Some Innovative Applications of DEA in the Interest of Investors: An

Efficiency Substantiate of Indian Banking Sector

“To those of you who study history, economics, sociology, literature and language I present the
challenge of the utilization of the enormous resources in our grasp to the problem of creating a
genuinely good life for yourselves and your children.”
- Polykarp Kusch
The Nobel Laureate German-American Physicist
(Jan 26, 1911-March 20, 1993)

1. Introduction:

It’s a truth that most of the developed countries of this world are not developed that they have
abundant amount of resources but they are affluent because of best utilization of resources they own.
In the same fashion, optimum use of available resources makes an organization efficient which leads
it to the ultimate success for the present and future as well. The efficiency for financial companies is
very crucial because it is highly sensitive to the various risks (such as bad debts or fraud
transactions) and they function with the funds provided by their investors. Therefore, efficient
operation of these organizations is the dire need of the time. Our research is a step towards the
understanding of efficiency of companies through most contemporary technique- Data Envelopment
Analysis and to apply these efficiency measures to assist the present and prospective investors in
decision making while they invest or disinvest.

In the contemporary business world, it is quit impossible that corporate sector can finance its mega
projects without the funds collected from its investors or shareholders. Investors/ shareholders of a
company allocate (or invest) money to the company with the expectation of future financial return.
Shareholders are the people or entities that legally own the stock certificates for a corporation.
Having made an investment in a business, shareholders are concerned with assessing the profitability
of their investment. Therefore, when shareholders look at the Annual Report of a company in which
they have invested, they will be mainly concerned with several measures, for example, historical
earnings per share, price/earnings ratio, debt/equity ratio, return on investment and dividend yield
etc. Undoubtedly, it is a challenging responsibility of management of business organization to use
the financial and non-financial resources to the optimum level, so that, a business entity can reach at
its objectives of shareholders’ well-being and build shareholders’ strong trust in company’s stability
and continue to hold it shares. Continued lack of confidence and stability can limit the company's
ability to grow and develop. So, measurement of performance of a company and analysis of
efficiency level is drastically needed for the betterment of stakeholders.

Generally, investors are more concerned with the profitability, solvency power, future growth
potentiality and expected risk-return trade offs for the alternative investments available. Our research
will be a humble initiative to understand these aspects and their relations with Data Envelopment
Analysis (DEA), a non-parametric technique of comparing and ranking several Decision Making
Units (DMUs) simultaneously. This research will also focus on some evolutionary and innovative
applications of DEA with classical models of Investment decisions. Before investing even a single
penny in a company, it is required that company’s performance must be analyzed on the basis of
various criterions.

The performance evaluation aims at regular and untiring monitor the efficiency and economy of an
organization operation and to provide the fundamental and crucial aspects to reach at better decision
making for best utilization of resources. Efficiency is a management concept that has a long history
in management science (Witzel M., 2002). Efficiency is typically related to the maximization of
outputs for a fixed level of input, or alternately the minimization of inputs to obtain the given level
of outputs. Over the past several years, substantial amount of efforts have been put into computing
the efficiency of companies and number of models are developed to get the right answer. Data
Envelopment Analysis (DEA) is one of the efficiency and productivity models and it is a non-
parametric mathematical programming based approach for measuring the relative efficiency of
Decision Making Units (DMUs) that may employ the multiple inputs and outputs.

Under non-parametric methods of efficiency, there are two main approaches: Data Envelopment
Analysis (DEA) and Free Disposal Hull (FDH) methods. However, in the field of research the most
commonly practiced approach is data envelopment analysis. As a non-parametric approach of
efficiency measurement, data envelopment analysis has been introduced by Charnes, Cooper and
Rhodes (CCR) in 1978 to measure the performance of various non-profit organizations, such as
educational and medical institutions, which were highly resistant to traditional performance
measurement techniques due to the complex and often unknown relations of multiple inputs and
outputs and non-comparable factors that had to be taken into account (Roman and Gotiu, 2017). In
contrast to the econometric approaches, non-parametric methods are based on the hypothesis that the
efficiency frontier is generated from the empirical results of the most efficient decision making unit
(DMU’s) or from the benchmarks. Aldamak et al., (2017) state that DEA non-parametric method
has attracted the attention of a number of researchers because of its unique ability to measure the
efficiency of multiple-input and multiple-output of DMUs without assigning prior weight to the
input and output. In this context, DEA is an effective non-parametric method for assessing the
relative efficiency of the decision-making units, which does not need the absolute functional form
between inputs and outputs approach (Toma et al., 2017). In our research we are taking Indian
Banking Industry as in recent years this sector has seen some drastic and crucial changes in policies
by government and its regulator. Banks have gone through many policy changes from last ten year
period and particularly the last five years is a metamorphosis time for Indian commercial banks.

2. Why Banking Sector? : Rationale of the Study

Now-a-days, banking sector acts as the backbone of modern businesses and development of any
country mainly depends upon the efficient banking system. A bank is a financial institution which
deals with deposits and advances and other related services. A bank deals in money and collects
savings in the form of deposits and it lends money to those who need it. The banking is one of the
most essential and important parts of the human life. In current faster lifestyle people cannot imagine
their daily transitions without proper and efficient banking network. The performance of the banking
sector is more closely linked to the economy than perhaps that of any other sector.
In fact economic development is a continuous process. The success of economic development
depends essentially on the extent of mobilization of resources and investment and on the operational
efficiency and economic discipline displayed by the various segments of the economy. Banks play a
positive role in the economic development of a country as they not only accept and deploy large
funds in a fiduciary capacity but also leverage such funds through credit creation (Ahooja, 2011).
Through mobilization of resources and their better allocation, commercial banks play an important
role in the development process of underdeveloped countries. It renders investment services such as
underwriters and bankers for its issue of securities to the public. Further, by offering attractive
saving schemes and ensuring safety of deposits, commercial banks encourages willingness to save
among the people. By reaching out to people in rural areas, they help convert idle savings into
effective ones. Commercial banks improve the allocation of resources by lending money to priority
sectors of the economy. These banks provide a meeting ground for the savers and the investors. All
in all, banking system of a nation gives a push to its economical activities and operates them
smoothly.

As per the RBI, India’s banking sector is sufficiently capitalized and well- regulated. Indian banks
are generally resilient and have withstood the global downturn well. The significance of banking
sector of India can be understand with the point that the digital payments system in India has
evolved the most among 25 countries with India’s Immediate Payment Service (IMPS) being the
only system at level 5 in the Faster Payments Innovation Index (FPII). The Indian banking system is
very vast consists of 20 public sector banks, 22 private sector banks, 46 foreign banks, 56 regional
rural banks, 1,562 urban cooperative banks and 94,384 rural cooperative banks, in addition to
cooperative credit institutions (RBI Website). Commercial Bank Survey published in RBI Bulletin’s
on Jun 11, 2019 revealed that the Aggregate Deposits of Residents on Apr. 26, 2019 was 123,080.0
billion INR and credit provided by these banks to the Government and to the Commercial Sector was
34,543.3 and 105,359.5 billion rupees respectively.

Scheduled Commercial Banks' Investments in shares issued by PSUs was Rupees 117.2 billion and
in Private Corporate Sector was rupees 692.3 billion. Business Standard published an article on April
21, 2019 which affirms that Mutual funds (MFs) are shuffling their sectoral exposures for 2019-2020
as they look to improve their portfolio returns. MFs’ exposure to the banking sector jumped about 50
per cent as of March 31, 2019, from a year ago, amid expectations of a sharp earning recovery. MFs’
exposure to banks stood at Rs 2.6 trillion at the end of FY19 compared to Rs 1.7 trillion as of March
31, 2018. Economic Times writes on Apr 17, 2019 that the classical pre-poll sectoral rotation by
investors to ‘domestic cyclical’ from ‘growth’ stocks has fuelled the recent rally in banking stocks.
This has amplified the price ratio between the Bank Nifty and the NSE Nifty to a record high. The
momentum in banking stocks has helped the Nifty touch a new high, given the 36 per cent weight
these stocks have on the NSE’s benchmark index. Out of the top 10 stocks that pushed the index to a
new high, seven are from the banking space and these have contributed 40 per cent of the index’s
last 1,000-point. This is the first time in almost three years when the Nifty’s gain has come primarily
due to support from banking stocks. The NSE Bank index is trading at 2.75 times its book value, the
highest since 2010 and against the 15-year average of 2.12. Undoubtedly, above empirical data
shows that banks’ stocks have a tremendous importance to the investors, traders, MFs and by and
large, to the economy.
With the potential to become the fifth largest banking industry in the world by 2020 and third largest
by 2025, India’s banking and financial sector is expanding rapidly. The Indian Banking industry is
so big that currently it has total market capitalization of worth Rupees 23,07,150 crore. The
Contribution of the banking sector to GDP ranges between 7 to 10 % of GDP which is higher than
national average. Indian banks are increasingly focusing on adopting integrated approach to risk
management. In the Union Budget presented on 5 th July, Finance Minister Nirmala Sitharaman
announced Rs 70,000 crore capital infusion to boost credit in debt-burdened state banks and credit
guarantees to support shadow lenders in a bid to boost lending and revive the economy. Banks have
already embraced the international banking supervision accord of Basel II, and majority of the banks
already meet capital requirements of Basel III, which had a deadline of 31 March 2019.

Despite its’ vastness, by the time, Indian banking sector faced many problems such as NPAs and
bad loans, cyber threats, bank frauds and henceforth, the last five years proved to be a huge shifting
period for this industry. Soumik Dey (April, 2019) in his published article in The Week states that
during the last five years, the loan interest rates have become visibly cheaper. From RBI's policy
repo rate of 8 per cent on May 2014, the interest rates have moved down to 6.25 per cent in March
2019. However, a lower interest rate regime has not yielded any greater cheer for the banking sector
or for its customers. In fact, it has been so bad that the last two years could still mark as the worst
years for the centuries old Bank of Calcutta, established in 1806, or today's State Bank of India, and
for most of the 19 other public-sector banks (remember Nirav-Modi PNB Scam). The government,
who is the majority shareholder in these banks, played its role in managing the banking crisis saga
and even it set the new limits in exercising its powers, to overrule and prevail over the banking
regulator's advice, several times in the last five years (Soumik Dey, 2019).

By the time Urjit Patel took over as the governor of RBI, banks’ bad loans ballooned 356 per cent,
from Rs 2.26 trillion in 2014 to Rs 10.39 trillion on March 2018 in spite of banks’ efforts to clean
their loan books in these years. Subsequently, with a government nod, the RBI under Patel
proceeded to set up a rescue fund for the banks and put banks under watch-list from lending
activities, just as his predecessor Rajan had prescribed. For depositors, visits to banks became a
desperate call during demonetization (Nov. 2016) despite large scale digitization in banking services.
Going ahead, as signals from the government are for even lower interest rates on bank loans, deposit
rates would only be allowed to fall further (Soumik Dey, 2019).

As banking is a service Industry and public money is the major input of the banks. Money has
opportunity cost and banks leave no stone unturned to attract people to invest their money with
them. To achieve this, banks must assure the public that money deposited with the banks is not only
safe and secure but also multiplying. With the onset of globalization, many new financial
instruments and institutions have come up. People today have multiple choice of investment. This
vast availability of alternative has made the banking industry rethink and they are forced to work
efficiently in order to survive and sustain themselves. Hence, the concept of efficiency is of
paramount importance for the banks. As said earlier, we will made our efforts to understand the
efficiency of banking sector from the prospective of shareholders and DEA will be used to
synthesize the decisional problems of shareholders or investors.
 3. Past Literature at a Glance

Although, DEA was developed in operations management, researchers have subsequently explored
DEA in a variety of applications across a wide range of academic disciplines and research pursuits.
Charnes et al. (1981) examine the success of an education program. Researchers have also used
DEA to measure the efficiency of higher education (Johnes, 2006). Other institutional establishments
include hospitals (Banker et al. 1986), telecommunications (Kim et al. 1999), and shipping ports
(Tongzon 2001). Studies also examine insurance (Kao and Hwang 2008). A number of studies
examine agricultural settings (Dhungana et al. 2004) and ecological performance (Korhonen and
Luptacik 2004). In the financial realm, researchers have used DEA to measure the efficiency of
banks (Sherman and Gold 1985; Avkiran 1999, 2012).

3.1 Banking Efficiency

Narasimham Committee, which submitted its report in 1991, suggested various measures to improve
the efficiency and health of the banking sector by making it more competitive and vibrant
(Ahluwalia, 2002) and enforced banks to operate their activities more efficiently. Now, after more
than two decades of banking sector reforms, it is a time to analyze that how the new banking policy
has affected the banking operations of the different banks and whether banks improved their
productivity or not.

Majority of the researchers applied traditional techniques to measure the banking industry
performance for instance Azhagaiah & Gejalakshmi (2012) enforced regression analysis; Sathye
(2005) used various accounting ratios; Rao (2007) formulated profitability indices; Kaur (2016)
applied CAMEL Model; Kalluru (2009) run T-test; Choudhary & Tandon (2010) exercised
compound Annual Growth rate and Coefficient of Variation; Dhanabhakyam & Kavitha (2012)
purposed different tools like ratio analysis, correlation and regression; Shanmugam K. R. and A. Das
(2004) worked with stochastic frontier production function model. Undoubtedly, there are number of
tools and techniques to measure the efficiency of banking industry and they are successfully
operated in past researches. However, DEA has an edge over these tools and techniques as it
simultaneously uses multiple heterogeneous inputs and outputs to reach at the answer which units
are performing most efficiently and inefficient units. DEA can indicate the causes of inefficiency in
the form of excess inputs which might be elapsed. Henceforth, in our study we are employing DEA
method to measure the productivity and efficiency.

3.2 Banks Efficiency and DEA: An Indian Perspective

In present time Indian academia is more concerned with evolutionary techniques of research and has
focused on emerging and innovative way of inquest investigation. Indian researchers are also
exploring DEA considering Indian business environmental requirements. Some studies are
embedded in our synopsis which have been executed in context of Indian banking sector and have
judged Cost, Technical, Scale and allocative efficiencies:

Debasish, (2006), in his paper entitled, Efficiency Performance in Indian Banking: Use of Data
Envelopment Analysis, employed DEA- CCR and Output Oriented Models to study Indian
Commercial banks, employed total assets and deposits as inputs and operating profits, total income
and advances as outputs. The result revealed that foreign owned banks were, on an average, most
efficient and new private banks were more efficient than the old banks. In the same year, Chatterjee
and Sinha (2006) evaluated the Cost Efficiency of 20 Public Sector Banks and 10 Private Sector
Banks from 1996-97 to 2002-03 by applying Data Envelopment Analysis (DEA). The results of the
study depicted that overall Indian Scheduled Commercial Banks had Cost Efficiency of 0.713 in
1996-97 which declined significantly to 0.662 in 2002-03. The results revealed that Private Sector
Banks had higher Cost, Technical, Allocative and Scale Efficiency as compared to Public Sector
Banks. Similarly, Sahoo et al. (2007) examined the Cost Efficiency of 81 Indian Commercial Banks
for the period 1997-98 to 2004-05 using Data Envelopment Analysis (DEA) with Wilcoxon-Mann-
Whitney t-test to check whether there exist significant differences across ownership. The ownership-
wise results of the study concluded that Foreign Banks were the most Cost Efficient while Private
Sector Banks were least Cost Efficient.

The results were quite opposite in the Ray and Das (2010) research which finds that Cost Efficiency
of State-Owned Banks were most efficient banks while Private-Banks were the least efficient
Further, Ray and Das (2010) also valuated Profit Efficiency of Indian Banks by using Data
Envelopment Analysis (DEA) during the Post-Reform Period from 1996-97 to 2002-2003. The
results showed that Cost Efficiency of Indian Banks was higher than 90% for each year from 1997 to
2003, but the Profit Efficiency of Indian Banks was lower than 50% in some years; though it was
higher with 64.4% in 2003. Profit Efficiency scores also depicted that State Banks and its associates
were more efficient followed by Foreign Banks. The results also concluded that banks with larger
amount of assets had higher profit and Cost Efficiency scores as compared to banks having lesser
worth of gross assets. Gulati (2011) extended the research with new variables and enquired the
impact of inclusion of non-interest income in the banks’ output on the Cost Efficiency of Indian
Banks from 1992-93 to 2007-08 by using Data Envelopment Analysis (DEA). For estimating the
Cost Efficiency scores, two models were adopted [one including (A) and other excluding (B) of non-
traditional activities in the output]. The entire study period was divided into two distinct sub-periods,
first (1992-93 to 1998-99), and second (1999-2000 to 2007-08). The year-wise mean Cost, Technical
and Allocative Efficiency scores showed that efficiency scores of Model A were higher than that of
Model B which highlighted that dropping non-interest income understated true efficiency of the
banks. The study also compared group-wise efficiency performance of Public Sector Banks (PSBs),
Private Sector Banks (PBs) and Foreign Sector Banks (FBs). The results of inefficiency depicted that
Cost Inefficiency among Indian Scheduled Commercial Banks was due to Allocative Inefficiency.
Further, Technical Inefficiency was due to Pure Technical Inefficiency. However, most of Indian
Scheduled Commercial Banks experienced Increasing Return to Scale (IRS).

Kumar (2013) analyzed the Cost Efficiency of Indian Public Sector Banks (PSBs) during 1992-1993
to 2007-2008. PSBs were further bifurcated into two groups namely, State Bank of India group (SBI
group) and Nationalized Banks group (NB group). The time period was split into two distinct sub-
periods representing different phases i.e. first phase of banking reforms (1992-1993 to 1998-1999),
and second phase of banking reforms (1999-2000 to 2007-08). The results showed that Indian Public
Sector Banks showed an upward trend in the Cost Efficiency. The average Cost Efficiency
(inefficiency) of Indian Public Sector Banks was 79.6% (25.6%). Further, the Cost Efficiency of SBI
group declined by 6.8% in second phase as compared to first phase, contrarily, it increased for NB
group in the second phase. The result of Kruskal-Wallis test showed that there was no significant
difference in average Cost Efficiency between the sub-periods for the two groups.

Kumar and Gulati (2008) used DEA-BCC model, to study 27 public sector banks. The study
employed physical capital, labor, loan-able funds as inputs and investments and advances as output.
The result found that the public sector banks operated at 88.5 per cent level of overall technical
efficiency. Kumar and Singh (2015) used CRR model and BCC model to estimate the technical and
scale efficiency of commercial banks in India during the periods 2006-2010. The results indicate that
deregulation of banking sector has led to an increase in the efficiency of commercial banks in India.
This increase in efficiency of banks in India is not only because of increase in pure technical
efficiency but also due to increase in its scale efficiency. The results show large spread of technical
efficiency between companies during the studied period. The estimated results also shows that
performance of private sector banks has been better than public sector banks during the period and
source of inefficiency is mainly due to its scale rather than pure technical inefficiency. Majid
Karimzadeh (2012) examined the efficiency of Indian commercial banks during 2000 – 2010 by
utilizing Data Envelopment Analysis (DEA). Based on the sample of 8 commercial banks, findings
reveal that the mean of cost (economic) efficiency, technical efficiency, and allocative efficiency are
0.991, 0.995, and 0.991 in VRS model while 0.936. 0.969 and 0.958 in CRR model, respectively,
using DEA approach. Inputs and outputs of this study were analyzed based on intermediation
approach. In addition, the results suggest that Bank of India and ICICI bank are more efficient as
compare to other banks in India and result confirmed that selected Public Sector Banks are more
efficient than Private sectors during the study period in India.
Mukta (2016) applied Data Envelopment Analysis (DEA) technique to measure the efficiency of a
sample of 57 banks. Data had been gleaned from the annual reports of banks on input variables
namely Capital, Total assets, Advances, Number of employees and Cost to income ratio and the
output variables namely Return on assets, Interest spread, Non-interest income, Deposits to advances
ratio and percentage decrease in non-performing assets. The study covered a period of four years
from 2009-2010 and 2012-2013. Results indicated an overall level of inefficiency in commercial
banks at 47%. This implies that the commercial banks have the scope of producing 1.88 times as
much output from the same inputs. In India, foreign banks are the most efficient while private Banks
operate at a higher level of efficiency compared with public banks. Mahendru et al. (2017) reviewed
the extent of overall technical efficiency, pure technical efficiency and scale efficiency (SE), with
data envelopment analysis (DEA). The result implicated that overall analysis of PSBs during the
time period of the study explained that a greater part of inefficiency among PSBs was attributed to
scale inefficiency. In addition, the number of banks operating at constant return to scale (CRS) came
down to 9 in 2011–2012 from 23 in 2007–2008. Simultaneously, there was a reduction in leaders
and increase in laggards. It is suggested that banks must optimize their scale of operations through
technological innovation.

Very recently, Goyal and Aggarwal, (2018) conducted their study to evaluate the intra-sector
efficiency in the Indian banking sector based on a cross-sectional data of 66 banks for the year 2015-
16. The authors employ directional distance function based meta-frontier DEA approach and the
results reveal that the Indian banking sector is 73.44% efficient. The research pursuits in many
indicative results as: First, the Indian banking sector is not fully efficient and the directional distance
function DEA results indicate the need for reduction of inputs and improvement in outputs so as to
induce efficiency in the banking sector. Second, the superior performance of the foreign banks over
the other ownership types highlight that they are more competitive, probably because of their
exposure to international markets and because of lower social obligations. Third, public sector banks
are the laggards in terms of the efficiency due to rising NPA levels and the corresponding
recommendations of the Asset Quality Reviews to enhance provisioning requirements that have
adversely affected their profitability.

Of course, researchers have made considerable efforts to know the banking productivity
effectiveness by means of Data Envelopment Analysis. However, as we need a keen observation on
banking interest and generalization of results which is still to be well-established, so, further
researches can be operated in the future with many potentialities. Further, the present study is quite
different and unique in its way as in the pursuance of research it will observe the efficiency of Indian
commercial banks during two different governments ruled during the time span of ten years (2009 to
2014- INC and 2014 to 2019- BJP).

3.3 Banks Efficiency and DEA: Studies outside of India

There have been numerous published researches executed outside of India on DEA applications to
measure the efficiency of banking system. The first paper was presented by Sherman and Gold
(1985) to study bank efficiency by applying CCR model of DEA. They claim that DEA results
provide meaningful insights which are a day dream by other techniques of efficiency. Subsequently,
it threw an ember on the hidden issues of productivity and efficiency. By the time, great efforts are
being done for the advancement of DEA by combining with other operational research
methodologies. Antreas D. Athanassopoulos (1997) employed DEA by incorporating tangible and
intangible variables to assess the operating efficiency; Dekker, D., & Post, T. (2001) worked with
the less restrictive microeconomic assumption of quasi-concavity; M.R.Alirezaee, M. Afsharian
(2007) improved DEA by taking referenced DMUs; Cook W.D and Hababou M.(2001) extended
DEA while combining other operational research technique such as sales performance measures; F.
Hosseinzadeh Lotfi et al. (2010) encompasses -objective linear programming (MOLP) and used
Zionts–Wallenius (Z–W) method to reflecting the Decision Maker’s preferences in the process of
assessing efficiency with combined; C Parkan (1987) analyzed operational inefficiencies with DEA;
M Oral et al. (1992) measured banks efficiency with their branches by various input-output
combinations; Paradi J.C., Rouatt S, Zhu H. (2011) analyzed profitability considering process of
bank branch to convert its expenses into revenues and the performance of operating units along
different dimensions (for line managers) and a modified Slacks-Based Measure model is applied for
the first time to aggregate the obtained efficiency scores from stage one and generate a composite
performance index for each unit; Paradi, J.C. et al (2010) developed Culturally Adjusted DEA model
to benchmark business units that operate under different cultural (business) environments to specify
the related inefficiencies, B. Golany and J. E. Storbeck (1999) incorporated service quality factors in
branch performance analysis into DEA model; McEachern, D. & Paradi, J.C. (2007) observed
Environments and technology impacts on branch performance; H. David Sherman, Timothy J.
(2006) valued the synergies after banking mergers. These references are out of some key papers and
of course-the representative studies, which can open new vistas for further study on DEA.

DEA model is originated and developed mainly in the countries where industrialization spreads its
claws thanks to the enhancement of productivity as the primal objective. So, a very vast research
reservoir of DEA is available from these leading countries. It is not feasible to compile a complete
research on DEA reconnoitered in developed countries. Henceforth, we prefer to discuss some of the
recent studies on DEA, so that, we can observe the new emerging trends in research of efficiency
management which is being synthesized in developed as well as developing countries (Except India).
Researchers have worked on extension of DEA methodologies and they contributed on more
accurate and feasible techniques of efficiency assessment. Rakhshan, et al. (2016) judged the
methodologies of 75 published research at the branch level since 1985 to early 2015, they found that
these models can be divided into four categories: standard basic DEA models, single level and multi-
level models, enriched (hybrid) models and special models. They inquired that the popularity of
multi-level models than the single level models are on the rise. Furthermore, as a result, they
concluded that from the perspective of performance measurement approaches applied to bank
branches, the production approach is more widely used than the others. Other researchers (J. Titko et
al., 2014; Martins, A. I., 2018) improved the methodology of measuring efficiency of banks by
including other variables and used different approaches. J. Titko et al. (2014) studied the Latvian
Banks by using input oriented DEA, with the assumption of Variable Returns to Scale was used to
calculate efficiency scores. Potential model variables were selected based on the intermediation and
profitability approach. Fourteen alternative models with different inputs-outputs combinations were
developed for the research purposes. To substantiate the variables selection for DEA model the
received data was processed, using methods such as correlation analysis, linear regression analysis,
analysis of mean values, and Kolmogorov-Smirnov test. The research results assisted the authors in
providing general recommendations about the variables selection for DEA application in the Latvian
banking sector. The research contributes to the existing analytical data on bank performance in
Latvia. Martins, A. I. (2018) in his study applied two-stage model and intermediation approach. The
application of regression for proportions, more appropriate than traditional linear and Tobit
regressions, to deal with the fractional nature of the DEA scores, allowed the identification of
efficiency determinant factors for the main banks operating in Portugal. The findings of the study
reveal that the variables which have a major influence on overall efficiency are internationalization,
size and type of ownership of capital.

Various researches included the contemporary issues in DEA efficiency of banks such as Novickytė
and Droždz (2018) purposed a study to examine the efficiency of the banks in Lithuania by
employing the DEA method and evaluate bank performance in a low interest rate environment.
Serrano et al. (2017) chose the methodology based on linear programming on the assumption of
variable returns to scale, with the objective of analyze independent units, comparatively. Inputs and
outputs of banking institutions were defined in the period between 2010 and the first half of 2016.
Results indicated that large banks tend to be more efficient while smaller banks use their resources
less efficiently. Novickytė and Droždz (2018) analyzed five alternative models with different input-
output combinations, based on production, profitability and intermediation dimensions. The main
bank profitability measure—the return on assets (ROA) ratio—was employed to validate the results
obtained using the DEA method. The Lithuanian bank’s efficiency analysis based on the VRS
assumption shows that better results are demonstrated by the local banks. Study observed that the
larger Lithuanian banks (subsidiaries) applied a more appropriate business model than smaller
(local) banks operating in Lithuania. Additionally, this research contributes to the scholarly literature
in the field of determinants of bank business performance in concentrated markets dominated by
foreign banks.
Some researches are based on ownership structure as Yin, Yu and Huang J, (2018); T. Mahbub et
al., (2019); Ouenniche, J. & Carrales, S. Ann Oper Res (2018) conducted their studies considering
control and holding rights. Yin, Yu and Huang J, (2018) took 93 Chinese commercial banks over the
period of 2005-2016, they first evaluate the banking efficiency by using the proposed data
envelopment analysis (DEA) model, which simultaneously incorporates a non-concave meta-frontier
technique, undesirable outputs, and super efficiency into a network slacks-based measure (NSBM)
model-NCMeta-USNSBM. Subsequently, the evolution of banking efficiency during the study
period is investigated on the basis of the Dagum Gini index and kernel density estimation methods.
The main empirical results show that the significant disparity/heterogeneity exists in banking
efficiency for overall efficiency, productivity efficiency, and profitability efficiency, the technology
gap ratio (TGR) is higher in case of stated-owned banks (SOB), joint-stock banks (JSB), and city
commercial banks (CCB) while there is more space for improvement, especially for SOB and JSB in
the profitability stage. T. Mahbub et al. (2019) examined the profit performance of family-dominated
banks in Bangladesh under competing hypotheses of bank-market structure using panel estimation
model with bootstrap method in DEA. In the study the principal drivers are costs, efficiency and
non-performing loans. Family-dominated banks are less efficient and less profitable because of
higher non-performing loans and higher costs, with indirect evidence of poor corporate governance.
Data Envelopment Analysis evaluated the efficiency with operational costs and fixed assets as inputs
and net interest revenue and non-interest revenue as outputs. The results are tantalizing in that for
banking firms, family ownership does not lead to enhanced performance in the case of Bangladesh.
Ouenniche, J. & Carrales, S. Ann Oper Res (2018) proposed a new DEA-based analysis framework
with a regression-based feedback mechanism and evaluated overall technical, pure technical and
scale efficiency of commercial banks of UK. Empirical results suggest that, on average, the
commercial banks operating in the UK—whether domestic or foreign—are yet to achieve acceptable
levels of overall technical efficiency, pure technical efficiency, and scale efficiency.

Wang et al. (2019) tried something unique and set the main objective of their research pursuit to
understand the efficiency of 18 large banks from all over the world during the period from 2013 to
2017. The performance was estimated by a dynamic slacks-based measure (SBM) model in data
envelopment analysis (DEA) to know the links of inputs and outputs of operational process. The
banks variables were considered as follows: Assets, capitalization, and liabilities as inputs; revenue
as output; and net interest income as a good link. The final empirical results exhibit the efficiency for
each term, and the overall score. The data analysis recommends a feasible solution to refine
inefficient terms based on the projections (slacks). This study visually observed the proficiency of
the banking industry to equip enterprises with the best choice for their finances and accompanied by
the growing economy, many banks are enhanced via sustainable development.

3.4 DEA: Banking Distress Prediction

The quest to detect the possibility of insolvency in its first stages has a tremendous importance. The
first contributions date back to Beaver (1966), Altman (1968) and Ohlson (1980), who examined
different methodological alternatives to the development of explanatory and forecast models for
these events. Wheelock and Wilson (2000) used a competing-risk model to identify the
characteristics resulting in U.S. banks’ failure or acquisition. Kolari et al. (2002) suggested that
computer-based Early Warning Systems (EWSs) can predict large U.S. bank failures. Canbas et al.
(2005) predicted Turkish commercial banks’ failure via multivariate statistical analysis of financial
structures. They combined four parametric models to construct an Integrated Early Warning System
(IEWS), which was reported to have high predictive ability for differentiation between sound and
troubled banks. In the era of dynamic research, DEA is being used even to estimate the failure
occurrence in corporate sector, beyond its conventional application. This model has been
incorporated into the prediction of corporate distress (or bankruptcy) in two different ways. Firstly,
DEA has been used to derive a classification algorithm to separate distressed firms from non-
distressed firms (Paradi et al, 2004). Secondly, the relative efficiency of firms has been computed
using DEA and this relative efficiency has been used as a feature of each firm in a subsequently
developed classification rule (Psillaki et al, 2010).

According to Avkiran, N. K (2012) credit-rating agencies (CRAs) can be criticized because of

improper timings of credit downgrades and for providing overly optimistic ratings for financial
institutions prior to financial crises. Nevertheless, timely and accurate credit ratings are particularly
significant in financial world. This study illustrates how the non-parametric technique Data
Envelopment Analysis (DEA) can be used as a forward-looking alternative method to flag bank
holding companies likely to become distressed in the near future. The results obtained generally
support DEA’s discriminatory and predictive power, suggesting that DEA can identify distressed
banks up to two years in advance. Robustness tests reveal that DEA has a stable efficient frontier and
that the technique’s discriminatory and predictive powers prevail even after data perturbations.

In addition, a large number of publications on banking systems demonstrate DEA’s appeal as a

benchmarking tool, but very few of them have attempted to use DEA to predict financial distress.
Among those papers applying DEA, some researchers have used DEA as a predictive tool of
financial distress. For example, Pille and Paradi (2002) developed DEA models along with the
modified Z-score model and equity-to-asset ratio to detect failure in credit unions in Ontario,
Canada, from 1991 to 1996. Some key studies are given below which are concentrated on estimating
the warning sign of financial problems in banking industry. Apart from using the worst-practice
DEA model, Paradi et al. (2004) used 1996 data as an evaluation sample and 1997 data as a
validation sample. Their initial pool of variables for input and output candidates was based on
literature. From a mixture of ten variables, they tried different combinations of inputs and outputs.
The authors show that one of those combinations provides the most correct prediction in the
estimation sample. They then apply the same set of inputs and outputs to a validation sample. The
results are convincing: when the best-practice DEA model is combined with the worst-practice DEA
model, the bankruptcy prediction is 100% after layer 3, i.e. after removal of banks on three
consecutive efficient frontiers. However, Tone (2002) used slack based model of DEA, where
inefficiencies (slacks) in all the variables are accounted for in computing an overall performance
score.

Indeed, financial distress does not necessarily lead to default or bankruptcy. This is also why DEA is
particularly attractive, as it can offer potential improvement guidelines for managers and regulators
before experiencing bankruptcy. Premachandra et al. (2009) who compared DEA to logistic
regression in terms of corporate bankruptcy prediction found that when applying logistic regression,
the parameters of the model can be estimated from an ex-post sample that includes failed and non-
failed firms. Through extrapolation from this sample, corporate failure can be predicted ex-ante.
They found that DEA has a lower overall level of correct prediction than Logit Regression, it is a
more accurate tool when predicting bankruptcy. The study especially shows that DEA can achieve
good results in predicting bankruptcy even with very large samples featured by a low value for the
relationship between the number of Bankrupt Firms (BR) and number of Non-Bankrupt Firms
(NBR) belonging to the sample. Roháčová and Kráľ (2015) used a very similar approach to the one
adopted by Premachandra et al. (2009) with the intention of demonstrating that D.E.A. is a
uncomplicated and effective tool for bankruptcy prediction using the technique that allows
identification of the Corporate Failure Frontier (CFF), in other words the frontier on which firms
about to fail are situated.

Researchers are trying to find out the accurate distress prediction through DEA. LIU and CHEN
(2012) observed that two stage model of DEA can work with negative profits and these negative
profits data can be utilized to the said purpose as in the real world, failed banks or firms often
produced negative profit for several years before they stuck into bankruptcy. To fit this situation
study introduces a two-stage worst-practice frontier DEA (WPFDEA) model that can deal with
negative profit data and effectively identify failed bank(s) in the worse-case scenario. The result is
compared with the result from a two stage best-practice frontier DEA model to show the adequacy of
WPF-DEA model for identifying failed bank(s) in the worst-case scenario. Li, Z et al. (2013)
investigated the predictive accuracy of corporate efficiency measures-the DEA along with standard
financial ratios in predicting corporate distress in Chinese companies. In contrast to previous
applications of DEA in credit risk modeling where it was used to generate a single efficiency
-Technical Efficiency, they assumed Variable Returns to Scale, and decompose Technical Efficiency
into Pure Technical Efficiency and Scale Efficiency. These measures are introduced into Logistic
Regression to predict the probability of distress, along with the levels of Returns to Scale. The
results show that the predictive power is improved by this corporate efficiency information using
DEA.

Although a large number of model building and prediction frameworks for corporate failure and
distress have been proposed, the relative performance evaluation of competing prediction models is
mono-criterion in nature because it has be explored further that to reduce conflicting rankings of
models. This methodological issue is addressed by Mousavi et al (2015) by proposing an orientation-
free super-efficiency data envelopment analysis model as a multi-criteria assessment framework. In
this research, researchers proposed a dynamic DEA framework to assess the relative performance of
an exhaustive range of distress prediction models and rank them accordingly. In addition, study
addressed several research questions including how robust is the out-of-sample performance of
dynamic distress prediction models relative to static ones with respect to sample type and sample
period length? And to what extent the choice of distress definition affects the ranking of competing
prediction models before, during, and after an important event? In their research, a dynamic
evaluation framework is also explored to assess the relative performance of these dynamic models in
predicting corporate distress using a sample of UK firms listed on the London Stock Exchange
(LSE).

Zakaria, S. (2017) studied the risk management efficiency of banking sector in uncertain business
climate, such as Asian Financial Crisis of 1997 and suggested that due to banking operations being
specifically affected by fluctuations in interest rates, which cause financial imbalances, banks are
now required to put in place an effective management structure that incorporates risk management
efficiency measures that help to mitigate the wide various risks they face. Such efficient risk
management measures are paramount in building robust and sound financial systems. This study
provides a new approach for measuring risk management efficiency levels in banks by offering a
more detailed insight into the data envelopment analysis (DEA) approach based on the usage of a
financial risk instrument. In the research, conducted by Mohammed A. SirElkhatim and Naomie
Salim (2015) comprehensively, reviewed the existing literature of prediction techniques that have
been used to assist on prediction of the bank distress. They categorized the review results on the
groups depending on the prediction techniques method, categorization started by firstly using time
factors of the past literature and marked the literature during the period (1990-2010) as history of
prediction techniques, and after this period until 2013 as recent prediction techniques and then
presented the strengths and weaknesses of both. Findings came out by the fact that only hybrid
techniques considered the most suitable methods in term of accuracy and reputation.

From the above literature, it can be said that the use of DEA to predict credit risk and, therefore,
financial distress in financial institutions has been attempted but this area needs a further
exploration. Nevertheless, to the best of our knowledge, application of DEA to predict financial
distress prior and during to this big change shift in Indian banking sector, has not been examined to
the deep, definitely it needs a great commitment to explore the research. Moreover, an issue with
some of the above studies that use DEA as predictive tools is the inadequate attention given to
justify the choice of variables and to convince readers that DEA results can withstand robustness
tests. This paper proposes various financial performance models to guide users in bringing together
performance variables, as well as test for robustness of DEA results.

3.5 DEA vs. Wealth Maximization: Is there a relation exist between efficiency and value?

Maximizing shareholders’ wealth maximization has become a new corporate paradigm, for a number
of years now, accounting measures such as earnings, return on assets, and return on equity have been
criticized and found wanting as performance indicators leading to greater shareholders’ wealth.
Currently, the majority of the shareholding of financial institutions is held by sophisticated and
institutional investors. These shareholders possess tools and resources to assess the value creation of
banks, which for them is essentially the evaluation of firm performance. Market forces and
shareholder pressure have driven banks to focus strategically on maximizing shareholder value. The
ultimate aim of any firm is the maximization of profit or maximization of the present value of the
firm. As per traditional value theory, survival of firms depends on the maximization of their
corporate performance and shareholder wealth. Efficiency and productivity determine the
operational efficacy of any firm, which means that operational efficiency and productivity should
ultimately result in the effectiveness of a bank’s operation too. Effectiveness is a measure of how
beneficial bank operations are. It can also be quantified in terms of the value to their investors. The
most efficient and productive banks should be effective also in terms of wealth maximization and
revenue generation. Wealth maximization further leads to increase in shareholder wealth creation
and is reflected in terms of market performance.

Although many researchers have investigated the relationship between bank efficiency and
shareholder value prior to Global Financial Crisis 2008-09 (Beccalli et al., 2006; Chu & Lim, 1998;
Fiordelisi, 2007; Fiordelisi & Molyneux, 2010), but a few studies study have initiated to explore this
relationship in the post-crisis period (Fu et al., 2014). Also, the methodology used by many prior
studies to estimate bank efficiency and its determinants has been criticized due to the problems of
correlation, biased estimators, and data-generating process with two-stage models (McDonald,
2009). Furthermore, the need to link bank efficiency with market-based indicators of performance is
identified in recent studies (Molyneux, 2018).

An initial study by Chu and Lim (1998) claimed to be first linking bank efficiencies with market
returns. This study used DEA to estimate cost and profit efficiencies of six big banks operating in
Singapore over the period of 1992-96. In a second stage, they interrogated efficiency scores with the
annual change in share price of banks and also examined correlations between the estimates. They
got motivation from earlier studies which claim that the stock price incorporates all relevant publicly
available information in efficient stock markets, while efficiency scores are estimated with published
accounting data. Therefore, a bank efficiency estimate should be relevant to the stock price in any
informationally efficient market. They reported that a change in profit efficiency was reflected in a
change in bank share prices, however, cost efficiency was not found to be related (Chu & Lim,
1998). Meanwhile, Fiordelisi (2007) introduced the concept of shareholder value efficiency by using
a parametric approach to estimate cost, profit, and shareholder value efficiency for listed banks of
four major EU economies over the period of 1997 to 2002. In the second stage, he regressed
shareholder value efficiency (SVE) against various bank efficiency variables in presence of some
control variables. He also explored the correlation among traditional bank efficiency estimates and
shareholder value efficiency. The results of this study found a weak relationship between cost, profit,
and shareholder value efficiency. However, all efficiency estimates were positively and significantly
associated with different measures of shareholder value creation. A later study by Fiordelisi and
Molyneux (2010) further extended the literature on the relationship between bank efficiency and
shareholder value, using similar estimation measures. Using correlation, regression, and Granger-
causality tests, they explored potential determinants of shareholder value and bank efficiency in
addition to looking at the relationship between them. They found a positive and significant
relationship between bank efficiency and shareholder value, while loan losses, financial leverage,
bank size, and liquidity were identified as major determinants of shareholder value.

According to Guzman and Reverte (2008), it means that banks with higher efficiency have higher
shareholder value. The findings of this study are in line with the findings of Pasiouras et al., (2008),
who studied Greek listed banks and Beccalli et al., (2006), who studied European banks. The results
show that bank efficiency has the greatest influence on stock prices as compared to the traditional
accounting measures of performance. They concluded that improvements in bank efficiency will
result in improvements in shareholder value creation through increase in stock prices. According to
Kasman and Kasman (2011), stocks of profit efficient banks tend to outperform their inefficient
counterparts and that managerially efficient banks should generate more profits and greater
shareholder value. Oberholzer, Merwe (2012), purposed the study using Data Envelopment Analysis
(DEA) to aggregate the overall performance (technical efficiency) of firms to convert scarce
resources into outputs that create wealth for shareholders, and secondly, to determine the degree to
which this mentioned performance is reflected in a number of profitability and market value ratios.
Annual financial statement data were used for 55 manufacturing companies listed on the JSE
Limited over a five-year period in a cross-sectional analysis. The study found that return on equity
has the most significant relationship with technical efficiency, followed by return on assets. The
market value ratios price/earnings and dividend yield have no significant relationship with technical
efficiency.
Study by O. Jakata & F. K. Mutasa ( 2014) investigated how stock prices and bank efficiency has
the greatest influence on stock prices as compared to the traditional accounting measures of
performance. The research of Singh & Vashisht (2017), verifies the effect of productivity growth, a
real missing link between EVA (Nopat net operating profit after taxees less the dollar cost of capital
required to create the profit) and firm’s financial health, on shareholders’ wealth maximization. The
study uses the firm level data from Indian Food Industry for the period of 1993-94 to 2005-2006 to
measure and decompose the Malquist Productivity Index into its different components, such as
technological change, pure efficiency change and change in scale efficiency by using DEA. Further
linkage between different components of productivity change and market value added, an indicator
of shareholders’ wealth maximization, was executed by using fixed effect regression model. They
find a positive linkage between TFP and MVA.

Fu et al. (2014) investigated the relationship between shareholder value and bank efficiency for the
274 commercial banks in 14 Asia-Pacific economies between 2003 and 2010. This study
investigated the impact of GFC 2008-09 on the relationship between accounting and market-based
shareholder value with bank efficiency. The authors observed a significant impact of post-crisis
regulatory and economic reforms on the banking environment; therefore, they expected a change in
the relationship between shareholder value and bank efficiency. They regressed accounting and
market-based shareholder value variables with cost and profit efficiency in presence of control
variables. The results indicated a positive relationship between profit efficiency and stock returns.
However, the impact of cost efficiency on shareholder value takes more time to be observed,
consistent with earlier studies by Liadaki, A. and Gaganis, C, (2010). Fernández et al. (2001),
studied economic efficiency in 142 financial intermediaries from eighteen countries over the period
1989-1998 and the relationship between efficiency , productivity change and shareholders, wealth
maximization. DEA Malmquist decomposition was used to separate efficiency change from
technical change that is economic efficiency relation with wealth maximization. A non-parametric
frontier analysis (DEA) is applied to estimate the relative efficiency of commercial banks of
different geographical areas. A Malmquist decomposition is then carried out in order to separate
efficiency change from technical change. They evaluated the relationship between economic
efficiency and wealth maximization. Results show different productivity patterns among three
geographical areas (North America, Japan and Europe) over the sample period. The estimates of
economic efficiency and productivity changes are consistent with the wealth maximization criterion.

Generous efforts are being executed to find the relation between efficiency and wealth maximization
objective of companies and the economic value added EVA (developed by Stern Stewart &
Company), a new financial tool that is being adopted to guess the relationship. However, evidence of
EVA as a predictor of shareholders’ wealth is mixed. It has been observed that competitive pressures
and market discipline in semi-strong markets exert pressure on banking operations to shift its focus
more on value creations to shareholders (Beccalli et al., 2006). Literature has claimed high
correlation between EVA and stock performance, and EVA is a driving force behind better stock
returns (de Medeiros, 2005). Considering the vital role of EVA and MVA in shareholder value
creation, some more relationship models as robustness test between EVA/MVA and bank efficiency
can be developed.

Our study can be in the context of an emerging economy to reduce the literature gap. Besides,
emerging markets and economies have been playing a vital role in the world economy due to
globalization of economies in recent years (Arora, Jain, & Das, 2009). Undoubtedly, India offers a
singular case study because it has an emerging and growing economic system and is quite
established in its banking sector. Indian banking industry has a diversified ownership and is well
regulated under the authority of the RBI (Sensarma, 2006). A considerable number of foreign banks
operate today in India and hence the efficiency benchmarks of the Indian banking industry are
significant for investors/shareholders and other economies that are looking forward to tapping the
Indian market. It has been observed that competitive pressures and market discipline in semi-strong
markets exert pressure on banking operations to shift its focus more on value creations to
shareholders (Beccalli et al., 2006). The present work will work on the emerging Indian economy
with the objective of determining an empirical linkage between technical efficiency and stock
market performance of the Indian banking industry.

3.6 DEA: Stock Picking Technique

Due to the advancement of Information Technology it has been easier for investors to invest
valuable money in portfolios. There has been availability of data various tool/techniques all the time
to predict the market in decision-making, for which some models have been developed. With the
rigorous research in this attractive topic some mathematical models have been developed & some
models from other industries have been considered. DEA model, originated from production
industry, may help in selecting securities for portfolio. In our study we will try to find the right ways
to select the worthwhile and profitable stocks from banking sector by using DEA technique. In the
past decade, an ample amount of researches have been done in portfolio building through traditional
and modern methods but DEA is an evolutionary technique in this area, which is employed by the
researchers in a few studies. Use of DEA in stock selection from a single industry on the basis of
sectoral indices is rarest of the rare, which present study is taking it as a research pursuit. Stock
evaluation process plays an important role in portfolio selection because it is the prerequisite for
investment and directly influences on the stock allocation (Zamani et al., 2014). Zamani et al.,
(2014) used A-P Super-Efficiency Model, to determine optimal portfolio stocks for investors in the
Mumbai stock exchange. Some literature related to the portfolio selection is discussed below:

Several portfolio management approaches have been developed for a successful portfolio selection.
In the traditional approach, portfolio risk is reduced by over variation and ignoring the correlation
among securities, whereas in the modern approach variation is provided by mean variance model
(Markowitz, 1952). Similar to Markowitz s-variance model, Roy also developed mean- variance
efficiency frontier by examining the relationship between the variance of the returns from the
securities of portfolio and the returns from the portfolio (Roy, 1952). By the time, several factors
such as borrowing, loaning, short term selling, transaction cost etc. are used to the original model,
based on the mean-variance model, to improve the portfolio management (Sharpe, 1963). The big
problem with Modern Portfolio Theory is the number of data to be calculated would increase
exponentially with the increase in the number of stocks as we have to calculate the estimated return,
standard deviations of all stocks and most importantly the covariance between these stocks.
Consequently, this would be a complicated process. CAPM model has some constraints such as its
rigidity, unrealistic assumptions etc., so, multi-factor models are evolved to relax the constraints, but
it is hard to determine the impact factors. Moreover, the traditional methods do not consider the
multi-variable for evaluation of stocks for portfolio, which is a very important factor in performance
evaluation. To solve these problems, DEA was introduced in portfolio performance evaluation. It
does not need the hypothesis of the effectiveness of capital markets and could avoid the impact of
the benchmark portfolio on evaluation result. So this approach led to the widespread concern in
recent years. Murthi, Choi and Desai (1997) first used DEA to take into account the investment costs
in defining a mutual fund performance. Afterward, Basso and Funari (2001) proposed a new mutual
fund performance indexes that take into account a variety of transaction costs and risk measure value
in DEA model. Xu and Zhang (2009) used the input oriented BCC DEA model. In portfolio
selection, Murthi et al. (1997), Basso & Funari (2001), Eilat et al. (2006), Amiri et al. (2010) used
DEA methodology in order to evaluation or choose assets, stocks, mutual funds etc.

Researchers are working on portfolio selection area with the help of using DEA. Wu (2017) used a
reliable method for evaluating the asset performance, and it was ranked from both self- and peer-
evaluation perspectives. Azadeh and Kokabi (2016) proposed a Z-number version of the CCR and
BCC models that can be applied to a portfolio selection problem to tackle uncertainties. Rȩbiasz et
al. (2017) used both probability distributions and fuzzy numbers to propose a framework that
enables to consider stochastic dependencies between model parameters and economic dependencies
between projects. Eilat et al. (2006) proposed a methodology that is based on DEA model and
Balanced Score Card approach to analyze portfolios of R&D projects with interactions. Mashayekhi
and Omrani (2016) combined DEA cross-efficiency and Markowitz mean-variance model to propose
a novel multi objective model for portfolio selection. Also, Lim et al. (2014) used DEA cross-
efficiency evaluation in portfolio selection.

A research carried by E. Pätäri et al. (2012) examines the applicability of data envelopment analysis
(DEA) as a basis of selection criteria for equity portfolios. It is the first DEA application for
constructing a combined equity investment strategy that aims to integrate the benefits of both value
investing and momentum investing. Allen & Singh, (2010); Alexandru C. Badoi, (2016),
investigated the Data Envelopment Analysis (DEA) CCR model as a value asset selection criterion
to construct portfolios and investigate their relative performance. Alexandru C. Badoi, (2016) used
the Fama French (2015) factor model to decompose the systematic risk into five components. The
CCR DEA is then applied to the risk factors and asset returns with the goal of selecting superior
investment options from a large US sample. Some of the complications with implementing the
CAPM model are also addressed through the implementation of Quantile Regression. These papers
provide the groundwork for future DEA research in the context of risk and return of financial assets.
Presented proposed study will also follow the tools and techniques of these papers.

The above review of studies makes it clear that a large number of researches exist on measuring the
efficiency performance of banks, but in spite of vast literature, lesser studies are available with
respect to India that analyze other different and unique Profit Efficiency parameters of banks (by
employing DEA) such as failure prediction or estimating solvency strength, wealth maximization of
shareholders/ investors and the most important- the selection of viable stocks to beat market return.
Moreover, the empirical studies covering these efficiency concepts simultaneously are completely
missing from the Indian Banking literature. Furthermore, there are lots of possibilities to study the
relation of these essential aspects from the point of view of investors and develop a consensus
technique for the better decision making while investors plan to employ their funds in financial
sector, mainly in banking industry. In the Efficiency parameters relating to DEA, this is a humble
effort to fill the research gap into the unexplored area of banking.
 4. Need of the Study: The Justification of the Study

Liberalization, Privatization and Globalization (LPG) reforms started in India in 1991 greeted
private and foreign banks to operate in Indian Banking Industry with the public sector banks. These
reforms initiated modification and reformulation in various laws and regulations of financial sector.
It made sea changes i.e. from fully regulated environment, Indian banks gradually moved into a
market driven competitive system. Moreover, intense competition in banking sector forced banks to
share of profitability and increase the work efficiency. The series of banking reforms as entry
deregulation, branch de-licensing, dismantling of administrated interest rate structure, reduction in
CRR and SLR and adoption of prudential norms etc were introduced in the Indian Banking Sector
with an intention to improve the efficiency of banks. Apart from the above the last five year period
(2014-19) was highly critical for banking industry to the abounding limits such as demonetization,
various govt. plans (Jan Dhan Yojana etc.), E-payments and cashless transactions, fraudulent and
corrupt activities and many more. Efficiency of banking system is one of the most important issues
in the financial market as it can affect the stability of banking industry and ultimately the
effectiveness of whole monetary system (Yilmaz, 2013). All stakeholders (especially shareholders)
are more concerned about how economically and efficiently banks convert their expensive inputs
into valuable financial products and services (Isik and Hassan, 2002). The prologue of various
banking remedial reforms encouraged the researchers and various academicians to examine the
performance of banks in terms of various efficiency measures. So along with profit efficiency of
Indian Commercial Banks, we need to evaluate other related efficiencies (they might be cost,
technical or allocative efficiencies) which can be evaluated through DEA. While investigating
efficiencies we should focus on its impact and relation with other crucial areas such as financial
distress prediction, shareholders’ value creation or wealth maximization and above all determining
the profitable portfolios by adding other sector’s securities along with banking sector stake. This
study will analyze the ten year span of time from 2009 to 2019 further splitting into two equal parts.
First five years period is a tableau of steps taken by Indian National Congress ruled government for
advancement of Indian banking industry while next five years period is indicative of changes made
by Bhartiya Janata Party led government in the policies and laws for the same industry to tackle the
giant sized problems surging into this sector.

5. Research Gap: Quest for Uniqueness

From the past discovered literature, researcher can reach at the opinion that ample amount of studies
have been done on evaluation of efficiencies of banking companies from the different-different
dimensions and these type of studies are also being executed in the present time, however some
other essential and significant aspects are still untouched or yet to be investigated. As the success or
failure of banks mainly depends on the extent to which they are managed efficiently and the study
must be helpful to all the stakeholders, particularly for investors to take well-informed investment
decisions. In the context of DEA, it can be observed some gaps or vacuum from the literature
review. They may be constituted as: First research gap in research of banking realm is the evaluation
of banking efficiency in the present scenario. The present study can be distinguished from other
studies with different angles. As Debasish, (2006), Sahoo et al (2007), Kumar (2013) and other
various researchers analyzed various efficiencies (Cost, Technical and Allocative) of Indian Banks
with DEA with the single approach either profitability or productive model approach but this
research will assess not only allover efficiency but also use different dimensions of productivity
changes such as profitability, intermediation, and a composite of both by taking a wider span of
time period to map the Indian Banking Sector efficiency. Apart from this the study will evaluate
other important but missed aspects in previous researches such as Deposit Mobilizing Efficiency
(DME), Fund Conversing Efficiency (FCE), Off-balance Sheet Activities Efficiency (OBE), Cost
Revenue Management Efficiency (CRE) through DEA-CRR technique. Second research vacuity is
to study the bank distress which must be at high priority. It is very important: (i) to the regulators to
predict potential crisis and enables them to manage timely, (ii) the early clarity and distinction
between troubled and sound banks allows for appropriate actions to prevent failure and to protect
healthy institutions and (iii) the crisis in the financial sector may generate other fatal situations in the
economy such as currency crisis. Literature on banking distress prediction in India is very rare.
However, Mukhopadhyay et al. (2012) tried the DEA methodology in bankruptcy prediction but
their study is not related to banking sector and up to our best knowledge there is no study available
which analyzes the banking sector distress prediction through DEA technique. This study will be
more important as the distress prediction level synthesized by DEA would be put into robustness test
with Altman’s Z- Score Model. Third research gap may be in the form of wealth maximization of
shareholders as creating value is one of the critical issues and problems that entrepreneurs face, and
it has a bearing on the financing of entrepreneurial ventures which is, of course, applicable on
banking companies as well. Goel S. (2013) studied indicators of shareholders’ wealth maximization
in Construction Industry; Goyal J. (2018) evaluated the shareholders’ value creation in food industry
with DEA methodology but efficiency relation with shareholders’ value in banking sector is rarely
available. We found just a single study (Sharma, 2018) to analyze this motive in Indian banking
sector and the present study is aligned with the study of Sharma (2018) but with new banking
regulations and current banking scenario. The fourth and a vital inquest investigation is fruitful stock
selection through DEA instead of traditional or older methods having many constraints. However,
Zamani et al. (2014), Singh et al. (2017) proposed the DEA methodology for portfolio selection and
optimization in Indian context but they did not consider the various risk beta coefficients which
arises some questions on the portfolios efficiency to face extraneous risks. This study is unique from
the point of view that it applies DEA tool on Fama- French (2015) five factor model to select the
stocks form only banking industry to build viable stocks which makes it more reliable technique.

6. Objectives of the Study: Visualization Outlined

On the basis of literature review and gaps identified; the following objectives have been framed for
the proposed study:
1. To evaluate the efficiency of Indian Banking sector.
2. To predict the financial distress of Indian Banks with DEA methodology and determine the
suitability of DEA as financial distress predictor.
3. To investigate the level of relationship of Banks’ DEA efficiency with Shareholders’ wealth
maximization.
4. To frame the methodology using DEA for selection or picking profitable stocks to build
viable portfolios with reference to banking sector.
 7. The Study Model: Data Envelopment Analysis (DEA)
This is a non-parametric approach of measurement of the efficiency relative to various Decision
Making Units (DMUs). DMUs are the homogeneous units and in our study DMUs are the sample of
commercial banks. Technical Efficiency Score is the total weighted sum of outputs divided by total
weighted sum of inputs. In this model, Technical Efficiency Score is measured by the ratio of
weighted outputs to weighted inputs. The efficiency of bank is measured in terms of how efficiently
a bank utilized its inputs by using the following formula (Charnes, Cooper and Rodhes, 1978). In the
model, the condition of efficiency is Ɵ*=1 and all the slack variables are assumed to be zero (O).
Under such conditions, DMU organizations must have to maximize values of outputs against its
inputs, which are the subject of evaluation. The efficiency score is a function of the weights of the
virtual input-output combination. The weights for the ratio are determined by pre-limiting those
ratios for every DMU as they have to be less than or equal to unity, thus reducing multiple inputs
and outputs to a single virtual output without requiring pre-assigned weights. The efficiency score of
a given DMUₒ is obtained by solving the following linear programming model:-
The original model by Cooper WW, Seiford LM and Tone K (2007) is given as:

∑ sr=1 ur y r u y +u y + …+u s y s
max h ₒ ( u , v )= rn = 1 1 2 2
∑ i=1 v i xi v 1 x1 + v 2 x +…+ v m x m

Where
y r = amount of output r
ur = weight assigned to output r
x i= amount of input i
vi = weight assigned to input i.
The wights may be fixed in advance or derived from the data. The former is sometimes referred as
an a priori determination (Cooper WW, Seiford LM and Tone K, 2007).

The above linear programming model is run to the n th times for identifying the efficiency score of all
the DMUs. Input weights are selected for each DMU which maximize its efficiency score.
Generally, the DMU which obtains a score of 1.00 is considered to be efficient (implying 100
percent efficiency) whereas a score of less than 1.00 makes it inefficient unit. It is to be noted that
the Technical Efficiency comprises pure technical efficiency and scale efficiency. This requires the
estimation of the two DEA models – one with Constant Returns to Scale (CRS) -CCR (Charnes,
Cooper and Rodhes, 1978) and the other with Variable Returns to Scale (VRS) – BCC (Banker RD,
Charnes and Cooper WW, 1984). If there is difference in the two technical efficiency scores of a
particular bank, it means that the bank’s scale is inefficient. Sherman and Zhu (2006) indicate the
main advantages of DEA as follow:
1. To calculate technical efficiency, it can work with multiple inputs and outputs.
2. With the efficiency levels it can determine the possible sources of inefficiencies.
3. It can provide a potential role model to the ‘peers’ which are not considered as efficient units
and they can improve their operations.
4. Except a linear form, DEA does not require a functional form relating to inputs and outputs.
(Researchers are also working with non-linear data in DEA analysis)
5. Inputs and outputs can have different units. In this situation, any hypothesis and converter are
not being required to use.
 8. Research Methodology: Methods and Phases of Study

In this proposed study, mainly secondary data will be used, collected from Annual Reports of Banks,
RBI Bulletins, various reputed journals and magazines, official websites or Govt. sources etc. A
dataset, (which may be a balanced or unbalanced panel from three ownerships namely Public Banks,
Private Banks and Foreign Banks depending upon objective of study and availability of data)
consisting a tentative sample of 50 Banks operating in India over the period of 10 years into two
segments of 1st April 2009 to 31st March 2014 and 1st April 2014 to 31st March 2019 representing
two different parties’ govt. agenda, concern and policy making for commercial banks, will be used to
actualize first two objectives. To work out on last two objectives, however, study period will be
same but data will be gathered from NSE’s official website endorsing the information of listed
banks.
A. To assess the first objective of the study DEA’s Constant Return to Scale and Variable
Return to Scale models (DEA-CRS and DEA-VRS) will be used for the two different periods
to evaluate the overall efficiency of the Banks with the sample of 50 banks operating in
India. This multifaceted study will take many steps to reach its objective such as running two
different DEA models, comparing the efficiency level of banks of two periods and studying
gain and loss of banks’ ranking and industry’s allover efficiency shift.
B. On the way of second objective a study with BSS (Barr, Seiford and Seims, 1993) approach
comprising DEA-CAMELS model would be executed by establishing a management quality
metric through DEA that will work as a proxy for the ‘M’ in the CAMELS rating. The BSS-
DEA model uses six inputs (namely as full time- equivalent employees, salary expenses,
premises and fixed assets, other non-interest expenses, total interest expenses and purchased
funds and three outputs (namely as core deposits, earning assets and total interest income).
After that by applying the CAMELS prediction of financial distress one year ahead and two
year ahead will be estimated with early warning signals. The variables of CAMELS model
will be regressed to explore the study to investigate which variable impacts more on the
stability of banking organizations. To test the suitability of this model Altman’s (2002)
modified version emerging market Z-Score Model will be applied in this analysis. The model
for companies from emerging markets (banks) the calculation formula takes the following
form:
Z=6.56*X1+3.26*X2+6.72*X3+1.05*X4+3.26
X1=Working Capital / Total Assets (WC/TA)
X2=Retained Earnings / Total Assets (RE/TA)
X3=Earnings before Interest and Taxes / Total Assets (EBIT/TA)
X4=Market Value Equity /Book Value of Total Liabilities (MVE/TL)
Z> 2.6 – ‘safe’ zone
1.1 <Z <2.6 – ‘grey’ zone
Z <1.1 – ‘distress’ zone
C. For the third objective, study will be conducted in four steps: In the first step, technical
efficiency score of Indian listed banks will be estimated by DEA intermediation approach.
Market performance along with annual average return of stock market prices will be
calculated in the second step. The third step would develop a statistical measure to testify the
bank’s performance and its efficiency relation with micro-economic banking industry factors
by using panel data regression. Finally the relationship of efficiency and shareholder value
will be analyzed by market value indicators, to fetch the results, Economic Value Added and
Market Value Added approaches would be regressed with dependent variables. Process is
briefed as follow:
1. Computation of Return Ri , t=log ⁡( Pi ,t /Pi , t−1)
Where, Ri , t is the return of stock i at time t, while Pi ,t and Pi ,t −1 are the prices of stocks at
time t and t-1, respectively. The annual returns were calculated by adding daily returns
because daily returns have smaller variations than monthly or annual returns (Beccali et al.,
2006).
2. Modeling stock market performance and banking efficiency: The stock market
performance and banking efficiency model may be defined as follows: Market Returni ,t =
α+ β Eff i , t+γ Bank i, t +δ i+ξ t +ε it
Where, Market Returni ,t is the dependent variable and represents annual stock returns of the i t h
bank in the t t h time period and Eff i , t includes change in technical efficiency scores based on
DEA. Bank i, t represents a vector of bank specific factors that include bank size in terms of
total assets, expense preference behavior in terms of noninterest expenses over total assets,
the bank’s diversification strategy in terms of non-interest income over total assets,
profitability in terms of the previous year’s return on assets and capital adequacy ratio and
non-performing assets. δ i captures firm effect, ξ t captures time effect and ε it are the error
terms for i = 1,2,…, N cross-sectional units (i.e., 22 banks) observed for the period t = 1,2,
…,T (i.e., 10 financial years, from 1 st April, 2009 to 31st March, 2014 and 1st April, 2014 to
31st March, 2019 ).
3. EVA is calculated as follows:
EVA = NOPAT − (WACC × invested capital)
Net Operating Profit (NOPAT) = (net profit + provisions and contingencies + interest on
borrowings) less (Taxes)
Invested Capital = total equity and reserves + total borrowings
Return on Invested Capital (ROIC) = NOPAT/ invested capital
Cost of Debt = (total interest expenses – interest on deposits) / total borrowings
Cost of Equity = as per capital asset pricing model (CAPM), Ke = Rf +β (Rm - Rf), where Rf
is risk-free rate of return, rate of 364 days treasury bills Rm is market return, based on
historical data for average return.
Beta (β) is monthly closing stock price of banks (respective percentage of change of prices)
vis-a-vis NSE nifty index (change in percentage in Nifty closing prices) using the following
formula:
Beta (β) = covariance of market return with the stock return /variance of market return =
covariance (stock return, market return)/ variance (market return) WACC = weighted cost of
equity + weighted cost of debt
MVA is closely linked to economic profit and is defined as the difference between a bank’s
market value and the capital invested. This measure gives the yield in terms of the economic
profit generated by any firm from its resources (IC). MVA is defined as follows:
MVA = market capitalization – invested capital, A positive value of MVA indicates that the
firm has created market value, whereas a negative value denotes the destruction in the market
value of the firm.
4. Robustness check relationship models are defined as follows:
EVA i ,t /IC i , t−1= α+βEff i , t+γBank i, t +δ i+ξ t +ε it, where IC= Invested capital.
D. This study will use the Fama-French five factor model and DEA methodology for
measurement the fundamental strength of stocks to achieve the fourth objective of the
presented study. The quantile regression approach will be used to get a most accurate
estimate of those risk coefficients. The variables in Fama-French will be used as inputs into
the Charnes (1978) DEA CCR Model while return will be treated as output in our study. In
our knowledge there is no study available on stock return efficiency in context of the sectoral
indices. For the further exploration of our study we will work on some new experiments with
sectoral indices and DMU’s stock return of the studied period. The Fama-French (2015)
model is given as: Rit - R Ft= a i+ b i( R Mt- R Ft)+ si SMB t+hi HMLt+rii RMW t +c i CMAt +e it

In the above given equation Rit is the return on security or portfolio i for period t, R Ft is the
risk free rate of return, R Mt is the return on the value-weight (VW) market portfolio, SMBt is
the return on a diversified portfolio of small stocks minus the return on a diversified portfolio
of big stocks, HMLt is the difference between the returns on diversified portfolios of high and
low B/M stocks, RMW t is the difference between the Returns on diversified portfolios of
stocks with robust and weak profitability, and CMA t is the difference between the returns on
diversified portfolios of the stocks of low and high investment firms, which we call
conservative and aggressive and e it is a zero-mean residual. In the exposures to the five
factors, a i, b i, si, hi , rii and c i capture all variation in expected returns. The fractional DEA
model presented above can now be expressed in linear form under the name CCR DEA in the
following way:
s
Max: e j = ∑ u r Y rj
r=1

s m
= ∑ u r Y rj - ∑ v i X ij ¿ 0
r=1 t =1
m

∑ v i X ij =1
t =1
ur , vi ¿ 0

9. TENTATIVE TIME SCHEDULE

Sr. No. ACTIVITY TIME IN

MONTHS
Updating of Review of Literature 4 – 5 months
1.

Finalizing secondary data sources 2 – 3 months

2.

Data Collection 4 – 6 months

3.
 Data Processing 5 – 6 months
4.

Data Analysis 6 – 8 months

5.

Report Writing 5 – 6 months

6.

10. TENTATIVE CHAPTER SCHEME

The tentative chapter wise study is as follows:
Chapter 1: Introduction
Chapter 2: Review of Literature
Chapter 3: Research Methodology
Chapter 4: Data Analysis and Interpretation
Chapter 5: Findings, Conclusions and Recommendations
Bibliography