Article

Readability of Annual Reports and IIM Kozhikode Society & Management Review
10(1) 20–30, 2021
Operating Performance of Indian © 2020 Indian Institute
of Management Kozhikode
Banking Companies Reprints and permissions:
in.sagepub.com/journals-permissions-india
DOI: 10.1177/2277975220941946
journals.sagepub.com/home/ksm

M. Jayasree1 and Rachappa Shette2

Abstract
Existing literature focuses on the evaluation of the readability of annual reports of non-banking companies. However,
banking companies’ opaque nature and a double motivation to abuse accounting discretion requires a separate study on
the readability of banks’ annual reports in association with their performance. We, therefore, attempt to explore firm
performance and readability of banking firms’ annual reports in India. Net interest margin (NIM) and Fog Index are used
as performance and readability variables respectively. We find that management discussion and analysis (MD&A) of the
Indian banks is difficult to read. However, when we compare it with existing literature, Indian banks’ MD&A is difficult
but not unreadable. Panel data regression analysis shows that firm performance would have a negative impact on the Fog
Index. Further analysis of good and weak performing banking firms shows that the effect of NIM on Fog Index is higher
in the case of weak performing banks. Empirical results affirm that firms with weak performance would structure their
annual reports to veil adverse information in unfavourable situations. Consistent with the opaque nature of banks and
incomplete revelation, managers of banks make MD&A harder to read to cover up the causes of weak performance.
Application of readability index in case of banking companies in an emerging economy in association with performance
is the contribution of this paper. An assessment of the readability of annual reports is an interesting topic for research
to better understand the recent negative developments in Indian banking industry such as high non-performing assets,
continuously declining return on assets, sharp increase in banking frauds and poor governance.

Keywords
Firm performance, banks, non-performing assets, readability, Fog Index, net-interest margin, non-interest ratio

Introduction more than forty years, I’ve studied the documents that pub-
lic companies file. Too often, I’ve been unable to decipher
The performance of banking companies and their financial just what is being said or, worse yet, had to conclude that
condition usually indicates the financial health of an econ- nothing was being said’. Similarly, Arthur Levitte (SEC,
omy. Corporate disclosures along with financial statements 1998), then the SEC chairman, notes, ‘Because many
are critical for investors, deposit holders and regulators investors are neither lawyers, accountants or investment
to understand the financial performance and condition bankers, we need to start writing disclosure documents in a
of banking companies. Corporate disclosures in annual language investor can understand’ (p.3). The analysis of
reports are in the form of Chairman’s Statement, Directors’ the readability of corporates disclosures in annual reports
Report, President’s Note, Management Discussion and thus becomes a key area of research. Readability refers to
Analysis (MD&A), audit report and footnotes to financial the ease of reading and understanding a written text (Harris
statements. However, in some cases, such disclosures are & Hodges, 1995). A document with long sentences, pas-
difficult to read. Warrant Buffet (SEC, 1998) notes, ‘… for sive voice, superfluous words and legal and financial

1
Associate Professor, GITAM Hyderabad Business School, Hyderabad, India.
2
Associate Professor, Indian Institute of Management Kozhikode, India.

Corresponding author:
M. Jayasree, Associate Professor, GITAM Hyderabad Business School, Hyderabad, India.
E-mail: [email protected]
Jayasree and Shette 21

jargon is difficult to read whereas a document with short (deposit holders), investors and other stakeholders of
sentences and easy to understand words is an easier docu- banks. We expect the readability of the MD&A of banking
ment to read. companies to be different as compared to that of industrial
Using readability indices, the existing literature evalu- companies due to the opaque nature of banks and the
ates the readability of different corporate disclosures double motive of earnings management.
(Hrasky, Mason, & Wills, 2009; Jones & Shoemaker, 1994; We select the MD&A of banks to evaluate their
Pashalian & Crissy, 1950; Soper & Dolphin, 1964). It also readability because it is an important source of information
relates positive earnings with easier to read corporate in understanding the positive and negative developments
disclosures and negative earnings with harder to read of banks and the effect these developments have on the
corporate disclosures (Bloomfield, 2002; Li, 2008). The financial performance and financial condition of the banks.
readability of corporate disclosures in different indus- In India, the recent negative developments in banking
tries could be diverse due to difference in the levels of industry include high non-performing assets (NPAs),
complexity of products and services (Li, 2008). This paper continuous decline in return on assets, sharp increase in
aims to analyse the effect of banking companies’ earnings banking frauds and poor governance. India’s NPAs are
performance on the readability of their MD&A. We have some of the worst in the world, ranking 5th with 9.98 per
selected the MD&A section of annual reports as it is the cent loans according to International Monetary Fund.3
most read and frequently used disclosure in annual reports According to a Report on Trend and Progress of Banking
(Tavcar, 1988). in India of Reserve Bank of India, the return on assets in
Unlike non-banking companies, banking companies banking industry was lowest among all emerging and
have two motives to abuse accounting discretion. Banks developing countries and it has declined continuously
abuse accounting discretion to meet the regulatory capi- during 2010 to 2017.4 The report also warns about the
tal requirements or earnings thresholds or both (Beatty, sharp increase in the number of frauds in various banks in
Chamberlain, & Magliolo, 1995; Collins, Shackelford, & India from `101,708 million during 2013–14 to `411,677
Wahlen, 1995). The opaque nature of banking companies million in 2017–18. One of the possible reasons for such
and the two motives for abusing accounting discretion are high NPAs and banking frauds resulting in decline in the
the reasons for selecting banking companies for this study. return on assets is the poor governance of banks. Discussion
The opaque nature (Dang, Gorton, Holmstr0m, & Ordonez, and understanding of such serious issues are vital to
2017; Howe & Haggard, 2012) is a unique feature of banks understanding the financial performance and financial
as compared to industrial companies. Consistent with the condition of banks. As MD&A also discusses these issues,
opaque nature of banks, banking companies were exempted the assessment of readability of the MD&A of banking
from full disclosures of financial reports until 1974. The companies enables the regulator, the deposit holder and
higher cost of improved transparency and quality of accou- the investor to better evaluate the banks. Therefore, the
nting information are causes of the opaque nature of banks. evaluation of readability of annual reports of Indian
If banks were fully transparent, it would reduce the ability banking companies is an interesting subject of research to
of banks to raise funds from deposit holders, thereby improve the understandability of disclosures and to reduce
raising concerns over liquidity. This is in contrast with the asymmetric information in banking industry.
positive effects of higher accounting quality and improved The readability analysis of MD&A in this paper is based
transparency in resolving the agency problem in non- on all 39 banking companies listed on National Stock
banking companies. The opaque nature of banks may Exchange (NSE) between 2013–14 and 2016–17. The list
impact disclosures in the annual reports adversely, making consists of 21 state-owned banks and 18 private banks.
annual reports difficult to read. The fact that such obscurity We applied Fog Index to measure the readability as used in
and secrecy is an acceptable attribute of banks, it may recent studies (Li, 2008; Lim, Chalmers, & Hanlon, 2018;
motivate the managers to write unreadable disclosures in Lo, Ramos, & Rogo, 2017). We also applied Flesch
annual reports. The present literature reveals the existence Readability Ease Index (FREI) as an alternative measure of
of earnings management practices in Indian banking readability to check the robustness of the results. We use
industry (Kumari & Pattanayak, 2017). As documented by net interest margin (NIM) as the performance indicator. We
Kumari and Pattanayak (2015), accounting for bad loans ran panel data regression to analyse the impact of the
and gains or losses on asset sales are two big items which current-year performance on the readability of MD&A of
depend on the managers’ discretion. Hence, they may banking companies.
indulge in writing difficult to read and sometimes, even We find mean Fog Index and FREI as 15.87 and 37.26,
unreadable disclosures in the annual report to hide such respectively, indicating that Indian banking firms’ MD&A
discretionary accounting judgments from the creditors are difficult to read. However, when we compare these
22 IIM Kozhikode Society & Management Review 10(1)

with firms from the USA and 42 other countries, which banks’ financial performance and financial condition are
have a Fog Index above 18 (Lang & Stice-Lawrence, 2015; the global economic scenario, macro environment, Indian
Li, 2008), Indian banking firms’ MD&A are difficult to banking industry, performance with a focus on operating
understand but not unreadable. Secondly, we find that NIM revenues and expenses, loans and advances, deposits, assets
has a significant negative effect on the Fog Index of quality including non-performing assets, capital adequacy
readability (higher Fog Index means difficult to read). We ratios, risk management and financial inclusiveness. The
also find that NIM has a significant positive effect on FREI MD&A of banking companies contains a detailed discus-
(lower FREI means difficult to read). This means that sion on these factors and their effect on the current-year
disclosures of poor (good) performing banking companies performance and financial condition of banks. MD&A
are harder (easier) to read. We provide evidence to support plays a vital role in better understanding the figures pre-
that Fog Index is influenced by firm performance, and that sented in the current­­ -year income statement, cash flow
firms structure their annual reports to veil adverse infor- statement and balance sheet as it is an important part of the
mation in unfavourable situations. This is in line with annual report and its readability is essential for the deposit
incomplete revelation hypothesis (Bloomfield, 2002). This holders and investors. The evaluation of readability of
paper makes two contributions which are as follows: annual reports is therefore vital in understanding the current
and future financial performance and position of a bank.
• First, understanding and exploring the relationship Using Fog Index as a readability measure, Li (2008)
between readability of annual reports with firm per- related the readability of annual reports of a large sample
formance, specifically in the context of the banking of American companies to their earnings performance in
sector within a developing economy. the current and subsequent years. The Fog Index of a
• Second, we ascertain that the readability of the document states the number of years of formal education
annual reports of Indian banking companies is diffi- required for an average reader to understand the document
cult but not unreadable. We suggest banking compa- in first reading. Li’s study finds a negative and significant
nies to reduce the difficulty level and improve the impact of current-year earnings performance on the read-
readability of these reports. ability index of disclosures. It also finds a negative and
significant impact of the current-year readability index of
The remainder of the paper is as follows. We review the disclosures on the earnings performance of the subsequent
literature and propose two hypotheses in Section 2. year. It means that the annual reports from firms with lower
Methodology is discussed in Section 3. Thereafter, we earnings in the current year are difficult to read and these
present and discuss the results in Section 4 and conclusions firms would have lower earnings in the subsequent year.
are in Section 5. Prior to Li’s paper, Subramanian, Insley and Blackwell
(1993) found that annual reports of profitable corporations
are easier to read as compared to those of unprofitable
Literature Review corporations. A better quality of readability is also linked
with greater capital investment efficiency (Biddle, Hilary,
Research on the readability of disclosures in annual reports & Verdi, 2009), with higher trading of shares by small
of firms has a long history. Pashalian and Crissy (1950), investors (Miller, 2010) and with higher investment
Soper and Dolphin (1964), Smith and Smith (1971), holdings by retail investors (Lawrence, 2013). Chipalkatti
Dolphin and Wagely (1977), Adelberg (1979), Lewis, (2005) relates more readable disclosures of Indian banking
Parker, Pound and Sutcliffe (1986) and Jones and Shoe- companies with a positive perception of investors.
maker (1994) analysed the readability of corporate annual Bloomfield (2002) related easier to read corporate
reports by undertaking analysis of the separate items of annual reports during positive earnings and harder to read
corporate annual reports such as Chairman’s Statement, corporate annual reports during negative earnings with
Directors’ Report, President’s Note, MD&A, audit reports ‘incomplete revelation hypothesis (IRH).’ IRH predicts
and footnotes to financial statements. that managers attempt to manipulate market prices by
These disclosures in annual reports provide relevant emphasizing good news and tucking bad news in footnotes.
information to the stakeholders of the companies in It implies that managers hide bad news related to earnings
understanding the reasons for current-year financial per- performance of the current year and subsequent year in
formance and financial position of the company. The difficult to read footnotes to financial statements and the
disclosures also guide them in estimating the future MD&A of annual reports. Consistent with IRH, managers
performance, financial position and cash flows at different make disclosures harder for the investors to uncover
levels of risk. A few important factors that influence Indian information if they do not want to affect their firms’ stock
Jayasree and Shette 23

prices negatively. Such harder to read disclosures may H1: MD&A of banking companies is unreadable.
relate to: H2: Earnings performance of the current year would
have a negative impact on the readability index of
• Managers’ decision to change the accounting meth- the MD&A of the current year.
ods to improve highly visible statistics such as
reported profit-after-tax, earnings-per-share, return- We expect the results to be different from the existing
on-assets, debt-to-equity ratio, leverage ratio, cur- studies such as Li (2008) and Lang and Stice-Lawrence
rent ratio, etc. (2015) due to the opaque nature of banking companies and
• Managers’ decision to conceal the expenses, capital- dual motives of earnings management.
ized revenue expenses, provisions, contingencies
and liabilities.
• Managers’ unreliable guidance about future sales Methodology
orders, capital expenditure projects, cash flows,
dividend, internal and external risk factors. Sample Firms
The major disclosures in a corporate annual report are The sample comprises of 39 banks, 21 state-owned and 18
MD&A, auditors’ report, notes to financial statements private, which were listed on the National Stock Exchange
including significant accounting policies and corporate (NSE) of India as of 31 March 2017. All these banks are
governance. Tavcar (1988) says that MD&A is arguably considered for this study, which covers the period between
the most read and most important component of the finan- 2013–14 and 2016–17.
cial section. Knutson (1993) and Rogers and Grant (1997)
observed that sell-side financial analysts in the USA most
frequently rely upon MD&A among all disclosure items of Proxies of Readability
the annual report. Similarly, the concept notes of the USA’s
We use Fog Index to measure the readability of the MD&A
Securities Exchange Commission (SEC) state:5
of the annual reports of banks. The computation and
The Commission has long recognized the need for a narrative interpretation of Fog Index are as follow:
explanation of the financial statements, because a numerical Fog = (words per sentence +
presentation and brief accompanying footnotes alone may (1)
be insufficient for an investor to judge the quality of earn- percent of complex words) # 0.4
ings and the likelihood that past performance is indicative Fog is a composition of a number of words per sentence
of future performance. MD&A is intended to give the inves-
and the percentage of complex words in that sentence.
tor an opportunity to look at the company through the eyes
of management by providing both a short-term and long- More complex words and long sentences would mean the
term analysis of the business of the company. The item asks MD&A is difficult to read. The words per sentence are
management to discuss the dynamics of the business and to calculated as a proportion of total number of words to the
analyse the financials. number of sentences in MD&A. A word is complex if it has
three or more syllables. Recent examples of studies using
This makes MD&A an important and useful disclosure for Fog Index for readability include Li (2008), Biddle et al.
investment and credit decisions. (2009), Miller (2010), Lehavy, Li and Merkley (2011),
For this paper, we have considered the MD&A section Dougal, Engelberg, Garcia and Parsons (2012), Lawrence
of the annual reports of banks for a readability analysis. (2013), Lo et al. (2017) and Lim et al. (2018). The
As per Securities and Exchange Board of India (Listing interpretation of the Fog Index is presented in Table 1. We
Obligations and Disclosure Requirements) Regulation, use Lingua Fathom software to calculate the readability
2015,6 the annual report shall contain MD&A as a separate scores.
item or part of Directors’ Report. There are many studies We used Flesch Reading Ease index (FREI) as an
on the readability of different items of corporate annual alternative measure of readability. The interpretation of
reports in association with earnings performance based on FREI is presented in Table 2. FREI is calculated as follows:
developed markets. However, there is no literature on the
Flesch Reading Ease
analysis of readability of MD&A in association with
earnings performance of banking companies, specifically = 206.8 - (1.015 # words per sentence)(2)
in a developing economy. - (84.6 # syllables per word)
Therefore, we propose to test the following two The annual reports of the sample banking companies
hypotheses: are downloaded from the websites of sample firms in PDF
24 IIM Kozhikode Society & Management Review 10(1)

Table 1.  Fog Index Description and Interpretation credit-related services, financial leasing services, trade
finance related services, payment and money transmission
Fog Index Reading Ease
services, fund management services, financial consultancy
≥18 Unreadable
and advisory services, underwriting services, clearing and
14–18 Difficult
settlement services, securities trading services and other
12–14 Ideal
financial services. In India, NII ratio is 28.973 per cent as
10–12 Acceptable
per World Bank database.7 It is a major source of income
8–10 Childish
for the improvement of overall performance of banking
Source: Li (2008, p. 225).
companies. Pennathur, Subrahmanyam and Vishwasrao
(2012) found that fee-based income significantly reduces
Table 2.  Flesch Reading Ease Index Scores Description (FRES) risk in profitability and default risk of Indian banking
and Interpretation companies. Thus, NII appears to benefit the banking
companies. Based on incomplete revelation hypothesis, we
FRES General reading ease scale
predict that banking companies with lower (higher) NII
Below 30 Very difficult
ratio may make their MD&A difficult (easy) to read.
30–50 Difficult
Therefore, we predict a negative correlation between NII
50–60 Fairly difficult
ratio and Fog Index of readability.
60–70 Standard
70- 80 Fairly easy Non-Performing Loan Ratio: Non-performing loan (NPL)
80–90 Easy ratio is the ratio of gross non-performing loans to total gross
90–100 Very easy loans. NPL is one of the biggest challenges of the bank-
Source: Flesch (1948). ing sector in India and other emerging countries. Chief
Economic Advisor Arvind Subramanian said, ‘the worsen-
ing problem of bad loans is India’s primary macroeconomic
format and MD&A from each annual report has been con- challenge’.8 Bad loans are deducted from the total income
verted into a word document. The word document of each of banking companies to arrive at the profit. Hence, increase
sample firm-year has been then edited to remove all tables, in bad loans deteriorates the performance. Increase in bad
figures, pictures, headers, footers, list of contents and loans and decrease in performance may make managers
abbreviations table from the MD&A before using the text indulge in writing the MD&A in a difficult language to
for computation of readability score. The financial data of avoid the attention of the reader. Therefore, we predict a
the sample was downloaded from the Prowess database of positive correlation between NPL and Fog Index of the
Centre for Monitoring Indian Economy (CMIE). MD&A.

Age: Age is defined as the number of years a sample com-

Net Interest Margin as Performance Indicator pany is operating from its incorporation. Older firms may
present different MD&A because there is less information
We used NIM as a performance variable. NIM is the asymmetry and less information uncertainty. If investors
difference between interest income and interest expense are aware of the performance and financial condition of the
as a percentage of the total assets in the beginning of the banks from the available historical information, the MD&A
year. Several studies in India and abroad use NIM as a of older banks will be easier to read. Therefore, we predict
performance variable in banking literature (Boubakri, that the older (younger) the bank, the easier (harder) the
Cosset, Fischer, & Guedhami, 2005; Boyacioglu, Kara, & MD&A would be to read.
Baykan, 2009; Chantapong, 2005; Drakos, 2003; Kao &
Liu, 2004; Saksonova, 2014; Saunders & Schumacher, Length: The length of MD&A is measured in the number
2000; Sensarma & Ghosh, 2004; Wong, 1997). of words it contains. According to Loughran and McDonald
(2014), when a small sample of partners of major account-
ing firms are polled and asked how they would legally
attempt to obscure some information whose disclosure is
Other variables
required, the partners immediately identified the strategy
Non-Interest Income Ratio: It is the ratio of non-interest of burying the awkward revelation in an overwhelming
income (NII) to the total income of a banking company. amount of uninformative text and data. Loughran and
NII comes primarily through collection of fees for various McDonald (2014) find that documents which are harder to
services such as deposit account management services, read are bigger and the documents which are easier to read
Jayasree and Shette 25

are smaller in size. Therefore, we predict that lengthier maximum and minimum of 7.95 and 0.015 per cent, respec-
MD&A would be difficult to read. tively. The Fog Index of the sample banks has a mean value
of 15.87 with a minimum value of 10.83 and a maximum
value of 20.69. The median of Fog is 15.73. The general
Panel Data Regression norm is if Fog Index is in the range of 14–18, the annual
reports are difficult to read, whereas a Fog Index above 18
We propose a panel data regression model to identify the would mean the annual reports are unreadable. An average
impact of the current-year performance of banking compa- of 15.87 would mean that the MD&A disclosed in the
nies on the readability of MD&A disclosed in their annual annual reports is difficult to read. The mean and median
reports. We have initially set return on assets, NIM, effi- Fog Index of MD&A of US companies is 18.23 and 17.28
ciency ratio, NII ratio, provisions and contingencies, respectively (Li, 2008). This indicates that the mean Fog
non-performing assets, capital adequacy, size, age and the Index of Indian banking companies is much lower as com-
number of words in the MD&A as explanatory variables pared to those of US companies. The mean Fog Index of
for Fog and FREI indices of readability. However, when annual reports of non-US companies based on a large sam-
we checked for variance inflation factors, we found multi- ple of 87,608 annual reports from 42 countries including
collinearity among the regressors. Therefore, certain vari- India and other emerging countries is 19.520 (Lang & Stice-
ables had to be dropped from the model. Finally, we took Lawrence, 2015). The Fog Index of the MD&A of Indian
NIM as a performance variable and NII ratio, NPL, age and
banking companies is lower and they are easier to read than
number of words as explanatory variables in our panel data
those of the US and non-US companies.
regression model.
The findings based on FREI, an alternative measure of
readability, are also similar to the findings of Fog Index.
Fog_indit = αi + β1NIMit + β2NII ratioit + β3NPLit +
We applied one sample t-test to examine whether the mean
β4Ageit + β5No_Wordsit + ɛit
of Fog Index is significantly lower than 18 and whether the
mean of FREI is significantly higher than 30. We observed
FREI_indit = αi + β1NIMit + β2NII ratioit + β3NPLit +
that the mean of Fog Index of Indian banks is significantly
β4Ageit + β5No_Wordsit + ɛit
less than 18 and the mean of FREI of Indian banks is
Further, to examine if the impact of NIM on Fog is different significantly greater than 30. Thus, the readability of the
for good performing banks and weak performing banks, we MD&A of our sample banks is difficult but they are not
used stepwise regression. For this purpose, we divided our unreadable. Therefore, we reject the first hypothesis (H1).
sample based on the median value of NIM. Banking firms The possible reasons for a lower Fog Index and higher
with NIM above the median are considered good and NIM FREI of banking companies could be due to the compre-
below the median are considered as weak performing firms. hensive legal system. In India, banks are regulated by
the Reserve Bank Act, 1934, Banking Regulation Act
1949, Banking Companies (Acquisition and Transfer of
Undertakings) Act of 1970 and 1980, Prevention of Money
Results and Discussion Laundering Act, 2002, Securitization and Reconstruction
of Financial Assets and Enforcement of Security Interest
Readability of MD&A of Banking Companies
Act, 2002 in addition to the new Companies Act, 2013 and
We start our analysis with descriptive statistics, which are the guidelines of the Securities Exchange Board of India.
shown in Table 3. The mean of NIM is 2.44 per cent with a Appointment of statutory auditors for banking companies

Table 3.  Descriptive Statistics

Variable Number of observations Mean Standard deviation Minimum Maximum Median

NIM 156 2.44 0.836 0.015 7.95 2.27
NII ratio 157 11.97 9.63 0.215 85.63 10.19
NPL 137 6.03 4.77 0.31 22.39 4.81
Age 160 73.97 37.88 0 152 83.50
Fog Index 142 15.87*** 1.86 10.83 20.69 15.73
FREI 142 37.26*** 6.77 10.83 52.28 38.22
Number of words 142 10896.38 63.7.95 693 35176 10440.50
Source: Authors’ calculation.
Note: *** Significant at one per cent. For testing the significance of mean of Fog Index and FREI for less than 18 and more than 30.
26 IIM Kozhikode Society & Management Review 10(1)

and the requirement of Long Form Audit Report (LFAR) which results in a difficult to read MD&A in their annual
by the Reserve Bank of India could be other reasons reports. In contrast to this, better performing banks disclose
for better readability of the MD&A of Indian banking a MD&A that is easier to read. This significant negative
companies. relation between performance and Fog Index is in line with
the incomplete revelation hypothesis (Bloomfield, 2002).
We have also used FREI as an alternative measure to check
Impact of Operating Performance on Readability the robustness of our findings.
of MD&A As predicted, we found negative impact of the age of
banking companies on Fog Index and it is significant at one
The correlation matrix as presented in Table 4 shows that per cent. This means the MD&A of new firms are harder to
NIM, NPL and age have a negative correlation with Fog read and less transparent. In other words, older banking
Index. The correlation matrix also shows NII and the companies present their MD&A in easier to read language
number of words in the MD&A have a positive correlation and a more transparent format. In India, most of the older
with Fog Index. This conveys that higher NII and higher banks are public sector banks and most of the younger
number of words result in a harder to read MD&A. banks are private sector banks. Therefore, we could say
We ran pooled regression to analyse the impact of the that MD&A of public sector banks is presented in easy
performance and other explanatory variables on Fog Index; language. It may be due to the transparency of public sector
the results are presented in Table 5. The p-values of the banks, policies of the Reserve Bank of India and the
regression show that the model is a good fit. The results compulsory audit conducted by the Comptroller and
show that NIM has a negative impact on Fog Index of the Auditor General of India, which are applicable only to the
MD&A of Indian banking companies, and it is significant public sector banks.
at one per cent. It means that the MD&A of underperforming Contrary to our prediction, we found a significant posi-
banking companies are harder to read, and the MD&A of tive impact of NII ratio on Fog Index at one per cent. This
better performing banking companies are easier to read. indicates that the MD&A of banking companies with higher
These results, based on statistics pertaining to Indian NII is harder to read and the MD&A of banking companies
banking companies, are consistent with the findings of with lower NII is easier to read. Generally, the NII ratio
existing studies (Lang & Stice-Lawrence, 2015; Li, 2008). would have many elements, hence, the readability would be
This also implies that when the performance of banks is difficult when compared to NIM, which includes interest
low, banks indulge in obfuscation of data and its analysis, received on loans and interest paid on deposits. As more

Table 4.  Correlation of the Variables

Variable Fog Index NIM NII Ratio NPL Age Number of Words
Fog Index 1.00
NIM −0.05 1.00
NII ratio 0.33*** 0.17** 1.00
NPL −0.30*** −0.47*** −0.15 1.00
Age −0.28*** −0.35*** −0.05 0.28*** 1.00
Number of words 0.32*** 0.15* 0.08 −0.51 −0.09 1.00
Source: Authors’ calculation.

Table 5.  Fog Index and Firm Performance

Fog Index Coef Std. Err t P>t

NIM −0.767*** 0.275 −2.79 0.006
NII ratio 0.056*** 0.014 3.92 0.00
NPL −0.116*** 0.035 −3.33 0.001
Age −0.015*** 0.004 −3.29 0.001
Number of words 0.00** 0.00 2.29 0.024
Cons 16.54*** 1.012 18.10 0.00
Source: Authors’ calculation.
Note: ***, **, and * represent significance level at the 1%, 5%, and 10% levels, respectively.
Number of observations: 123; F = 11.21; Prob > F = 0.00; R-squared = 0.3238; Adj. R-square = 0.2950
Jayasree and Shette 27

fee-based services are offered, the technical terminology the groups and year effects by using fixed effects model.
used in the annual reports may make them more difficult to We also ran random effects model, which assumes that the
read. Contrary to our prediction, we find a significant nega- error term is not correlated with the regressors and that
tive impact of NPL on Fog Index. It means that the MD&A the variation across entities is random. Later, we apply
of banking companies with lower NPL are harder to read Hausman test to understand if the fixed effects model is or
and vice versa. We could not find any impact of number of the random effects model is better; our test results indicate
words of MD&A on Fog Index as the coefficient is zero. that the fixed effects model is better. The results of the
Although the regression results are significant, these fixed effects model are presented in Table 7. These results
results could be affected due to differences in composition are also similar to the pooled regression presented in Table
of sample consisting of public sector and private sector 5. In this model, too, we find NIM to be significant at one
banks. The Wilcoxon rank sum test showed that there is a per cent. We further use FREI as an alternative measure to
significant difference in the medians of the Fog Index of check the robustness of our findings. For FREI also our
private and public sector. The results are presented in Table Hausman test results suggest that fixed effects model is
6. In the pooled regression, only the time variable is better. As presented in Table 8, the findings based on FREI
invariant, hence, we propose to control for differences in are also similar to Fog Index.

Table 6.  Wilcoxon Rank Sum Test

For Fog Index For Flesch Index

Company Type No. of Observations Rank Sum Expected No. of Observations Rank Sum Expected
Private 58 5257 4147 58 3080 4147
Public 84 4896 6006 84 7073 6006
Combined 142 10153 10153 142 10153 10153
Source: Authors’ calculation.
Significant at 1% z = −4.428; there is a significant difference in median of the Fog Index for public and private sector banks

Table 7.  Fog Index and Firm Performance Fixed Effects

Fog Index Coef Std. Err t P>z

NIM −0.986*** 0.248 −3.96 0.000
NII ratio 0.039*** 0.013 2.96 0.004
NPL −0.059 0.043 −1.37 0.174
Age −0.002 0.004 −0.56 0.578
Number of words 0.0001*** 0.00 4.37 0.000
Cons 17.031*** 0.980 17.36 0.000
Year fixed effects Yes
Source: Authors’ calculation.
Note: ***, **, and * represent significance level at the 1%, 5%, and 10% levels, respectively.
Number of observations = 123; R2 within groups = 0.3420; Number of Groups = 2; Between = 1.000; Overall = 0.1972; Prob > F = 0.000; Observations
per group min = 52; Average = 61.5; Max = 71

Table 8.  Flesch Readability Ease Index and Firm Performance Fixed Effects

Flesch Readability Ease Index Coef Std. Err t P>z

NIM 3.347*** 0.978 3.42 0.001
NII ratio –0.128*** 0.518 –2.48 0.014
NPL 0.314 0.171 1.84 0.680
Age 0.050* 0.186 2.72 0.080
Number of words –0.002*** 0.00 –2.260 0.026
Cons 28.425*** 3.590 7.37 0.000
Year fixed effect Yes
Source: Authors’ calculation.
Note: ***, **, and * represent significance level at the 1%, 5%, and 10% levels, respectively.
Number of observations = 123; R2 within groups = 0.2677; Number of groups = 2;
Between = 1.000; Overall = 0.2945; Prob > F = 0.000; Observations per group min = 52, Average = 61.5, Max = 71
28 IIM Kozhikode Society & Management Review 10(1)

Table 9.  Step Wise Regression of Good and Weak Performing Banks

Panel A: Good Performing Banks

Model 1 Model 2 Model 3 Model 4 Model 5
Variables Fog Fog Fog Fog Fog
NIM −0.937***
NII ratio 0.057*** 0.039*** 0.04*** 0.37*** 0.38***
NPL −0.271*** −0.22*** −0.238*** −0.259***
Age −0.008 −0.002 −0.006
Number of Words 0.000* .000***
Cons 15.12*** 16.60*** 16.92*** 15.98*** 19.00***
No. of observations 67 62 62 62 62
R-square 0.175 0.35 0.37 0.41 0.47
F 13.83 16.1 11.53 9.84 9.72
Panel B: Weak Performing Banks
Model 1 Model 2 Model 3 Model 4 Model 5
Variables Fog Fog Fog Fog Fog
NIM −1.42
NII ratio 0.149** 0.19** 0.19*** 0.21*** 0.21***
NPL −0.10** −0.11** − 0.09** −0.11**
Age −0.024*** − 0.024*** −0.02**
Number of words 0.00** 0.00*
Cons 14.27*** 14.55*** 16.54*** 15.59*** 18.09***
No. of observations 60 60 60 60 60
R-square 0.06 0.12 0.23 0.29 0.32
F 4.86 4.21 5.87 5.73 5.01
Source: Authors’ calculation.
Note: ***, **, and * represent significance level at the 1%, 5%, and 10% levels, respectively.

Further, to show the explanatory power of NIM for companies from the USA and 42 other countries, Indian
good and weak performing banks, we carried out stepwise banking firms’ MD&As are found difficult to read but not
regression. The results are in Table 9. We observe from the unreadable. Our results show that a firm’s performance
results that in the case of good performing firms, NIM influences the readability of its MD&A. The results are sta-
shows a negative significant impact on Fog Index and we tistically significant, showing that firm performance would
see an increase in the R-square from 0.41 to 0.47. In the have a negative impact on Fog Index and a positive impact
case of weak performing firms, the effect of NIM is also on the FREI of readability. Based on our results, we affirm
negative and we observe an increase in the R-square from that firms with poor performance would structure their
0.29 to 0.32, showing that NIM can explain Fog index. We annual reports to veil adverse information in unfavourable
also observed that the beta of NIM of weak performing situations. The results support IRH and the opaque nature
banks is higher than good performing banks. Hence, we of banking companies. Application of readability index in
conclude that NIM has a negative impact on Fog Index and case of banking companies of an emerging economy in
the impact of NIM on weaker firms is higher. association with their performance is the contribution of
this paper to the existing literature.

Conclusion Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect
This paper is the first study of the readability of MD&A to the research, authorship and/or publication of this article.
and performance of the banking firms and determines that
the performance variable influences the readability indices. Funding
We find that Indian banking firms’ MD&A are difficult to The authors received no financial support for the research, author-
read. However, when we compare them with the MD&A of ship, and/or publication of this article.
Jayasree and Shette 29

Notes Harris, T. L., & Hodges, R. E. (1995). The literacy dictionary:

The vocabulary of reading and writing. Delaware: IRA
1. https://www.thehindubusinessline.com/opinion/indias-npas-
(International Reading Association).
and-the-global-scenario/article24145872.ece (accessed 13 July
Howe, J. S., and Haggard, K. S. (2012). Are banks opaque?
2019)
International Review of Accounting, Banking and Finance,
2. https://rbidocs.rbi.org.in/rdocs/Publications/PDFs/0RTP
4(1), 51–72.
2018_FE9E97E7AF7024A4B94321734CD76DD4F.PDF
(accessed 13 July 2019) Hrasky, S., Mason, C., & Wills, D. (2009). The textual complexity
3. https://www.sec.gov/rules/interp/33–6835.htm of annual report narratives: A comparison of high- and low-
4. https://www.nseindia.com/content/equities/eq_list_ performance companies. New Zealand Journal of Applied
regulations.zip Business Research, 7(2), 31–45.
5. http://databank.worldbank.org/data/reports.aspx?source= Jones, M. J., & Shoemaker, P. A. (1994). Accounting narratives:
global-financial-development&preview=on# A review of empirical studies of content and readability.
6. https://www.moneycontrol.com/news/business/economy/ Journal of Accounting Literature, 13, 142–184.
npa-issue-is-indias-no-1-macroeconomic-challenge-arvind- Kao, C., & Liu, S. T. (2004). Predicting bank performance with
subramanian-2248985.html financial forecasts: A case of Taiwan commercial banks.
Journal of Banking & Finance, 28(10), 2353–2368.
References Knutson, P. H. (1993). Financial reporting in the 1990s and
beyond. Charlottesville, VA: Association for Investment
Adelberg, A. H. (1979). Narrative disclosures contained in Management and Research.
financial reports: Means of communication or manipulation? Kumari, P., & Pattanayak, J. K. (2017). Linking earnings
Accounting and Business Research, 9(35), 179–190. management practices and corporate governance system
Beatty, A., Chamberlain, S. L., & Magliolo, J. (1995). Managing with the firms’ financial performance: A study of Indian
financial reports of commercial banks: The influence of commercial banks. Journal of Financial Crime, 24(2), 223–
taxes, regulatory capital, and earnings. Journal of Accounting 241.
Research, 33(2), 231–261. Kumari, P., & Pattanayak, J. K. (2015). Earnings management
Biddle, G. C., Hilary, G., & Verdi, R. S. (2009). How does and firm performance: an insight into Indian commercial
financial reporting quality relate to investment efficiency?
banks. Journal of Scientific Research and Development,
Journal of Accounting and Economics, 48(2–3), 112–131.
2(11), 76–84.
Bloomfield, R. J. (2002). The ‘incomplete revelation hypothesis’
Lang, M., & Stice-Lawrence, L. (2015). Textual analysis and
and financial reporting. Accounting Horizons, 16(3), 233–
international financial reporting: Large sample evidence.
243.
Journal of Accounting and Economics, 60(2–3), 110–135.
Boubakri, N., Cosset, J. C., Fischer, K., & Guedhami, O. (2005).
Lawrence, A. (2013). Individual investors and financial disclosure.
Privatization and bank performance in developing countries.
Journal of Accounting & Economics, 56, 130–147.
Journal of Banking & Finance, 29(8–9), 2015–2041.
Lehavy, R., Li, F., & Merkley, K. (2011). The effect of annual
Boyacioglu, M. A., Kara, Y., & Baykan, Ö. K. (2009). Predicting
report readability on analyst following and the properties
bank financial failures using neural networks, support vector
of their earnings forecasts. The Accounting Review, 86(3),
machines and multivariate statistical methods: A comparative
analysis in the sample of savings deposit insurance fund 1087–1115.
(SDIF) transferred banks in Turkey. Expert Systems with Lewis, N. R., Parker, L. D., Pound, G. D., & Sutcliffe, P.
Applications, 36(2), 3355–3366. (1986). Accounting report readability: The use of readabi-
Chantapong, S. (2005). Comparative study of domestic and lity techniques. Accounting and Business Research, 16(63),
foreign bank performance in Thailand: The regression 199–213.
analysis. Economic Change and Restructuring, 38(1), 63–83. Li, F. (2008). Annual report readability, current earnings, and
Chipalkatti, N. (2005). Do investors reward bank disclosure earnings persistence. Journal of Accounting and Economics,
transparency? Evidence from India. Journal of International 45(2–3), 221–247.
Accounting Research, 4(2), 25–52. Lim, E. K., Chalmers, K., & Hanlon, D. (2018). The influence
Collins, J. H., Shackelford, D. A., & Wahlen, J. M. (1995). Bank of business strategy on annual report readability. Journal of
differences in the coordination of regulatory capital, earnings, Accounting and Public Policy, 37(1), 65–81.
and taxes. Journal of Accounting Research, 33(2), 263–291. Lo, K., Ramos, F., & Rogo, R. (2017). Earnings management
Dang, T. V., Gorton, G., Holmström, B., & Ordonez, G. (2017). and annual report readability. Journal of Accounting and
Banks as secret keepers. American Economic Review, 107(4), Economics, 63(1), 1–25.
1005–29. Loughran, T. & McDonald, B. (2014). Measuring the readability
Dolphin, R., & Wagley, R. A. (1977). Reading annual report. of financial disclosure. The Journal of Finance, 69(4), 1643–
Financial Executive, 45(6), 20–22. 1671.
Dougal, C., Engelberg, J., Garcia, D., & Parsons, C. A. (2012). Miller, B. P. (2010). The effects of reporting complexity on small
Journalists and the stock market. The Review of Financial and large investor trading. The Accounting Review, 85(6),
Studies, 25(3), 639–679. 2107–2143.
Drakos, K. (2003). Assessing the success of reform in transition Pashalian, S., & Crissy, W. J. (1950). How readable are corporate
banking 10 years later: an interest margins analysis. Journal annual reports? Journal of Applied Psychology, 34(4), 244–
of Policy Modelling, 25(3), 309–317. 248.
Flesch, R. (1948). A new readability yardstick. Journal of Applied Pennathur, A. K., Subrahmanyam, V., & Vishwasrao, S. (2012).
Psychology, 32, 221–233. Income diversification and risk: Does ownership matter?
30 IIM Kozhikode Society & Management Review 10(1)

An empirical examination of Indian banks. Journal of Sensarma, R., & Ghosh, S. (2004). Net interest margin: does
Banking & Finance, 36(8), 2203–2215. ownership matter? Vikalpa, 29(1), 41–48.
Rogers, R. K., & Grant, J. (1997). Content analysis of information Smith, J. E., & Smith, N. P. (1971). Readability: A measure of
cited in reports of sell-side financial analysts. Journal of the performance of the communication function of financial
Financial Statement Analysis, 3, 17–31. reporting. The Accounting Review, 46(3), 552–561.
Saksonova, S. (2014). The role of net interest margin in Soper, F. J., & Dolphin, R. (1964). Readability and corporate
improving banks’ asset structure and assessing the stability annual report. The Accounting Review, 39(2), 358–362.
and efficiency of their operations. Procedia – Social and
Subramanian, R., Insley, R. G., & Blackwell, R. D. (1993).
Behavioural Sciences, 150, 132–141.
Performance and readability: A comparison of annual reports
Saunders, A., & Schumacher, L. (2000). The determinants of
bank interest rate margins: an international study. Journal of of profitable and unprofitable corporations. The Journal of
International Money and Finance, 19(6), 813–832. Business Communication, 30(1), 49–61.
Securities and Exchange Commission, Office of Investor Tavcar, L. R. (1998). Make MD&A more readable. The CPA
Education, & Assistance (SEC). (1998). A plain English Journal, 68(1), 10–11.
Handbook: How to create clear SEC disclosure documents. Wong, K. P. (1997). On the determinants of bank interest margins
Washington, DC: The Office of Investor Education and under credit and interest rate risks. Journal of Banking &
Assistance, U.S. Securities and Exchange Commission, N.W. Finance, 21(2), 251–271.