy activities. Study found tax
increases to be highly contractionary. Results were more significant, robust and much larger than the estimates
attained by using wider measures of tax. Higher tax on investment creates larger effects on GDP than other taxes.
Conclusion of the study was that reductions in budget deficit through legislated tax increases create less output
costs than other type of taxes. Babalola & Aminu, (2010) employed co-integration technique to assess the
association between public policy and Nigerian’s economy development. He found an encouraging and long run
productive expenditure impact on growth during the studied period. He suggested enhancing government
productive expenditures on schooling, health and economic amenities to increase economic progress.
On the other side Kalle Kukk (2007) found no significant influence of fiscal policy on economic progression or
development in short run. For example if direct taxes are increased, they will reduce private consumption and
increase net exports by the same amount. Thus this tax change does not affect total GDP. While on the other side
207
in long run fiscal policy had significant impact on real GDP growth. Tax revenues had positive and expenditures
categories had negative impact on GDP but non tax revenues were not significant in general. Only expenditures
which had positive impact on GDP growth was spending on public investment. The effect of interest and grant
spending was insignificant in many variable combinations. Study recommended government to recognize the
fact that change in the composition and categories of revenues and taxes might have same impact on budget
balance, total government revenues and expenditures but will have different impact on economic growth.
Shaheen and Turner (2008) also proved the same impact of fiscal policy on macro variables by employing SVAR
methodology for the period 1973 to 2008. Their estimates from recursive approach proved insignificant impact
of government expenditure shocks on output and inflation. But the results were different as obtained from the
Blanchard and Perotti (2002) approach which gave significant impact of government expenditures shocks and
taxes on output and inflation in Pakistan. Government expenditures shocks had progressive impact in short run
and adverse in the long run and had a tendency to increase over five year period. Interest rate was also increased
in sort run due to government shock.
Coming to the time frame of the fiscal policy impact, different studies has identified different time frames for the
turnout of the fiscal policy. Gemmell et al., (2011) used panel data of OECD countries since 1970s and found
that in long run fiscal structural changes between different forms and types of taxes and spending affect GDP
which is realized quite quickly following a few years. Moreover earlier evidence of long run effect seemed to
pick up short run and long run effects and long run normally persisted from one to five years of time period.
Type of tax and expenditures which is changed determines the impact of changing spending and revenues. For
example if the infrastructure spending is financed through highly distortionary taxes, then they will have a
negligible influence on GDP growth in long time frame.
1.3 Hypothesis Development
Kakar (2011) has discussed that fiscal policy can be used as a tool to boost fiscal sustainability and
macroeconomic stability. Pakistan is facing decreasing growth and rising poverty levels due to macro-economic
imbalances. Akram et al, (2011) claimed that Pakistan in 1990s had to decrease its budget deficit volume to less
than five percent of gross domestic product to meet its rising debt servicing requirements and to fulfill its
commitment with IMF under structural adjustment program. All successive governments being failed in revenue
generation efforts continued reducing development public spending to resolve budget deficit problem. After 9/11
Pakistan received immense funds under debt relief and debt rescheduling facility which helped Pakistan to
increase its development expenditures and consequently experienced a moderate improvement in GDP growth
rate, poverty reduction and in social indicators. So we can draw the hypothesis for our study about the fiscal
policy effects on GDP growth.
KAKAR, (2011) by employing cointigration technique indicated that fiscal policy significantly affected long run
economic development of Pakistan. Moreover he identified that public policy techniques were more important in
long run than in short span of time. He suggested controlling interest rate and government expenditures for
positive short run outcomes.
On the basis of above quoted study we draw two hypotheses:
H0: Fiscal policy has significant impact on GDP growth of Pakistan.
Our second hypothesis is that
H0: Fiscal policy has both long and short run implications for growth of gross domestic product.
Fiscal policy by enhancing output or productivity affects economic activity and lead to better economic growth
and development. The study will use fiscal variables (government consumption expenditures and per capita real
total revenues) to identify their impact on GDP growth of Pakistan. We will use discount rate, trade openness and
gross fixed capital formation as control variables along with fiscal variables as contributing towards enhancing
economic activity. Annual time series data is obtained for the period of 1980-2012 from different sources
including WDI and Pakistan Economic Survey. As non-stationarity is a general problem with time series data so
Augmented Dickeyfuller test will be used to identify stationarity of the variables. On the basis of ADF test
results Johanson cointigration test or ARDL test will be employed to determine long run impact of government
fiscal policy on gross domestic product of Pakistan. Moreover error correction model is also used to conclude
about short run effects of fiscal policy. Following econometric model is used for empirical analysis.
InGPCC= α + β1 (InDR) + β2 (InTO)) + β3 (InREV) + β4(InGCE) + β5(InGFCF ) + μi (1)
208
Here:
GPCC =Annual GDP Per Capita (in US$)
GCE = Government Consumption Expenditure (Measured in Billion Rupees)
REV = Per Capita Real Total Revenue at constant prices of 1999-2000
TO =Trade Openness (Measured as ratio of Exports + Imports to GDP)
GFCF= Gross Fixed Capital Formation (Measured in Billion Rupees)
DR =Discount Rate (% per Annum)
μi = Error- Correction –Term
3. Empirical Findings and Discussion
The ADF test is conducted first of all at level and at first difference. Results of ADF test are showing the
stationarity of variables at 1st difference. It means that variables of our model are integrated at the same order
(i.e., I (1)).
Table 1. ADF unit root test

Variables Level 1st Difference
C C&T C C&T Result
LNGPCC 0.076609 -1.630809 -4.905303* -5.047286* I(1)
LNGCE -1.852712 -2.317313 -5.467984* -5.770146* I(1)
LNREV -0.185127 -2.736503 -6.010841* -5.982361* I(1)
LNTO 0.100775 -1.206994 -5.008329* -5.49253* I(1)
LNGFCF 0.308614 -2.144261 -4.527158* -4.478101* I(1)
LNDSR -2.381442 -2.322422 -3.776731* -3.690457* I(1)
Notes: the series is stationary at 5% level of significance at critical value "-2.976263" for constant and critical value "-3.580623" for constant
and trend component.
Table 2. VAR lag order selection criteria

Endogenous variables: LNGPCC LNGFCF LNPEX LNREV LNTO LNDSR
Lag Log L LR FPE AIC SC HQ
0 18.83368 NA 1.66e-08 -0.88508 -0.60219 -0.79648
1 168.5870 227.2119* 6.90e-12* -8.730136* -6.749914* -8.109955*
2 203.3519 38.36133 1.07e-11 -8.64496 -4.96741 -7.4932
Note: * indicates lag order selected by the criterion. LR: sequential modified LR test statistic (each test at 5% level); FPE: Final prediction
error; AIC: Akaike-information criterion; SC: Schwarz-information criterion; HQ: HannanQuinn information criterion.
One lag has been selected by all the criterias. Maximum one lag has been selected by sequential modified LR
test statistic (each test at 5% level), Schwarz information criterion, Akaike information criterion, Final prediction
error and Hannan-Quinn information criterion. Instead of Engle-Granger single equation based cointegration
test, Johanson (1988, 1991) and extended by Johanson and Juselious (1990) is applied. Johanson and Juselious
(1990) have derived two tests (Trace test and Maximum Eigen Value test) for cointegration. The result of Trace
and Maximum Eigen Value test statistic are shown in table 3 and 4.
Table 3. Unrestricted co-integration rank test (trace)

Hypothesized at Trace 0.05
No. of CE(s) Eigenvalue Statistic Critical Value Prob.**
None * 0.754744 116.4283 83.9371 0.0000
At most :1 * 0.705895 75.67022 60.0614 0.0014
At most :2 * 0.486655 40.17946 40.1749 0.0500
At most :3 0.380402 20.84205 24.2759 0.1276
At most :4 0.173676 6.960194 12.3209 0.3291
At most :5 0.048046 1.427916 4.1299 0.2718
Notes: Trace test indicates 3 cointegrating eqn(s) at the 0.05 level. * denotes rejection of the hypothesis at the 0.05 level; **Mackinnon
HaugMichelis (1999) p-values.
209
Table 4. Unrestricted co-integration rank test (maximum eigenvalue)

Hypothesized at Max-Eigen-value 0.05
No. of CE(s) Eigenvalue Statistic Critical Value Prob.**
None * 0.754744 40.75809 36.6301 0.0156
At most: 1 * 0.705895 35.49075 30.4396 0.0108
At most :2 0.486655 19.33742 24.1592 0.1970
At most :3 0.380402 13.88186 17.7973 0.1764
At most: 4 0.173676 5.532278 11.2248 0.4060
At most :5 0.048046 1.427916 4.1299 0.2718
Notes: Max eigenvalue text indicates 2 cointegrating eqn(s) at the 0.05 level. * denotes rejection of the hypothesis at the 0.05 level;
**MacKinnon HaugMichelis (1999) p-values.
The above table shows that Trace and Maximum Eigen Value statistics are above the critical value of 5 percent.
Hence we reject the null hypothesis of no co-integration. Therefore, the results confirm the long run relationship
of economic growth with fiscal policy variables.
Table 5. Normalized co-integrating coefficients

LNGPCC LNGFCF LNGCE LNREV LNTO LNDSR
1 1.278568 -1.245848 0.507 1.05375 0.1247
S.E 0.11434 0.12608 0.03111 0.15280 0.06058
t Ratios 11.18217597 -9.881409 16.29701 6.89627 2.058435
From the results of the normalized cointegrating coefficients we can clearly conclude that fiscal policy has
significant impact on the GDP growth of Pakistan. Gross fixed capital formation, trade openness and discount rate
have positive and significant impact on the long run growth path of Pakistan. The analysis indicates a significant
impact of public expenditures and revenues on the GDP as well. Government consumption expenditures have
negative impact on the GDP per capita while the real total revenues have positive signed coefficient which means
if the government is able to acquire greater revenues then it will be able to spend more on all the expenditures
categories including spending on infrastructure, education, health and public welfare works which will lead to
improve the GDP growth of the country.
Table 6. Error-correction model

Error-Correction: D(LNGPCC) D(LNDSR) D(LNGCE) D(LNREV) D(LNTO) D(LNGFCF)
CointEq-1 -0.3751 -0.4653 -0.30269 -0.37526 0.575456 0.013882
(0.15666) (0.32301) (0.27147) (0.09338) (1.82279) (0.21429)
[-2.39437] [-1.44050] [-1.11501] [-4.01858] [ 0.31570] [ 0.06478]
The error correction coefficient is -0.3751 with t-statistic value of -2.39437. It shows the error correction term is
significant. The error correction sign should be negative. It indicates that approximately 37% of the
disequilibrium error is corrected in each period and it suggests a good speed of convergence towards equilibrium
if there appears any disequilibrium. The negative sign of error correction term is in accordance with the theory.
Public expenditures, discount rate and public revenues have also tendency to converge to their long run
equilibrium.
4. Conclusion and Policy Implications
Impact of fiscal policy variables on short and long-run growth of Pakistan was the main objective of the study.
So it has been tested on the basis of theoretical literature and empirical testing of data on Pakistan’s economy for
the period 1980-2012. For theoretical determination we have reviewed several empirical studies. From all of the
literature reviewed we conclude that fiscal policy is very important for macroeconomic stability of the economy
and for attaining sustained economic growth. For empirical determination of the importance of fiscal policy for
Pakistan we have employed Johansen co-integration test for anticipating long-run performance of the fiscal
variables, and vector error-correction model is operated to identify the existence of error correction term in our
model. Empirical findings show that public policy is very vital for meaningful economic progress and results
also point to the fact that public policy techniques are more meaningful for long run growth then in short time
frame in Pakistan. Moreover from earlier studies we conclude that the structure, efficiency and composition of
210
the expenditure and revenue system are most important to make fiscal policy effective. The study has established
a negative relation of government consumption expenditures with GDP and public revenues has a progressive
impact on economic activity of Pakistan. Less consumption expenditures and effective revenue structure can
help Pakistan to boost its lethargic growth rate. Institutional structure which governs fiscal policy making and the
system in which fiscal policy is implemented also play an important role in this regard. Policy must be credible
means people must have confidence in government decisions. There should be more productive expenditures and
less distortionary taxes. With proper implementation and continuity of the fiscal decisions, fiscal policy can
better lead to sustained economic growth along with other macro policy measure.
References
Afonso, A., & Sousa, R. M. (2009). The Macroeconomic Effects of Fiscal Policy. Working Paper Series, 991.
Afonso, A., Ebert, W., Schuknecht, L., & Thöne, A. M. (2005). Quality of Public Finances and Growth. Working
Paper Series, 438.
Akram, N., et al. (2011). Fiscal Situation in Pakistan and its consequences for Economic Growth and Poverty.
American International Journal of Contemporary Research, 1(1).
Babalola, S. J., & Aminu, U. (2013). Fiscal Policy and Economic Growth Relationship in Nigeria. International
Journal of Business and Social Science, 2(17).
Berkmen, S. P. (2011). The Impact of Fiscal Consolidation and Structural Reforms on Growth in Japan. IMF
Working Paper Asia and Pacific Department, WP/11/13.
Committee, D., et al. (2007). Fiscal Policy For Growth And Development: Further Analysis And Lessons From
Country Case Studies. Internat Ional Bank For Reconstructio N And Development.
Easterly, W., & Rebelo, S. (1993). Fiscal Policy And Economic Growth: An Emperical Investigation. National
Bureau Of Economic Research, Working Paper, No. 4499.
Fueki, T., Fukunaga, I., & Saito, M. (2011). Assessing The Effects Of Fiscal Policy In Japan With Estimated And
Calibrated DSGE Models. Bank Of Japan Working Paper Series, No.11-E-9.
Gemmell, N., Kneller, R., & Sanz, I. (2006). Fiscal Policy Impacts on Growth in the OECD: Are They Long- or
Short-Run?
Gemmell, N., Kneller, R., & Sanz, I. (2011). Fiscal Policy: New Cross-Country Evidence Of The Impact On
Short- And Long-Run Growth. The Economic Journal.
Gerson, P. B. P. (1998). The Impact of Fiscal Policy Variables on Output Growth. IMF Working Paper, 1(98).
Gupta, S., Clements, B., Baldacci, E., & Granados, C. (2005). Fiscal policy, expenditure composition,and growth
in low-income countries. Journal of International Money and Finance, 24, 441-463.
http://dx.doi.org/10.1016/j.jimonfin.2005.01.004
Jawaid, S. T., Qadri, F. S., & Ali, N. (2011). Monetary-Fiscal-Trade Policy And Economic Growth In Pakistan:
Time Series Empirical Investigation. International Journal Of Economics And Financial Issues, 1(3),
133-138.
Kakar, Z. K. (2011). Impact Of Fiscal Variables On Economic Development of Pakistan. Romanian Journal of
Fiscal Policy, 2(2), 1-10.
Knellera, R., Bleaneyb, M. F., & Gemmellb, N. (1999). Fiscal policy and growth: evidence from OECD
countries. Journal of Public Economics, 74, 171-190. http://dx.doi.org/10.1016/S0047-2727(99)00022-5
Long, J. B. D., & Summers, L. H. (1992). Macroeconomic Policy and Long-Run Growth.
López, R. E., Thomas, V., & Wang, Y. (2010). The Effect of Fiscal Policies on the Quality of Growth. The World
Bank. Washington, D.C.
Mccracken, C. (2006). Whether State Fiscal Policy Affects State Economic Growth.
Nurudeen, A., & Usman, A. (2010). Government Expenditure And Economic Growth In Nigeria, 1970-2008: A
Disaggregated Analysis. Business and Economics Journal, BEJ-4.
Ogbole, O. F., Amadi, S. N., & Essi, I. D. (2011). Fiscal policy: Its impact on economic growth in Nigeria 1970
to 2006. Journal of Economics and International Finance, 3(6), 407-417.
Romer, C. D., & Romer, D. H. (2007). The Macroeconomic Effects Of Tax Changes: Estimates Based On A New
Measure Of Fiscal Shocks.
211
Shaheen, R., & Turner, D. P. (2010). Measuring the dynamic Effects of Fiscal Policy shocks in Pakistan.
Thomas, V., et al. (2010). The Effect of Fiscal Policies on the Quality of Growth. The World Bank. Washington,
D.C.
Zee, V., & Tanzi, V. (1997). Fiscal Policy and Long-Run Growth. Palgrave Macmillan Journals, 44(2), 179-209.
Notes
Note 1. http://info.worldbank.org/governance/wgi/index.asp
Note 2. http://www.wvsevsdb.com/wvs/WVSData.jsp
212
ISSN 1916-971X E-ISSN 1916-9728
The Role of Corporate Governance in Reducing the Negative Effect of

Earnings Management
Nopphon Tangjitprom1
1
National Institute of Development Administration & Assumption University, Bangkok, Thailand
Correspondence: Nopphon Tangjitprom, National Institute of Development Administration & Assumption
University, Bangkok, Thailand. Tel: 66-8-5815-6177. E-mail: [email protected]
Abstract
This paper aims to examine the role of corporate governance in reducing the negative effect of earnings
management. The accounting data for U.S. firms during 2002-2010 were collected from WorldScope database
and the corporate governance data were from ASSET4, which is an affiliate of Thomson Reuter. Earnings
management can be harmful to firm value if it arises from managerial opportunism, whereas it can also be
beneficial if managers intend to convey some information about future earnings or reduce the volatility of
reported earnings. The empirical evidence has shown that earnings management has a negative effect on firm
value. However, the negative effect of earnings management is neutralized by the role of corporate governance,
which helps to reduce managerial opportunism. Firms with a lower CG score face the negative effect of earnings
management, whereas firms with a higher CG score face a less-negative effect from earnings management. In
other words, managerial opportunism with earnings management is lower in good-governance firms. Therefore,
corporate governance provides a crucial role in reducing the negative effect of earnings management.
Keywords: earnings management, corporate governance, accruals
1. Introduction
Earnings Management usually refers to the efforts of firm managers or executives in manipulating the earning
figures in financial reporting. Even though these activities have complied with regulations, these activities can
arise from managerial opportunism to take advantage of compensation plans (Healy, 1985; Baker, Collins, and
Reitenga, 2003; Bregstresser and Philipon, 2006; Kuang, 2008). For example, managers can overstate the
reported profit in order to demonstrate the firm’s outperformance and obtain incentive payments such as bonuses
or understate the earnings to reduce the current share price in order to obtain more benefits from the employee
stock options plan. However, some may argue that managers can use earnings management techniques to
communicate or convey certain information (Dutta and Gigler, 2002) and to smooth the earnings to reduce its
volatility (Magrath and Weld, 2002).
Therefore, earnings management can be both beneficial and harmful to firm value based on how managers
employ. There is no simplified way to determine which component of earnings management is harmful or
beneficial. However, if there were a mechanism that mitigated managerial opportunism, it would be helpful in
reducing harmful earnings management. In this paper, I propose the Corporate Governance mechanism as the
useful tool to restrain the opportunistic behaviors of managers that may deteriorate firm value.
Corporate Governance can act as a mechanism to reduce managerial opportunism. Firms with poor corporate
governance are more vulnerable to managerial opportunism, and earnings management will be harmful for the
firm’s value. Meanwhile,, firms with good corporate governance can mitigate these problems, and earnings
management will be less harmful or even beneficial for firm value.
The empirical evidence based on U.S. firms supports the above argument. In general, earnings management is
negative for firm value. However, after including the effect of corporate governance, this negative relationship
becomes neutral. In other words, firms with a lower CG score face the significantly negative effect of earnings
management, whereas firms the negative effect of higher CG firms is not significant at the convention level.
Therefore, corporate governance provides a crucial role in reducing the negative effect of earnings management.
The outline of this paper is as follows; Section 2 provides a discussion of earnings management and corporate
213
governance in the literature; Section 3 discusses the data and methodology; Section 4 discusses the empirical
evidence and statistical testing; and Section 5 is the conclusion of the study.
2. Previous Studies and Hypothesis Development
Earnings management usually refers to the efforts of firm managers in manipulating earnings reporting. In
general, this earnings management is perceived as negative because it may stem from managerial opportunism.
Managers use their discretion in earnings management to expropriate private benefit, for example, to get more
benefits from their compensation (Healy, 1985; Baker, Collins, and Reitenga, 2003; Bregstresser and Philipon,
2006; Kuang, 2008).
Meanwhile, some studies have pointed out that earnings management is not actually destructive for firms. If the
accounting rules are too restrictive, it will discourage the potential benefits of earnings management as a way for
managers to communicate trustful forecasting (Dutta and Gigler, 2002). Magrath and Weld (2002) have proposed
the idea that managers can use earning management to reduce earnings volatility. This could be beneficial for
firm value, as it can help to reduce the risk perceived by investors. Therefore, earnings management may not be
destructive for firm value but be consistent with firm value maximization principal.
Earnings management can be both beneficial and negative for firm value. Even though it is not easy to classify
which earnings management is beneficial or negative, it is proposed here that if there exists a mechanism that
will help to reduce managerial opportunism in terms of expropriating private benefits from firm value, that
mechanism should be able to discourage negative earnings management but not discourage beneficial ones.
Therefore, the effect of earnings management should be less negative (or more positive) if such a mechanism
existed. Corporate governance mechanism that included how shareholder rights are protected should be that
mechanism to reduce managerial opportunism.
Jensen and Meckling (1976) have discussed the agency problem in the firm, as managers can consume more
perquisites for their private benefit as the expenses of firm if they hold a smaller fraction of ownership in that
firm. With the separation of ownership and control, managerial decisions may not follow the firm value
maximization principal, as they do not share the wealth with shareholders. Therefore, an effective controlling
mechanism such as corporate governance is essentially required (Fama and Jensen, 1983). La Porta et al. (2000)
have discussed the notion that the legal investor protections act like external corporate governance to protect
outside shareholders (or minority sharedholers) from the expropriation of private benefits by controlling
shareholders.
Gomper, Ishii, and Metrick (2003) developed the G-index (Governance Index) based on firm provisions
concerning shareholder rights. They found that the firm with weaker shareholder rights will have lower value,
measured by the Tobin Q ratio, compared to the firm with stronger shareholder rights. Johnson et al. (2000)
studied corporate governance and managerial expropriation during the Asian financial crisis and found that firms
in the countries with weaker corporate governance and shareholder rights protection could face more problems
related to the managerial expropriation of private benefits.
There has been a number of studies about the role of corporate governance and investor protection in reducing
the level of earnings management. Cornett, McNutt, and Tehranian (2009) have studied the corporate governance
mechanism and the level of earnings management. Board independence can reduce earnings management
because managers cannot influence the board entirely, whereas performance pays can motivate managers to
manipulate earnings reporting in order to obtain higher compensation. Wang, Sheu, and Chung (2011) have
shown evidence that the level of earnings management was reduced after the implementation of Sarbanes-Oxley
Act. Additionally, Hazarika, Karpoff and Nahata (2012) showed evidence that the number of CEO-forced
turnovers were positively related to the level of earnings management. This implies the role of the board of
directors in preventing aggressive earnings management before there are external consequences.
Leuz at al. (2003) used country-level analysis to show that the aggregate earnings management level is lower in
the country with stronger investor protection. Investor protection is based on investor rights and legal
enforcement based on La Porta et al. (1998). Defond, Hung, and Trezevant (2007) have shown that earnings
announcement is more informative in the country with stronger investor protection.
Based on the above previous studies, it can be seen that earnings management is more likely to be negative for
firm value. However, if there is a mechanism such as corporate governance that helps to reduce managerial
opportunism, such a mechanism should also be able to discourage negative earnings management. Therefore, the
earnings management for a firm with good governance should be less negative (or even positive) than a firm
with poor governance. The main research hypothesis is as follows.
214
Research Hypothesis: The effect of earnings management on firm value is less negative for the firm with better
corporate governance compared to the firm with lower corporate governance.
The data used in this research include all firms traded on the New York Stock Exchange (NYSE) or NASDAQ
during 2002-2010. The accounting data were from the WorldScope Database. The Corporate Governance Score
was provided by ASSET4, an affiliate of Thomson Reuters, that provides Environmental, Social, and
Governance (ESG) data for worldwide firms. The firms whose information as not available in WorldScope or
Asset4 were excluded from the analysis. Finally, there were 5,153 firm-year observations included in the
analysis.
First, the relationship between firm value and earning management will be examined. Firm value was measured
by Tobin’s Q, which is the ratio between the market value of a firm and the book value of a firm. In other words,
Tobin’s Q is the market value of a firm scaled down by its book value. More value of this ratio can be interpreted
as the firm creating more value added to its book value. Earning management consist of the absolute accruals
scaled down by operating cash flows, which is one component of aggregate earnings management measurements
based on Leuz et al. (2003). The accruals are calculated by
Accruals i  CAi  Cashi   CLi  STDi  TPi   DEPi (1)
where CA is the total current assets and Cash represents total cash and equivalents. CL is total current liabilities,
STD is total short-term debts, TP is total taxes payable, and DEP is total depreciation expenses.
The effect of earnings management to firm value is examined based on the following regression model.
m
FVi   0  1EM it    j CONTROLit   it
j 1
FVi   0  1EM it   1DRit   2 SIZEit   3GROWTHit   it (2)

where FV is the value of the firm i at time t, which is measured by Tobin’s Q ratio or the ratio between the
market value of the firm and its book value. EM is the earnings management measurement, which is the absolute
accruals from equation 1 scaled down by operating cash flow. CONTROL is a set of control variables, which are
DR, SIZE and GROWTH, in the analysis. DR is the firm’s debt ratio representing the leverage or capital structure
policy. SIZE is the firm size measured by the logarithm of total assets. GROWTH is the firm’s growth rate of
total assets.
According to equation 2, the effect of Corporate Governance is included as follows.
m
FVi   0  1EM it   2 EM it  CGit    j CONTROLit   it (3)
j 1
where CG is the Corporate Governance Score from ASSET4. The interaction effect between earnings
management and corporate governance is expected to be positive. If, the coefficient of interaction term (β2) is
significantly positive, it means that the effect of earnings management on firm value for the firms with a higher
CG score is less negative (or even become positive) than for firms with a lower CG score.
In general, investors may not be concerned about the level of firm corporate governance as a continuous
measurement. They usually perceive firms in terms of good or poor governance. Therefore, another way to
examine the effect of CG on the relationship between firm value and earnings management is to categorize all
firms based on the CG score. Each firm is classified into two groups based on the median. Firms having a CG
score above the median are named high-CG or good-governance firms, whereas firms with a CG score lower
than median are classified as low-CG or poor-governance firms.
m
FVi   0  1EM it   2 EM it  DHIGHit    j CONTROLit   it (4)
j 1
where DHIGH is the dummy variable that equals one if that firm has a CG score above the median in that fiscal
year or zero otherwise. The coefficient of the interaction term (β2) can be interpreted as the difference between
high-CG firms (CG score higher than median) and low-CG firms (CG score lower than median) in the
relationship between their firm value and earnings management. This coefficient is expected to be positive to
show that the effect of earnings management is less negative (or even positive) for high-CG firms.
215
4. Analysis and Results

Table 1 reports on the data for the firms in the sample during the period of analysis. The ratio between market
value and book value of firm assets or Tobin Q was around 2 times during the analysis period. This market value
ratio was lowest in 2008 as a result of the credit crunch from the subprime crisis. The earnings management
measures, which are the accounting accruals scaled down by operating cash flow, varied. It is interesting that this
figure was highest during 2008, which was the period when the U.S. economy most suffered from the crisis.
Moreover, U.S. firms, in average, use debt at around 60% of their total assets. The average annual growth rate of
their assets is 13.69%. Before 2008, the growth rates increased year-by-year and have drop dramatically to lower
than 10% during 2008-2009.
Table 1. Descriptive statistics on sample firms

Earnings Debt Asset
Year Tobin Q
Management Ratio Growth (%)
2002 2.0103 1.8145 0.5865 12.7783
2003 2.2452 1.2301 0.5793 15.3588
2004 2.3249 0.8201 0.5801 16.0270
2005 2.2649 0.9714 0.5730 16.9119
2006 2.1804 0.8518 0.5954 17.4062
2007 2.1372 1.3431 0.6040 20.9936
2008 1.4783 2.5777 0.6288 9.0601
2009 1.7590 1.0765 0.5887 8.6067
2010 1.9426 1.6249 0.5837 12.6520
Overall 1.9718 1.4189 0.5933 13.6944
The effect of earning management and firm value is reported in table 2. In model 1, earnings management is a
sole regressor and the coefficient is negative and significant at the convention level. This means that the effect of
earnings management on firm value, in general, was negative. Model 2 includes a set of control variables that
may have an impact on firm value. This cannot alter the earlier result, as the coefficient of earnings management
is still significantly negative. Model 3 includes the interaction term between earnings management and corporate
governance score to show the effect of corporate governance on the relationship between earnings management
and firm value. The coefficient of the interaction term was positive and significant at the convention level, as
expected. It can then be interpreted that the effect of earning managements is usually negative. However, the
firm with a higher CG score will have a less negative effect of earnings management. In other words, for firms
with a higher CG score, earnings management will deteriorate the firm value less than those with a lower CG
score.
Table 2. The effect of earnings management to firm value

Coefficient Model 1 Model 2 Model 3
Constant 1.9800** 6.7036** 6.7836**
(105.28) (33.78) (34.09)
EM -0.0057** -0.0040** -0.0268**
(-3.64) (-2.74) (-4.83)
Debt Ratio -1.1306** -1.1193**
(-19.49) (-13.42)
Size -0.2550** -0.2607**
(-19.49) (-19.85)
Growth -0.0656** -0.0333
(-2.05) (-1.01)
EM x CG 0.0326**
(4.25)
F-statistics 13.27** 217.43** 217.43**
Adjusted R2 0.0024 0.1439 0.1467
Notes: Number in parenthesis is t-stat ** indicate significant at 5% * indicate significant at 10%.
216
For general investors, the measurement of corporate governance may not be continuous as with the CG score,
but rather in discrete form, as in the high or low group. Therefore, instead of using the CG score, each firm will
be classified into a high-CG firm or low-CG firm. In each year, firms with a CG score above the median are
classified as high-CG and those with a CG score below the median are classified as low-CG. Panel 1 of table 3
reports the effect of earnings management on firm value between high-CG and low-CG firms. As expected, the
coefficient of earnings management for low-CG firms was -0.0074, which is negative and significant. However,
for high-CG firms, the coefficient was only -0.0010, which is less much less negative and not significant at the
convention level. Panel 2 of table 3 reports the results of the regression with the interaction term between
earnings management and dummy for high-CG. The result is consistent with earlier, as the coefficient of
earnings management was negative while the coefficient of interaction term was positive and significant. This
means that the negative effect of earnings management was lower for firms in the high-CG group.
Table 3. The effect of earnings management between High-CG firms and Low-CG firms
Panel 1
Coefficient Panel 2
Low-CG High-CG
Constant 7.4373** 6.3110** 6.7269**
(24.76) (24.14) (33.89)
EM -0.0074** -0.0009 -0.0087**
(-3.10) (-0.57) (-3.93)
Debt Ratio -0.9583** -1.1634** -1.1270**
(-7.99) (-10.15) (-13.51)
Size -0.3085** -0.2322** -0.2566**
(-15.24) (-13.70) (-19.60)
Growth -0.2061** 0.6365** -0.0574*
(-5.39) (8.70) (-1.78)
EM x DHIGH 0.0083**
(2.81)
F-statistics 116.87** 134.59** 175.76**
Adjusted R2 0.1527 0.1715 0.1450
From the previous literature, it was seen that corporate governance is a mechanism for reducing some managerial
opportunism. With the separation between ownership and control, managers that may very small or none of the
firm ownership may make decisions for their own private benefit rather than maximizing firm value (Fama and
Jensen, 1983). Corporate governance is a mechanism for controlling these problems. Protecting the rights of
shareholders, especially minority shareholders, is one component of corporate governance. Regarding the
ASSET4 CG score, it has included many perspectives of corporate governance, such as board structure,
compensation policy, board functions, management vision and strategy, as well as shareholders’ rights.
Table 4 shows the results of the analysis using the shareholders rights compared to using the overall CG score
reported earlier. The score on shareholders rights was calculated by the average score on the components related
to shareholders’ rights. The results of the statistical tests using shareholders rights instead of the overall CS score
were quite similar. Firms with higher shareholder rights protection will be less vulnerable to managerial
opportunism. Therefore, the effect of earnings management is less negative for firms with higher scores on
shareholders rights.
Table 5 reports on additional tests on the model in order to ensure the robustness of the results. First, the data
used in this paper were collected across various firms in the U.S. for many years. The effect of earnings
management on firm value may vary year-by-year. Therefore, the regressions based on equation 3 were
estimated year-by-year. Then, a test of significance of the regression coefficient was conducted by t-test based on
Fama and MacBeth (1973). The figures are reported in table 5 and show the same results as discussed earlier.
Another robustness check was carried out by excluding the banking industry because this industry may have
different characteristics from those of non-banking firms, for example, leverage and asset types. Therefore, the
previous regression was re-examined by excluding all observations from the banking industry. There were 148
firm-year observations excluded and 5,005 firm-year observations left in the analysis. The regression analysis,
excluding the banking industry, is reported in table 5. The results are also quite similar to those reported earlier.
217
Table 4. The role of shareholder rights component of corporate governance

With Interaction Terms With High-CG Dummy
Coefficient
Overall CG Score Shareholder Rights Overall CG Score Shareholder Rights
Constant 6.7836** 7.4780** 6.7269** 6.7195**
(34.09) (36.88) (33.89) (33.84)
EM -0.0268** -0.0317* -0.0087** -0.0067**
(-4.83) (-1.85) (-3.93) (-3.27)
Debt Ratio -1.1193** -0.9979** -1.1270** -1.1299**
(-13.42) (-12.07) (-13.51) (-13.54)
Size -0.2607** -0.3106** -0.2566** -0.2560**
(-19.85) (-23.06) (-19.60) (-19.55)
Growth -0.0333 0.2560** -0.0574* -0.0650**
(-1.01) (6.5) (-1.78) (-2.01)
EM x CG 0.0326** 0.0467*
(4.25) (1.68)
EM x DHIGH 0.0083** 0.0054*
(2.81) (1.87)
F-statistics 217.43** 216.58** 175.76** 174.72**
Adjusted R2 0.1467 0.1732 0.1450 0.1443
Table 5. Additional test of the effect of earnings management

Coefficient Full Sample Fama-MacBeth Regression Exclude Banking
Constant 6.7836** 7.5001** 6.8846**
(34.09) (42.24) (32.82)
EM -0.0268** -0.1147** -0.0271**
(-4.83) (-3.25) (-4.78)
Debt Ratio -1.1193** -1.3396** -1.0990**
(-13.42) (-13.44) (-12.80)
Size -0.2607** -0.2942** -0.2677**
(-19.85) (-36.69) (-19.47)
Growth -0.0333 0.2076** -0.0369
(-1.01) (4.76) (-1.10)
EM x CG 0.0326** 0.2133** 0.0330**
(4.25) (10.79) (4.23)
In this paper, earnings management was measured by accrual-based earning managements based on Leuz et al.
(2003). Jones (1991) has proposed the discretionary accruals to as the measurement of earnings management
instead of using total accruals. Discretionary accruals are the difference between total accruals and expected
accruals, whereas expected accruals are estimated by the following regression model.
1 REVt PPEt
Accruali   0  1  2   it (5)
Assett 1 Assett 1 Assett 1
where REV is total revenue and PPE is gross property, plant, and equipment. Accrual is the total accrual scaled
down by total assets at the beginning of the year. The residuals from the regression model in equation 5 are a part
of accruals that cannot be explained by change in revenue or non-current assets. Jones (1991) estimated the
above regression using time-series and used the residuals as discretionary accruals. However, estimating the
above regression with time-series is under the assumption that the parameters were stable over time. DeFond and
Jiambalvo (1994) estimated the above regression model by cross-sectional with all firms in the same industry
and same period. Therefore, the earnings management used in the previous analysis was recomputed according
to Jones’ model but estimated cross-sectionally with all firms within each industry. However, the accruals were
based on Leuz et al. (2003) so that they were comparable with previous estimations.
Table 6 reports the role of corporate governance in the relationship between earnings management and firm
value, whereas absolute discretionary accruals were used instead of total accruals, as in table 2 and 3. The results
218
are also qualitatively similar. More discretionary accruals can lead to lower firm values. However, for the firms
with better governance, this negative effect will be apparently lower. Therefore, all of this empirical evidences
supports the role of corporate governance in reducing managerial opportunism in earnings management. Firms
with lower CG scores face the negative effect of earnings management, whereas the negative effect of earnings
management for firms with higher CG scores is insignificant.
Table 6. The effect of earnings management using discretionary accruals

Coefficient Model 1 Model 2 Model 3
Constant 7.3643** 7.3803** 7.3771**
(35.27) (35.32) (35.33)
DA -0.0024 -0.0367* -0.0106**
(-1.29) (-1.80) (-2.55)
Debt Ratio -1.0374** -1.0339** -1.0333**
(-12.20) (-12.16) (-12.15)
Size -0.3019** -0.3028** -0.3026**
(-21.73) (-21.78) (-21.78)
Growth 0.2179** 0.0398** 0.2176**
(5.47) (5.46) (5.46)
DA x CG 0.0434
(1.69)*
DA x DHALF 0.0102**
(2.21)
F-statistics 247.62** 198.61** 199.47**
Adjusted R2 0.1717 0.1719 0.1725
5. Conclusion
In general, firm managers can use accounting accruals to manage earning reports for their own benefits. This will
deteriorate firm value and the effect of earnings management will be negative for firm value. However, some
argue that managers can use accounting accruals as a signal in earning reports. From this perspective, earnings
management can be positive for firm value. Corporate governance can play a crucial role in restraining
managerial opportunism. Previous studies have documented that corporate governance can reduce the level of
earning managements. However, if corporate governance is important for reducing the conflict of interests that
may be negative for the firm’s value, it should help to reduce only harmful earnings management, not beneficial
management.
The results in this paper provide evidence to support the role of corporate governance in reducing harmful
earnings management rather than beneficial earnings management. The evidence in this research shows that the
effect of earnings management on firm value is less negative for good-governance firms. Firms with poor
governance face the negative effect of earnings management because these firms are more vulnerable to
managerial opportunism. However, the negative effect of good-governance firms is not significant at the
convention level. Strong corporate governance can reduce managerial opportunism, and earnings management is
more likely to be used in a beneficial way. This means that corporate governance provides a crucial role in
reducing the negative effect of earnings management.
The accounting professions have discussed and improved accounting standards to provide more meaningful
accounting figures and to reduce managerial discretions in the report of earnings. The evidence from this paper
shows that managerial discretion, such as earnings management, is not actually unfavorable under the condition
that there is a good corporate governance mechanism. If firms maintain good governance and shareholders rights
are well-protected, they will be less vulnerable to opportunistic earnings management. In this case, earnings
management will be used in a favorable way, for example, to reduce earnings volatility or to convey information
about future earnings by managers. Therefore, encouraging good governance is as important as improving
accounting rules and standards.
References
Baker, T., Reitenga, A., & Collins, D. (2003). Stock option compensation and earnings management incentives.
Journal of Accounting, Auditing & Finance, 18(4), 583-593.
219
Bergstresser, D., & Philippon, T. (2006). CEO incentives and earnings management. Journal of Financial
Economics, 80(3), 511-529. http://dx.doi.org/10.1016/j.jfineco.2004.10.011
Cornett, M. M., McNutt, J. J., & Tehranian, H. (2009). Earnings Management at Large U.S. Bank Holding
Companies. Journal of Corporate Finance, 15, 412-430. http://dx.doi.org/10.1016/j.jcorpfin.2009.04.003
DeFond, M. L., & Jiambalvo, J. (1994). Debt covenant violation and manipulation of accruals. Journal of
Accounting and Economics, 17(1-2), 145-176. http://dx.doi.org/10.1016/0165-4101(94)90008-6
DeFond, M., Hung, M., & Trezevant, R. (2007). Investor protection and the information content of annual
earnings announcements: International evidence. Journal of Accounting and Economics, 43(1), 37-67.
http://dx.doi.org/10.1016/j.jacceco.2006.09.001
Dutta, S., & Gigler, F. (2002). The effect of earnings forecasts on earnings management. Journal of Accounting
Research, 40(3), 631-655. http://dx.doi.org/10.1111/1475-679X.00065
Fama, E. F., & Jensen, M. C. (1983). Separation of Ownership and Control. Journal of Law and Economics,
Gompers, P., Ishii, J., & Metrick, A. (2003). corporate governance and equity prices. Quarterly Journal of
Economics, 118(1), 107-156. http://dx.doi.org/10.1162/00335530360535162
Hazarika, S., Karpoff, J. M., & Nahata, R. (2012). Internal corporate governance, CEO turnover, and earnings
management. Journal of Financial Economics, 104, 44-69. http://dx.doi.org/10.1016/j.jfineco.2011.10.011
Healy, P. M. (1985). The effect of bonus schemes on accounting decisions. Journal of Accounting and
Economics, 7(1-3), 85-107. http://dx.doi.org/10.1016/0165-4101(85)90029-1
Jensen, M. C., & Meckling, W. H. (1976). Theory of the firm: Managerial behavior, agency costs and ownership
structure. Journal of Financial Economics, 3(4), 305-360. http://dx.doi.org/10.1016/0304-405X(76)90026-X
Johnson, S., Boone, P., Breach, A., & Friedman, E. (2000). Corporate governance in the Asian financial crisis.
Journal of Financial Economics, 58(1-2), 141-186. http://dx.doi.org/10.1016/S0304-405X(00)00069-6
Jones, J. J. (1991). Earnings management during import relief investigations. Journal of Accounting Research,
Kuang, Y. F. (2008). Performance-vested stock options and earnings management. Journal of Business Finance
and Accounting, 35, 1049-1078. http://dx.doi.org/10.1111/j.1468-5957.2008.02104.x
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. W. (1998). Law and finance. Journal of Political
Economy, 106(6), 1113-1155. http://dx.doi.org/10.1086/250042
La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. W. (2000). Investor protection and corporate
governance. Journal of Financial Economics, 58(1-2), 3-27.
http://dx.doi.org/10.1016/S0304-405X(00)00065-9
Leuz, C., Nanda, D., & Wysocki, P. D. (2003). Earnings management and investor protection: an international
comparison. Journal of Financial Economics, 69(3), 505-527.
http://dx.doi.org/10.1016/S0304-405X(03)00121-1
Wang, J. L., Sheu, H. J., & Chung, H. (2011). Corporate governance reform and earnings management.
Investment Management and Financial Innovations, 8(4), 109-118.
220
ISSN 1916-971X E-ISSN 1916-9728
An Investment Strategy Based on Stochastic Unit Root Models

Mamadou A. Konté1
1
Université Gaston Berger, UFR Sciences Economiques et de Gestion, BP 234 Saint Louis, Sénégal
Correspondence: Mamadou A. Konté, Université Gaston Berger, UFR Sciences Economiques et de Gestion,
BP 234 Saint Louis, Sénégal. Tel: 221-77-622-1350. E-mail: [email protected] or
[email protected]
Received: September 29, 2012 Accepted: January 17, 2013 Online Published: February 26, 2013
Abstract
An algorithm is presented that locally approximates the nonlinearity of stochastic unit root (STUR) models by n
linear models. The previous integer n is chosen so that the Hadamard matrix of order n can be defined. The
strategy STUR(n), then consists in creating n linear models from this Hadamard matrix and taking their average
forecast. A purchase (sell) signal is made if the obtained average forecast is positive (negative). Subsequently, a
comparison is made with respect to competing models (Moving average strategies) to assess their ability to
forecast the variation of five international indexes. It is found, after taking account transaction costs, that STUR(n)
generates generally the highest profitability in the out-of-sample data.
Keywords: forecasting, trading rules, random coefficient autoregressive models, efficiency market hypothesis
1. Introduction
The question of Efficiency Market Hypothesis (EMH) has been studied for many years by both academics and
market participants. The aim is to see if the assumptions of market frictionless and traders rationality are a good
description of real markets where microstructure (transaction costs, information asymmetry, etc.) and noise
traders are present. This is an ongoing debate and there has been no consensus. That is why some authors have
tried to reconcile the EMH and Behavioral finance arguments through dynamic systems, see for example Lo
(2005) and Konté (2010). The empirical studies of this hypothesis are based generally on three classes. The first
is traditional regression models. Their aim is to test the validity of the EMH in its weak form through traditional
time series forecast such as Auto Regressive Moving average models ARMA(p,q). If the market is supposed to be
a nonlinear dynamic system, one may consider nonlinear models such as the Random Coefficient Autoregressive
RCA(p) or regime switching models among others. The traditional regression models also contain analysis tools
based on firms’ fundamental (dividend, Book-to-Market, etc.). In this case, the objective is to test the EMH in its
semi-strong form (fundamental analysis). We refer to Ou and Penman (1989) and references therein.
The second class uses Technical Analysis tools such as Moving Average, Support and Resistance methods. This
approach is widely applied by traders to detect trends or reversal effects by using information such that prices,
trading volume, etc. Here, the validity of EMH is tested through its weak form, see for example Sullivan et al.,
(1999).
The last class, based on Machine Learning (Genetic Algorithms, Neural Networks methods) investigates the
EMH in its weak and semi-strong form as for the class of traditional regression models. Their difference is that
Machine learning models are self-adaptive methods in that there are few a priori assumptions about the
relationship between inputs while the traditional regression models make strong assumptions (parametric
approach).
The paper belongs to the first class where the nonlinearity of financial asset prices is modeled by RCA(p). This
econometric model generates the main stylized facts of financial time series, see Yoon (2003). It may be also
related to an Agent Based Model with a switching phenomenon between fundamentalists and noise traders, see
for example Konté (2011). There are many methods proposed in the literature to estimate its parameters for
trading or forecast purpose. For example Nicholls and Quinn (1981) employed the traditional least squares and
the maximum likelihood methods, see also Granger and Swanson (1997). Wang and Ghosh (2002) use Bayesian
approach while Sollis et al. (2000) work with Kalman filter. We follow here another approach consisting to
approximate the RCA(1) model by n simple linear models where n is any integer such that the Hadamard matrix
221
H of order n can be defined. The latter is an n × n matrix with all its elements being either −1 or 1, and such that
HHT=n*In where HT is the transpose of H, In is the identity matrix of order n. Therefore, the Hadamard
matrix columns is an orthogonal binary basis of Rn explaining why it is widely used in physics particularly in
the field of signal transmission. The integer n, in this study, must satisfy the constraint n , n/12 or n/20 is a power
of 2. The prediction is then made by taking the average forecast of these n linear models extracted from the
Hadamatrix since many researchers agree that combining multiple forecasts leads to increased accuracy, see
Granger and Ramanathan (1984).
The paper contributes in two ways to the literature. First, contrarily to other forecasting methods, the estimation
procedure is made locally to capture traders’ feedback or interaction since the variance and other higher
moments of STUR model do not exist. Only n data are used in the linear regression models where 8≤n≤50. This
constraint gives us exactly height (08) strategies STUR(n) with n {8, 12, 16, 20, 24, 32, 40, 48}. The second
contribution shows an application of the Hadamard basis to reduce the complexity of a problem (from
exponential to linear) for trading purposes.
The paper is divided into four additional sections. Section 2 presents our methodology and its competing
strategies to forecast the variation of asset prices of five international indexes (CAC 40, DAX 30, FTSE 100,
Nikkei 225, S&P 500). Section 3 describes the data and the methodology used in the empirical application.
Section 4 presents the empirical results and the last section concludes.
2. Some Forecasting Rules
2.1 Our Methodology
Consider the following stochastic unit root STUR(1) model defined by:
yt  (1  bt ) yt 1   t (1)
E (bt )  E ( t )  0, E (bt2 )   2 , E ( t2 )   2 , cov(bt ,  t )  0

where (εt) is an i.i.d Gaussian process and yt=logSt (log of asset prices).
The properties of eq. (1), to replicate financial times series, have been studied by (Yoon, 2003). The econometric
model is also related to an agent based model with interaction between fundamentalist and noise traders, see
(Konté, 2011). It is a special case of the Random Coefficient Autoregressive RCA(1) model which is defined by:
yt = (φ+bt)yt−1 + εt
E (bt )  E ( t )  0, E (bt2 )   2 , E ( t2 )   2 , cov( bt ,  t )   .

(Nicholls and Quinn, 1982) have shown that the RCA(1) process (yt) is a finite second-order stationary moment
if the condition φ2+ω2<1 is satisfied. Since φ=1 in our case, the stationary condition is violated (Note 1). That
means conventional methods based on this assumption such as Maximum likelihood method cannot perform, see
Yoon (2006) (Note 2). Consequently, we propose a methodology that locally approximates the nonlinearity of
asset prices by n linear models. For this purpose, bt is supposed to take only two values α and −α at any time.
Therefore, it may be rewritten as bt=αXt−1 where for any t, Xt−1=1 or Xt−1=−1. The equation (1) becomes
rt  log S t  log S t 1  yt  yt 1  c  X t 1 yt 1   t (2)

where a constant c is added, as usual, in the regression model.
For the moment, the estimation cannot be proceed because the variable Xt−1 is not known. To circumvent this
problem, regressions models are used conditional on the path of (Xt). For example, in the equation (2), if it is
decided to use n data for the estimation process, we will have 2n paths for (Xt) , t=1,⋯,n since Xt takes −1 or
1 at any time. Each trajectory generates a linear model with three input variables Xt−1, yt−1 and the constant
variable c. Therefore, the nonlinearity is approximated by 2n linear models (Note 3). This feature comes at a
cost, as we need to store a binary matrix of size n×2n−1 to make all linear regressions (Note 4). Generally, we
need the parameter n to be big for estimation precisions but not too much to keep a local approximation. In the
application, it is taken 8≤n≤50. To solve the dimensionality problem, techniques similar to Component Principal
222
Analysis (CPA) in exploratory data analysis are used. Namely, we extract "n orthogonal linear models". Here,
i
the orthogonality of two models i and j is defined by the orthogonality of their corresponding paths (Xt−k) and
j
(Xt−k), k=1,⋯,n. If n is constrained to be an integer such that n/2, n/12 or n/20 is equal to 2k, k N, then
Hadamard matrices exist. For example for n=2, the Hadamard matrix is
1  1
H2   
1 1 
that is a basis of R2. Recursively, we can define the matrix H4, H8, ⋯, by using the following formula
H  Hn 
H 2n   n
H n H n 
This approach allows to pass from exponential (2n) to linear (n) complexity since now for n data used in the
equation (2), n linear models are also employed where their paths correspond to the columns of the Hadamard
matrix of order n. For each model, determined by the path of (Xt), the parameters α and c are estimated by the
Ordinary Least Square method. Then a forecast is made at time t+1 through the equation
rˆt 1  yˆ t 1  y t  cˆ  ̂X t y t (3)
A recursive regression is applied. At any time, the previous n data are used in the regression model to
determine the new estimated parameters. We denote by SUR (n) the strategy that consists to take the average
forecasts of all n "orthogonal linear models". The procedure to create the buy and sell signals is then simple: a
buy (sell) signal is produced if the average forecast, denoted by râ t 1 , is positive (negative). To reduce the
number of transaction costs, we enhance the strategy by allowing static positions in the case where the forecast
signal is not significant. In other words, the following strategy is applied for SUR (n) .
 sign (rât 1 ) if | rât 1 ) | c; c  (max( Rt )  min( Rt )) / 30



position (t , n)  
 position (t  1, n) otherwise.

where Rt=logSt−logSt−1 represents the index return at time t (Note 6). The sign function is defined by
sign(x)=1 if x>0, sign(x)=−1 if x<0 and sign(0)=0.
2.2 Competing Trading Rules
If the market is supposed to be efficient, an optimal strategy is to buy and hold an index. The strategy B&H
consists therefore to be long on the index at any time and consequently there are no transaction costs. We
consider also simple and exponential moving average strategies that have been widely used by traders to capt
momentums or reversal effects. The idea is to consider two moving average series M(n,t) and M(m,t) with
different lengths n and m. If we denote by (St) the asset price process, the simple and exponential moving
average are defined, for a given length k>0, by respectively the equations (4) and (5).
1 k 1
M (k , t )   S t i
k i 0
(4)
M (k , t )  (1   ) M (k , t  1)  S t , with M (k , 0)  S1 (5)
If m<n then M(m,t) (resp. M(n,t)) is called the short-term moving average (resp. the long-term moving average).
The decision rule for taking positions is specified as follows. If the short-term moving average M(m,t) intersects
the long-term moving average M(n,t) from below, a long position is taken. Conversely, if the M(n,t) is
intersected from above, a short position is taken. The moving average strategies are implemented by using the
Matlab function movavg. Note that in these strategies, transaction costs appear only when an intersection appears
between M(m,t) and M(n,t). In the decision making process of traditional regression models, if a threshold is not
used, the number of transactions may be very high.
223

In this paper, we consider the daily closing prices of five international indexes CAC 40, Dax 30, FTSE 100,
Nikkei 225 and S&P 500 obtained from Yahoo Finance website. All time series have the same length of data as
shown by Table 1.
Table 1. Interval of study of the five international indexes

Index CAC 40 DAX 30 FTSE 100 NIKKEI 225 SP 500
in-sample 30 Jun 2000 09 Aug 2000 09 May 2000 04 Jan 2000 20 Apr 2000
19 Feb 2009 13 Feb 2009 30 Jan 2009 25 Dec 2008 30 Jan 2009
out-of-sample 20 Feb 2009 16 Feb 2009 02 Feb 2009 26 Dec 2008 02 Feb 2009
30 Dec 2011 30 Dec 2011 30 Dec 2011 30 Dec 2011 30 Dec 2011
Total data 2943 2943 2943 2943 2943
All the series end to 30 December 2011, totaling N=2943 trading days. Their difference appears only on the
beginning period where the latter is chosen so that to have the same length of data than the Nikkei index.
Each data is after divided into two periods: the first period (in-sample data) contains 2207 (0.75*N) trading days.
The remaining data (736 or approximatively 0.25*N) is retained for the second period (out-of-sample data). The
use of many geographic zones (Asia, Europa, United States) is to test the robustness of the different algorithms.
The methodology is the following. For each class of trading rules, here STUR(n), Simple Moving Average
SMA(m,n) and Exponential Moving Average EMA(m,n), a training period (in-sample data) is used to find its best
model in terms of the Sharpe Ratio which is an economic gain adjusted for risk. If we let Rt=logSt−logSt−1, the
log return of the index at time t, then the Sharpe ratio (SR) is defined for any strategy, say k, by
SR (k ) 
RM (k ) (6)
 (k )
where
1 T
RM (k ) 
T
z, t z t  position(t , k )  Rt 1
t 1 (7)
1 T
 (k ) 
T t 1 [ (
z t  RM ( k ) 2 0 , 5 )]
Here position(t,k) takes 1 (−1) if the strategy k is long (short) at time t and T represents the number of
predictions.
The second part consists to compare the performance of the best in-sample models with respect to the
out-of-sample data. The comparison is based on many criteria such as the Sharpe Ratio, the winning up periods
(W.U.P), the winning down periods (W.D.P), the correct directional changes (C.D.C) and the Maximum
~
Drawdown (M.D). Let R t and R t be respectively the daily trading profit and actual return at time t ,
Qt  1{ Rt 0} and Ft  1{ Rt  0} the indicator functions for rise and fall at time t , and finally U t  1{ R  0, R~  0} and
t t
Dt  1{ Rt 0, R~t 0} the indicator functions for winning up and winning down periods, then the above expressions can
be defined by
Tt11{ R~t 0}
~ ~ ~
( ~
)
t
Rtc   Ri , M .D  min Rtc  max ( Ric ), , C.D.C  (8)
t 1,, T i 1,, t T
i 1
Tt1U t Tt1 Dt (9)

W .U .P  100  , W .U . D  100 
Tt1 Qt Tt1 Ft
~ ~
where Rtc  ti 1 Ri is the cumulative trading returns up to time t (Note 5).
224
Finally, we integrate the transaction costs in the analysis. Namely, it is supposed that any transaction implies a
constant cost of 20 basis points.
4. Results
We recall that the in-sample data contains approximatively 9 years of data for each index. The STUR(n) class,
with the constraint 8≤n≤50 and n, n/12 or n/20 is a power of 2, contains height (08) admissible strategies
characterized by the integer n valued in {8, 12, 16, 20, 24, 32, 40, 48}. The simple and exponential moving
average classes are parametrized by two integers m and n, representing respectively the lead and lag parameter.
In this study, sixteen (16) strategies are proposed for each Moving Average class with their parameters given by
m {1, 5, 10, 15,} and n {50, 100, 150, 200}. All these algorithms need some initial data to start the
forecasting procedure. For example, the STUR(n) strategy needs n+1 data to make the first forecast. For these
initial data, the agent decision is supposed to be always 1. The cost of one transaction is taken to be 20 basis
point i.e 0.2%.
Table 2 shows the performance of the best strategies in each class through the different indexes and through their
respective in-sample data given in Table 1.
For the STUR class, the best strategy is given by the parameter n=16 for the CAC, NIKKEI and S&P indexes
and by n=20 and 24 for the DAX and FTSE indexes, respectively. Overall, it is seen for the STUR class, the
approximation needs to be local or to have less data (n≤24) to generate good results.
For the Exponential Moving Average class, the lag parameter of the best strategy is always equal to n=150 for
the different indexes and the lead parameter lies to the set {10, 15}. For the Simple Moving Average class, the
lag parameter varies through indexes where the parameter n=150 is more frequent. The same remark applies also
for the lead parameter m where the mode is given by m=15. We also remark that for both moving average classes,
a small lead (m=1 or 5) does not give satisfactory in-sample results. All best competing models (STUR, EMA,
SMA), in the in-sample evaluation, generate economic gains or a positive Sharpe Ratio. Furthermore, except in
the FTSE index, the optimal strategy of the EMA class outperforms the other best models.
Table 2. Sharpe ratio of the best trading rules in each class (In-sample)
Class STUR(n ) EMA (m,n) SMA(m,n) Buy and Hold
CAC n=16 m=10, n=150 m=15, n=150
Sharpe Ratio 0.45 0.741 0.730 -0,37
Dax n=20 m=15, n=150 m=15, n=150

Sharpe Ratio 0.627 0.7825 0.5974 -0.218
FTSE n=24 m=15, n=150 m=15, n=150
Sharpe Ratio 0.4938 0.451 0.492 -0.210
Nikkei n=16 m=10, n=150 m=10, n=50
Ratio 0.222 0.611 0.528 -0.355
S&P n=16 m=15, n=150 m=15, n=200
Sharpe Ratio 0.274 0.4728 0.4436 -0.2916
On the other hand, the Buy and Hold Strategy has a negative mean in the in-sample data of all geographical
zones showing consequently a negative Sharpe ratio. This may be explained by the fact that all five indexes are
highly correlated and therefore the probability to have the same sign performance in the five indexes is very
high.
After getting the best strategy in each class, we make a comparison between them. Namely, three trading rules
are investigated for each index in their out-of-sample data given in Table 1. The aim is to see if it is possible to
do better than the benchmark strategy after taking into account transaction costs. To reduce the chance feature, a
long time series of out-of-sample is considered as containing around three years of data. The results are shown in
the Table 3 and Table 4 .
225
Table 3. Out-of-sample performance of the best trading rules in each class (Part I)
CAC STUR(16) EMA (10,150) SMA(15,150) Buy and Hold
Sharpe Ratio 0.135 -0.72 -0.16 0.125
Transactions 4 20 6 0
M.D -0.35 -0.83 -0.52 -0.40
DAX 30 STUR(20) EMA (15,150) SMA(15,150) Buy and Hold
Sharpe Ratio 0.53 0.13 0.32 0.40
M.D -0.33 -0.42 -0.30 -0.39
FTSE 100 STUR(24) EMA (15,150) SMA(15,150) Buy and Hold
Sharpe Ratio -0.11 -0.29 0.30 0.49
M.D -0.37 -0.34 -0.29 -0.20
Nikkei 225 STUR(16) EMA (10,150) SMA(10,150) Buy and Hold
Sharpe Ratio 0.03 -0.46 -0.77 -0.01
M.D -0.32 -0.59 -0.61 -0.33
S&P 500 STUR(16) EMA (15,150) SMA(15,200) Buy and Hold
Sharpe Ratio 0.09 -0.73 -0.20 0.63
M.D -0.41 -0.66 -0.40 -0.21
Description: This table presents the out-of-sample values of the Sharpe ratio, the number of transactions and the Maximum Drawdown (M.D)
for each best strategy.
Table 4. Out-of-sample performance of the best trading rules in each class (Part II)
CAC 40 STUR(16) EMA (10,150) SMA(15,150)
C.D.C 50.41% 47.83% 50.82%
W.U.P 34.32% 52.82% 57.91%
W.D.P 66.94% 42.70% 43.53%
DAX 30 STUR(20) EMA (15,150) SMA(15,150)
C.D.C 52.58% 51.90% 51.77%
W.U.P 73.26% 77.12% 76.61%
W.D.P 29.39% 23.63 % 23.92%
FTSE 100 STUR(24) EMA (15,150) SMA(15,150)
C.D.C 51.90% 50.27% 50.82%
W.U.P 38.60% 63.73% 60.10 %
W.D.P 66.57% 35.43% 40.57%
Nikkei 225 STUR(16) EMA (10,150) SMA(10,150)
C.D.C 51.90% 52.58% 51.36 %
W.U.P 43.16% 48.16 % 51.32%
W.D.P 61.24% 57.30% 51.40%
S&P 500 STUR(16) EMA (15,150) SMA(15,200)
C.D.C 50.82% 51.63% 53.53%
W.U.P 46.96 % 66.67% 65.21%
W.D.P 55.69% 32.62% 38.77%
Description: This table presents the out-of-sample values of correct directional change (C.D.C), the winning up periods (W.U.P) and the
Winning down periods (W.D.P) for each best strategy.
Table 3 shows that for the Sharpe Ratio criterion, the STUR strategy gives overall the best results, namely three
over the five indexes. Then it is followed by the B&H strategy which performs two times over the five cases.
The results of SMA and EMA trading rules are not satisfactory in the out-of-sample data.
For the Maximum Drawdown (M.D) measure, it is found over all that the two best strategies are also given by
the STUR class and the Buy and Hold Strategy ( 2 over 5 indexes for each). Precisely, the STUR trading rule
obtains the good results from the CAC and Nikkei indexes while B&H does better in the FTSE and S&P indexes.
226
The more M.D, the littlee is the downside risk. The ccumulative retu
turns of Figuree 1 illustrates bboth the conceept of
Maximumm downside andd profitability ((Sharpe Ratio)) over the out-oof-sample dataa for the differrent indexes.
MA and EMA in four among
It is notedd that the riskieest strategies ((downside riskk) are usually reached by SM g five
indexes. FFor the profitabbility measure,, it is also seenn that the STUUR and B&H ccumulative retturns end at the top
of the otheer strategies foor the CAC, D DAX, Nikkei aand S&P indexxes. The excepption appears only for the FTSE
F
index wheere SMA gives some interesting results.
Now, we aare interested in the percenttage of correctt directional chhanges (C.D.C C), and the percentage of co
orrect
directionall changes in thhe rise and fall periods (W.U..P and W.D.P)), see the eqs. ((8) and (9) andd Table 4.
For the C.D.C criterionn, the two besst strategies arre given by S SMA and STU UR. The simplle moving ave erage
trading rulle gives betterr results for thee CAC and S&&P indexes whhile the STUR R strategy perfo
forms for DAX X and
FTSE. Thee EMA, with a C.D.C of 52.58% only outpperforms the other trading ruules for the Nikkkei index. Bu ut, for
the W.U.PP criterion, it iss EMA the besst trading rules in three over the five indicees and then it is followed by SMA.
However, in financial markets,
m investtors are more concern to deetect falling peeriods. Conseqquently, the W.D.P
W
measure w will play a majjor role for theem. It is notedd that, for all ffive indexes, tthe STUR straategy gives the
e best
results andd sometimes thhe difference iss very significaant as in the caase of CAC annd S&P indexes.
Figure 1.
1 The cumulattive return pathh of the differeent trading rulees for each inddex
STUR strategyy outperforms the moving avverage trading rules

We can resume the analyysis, by sayingg overall, the S
since it beelongs to the best
b strategies for many criteeria such as thhe Sharpe ratioo, the Maximuum Drawdown n, the
Correct dirrectional channge and the Wiinning down pperiods. Its pooor results in thhe out-of-sampple only come from
Winning U Up period criteerion.
5. Conclussion
In this paaper, the stochhastic unit rooot model is uused to forecaast the directiion change off asset prices. The
contributioon consists inn approximatting locally thhe nonlinearitiies of financiial time seriess by simple linear
models. For this purposse, two simpliifications are made to facillitate the estim mation of the parameters. These
T
simplificattions are also relevant to reduce the execution time andd the memoryy cost involvedd by the algorithm.
The first is to constraintt the stochasticc parameter too take, at any ttime, two posssible values. C
Consequently, for n
data used in the regressiion, the nonlinnearities are appproximated byy 2n linear m models. The seccond simplification
is to solvee the dimensionnality problem
m by extractingg only the mosst significant liinear models. WWe follow thee idea
of the prinncipal componeent analysis (PPCA) by takingg n orthogonall binary vectorrs (columns off Hadamard ma atrix)
of Rn andd then associatte to each vecttor, a simple liinear model. T Then, the strateegy STUR(n) iss defined by ta
aking
the averagge forecast of these
t n modelss and the decission making iss simply to buyy (sell) if this average forecast is
positive (nnegative). To diminish the transaction coosts for profitaability reasonss, but also to capture signifficant
227
informations, an endogenous threshold c is used to activate a decision. Namely, the trader will transact if the
average forecast return is superior in absolute value to the threshold, elsewhere the previous position is
conserved. It is found that the strategies from STUR class dominate overall the moving average trading rules
(simple and exponential) and also the Buy and Hold strategy for the Sharpe criterion.
These interesting results may be explained by two facts. First, it is known that random coefficient autoregressive
models are able to fit well financial asset prices. So it is expected to have satisfactory results when this
econometric model is used for forecasting. The second reason is due to our estimation procedure which is local
and allows to capture feedback or interaction of traders rather using methods based on stationarity assumptions
in variance or higher moments that are violated in our case of stochastic unit root model.
References
Dunis, C. L., Williams, M. (2003). Applications of Advanced Regression Analysis for Trading and Investment. C.
In L. Dunis, J. Laws & P. Naïm (Eds.), Applied Quantitative Methods for Trading and Investment. John
Wiley & Sons.
Granger, C. W. J., & Ramanathan, R. (1984). Improved Methods of Combining Forecasts. Journal of
Forecasting, 3, 197-204. http://dx.doi.org/10.1002/for.3980030207
Granger, C. W., & Swanson, N. R. (1997). An introduction to stochastic unit-root processes. Journal of
Econometrics, 80, 35-62. http://dx.doi.org/10.1016/S0304-4076(96)00016-4
Konté, M. A. (2011). A link between Random Coefficient AutoRegressive Models and some kind of Agent Based
Models. Journal of Economic Interaction and Coordination, 6(1), 83-92.
Konté, M. A. (2010). Behavioural Finance and Efficent Markets: Is the joint hypothesis really the problem? IUP
Journal of Behavioral Finance, 7, 9-19.
Lo, A. (2005). Reconciling Efficient Markets with Behavioral Finance: The Adaptive Markets Hypothesis.
Journal of Investment Consulting, 7(2), 21-44.
Nicholls, D. F., & Quinn, B. G. (1982). Random Coefficient Autoregressive Models: An Introduction. Lecture
Notes in Statistics, 11. New York: Springer-Verlag. http://dx.doi.org/10.1007/978-1-4684-6273-9
Nicholls, D., & Quinn, B. (1981). The Estimation of Random Coefficient Autoregressive Models. II. Journal of
Time Series Analysis, 2, 185-203. http://dx.doi.org/10.1111/j.1467-9892.1981.tb00321.x
Ou, J., & Penman, S. (1989). Accounting measurement, Price-Earnings Ratio, and the information content of
security prices. Journal of Accounting Research, 27, 111-152. http://dx.doi.org/10.2307/2491068
Sollis, R., Leybourne, S., & Newbold, P. (2000). Stochastic unit roots modelling of stock price indices. Applied
financial economics, 10, 311-315. http://dx.doi.org/10.1080/096031000331716
Sullivan, R., Timmermann, A., & White, H. (1999). Data-snopping, technical trading rule performance, and the
boostrap. The journal of finance, 54, 1647-1691. http://dx.doi.org/10.1111/0022-1082.00163
Wang, D., & Ghosh, S. K. (2002). Bayesian Analysis of Random Coefficient Autoregressive Models. Model
Assisted Statistics and Applications, 3(2), 281-295.
Yoon, G. (2003). A simple model that generates stylized facts of returns. Department of Economics, UCSD.
Paper 2003-04.
Yoon, G. (2006). A note on some properties of STUR processes. Oxford Bulletin of Economics and Statistics, 68,
253-261. http://dx.doi.org/10.1111/j.1468-0084.2006.00161.x
228
Call for Manuscripts
International Journal of Economics and Finance (IJEF) is a double-blind peer-reviewed international
journal dedicated to promoting scholarly exchange among teachers and researchers in the field of
economics, financial economics and finance. The journal is published monthly in both print and online
versions by the Canadian Center of Science and Education.
The scope of IJEF includes the all fields in the economics and finance. Authors are encouraged to
submit complete, unpublished, original, and full-length articles that are not under review in any other
journals. The online version of the journal is free access and download. If you want to order print
copies, please visit: www.ccsenet.org/store
We are seeking submissions for forthcoming issues. All manuscripts should be written in English.
Manuscripts from 3000–8000 words in length are preferred. All manuscripts should be prepared in
MS-Word format, and submitted online, or sent to: [email protected]
Paper Selection and Publishing Process
a) Upon receipt of a submission, the editor sends an e-mail of confirmation to the submission’s author
within one to three working days. If you fail to receive this confirmation, your submission e-mail may
have been missed.
b) Peer review. We use a double-blind system for peer review; both reviewers’ and authors’ identities
remain anonymous. The paper will be reviewed by at least two experts: one editorial staff member and
at least one external reviewer. The review process may take two to three weeks.
c) Notification of the result of review by e-mail.
d) If the submission is accepted, the authors revise paper and pay the publication fee.
e) After publication, the corresponding author will receive two hard copies of the journal, free of charge.
If you want to keep more copies, please contact the editor before making an order.
f) A PDF version of the journal is available for download on the journal’s website, free of charge.
Requirements and Copyrights
Submission of an article implies that the work described has not been published previously (except in
the form of an abstract or as part of a published lecture or academic thesis), that it is not under
consideration for publication elsewhere, that its publication is approved by all authors and tacitly or
explicitly by the authorities responsible where the work was carried out, and that, if accepted, the
article will not be published elsewhere in the same form, in English or in any other language, without
the written consent of the publisher. The editors reserve the right to edit or otherwise alter all
contributions, but authors will receive proofs for approval before publication.
Copyrights for articles are retained by the authors, with first publication rights granted to the journal.
The journal/publisher is not responsible for subsequent uses of the work. It is the author's responsibility
to bring an infringement action if so desired by the author.
More Information
Website: www.ccsenet.org/ijef
Paper Submission Guide: www.ccsenet.org/submission
Recruitment for Reviewers: www.ccsenet.org/reviewer

Ijef v5n3 All

Uploaded by

Document Informationclick to expand document information

Document Informationclick to expand document information

Copyright:

Available Formats

Ijef v5n3 All

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Ijef v5n3 All

Uploaded by

Copyright:

Available Formats

Editorial Board

Editorial Board Members

Using Morningstar’s Analytical Measures to Evaluate Investment

Table 1. Morningstar return rankings

Table 2. Morningstar risk rankings

Table 3. Morningstar rating rankings

Average 3.32 3.47 3.57

Smiley or Frowney: The Effect of Emotions and Empathy Framing in

Table 1. Selected strategies and payoffs of the players

3.2 Experimental Treatments

referred too as Player 1, 2,

Figgure 2. Conserrvation Tillagee (CT) chosen bby the farmerss in round 1

rank-sum test. SELF-IN NTEREST FRAM

4.3 Effectiiveness of Emootions: The UF

Log-Likelihoood -2455.8 -241.3 -3772.4 -3770.1 -3382.4 -3

Figuure 6. Averagee lake cleanlineess by round

On a largeer scale the eff

The Stock Market/Unemployment Relationship in

Table 1. Augmented Dickey-Fuller (ADF) unit root tests in USA 1970.1—2011.3

Table 2. Augmented Dickey-Fuller (ADF) unit root tests in Japan 1970.1—2011.3

Table 3. Augmented Dickey-Fuller (ADF) unit root tests in China 2002.1—2011.3

Table 4. Cointegration test between stock prices and unemployment

Table 5. Granger causality tests

Saudi Financial Structure and Economic Growth:

Table 1. Variables definitions

lnGDPt  0  1 ln ISTM t   2 ln M 3t  3 ln BCPSt   4 ln EX t  5 ln BDt  6 lnCCPSt  7 FD2002t 8 FD2007t  9 FD2010t  t

Table 2. Augmented Dickey-Fuller for stationary unit root tests

Table 3. Regression analysis

Table 4. Johansen co-integration test

Table 5. Vector Error Correction Model (VECM)

Managerial Incentives and the Risk-Taking Behavior of

WT  aS T  (1  a)mST  k maxS T  H T ,0 (6)

Investors wealth is expressed as follows:

where 1k is an indicator variable such that:

 (c3) impplies that xe ≤ xc3 with

Figure 3. Values of the

x e*  maxx c1 , min( x v , x c 2 , x c 3 ) (17)

We can theen obtain the optimal

Figuree 5. Optimal weight

EWT   aEST   (1  a)mEST   kEmaxST  H T ,0

where 1k is an indicator variable such that:

By the same way we have:

A.2. Derivatives of xv with Respect to Parameters k, m, L0, S0 and a

Behaviour of Bank Share Prices and Their Impact on National Stock

Table 1. Selected countries, stock markets and banks

3.2 Data Analysis

Table 2. Correlations between indices across stock markets in different countries

The Determinants of Informal Capital in the Financing of Small Firms

before work experience education employees capital form affiliation

N 2510 2510 2510 2510 2510 2510 2510 2510

N 304 304 304 304 304 304 304 304

N 2814 2814 2814 2814 2814 2814 2814 2814

Sig. 0.07 0.11 0.10 0.38 0.000 0.22 0.63 0.03

Table 2. Sources of start-up funds by ethnic background

Table 3. Mean, standard deviation and number of dependent variables

9.3 Test of Statistical Significance: ANOVA Test

Table 4. ANOVA, Levene, and Welch tests of the dependent variables

Total 150.016 2813

i* = (qi)2, and u = (2a  c1  c2 )