Is Better Banking Performance Associated With Financial Inclusion and Mandated CSR Expenditure in A Developing Country?

Accounting & Finance
Is better banking performance associated with financial

inclusion and mandated CSR expenditure in a
developing country?
Asit Bhattacharyyaa , Sue Wrightb , Md Lutfur Rahmana

a
Accounting and Finance Discipline, University of Newcastle, Callaghan,
b
UTS Business School, UTS, Sydney, NSW, Australia
Abstract
Motivated by legislation mandating CSR expenditure to improve social

equality and economic development in India, we examine the association of
CSR expenditure and financial inclusion with the performance of banking firms
in the period after introduction of the legislation. We study whether mandated
CSR expenditure and/or financial inclusion measures are associated with better
financial performance, using both accounting and stock market measures of
performance, for Indian banks during 2015–2017. Our results demonstrate that
level of CSR expenditure and degree of financial inclusion is not associated
with banks’ financial performance when performance is measured in account-
ing terms. However, a significant negative association is found when perfor-
mance is measured by stock market return. These results suggest that the
current design of the legislation is unlikely to achieve its purpose. This is the
first study to present clear evidence on the associations of mandatory CSR
spending and firm-level financial inclusion with accounting-based and market-
based bank performance.
Key words: Mandated CSR spending; Financial inclusion; Performance; Bank
JEL classification: M14, M40, M41, M48
doi: 10.1111/acfi.12560
The authors acknowledge the financial support of AFAANZ research grant, 2017
(No. G1701094) of $5000.
Please address correspondence to Sue Wright via email: sue.wright@uts.edu.au
© 2019 Accounting and Finance Association of Australia and New Zealand

2 A. Bhattacharyya et al./Accounting & Finance
1. Introduction
In emerging countries, poverty, poor working conditions and abuse of

human rights are critical issues. In 2013, the Indian government introduced a
novel approach to improving social conditions by mandating corporate social
responsibility (CSR) expenditure by companies over a certain threshold size.
Businesses in India are now expected to contribute to sustainable economic
development and the establishment of an equitable society. However, the
legislation provides firms with the option of avoiding the mandated CSR
expenditure by disclosing in their annual reports why it has not been spent.
Therefore, the government’s approach will only be successful if firms choose to
undertake CSR activities, a choice that will largely depend on how those
activities impact their financial performance.
For banks in India, a strategic growth initiative that also achieves the desired
social goals of reduced inequality and poverty is greater financial inclusion,
which is defined as a course of action that allows access, availability and usage
of formal financial services to all individuals and firms in an economy (World
Bank, 2014). Focusing on financial inclusion in order to comply with the
mandated CSR legislation may be a win-win strategy for Indian banks, because
an increase in financial inclusion activities could lead to an improvement in
their financial performance via improved reputation and improved legitimacy.
Despite its growing importance, however, research on financial inclusion has
been limited largely to measuring and investigating the country-level institu-
tional factors that influence the level of financial inclusion (Burgess and Pande,
2005; Beck et al., 2007; Sarma and Pais, 2011). The literature typically ignores
the context of firm-level research, although firm-level engagement can play a
pioneering role in augmenting country-level financial inclusion (Bose et al.,
2016). Further, prior studies (Kempson et al., 2004; Sarma and Pais, 2011; Bose
et al., 2016) argue that because financial inclusion is multi-dimensional,
country-level and institutional indicators have limited ability to fully measure
overall financial inclusion. By examining the association between the firm-level
financial inclusion factors of banks and their performance, in the context of
legislation designed to increase CSR expenditure by banks, our study addresses
this methodological problem. In this study, we examine the association of CSR
spending mandated by the legislation and financial inclusion activities with
banking performance in the period after its introduction. Our findings will
inform legislators and regulators in emerging countries about the link between
mandatory CSR expenditure and bank performance as they argue for similar
legislation designed to involve the corporate sector in sustainable economic and
social development.
Our study is able to examine the effects of both mandatory CSR expenditure
and bank-level financial inclusion on bank performance because of the
availability of relevant data in India following the introduction of the
mandatory CSR expenditure regulation. The Indian government has embraced

A. Bhattacharyya et al./Accounting & Finance 3
the global vision of the United Nation’s Millennium Development Goals

(MDGs) (United Nations Conference on Trade and Development, 2014), and
is the first national government to mandate CSR expenditure and financial
inclusion activities, through section 135 of the Indian Companies Act, 2013.
This legislation requires that in any financial year, a company with an annual
turnover of at least $US180 million, or a net worth of $US90 million, or a net
profit in excess of $US900,000, must spend at least 2 percent of its net profit
(averaged over the previous 3 years) on prescribed CSR activities.1
This study addresses the question: are mandated CSR expenditure and
financial inclusion associated with better bank performance in the context of a
developing country? Using two measures of financial performance for Indian
banks listed on two major stock exchanges in India, the Indian National Stock
Exchange and the Bombay Stock Exchange, we examine their associations with
mandated CSR expenditure and with (firm-level) financial inclusion. Our key
findings are that the level of CSR expenditure and the degree of financial
inclusion are not associated with banks’ performance when measured in
accounting terms (return on assets). However, a significant negative association
is present when performance is measured as stock market return, and when the
performance–financial inclusion relation is moderated by the mandated CSR
expenditure. The results indicate that CSR expenditure and financial inclusion
are not associated with higher financial performance by banks, which may be
attributed to investors’ negative perception towards mandatory CSR expen-
diture and financial inclusion.
The study is located in India because it is the only country that has legislated
CSR spending, and because the impact of financial inclusion is expected to be
more apparent in emerging and developing economies (World Bank, 2014;
Park and Mercado, 2015). We study the banking industry because of the role of
financial inclusion in the relation between CSR spending and performance in
that industry, and because the lack of available data in the past has excluded it
from previous capital market studies.
The primary contribution of this research is the presentation of clear evidence
on the association of mandatory CSR spending and firm-level financial
inclusion with accounting-based and market-based bank performance. The
future actions of regulators, banks and accountants and auditors will be
impacted by the findings of this study. Our findings contribute to the wider
policy debate on voluntary versus mandatory CSR expenditure. Regulators of
other emerging countries may be able to use our findings to better design
mandated CSR spending legislation or regulations to encourage financial
inclusion activities. The results of our study will be of interest to all country-
level regulators and various international organisations (e.g., the International
Monetary Fund (IMF), the Alliance for Financial Inclusion (AFI) and the
1
Firms have the alternative of disclosing in their annual reports why the stipulated 2
percent has not been spent on the prescribed schedule VII CSR activities.

World Bank). We demonstrate that performance improvements and gains in

market share for banks that are considering investing in financial inclusion
activities and CSR activities are unlikely. Our findings will also be useful to
accountants and auditors who need to assess the adequacy or otherwise of a
company’s CSR practices (International Auditing and Assurance Standards
Board, 2007; Association of Chartered Certified Accountants, 2008).
Our study extends the existing literature in four important ways. First, we
explore the association between CSR expenditure and firm performance in a
mandatory setting, whereas prior studies have been limited to voluntary
settings. Second, we are the first to examine this relation using actual CSR
spending rather than a CSR proxy score for voluntary CSR spending (see
Servaes and Tamayo, 2013; Lys et al., 2015), and the first to address the
association of firm-level financial inclusion activities and banks’ performance.
Previous studies have used voluntary CSR expenditure, and have assumed that,
on average, banks’ financial inclusion, CSR expenditure and activities are at
their optimum level (Himmelberg et al., 1999; Servaes and Tamayo, 2013).
Third, we address methodological issues such as model misspecification and the
omitted variable bias, which may affect the existing empirical evidence on
whether firm-level financial inclusion factors, CSR expenditure and disclosures
improve a bank’s financial performance (Margolis et al., 2009; Servaes and
Tamayo, 2013). Fourth, we control for a bank’s ownership structure (see Shane
and Spicer, 1983; McWilliams and Siegel, 2001), a factor that has been largely
ignored by previous studies.
The rest of the paper advances as follows. Section 2 presents the theory,
literature and hypotheses. Methodological aspects are described in Section 3.
Section 4 outlines the data and descriptive statistics. Empirical results and
analyses are presented in Section 5. This section also includes robustness
checks. Section 6 concludes by providing a summary of the paper.
2. Theory, literature and hypotheses
According to Freeman’s stakeholder theory, a firm should consider the

interests of all stakeholders who significantly affect or are affected by its
operations. Therefore, stakeholders’ social, environmental and ethical perfor-
mances can encourage overall CSR activities (Baron, 2001; McWilliams and
Siegel, 2001). As strategically motivated CSR activities can be cost effective, the
management literature describes stakeholder theory as ‘doing well by doing
good’. CSR activities are expected to be profitable because they reduce
transaction costs related to dealings with stakeholders, and the net benefits are
channelled to the business. The theory of the firm, progressed by Coase (1937)
and Jensen and Meckling (1976), interprets a business as a network of explicit
and implicit agreements between shareholders and other stakeholders. Critical
resources of the firm have been supplied by each group of stakeholders due to

these contracts. Through CSR activities, firms intend to create and maintain a
standing for their obligations related to implicit contracts.
Supporting this theoretical concept, there is a stream of literature demon-
strating how CSR increases firms’ performance and stakeholder value. The
studies of Servaes and Tamayo (2013) and Fombrun (2005) in general find a
positive relation between CSR and firms’ stakeholder value. Barnett (2007) also
advocates that CSR spending has an effect on firm value that is determined by
the ability of CSR to influence the stakeholders of the firms. Further, several
studies document an explicit linkage between CSR expenditure and financial
performance. For example, higher levels of CSR initiatives and spending on
strategic CSR activities help firms to (i) attract and retain high quality skilled
employees (Greening and Turban, 2000), (ii) improve product and service
promotion efficiency (Fombrun, 2005), providing them with greater rights to
use scarce resources (Cochran and Wood, 1984), (iii) increase demand for
goods and services (Navarro, 1998) and (iv) create goodwill that can restore
performance in the event of a violation of a legal requirement by reducing
penalties (Godfrey, 2005). In contrast, a negative relation between increased
CSR undertakings or ratings with a firm’s return on assets (ROA) and stock
return is reported by Di Giuli and Kostovetsky (2014).
In addition, the impact of CSR spending may extend beyond firms’ financial
performance measures (Lys et al., 2015). For example, higher institutional
investor ownership, broader analyst coverage and reductions in firms’ cost of
capital may be achieved by voluntary CSR activities (Dhaliwal et al., 2011).
Firms with higher consumer responsiveness document a positive association
between spending on CSR activities and their financial performance although
firms with lower consumer responsiveness show no association (Servaes and
Tamayo, 2013). A firm’s legitimacy and reputation may be maintained through
CSR expenditure (Peloza, 2006) and possible legislative action or negative
regulatory views may be moderated (Hillman and Keim, 2001).
In contrast, Lys et al. (2015) assert that CSR activity may be undertaken by a
firm because it anticipates sustainable positive financial performance, rather
than the CSR activity itself causing this positive performance. The methods
used in prior studies do not allow for conclusions to be drawn about the direct
effect of CSR expenditure on firms’ current financial performance. A further
limitation of this literature is its mixed findings. Reviewing the literature on the
association between CSR activities and firms’ financial performance, Margolis
et al. (2009) note that a few studies report significant positive relations between
CSR activities and firm performance, while others report negative associations.
The context of this study is mandatory CSR expenditure in India, which may
have a detrimental effect on shareholders’ wealth rather than a beneficial one,
for a number of reasons. Consistent with Friedman (1970), if firms choose not
to spend on CSR, then mandating CSR activities and spending will reduce their
value. Furthermore, by directing CSR spending in a particular way, business
choices over the type of CSR activity they undertake are restricted. Karnani

(2003) argues that section 135 basically imposes an additional 2 percent tax on
companies, which may make India less attractive to foreign investors and large
multinational companies. If they take the option of disclosing why the
stipulated 2 percent has not been spent as directed, firms are likely to
implement a number of costly face-saving strategies in order to avoid any
potential adverse legal consequences. Associated compliance costs also are an
integral part of a mandated CSR policy.
Although mandated CSR expenditure regulations may impose certain costs on
shareholders, there may also be some benefits. A firm’s CSR policies become
more formalised and visible under a mandatory CSR expenditure rule, providing
a stronger signal to shareholders. Strategic CSR reporting and spending involves
more transparent CSR and financial disclosures, which have the potential to
improve a company’s reputation both locally and nationally. The literature
reports enhanced firm values resulting from the 1999 Korean governance reforms
(Black and Kim, 2012), significant increases in CSR performance after the
introduction of section 299(1)(f) of the Australian Corporations Act 2001 (Frost,
2007), and a positive stock market reaction to Indian governance reforms
(Clause 49, adopted in early 2000) (Black and Khanna, 2007).
In summary, the majority of the literature reports that CSR can have a
positive effect on firm performance. Thus we hypothesise for banks:
H1: In a mandatory setting, CSR expenditure is positively associated with financial
and market performance.
In developing countries, only a small proportion of the population can afford
banking products and services, thus excluding the majority who do not have
access to them. However, the unbanked population has the same entitlement to
banking services as those who can afford them, and banking services and
products could have a significant impact on their lives. Using the argument of
Brown and Deegan (1998) that there is an implicit contract between the firm
and those who are affected by its operations, we recognise that there is a social
contract between banks and the unbanked population. Hawkins (2006) and
Jamali (2008) find that the success of catering for the needs of primary
stakeholders depends on whether firms are fulfilling the needs of other
stakeholders. Following stakeholder theory, Freeman (2010) and Jo et al.
(2015) point out that a firm can enhance its reputation and performance
positively by attending to the implicit rights of all stakeholders. Financial
inclusion activities widen the financial system’s resource base by engaging with
all of the country’s populace, thus playing a pivotal role in the process of
economic development.
The literature uses well-established country-level supply side measures of
financial inclusion, such as branch density, the number of ATMs, market
penetration of mobile phones, or the share of households or firms with access
to financial services. In addition, international organisations regularly survey
financial inclusion (IMF Financial Access survey, Fidex database).

There is substantial evidence on the benefits of greater financial inclusion at

the country level. Beck et al. (2005) argue that in low-income countries,
financial inclusion provides a critical stimulus for competition, industrial
structure and performance through the distribution of credit across segments.
A more inclusive financial system diminishes informal avenues for credit, thus
improving the distribution of resources (Sarma and Pais, 2011). Recent
empirical evidence (Dupas and Robinson, 2009; Ardic et al., 2011) reports that
financial inclusion can substantially improve the underprivileged’s living
standards. Measuring access to finance by the density of branches, Abdmoulah
and Jelili (2013) explain the indirect relation between growth and financial
development. Cross-country studies show that by reducing individual income
inequality and other poverty indicators, financial development contributes to
economic development (Beck et al., 2007). In low-income countries, financial
inclusion is a critical factor needed to accelerate the growth of the poorest
segments of the population (Burgess and Pande, 2005). Based on this type of
evidence, policy makers and global development organisations place financial
inclusion at the centre of international development programs (Ardic et al.,
2011). For example, Burgess and Pande (2005) examine the consequence of
financial inclusion on rural poverty reduction. Beck et al. (2007) explore the
determinants of the financial sector’s outreach, and Sarma and Pais (2011)
examine the relation between the level of financial inclusion and country-
specific factors such as human development, literacy, income inequality and
infrastructure for connectivity.
This substantial body of research focuses on country-level measures and,
with the exception of Bose et al. (2016), does not investigate financial inclusion
from the firm perspective. In addition, these prior studies use a single indicator
to measure financial inclusion, which may not be able to adequately capture the
nature and extent of financial inclusion. For example, it is not enough to have a
bank account to be included in the financial system; banking services must be
used effectively (Kempson et al., 2004). Drawing on the positive association
between financial inclusion and various economic measures, and extending this
literature by using a firm-level measure that captures different aspects of
financial inclusion, we hypothesise the following in the context of Indian banks:
H2: Firm-level financial inclusion activities are positively associated with financial
and market performance.
Various stakeholders, including governments, expect that firms will engage in
financial inclusion activities to discharge their corporate social responsibility.
Firms’ social contributions signal that their corporate leaders are genuinely
engaging with their stakeholders, and their performance can be significantly
impacted by activities that satisfy their powerful stakeholders, including
government (Frooman, 1999; Wang and Qian, 2011). However, it is also
argued that ‘socio-political’ legitimacy has no explicit effect on the performance
implications of financial inclusion as neither key stakeholders nor the

government controls the resources that are vital for the firms’ sustained
operations and success (Pfeffer and Salancik, 1978). As a firm’s engagement
with financial inclusion activities and mandated CSR both fulfil the govern-
ment’s inclusive growth and poverty reduction goals and satisfy the demands of
all stakeholders, we argue that mandated CSR activity and spending will
moderate the association of financial inclusion with financial performance.
Therefore, we hypothesise that:
H3: Mandated CSR spending moderates the association of financial inclusion with
market performance.
The relations between the three hypotheses proposed in this study are shown
in Figure 1.
3. Method
3.1. Models
We examine whether the levels of financial inclusion and CSR expenditure

are contemporaneously related to bank performance. Three sets of regression
models are estimated whereby measures of bank performance are regressed on
three sets of explanatory variables: (i) CSR expenditure, (ii) financial inclusion
measures and (iii) both CSR expenditure and financial inclusion measures. In
order to examine if the levels of financial inclusion and CSR expenditure can
explain bank performance, we estimate the following models:
X
Performancet ¼ aa þ b1a CSRt þ li;a Control variablesi;t þ e ð1Þ
X
Performancet ¼ ab þ b2a FIt þ li;b Control variablesi;t þ e ð2Þ
CSR
Expenditure
H1
Financial H3 Firm
Inclusion H2
Performance
Figure 1 Summary of hypotheses tested.

X
Performancet ¼ ac þ b1b CSRt þ b2b FIt þ li;c Control variablesi;t þ e
ð3Þ
The above regression models are estimated using ordinary least squares
(OLS) as it is the most commonly used approach for examining a causal
inference in the case of a repeated cross-section sample (see, for example, Bose
et al., 2017; Chauvet and Jacolin, 2017; Alhaj-Ismail et al., 2019). The literature
also uses two-stage least squares (2SLS) to examine causal inference between
financial variables, particularly to address reverse causality and endogeneity
problems (Richardson et al., 2016; Ahmed and Ali, 2017; Ahmed et al., 2017).
Moreover, in order to address endogeneity, simultaneity and unobserved
heterogeneity, researchers also use generalised method of moment (GMM)
estimation in the corporate finance/governance literature (Sufian and Habibul-
lah, 2010; Ahmed et al., 2017; Safiullah and Shamsuddin, 2018). Since our
initial estimation shows that CSR expenditure or financial inclusion activities
have an insignificant (negative and significant) association with the accounting
performance (stock return) measure, we consider that our regression modelling
is unlikely to be characterised by reverse causality or endogeneity problems.
Although we do not address them in our baseline analysis,2 as a robustness
check, we estimate the above-mentioned regression models using both 2SLS
and GMM approaches.
Statistical inference with regard to CSR expenditure’s and financial inclusion’s
causal relation with firm performance is derived from both standard t-statistics
and heteroscedasticity-consistent robust t-statistics following the Huber–White
procedure. Since we find qualitatively similar results, only regression estimates
based on standard t-statistics are reported. However, we present our main results
based on Huber–White robust t-statistics in Appendix Table A1. We also
examine potential multicollinearity by calculating the variance inflation factor
(VIF) of each independent variable. Both the centred and uncentred VIFs are
below 5, indicating the absence of a high degree of multicollinearity (Wooldridge,
2015). In estimating the regression models, we include year-fixed effects to
remove the potential impact of any time trend.
Our entire analysis concentrates on the relation of banking firm performance
with CSR expenditure and financial inclusion after the introduction of the
mandatory CSR legislation. We are unable to examine the causal relation in
both the pre- and post-legislation periods because actual CSR expenditure data
before the introduction of the mandatory CSR legislation is mostly unavailable
for banking firms in India. Additionally, because the number of banking firms
listed on the Indian stock exchanges is not very large, segregating them into
treatment and control groups does not provide a statistically valid sample to
conduct a difference-in-difference analysis.
2
This issue is further discussed in Section 5.2.3.

3.2. Variables
Two measures of firm performance, return on assets (ROA) and stock return
(SR) are used. The former is an accounting-based measure that captures
historical financial performance whereas the latter is a market-based measure
that represents the market perception of future performance. ROA is the most
widely used indicator of firm performance with regard to accounting profit.
Stock returns are expected to capture all possible ways that the level of CSR
expenditure and financial inclusion can influence firm performance. According
to the semi-strong interpretation of the efficient markets hypothesis, stock
prices reflect all publicly available information, thereby providing a more
comprehensive picture of firm performance relative to the accounting-based
measures. These measures are also used in the literature. For example, Lys et al.
(2015) use ROA, cash flow from operations (CFO) and SR while Cheng et al.
(2015) use ROA and SR as indicators of firm performance.
To measure financial inclusion at the firm level, we consider five financial
inclusion indicators. They are the number of bank branches, the number of
ATMs, the number of accounts, the amount of bank credit and the amount of
bank deposits. The choice of these financial inclusion variables is consistent
with the previous literature (see, for example, Beck et al., 2007; Sarma and Pais,
2011).
To mitigate the omitted variable problem, we include several control
variables in the regression models. We include control variables that capture
board independence because the Companies Act 2013 includes several key
provisions on corporate governance. However, prior studies by Lys et al.
(2015), Chauvet and Jacolin (2017), Manchiraju and Rajgopal (2017) and
Mukherjee et al. (2018) used only firm-related control variables, ignoring board
independence. We control for firm-related effects for size, risk and profit
margin. Larger firms may have additional resources for CSR spending and,
therefore, may be subject to more pressure to engage in CSR-related activities.
In general, firms with lower risk are more likely to spend on CSR activities
(Orlitzky and Benjamin, 2001).
Variable definitions for measures used in this study are provided in Table 1.
4. Data and descriptive statistics
4.1. Sample
Banking firms and their industry groupings are chosen from Thompson
Reuters DataStream. The main source of data used in this paper is Prowess,
which is the largest database of financial performance of Indian business
entities. This database is maintained by the Centre for Monitoring Indian
Economy (CMIE), an independent Indian think-tank and business

Table 1
Variable description
Variable Description
Panel A: Firm performance measures
ROA Profit before tax scaled by total assets at the end of financial year t
SR Annual stock return
Panel B: CSR expenditure variable
CSR Natural logarithm of actual CSR expenditure
Panel C: Financial inclusion variables
BRANCH Natural logarithm of number of branches

ATM Natural logarithm of number of ATMs
ACCOUNT Natural logarithm of number of accounts
CREDIT Natural logarithm of total loan advances
DEPOSIT Natural logarithm of total deposits accepted by commercial backs
FI This is a composite financial inclusion variable that captures the common
components of individual financial inclusion variables. This variable is
estimated from first principal component analysis of five proxies of financial
inclusion
Panel D: Control variables
SIZE Natural logarithm of total assets at the end of a given financial year t
LEVERAGE Total liabilities scaled by total assets at the end of a financial year t. Total
liability is calculated as the sum of current liabilities and non-current
liabilities
ATO Interest income scaled by total assets at the end of a given financial year t
PM Profit before tax scaled by interest income for the financial year t
DIRECTOR Number of independent directors as a proportion of total directors
information company.3 The studies of Gupta (2005), Gopalan et al. (2007),

Vig (2013), and Manchiraju and Rajgopal (2017), among others, have used
this database in their firm-level analysis of Indian companies. By necessity,
data related to firm-level financial inclusion factors is hand-collected from
annual reports and relevant websites.
The data were collected for a sample of 149 banking firm-years over the
period 2015–2017. Our data set is cross-sectional with variables obtained from
(almost) the same set of firms over the sample period. The choice of the sample
period is dictated by the availability of CSR expenditure data. Compliance with
Section 135 of the Indian Companies Act 2013 was made mandatory in 2014,
so CSR expenditure data is mostly available from 2015.
3
More information about this database can be obtained from https://www.cmie.com/
kommon/bin/sr.php?kall=wapps&tabno=7030&page=about_us

To address the concern that a three-year sample period may not be enough to
capture the impact of CSR expenditure and financial inclusion on firm
performance, we conduct a robustness check. In addition to estimating the
models (Eqns 1–3) for the whole sample period (2015–2017) that generates our
main results, we also estimate the models for yearly sub-periods. We find that
the cross-sectional relation between mandatory CSR expenditure and financial
inclusion with firm performance is remarkably similar across the years and with
the whole sample period. Additionally, we observe that CSR expenditure as a
proportion of income, financial inclusion variables and our measures of firm
performance do not vary significantly across the years (t-statistics for mean
difference of the variables across the years are statistically insignificant). The
stability of our results through our sample period implies that our baseline
results are not sensitive to the length of sample period.4
Our initial sample includes 61 banking firms publicly listed on the Indian
stock exchanges across the sample period. However, we exclude those banks for
which none of the required variables are available. The number of excluded
banks due to data omission is not the same across the years. Therefore,
although our initial sample remains the same (61 banking firms) over the
sample period, in the final sample, the number of sample firms differs in
different years. More specifically, the final sample is 149 banking firm-years and
comprises 49, 52 and 48 banks, respectively, in the years 2015, 2016 and 2017.5
4.2. Descriptive statistics and correlation matrix
Descriptive statistics are presented in Table 2. Panels A, B, C and D,

respectively, display bank performance measures, CSR expenditure variable,
financial inclusion (FI) variables, and control variables.6 Panel A shows that
annualised average ROA and SR are, respectively, 0.87 and 0.74 percent. SR
appears to be a more volatile measure than ROA, as indicated by SR’s higher
standard deviation. The 25th percentile of SR is negative while that of ROA is
marginally positive. The average amount spent on CSR activities by the banks
4
The results pertaining to these analyses are available from the corresponding author on
request.
5
However, we have also undertaken a robustness test of the results on a uniform sample
(44 firms over the period 2015–2017, total 132 firm-year observations). A few
observations were missing from our workable sample that we have imputed using
Markov Chain Monte Carlo assuming a multivariate normal distribution. This data
imputation process is particularly suitable in the case of a non-monotonous data pattern
which matches our case.
6
Although we take the natural logarithm of the actual amount of the CSR expenditure
variable (CSR), FI variables (BRANCH, ATM, ACCOUNT, CREDIT and DEPOSIT)
and one of the control variables (SIZE) for our empirical analysis, descriptive statistics
(reported in Table 2) are based on actual data before taking their natural log.

Table 2
Descriptive statistics
Mean Median Standard deviation 25th percentile 75th percentile
Panel A: Firm performance measures
ROA (%) 0.87 0.73 1.59 0.14 1.40

SR (%) 0.74 0.54 2.21 0.56 1.60
Panel B: CSR expenditure variable
CSR 192 27 449 5 133
Panel C: Financial inclusion variables
BRANCH 2,258 1,457 2,734 505 3,210

ATM 3,938 1,667 7,365 820 3,725
ACCOUNTS 5,332,143 1,219,000 11,151,827 167,760 4,940,000
CREDIT 1,490,156 786,420 2,261,194 237,289 1,664,379
DEPOSIT 1,932,349 1,117,195 2,832,681 301,157 2,245,143
Panel D: Control variables
SIZE 2,440,283 1,315,684 3,734,358 353,184 2,744,731

LEVERAGE (%) 6.43 2.93 14.61 2.33 4.37
ATO (%) 8.54 8.24 2.14 7.63 8.83
PM (%) 8.88 8.58 14.73 1.72 15.60
DIRECTOR (%) 29.69 28.57 23.50 6.25 50.00
ROA, return on assets; SR, stock return; CSR, actual amount spent on CSR activities during
the year; BRANCH, number of bank branches; ATM, number of ATMs; ACCOUNTS,
number of accounts; CREDIT, amount of credit provided by each bank; DEPOSIT, amount
of deposit received by each bank; SIZE, firm size; LEVERAGE, level of debt in capital
structure; ATO, asset turnover; PM, profit margin; DIRECTOR, proportion of independent
directors. All variables are explained in Table 1.
is INR192 million or $US2.86 million (Panel B). However, this may be affected
by outliers as we can see that the 25th and 75th percentiles of CSR expenditure
are, respectively, INR5 million ($US0.075 million) and INR133 million
($US1.98 million). From Panel C, it is observed that in general, the banks
have large networks of branches and ATMs. The average number of branches
and ATMs is 2,258 and 3,938, respectively. This large network of branches has
enabled the banks to access a large customer base as indicated by the number of
accounts (an average of 5,332,143), and the amount of credit provided (an
average of 1,490,156) and deposits received (an average of 1,932,349). With
regard to the control variables (Panel D), it can be seen that banks are more
reliant on equity than debt, as on average only 6.43 percent of total capital
comes from debt. Average ATO and PM are about 8 percent each. The
governance variable (DIRECTOR) implies that on average about 30 percent of
the directors on the banks’ boards are independent.

14
Table 3
Correlation matrix
ROA SR CSR BRANCH ATM CREDIT DEPOSIT ACCOUNTS SIZE LEVERAGE ATO PM DIRECTOR
Measures of firm performance

ROA 1.00
SR 0.11 1.00
CSR expenditure
CSR 0.05 0.50*** 1.00
Financial inclusion variables
BRANCH 0.32*** 0.45*** 0.57*** 1.00
ATM 0.22*** 0.49*** 0.71*** 0.93*** 1.00
CREDIT 0.29*** 0.52*** 0.75*** 0.93*** 0.92*** 1.00
DEPOSIT 0.27*** 0.47*** 0.68*** 0.93*** 0.95*** 0.96*** 1.00
ACCOUNTS 0.21** 0.35*** 0.52*** 0.83*** 0.88*** 0.78*** 0.91*** 1.00
Control variables
SIZE 0.33*** 0.50*** 0.73*** 0.92*** 0.91*** 0.99*** 0.95*** 0.78*** 1.00
LEVERAGE 0.64*** 0.18** 0.34*** 0.28*** 0.29*** 0.37*** 0.30*** 0.18* 0.44*** 1.00
ATO 0.58*** 0.26*** 0.38*** 0.41*** 0.39*** 0.48*** 0.40*** 0.23*** 0.53*** 0.68*** 1.00
PM 0.94*** 0.05 0.09 0.27*** 0.14* 0.19** 0.20** 0.19** 0.22** 0.43*** 0.37*** 1.00
DIRECTOR 0.23*** 0.08 0.20** 0.24*** 0.08 0.10 0.12 0.14 0.11 0.04 0.00 0.31*** 1.00
Entries are the Pearson’s bivariate correlation between variables. ROA, return on assets; SR, stock return; CSR, actual amount spent on CSR
A. Bhattacharyya et al./Accounting & Finance
activities during the year; BRANCH, number of bank branches; ATM, number of ATMs; ACCOUNTS, number of accounts; CREDIT, amount
of credit provided by each bank; DEPOSIT, amount of deposit received by each bank; SIZE, firm size; LEVERAGE, level of debt in capital
structure; ATO, asset turnover; PM, profit margin; DIRECTOR, proportion of independent directors. All variables are explained in Table 1. ***,
** and * indicate statistical significance at the 1, 5 and 10 percent level, respectively.

Table 3 presents bivariate correlations between the variables considered in

this study. We observe that the firm performance variables are not highly
correlated, with the correlation coefficient being statistically insignificant at the
conventional level. This result firstly indicates the heterogeneous aspects of
performance captured by these variables, and secondly justifies the use of
multiple performance measures. CSR’s relation to SR (r7 = 0.50) is negative
and statistically significant. However, no statistically significant relation is
found between CSR and ROA. All the FI variables are negatively correlated
with ROA (r ranges between 0.21 to 0.32) and SR (r varies between 0.35 to
0.52). However, CSR is positively correlated with FI variables. All these
correlation coefficients are statistically significant. As expected, FI variables are
highly positively correlated to each other. For example, BRANCH’s correlation
coefficient with ATM, CREDIT and DEPOSIT is 0.93, while with
ACCOUNTS, it is 0.83. We find similar results for other pairs of the FI
variables. Control variables are moderately correlated with each other and also
with FI variables and CSR except for SIZE. SIZE is highly correlated with
CSR (r = 0.73) and FI variables (r ranges from 0.78 to 0.99).
5. Empirical results
5.1. CSR expenditure and firm performance
We first examine the association between firm performance and mandated

CSR expenditure. The results are reported in Table 4. Columns 1 and 2 show
results derived from models with dependent variables of ROA and SR,
respectively. Because of the high correlation of SIZE with CSR, and to avoid
multicollinearity concerns, we orthogonalize the SIZE variable. Specifically, we
regress SIZE on CSR, using the resulting residual as a proxy for SIZE.
We find that ROA is invariant to the level of CSR expenditure, whereas CSR has
a statistically significant (at the 1 percent level) negative association with
contemporaneous SR. This indicates that mandated CSR expenditure is perceived
negatively by investors, a result that is consistent with Manchiraju and Rajgopal
(2017) who find a negative relation between stock prices and mandated spending
on CSR activities. However, our results are somewhat at odds with Lys et al.
(2015) who find an insignificant association between CSR expenditure and future
firm performance. Overall, our result implies that mandated CSR expenditure
does not contribute to improving the operating performance of the banks (as
measured by ROA), and is not favoured by shareholders.
With regard to the control variables, we find that LEVERAGE, ATO and
PM have positive associations with ROA, while SIZE’s coefficient is not
statistically significant. SR is found to be insensitive to all the control variables
except SIZE, which exhibits a negative coefficient that is marginally significant.
7
The term r represents bivariate correlation.

Table 4
CSR expenditure and firm performance
(1) (2)
Variable ROA SR
C 0.01*** 1.28
(6.94) (1.38)
CSR 0.00 0.38***
(0.23) (6.64)
SIZE 0.00 0.27*
(1.53) (1.71)
LEVERAGE 0.02*** 1.48
(8.89) (0.99)
ATO 0.13*** 10.78
(7.74) (1.03)
PM 0.09*** 0.66
(43.56) (0.51)
DIRECTOR 0.00 0.83
(0.31) (1.15)
Adjusted R2 0.97 0.33
Year fixed effect Included Included
Dependent variable: Firm performance measured by ROA and SR. Entries are the regression
coefficients and t-statistics (standard) within parentheses derived from estimating Equation
(1) using the OLS method. The sample includes 149 banking firm-years over the period 2015–
2017. CSR, actual amount spent on CSR activities during the year; ROA, return on assets;
SR, stock return; SIZE, firm size; LEVERAGE, level of debt in capital structure; ATO, asset
turnover; PM, profit margin; DIRECTOR, proportion of independent directors. All variables
are explained in Table 1. ***, ** and * indicate statistical significance at the 1, 5 and 10
percent level, respectively.
Next, we investigate the relation between financial inclusion and firm

performance. In line with our previous analysis, we use an indicator of financial
inclusion to explain the cross-sectional variation in bank performance as
measured by ROA and SR. We consider five indicators of financial inclusion:
the numbers of branches, ATMs, bank accounts, and the amounts of credit
provided and deposits received. Since these variables are highly correlated, we
do not include them in the same regression model. Instead, the first principal
component analysis is used to isolate the common component among the
variables, and a composite financial inclusion variable (FI) is formed to capture
the common components among the five indicators.8 We observe that FI’s
correlation with individual financial inclusion variables is 0.95 or above,
8
A similar approach is used by Baker and Wurgler (2006). While the principal
components of a list of variables are derived from estimating the eigenvalue
decomposition of observed variance, the first principal component is the original
variables’ unit-length linear combination with maximum variance (for more details, see
Johnson and Wichern, 2013).

Table 5
Financial inclusion and firm performance
(1) (2)
Variable ROA SR
C 0.01*** 0.55
(7.44) (0.62)
FI 0.00 0.56***
(0.33) (5.56)
SIZE 0.00 0.48***
(1.00) (2.33)
LEVERAGE 0.02*** 0.00
(9.00) (0.00)
ATO 0.13*** 18.81*
(7.91) (1.77)
PM 0.09*** 0.39
(44.46) (0.30)
DIRECTOR 0.00 0.75
(0.32) (1.00)
2017. ROA, return on assets; SR, stock return; FI, financial inclusion; SIZE, firm size;
LEVERAGE, level of debt in capital structure; ATO, asset turnover; PM, profit margin;
DIRECTOR, proportion of independent directors. All variables are explained in Table 1. ***,
indicating that very little information is lost in the aggregating process.9

Relevant results are presented in Table 5. Columns 1 and 2 display results for
the models with dependent variables ROA and SR, respectively. We find that
SR has a statistically significant negative association with FI, but we do not find
any statistically significant relation between ROA and FI. These results are
economically rational, because our sample is dominated by government banks.
The government banks’ financial inclusion programs are driven less by a profit
motive and more to extend financial services to the large underprivileged
section of the population. Despite increasing the banking network, financial
inclusion is not associated with increased bank profitability (measured by
ROA). This result indicates that FI mostly achieves social and political
outcomes rather than economic ones. This view is consistent with the investors’
negative interpretation of financial inclusion, as shown by our results.
9
The results pertaining to these analyses are available from the corresponding author on
request.

Table 6
CSR expenditure, financial inclusion and firm performance
(1) (2)
Variable ROA SR
C 0.01*** 1.75
(5.70) (1.59)
CSR 0.00 0.49***
(0.18) (3.38)
FI 0.00 0.20
(0.30) (0.81)
SIZE 0.00 0.53
(0.40) (1.47)
LEVERAGE 0.02*** 1.88
(8.33) (1.19)
ATO 0.13*** 9.29
(7.55) (0.87)
PM 0.09*** 0.88
(42.54) (0.66)
DIRECTOR 0.00 0.83
(0.31) (1.14)
2017. ROA, return on assets; SR, stock return; CSR, actual amount spent on CSR activities
during the year; FI, financial inclusion; SIZE, firm size; LEVERAGE, level of debt in capital
directors. All variables are explained in Table 1. ***, ** and * indicate statistical significance
at the 1, 5 and 10 percent level, respectively.
The associations of control variables with firm performance in the presence

of financial inclusion are somewhat consistent with those presented in Table 4
for CSR. ROA has a positive association with LEVERAGE, ATO and PM.
Although SIZE has a statistically insignificant relation to ROA, its coefficient is
highly significant when SR is the dependent variable. ATO has a positive (but
marginally significant) association with SR. The governance variable (DIREC-
TOR) has a statistically insignificant coefficient across both the measures of
firm performance.
In the next stage, we analyse the cross-sectional variation in firm performance
using both CSR and FI as explanatory variables in the same regression model.
Results are presented in Table 6. In line with our previous results presented in Tables
4 and 5, Column 1 shows that neither CSR nor FI has any statistically significant
relation with the operating performance (measured by ROA) of the commercial
banks in India. The coefficient of neither variable is statistically different from zero.
This result supports our findings that CSR expenditure, being mandated by

legislation rather than being voluntarily chosen by the firms, is not associated with
better firm performance, and that FI inclusion, being more a strategy for societal
benefit than corporate profit, also is not associated with better firm performance.
Column 2 shows that CSR expenditure has a statistically significant negative
association with SR, the market-based measure of firm performance. This
result is consistent with our previous findings presented in Table 4, and
supports the interpretation that investors perceive mandatory CSR to be an
additional and unwelcome cost imposed by legislation. In contrast to our
previous findings of a statistically significant negative relation between SR and
FI presented in Table 5, we observe that SR is not associated with FI. This
result indicates that mandated CSR expenditure has a moderating effect on the
relation between SR and FI.10
As indicated earlier, relevant results in the previous literature may be subject to
model misspecification bias. This bias may arise from several sources, such as (i)
the true functional form of the relation between economic and financial variables
being generally unknown, (ii) a single econometric model being unable to
incorporate all the relevant explanatory variables, and (iii) variables themselves
being subject to measurement error. We address the model misspecification bias in
several ways. First, as indicated already, the causal inference of CSR expenditure
and financial inclusion with firm performance is measured by running three
separate regression models (Eqns 1, 2 and 3) with different model specifications.
Second, in estimating each regression model (Eqns 1, 2 and 3), we check the
robustness of results derived from different model specifications such as (i)
including only point-of-interest variable(s) (for example, CSR, FI or both), (ii)
including both point-of-interest and control variables, and (iii) incorporating/
ignoring year-fixed effects.11 As we find consistent results across the models, it may
indicate the absence of model misspecification bias. Third, we use a set of control
variables supported by economic arguments and the previous literature. Fourth,
we use different definitions of the explanatory variables. For example, financial
inclusion is represented by five individual measures and a composite measure of
them. Moreover, although SIZE and LEVERAGE are, respectively, defined as
natural logarithm of total assets and total liabilities scaled by total assets, we also
measure them as the natural logarithm of market capitalisation and non-current
liabilities scaled by non-current assets.12 Finally, to check the general functional
10
While our baseline results are derived from a sample of 149 banking firms (49, 52 and
48 banks respectively in the years 2015, 2016 and 2017), we also estimate our baseline
regression models using a common sample of 44 banking firms each year. We find
remarkably similar results for these two samples. The regression results based on
common samples are presented in Appendix Table A2.
11
Results relating to all these model specifications are presented in Appendix Table A3.
12
Results pertaining to these alternative measures of SIZE and LEVERAGE are not
collated in this paper. However, they can be obtained from the corresponding author on
request.

form of misspecification, we conduct the regression specification error test

(RESET) of Ramsey (1969). In the case of Equations (1), (2) and (3), the F-statistics
(p-values) are, respectively, 0.05 (0.83), 1.3 (0.25) and 0.10 (0.76) indicating that the
null hypothesis (the model is correctly specified) cannot be rejected.13
5.2. Robustness checks
5.2.1. Individual financial inclusion variables
In the previous subsection, we reported the association between FI and firm

performance where FI is a composite indicator of financial inclusion variables
derived from first principal component analysis of five financial inclusion
indicators. In this subsection, as a robustness check, we report the relation
between firm performance and individual financial inclusion indicators. In
Table 7, only financial inclusion variables are included, while Table 8 includes
both CSR expenditure and financial inclusion measures as explanatory
variables in the regression models.
Panels A and B of Table 7 show the results from a model with dependent
variables of ROA and SR, respectively. In Panel A, we observe that none of the
individual financial inclusion measures (BRANCH, ATM, CREDIT,
DEPOSIT and ACCOUNTS) has a statistically significant positive association
with firm performance as measured by ROA. In Panel B, we show that financial
inclusion variables have a statistically significant (at the 1 percent level)
negative association with firm performance as measured by SR. These results
are consistent with our previous findings presented in Table 5. The signs and
magnitudes of the control variables’ coefficients are also consistent across the
financial inclusion variables and with our previous results.
In Table 8, we observe that the relations of firm performance with CSR
expenditure and individual financial inclusion indicators are less clear.
However, on the whole, they are not inconsistent with our previous results.
In Panel A, we find a statistically significant coefficient for CREDIT, indicating
a negative relation between ROA and CREDIT. However, the coefficient is
estimated to be 0.00, indicating little (if any) economic significance in this
result. In Panel B, we find that CSR has a negative association with SR only
when financial inclusion is measured as BRANCH, DEPOSIT and
ACCOUNTS.
13
The RESET test, a general specification test, examines whether the response variables
can be explained by non-linear combinations of the fitted values. This test is particularly
applicable for the linear regression model and based on the notion that a model is mis-
specified when non-linear combinations of explanatory variables exhibit explanatory
power with regard to the response variable. Under this circumstance, a polynomial or
other non-linear functional form may better explain the data-generating process.

Table 7
Financial inclusion and firm performance
Model 1 Model 2 Model 3 Model 4 Model 5
Panel A: Dependent variable: ROA
C 0.01*** 0.01*** 0.01** 0.01*** 0.01

(4.54) (5.70) (2.74) (3.33) (6.32)
BRANCH 0.00
(0.59)
ATM 0.00
(0.59)
CREDIT 0.00
(0.32)
DEPOSIT 0.00
(0.08)
ACCOUNTS 0.00
(0.75)
Control variables
SIZE 0.00 0.00 0.00 0.00 0.00
(0.76) (1.00) (1.41) (1.25) (0.87)
LEVERAGE 0.02*** 0.02*** 0.02*** 0.02*** 0.02***
(8.99) (9.02) (8.94) (8.98) (9.01)
ATO 0.13*** 0.13*** 0.12*** 0.13*** 0.13***
(7.94) (7.95) (7.63) (7.83) (8.01)
PM 0.09*** 0.09*** 0.09*** 0.09*** 0.09***
(44.74) (44.23) (44.18) (44.41) (44.92)
DIRECTOR 0.00 0.00 0.00 0.00 0.00
(0.29) (0.35) (0.26) (0.28) (0.36)
Adjusted R2 0.97 0.97 0.97 0.97 0.97
Year fixed effect Included Included Included Included Included
Panel B: Dependent variable: SR
C 6.08*** 3.77*** 11.34*** 8.20*** 1.39

(3.69) (3.03) (5.39) (4.21) (1.16)
BRANCH 0.91***
(5.23)
ATM 0.60***
(5.62)
CREDIT 0.86***
(6.51)
DEPOSIT 0.65***
(5.40)
ACCOUNTS 0.20***
(3.44)
Control variables
SIZE 0.53** 0.34* 0.55 0.47** 0.17
(2.44) (1.78) (2.78) (2.26) (0.81)
LEVERAGE 0.40 0.13 0.51 0.07 0.30
(0.26) (0.08) (0.34) (0.05) (0.18)
(continued)

Table 7 (continued)
ATO 18.12 18.08* 12.18 19.07* 28.40**

(1.68) (1.70) (1.16) (1.78) (2.560
PM 0.78 0.16 0.18 0.32 1.17
(0.59) (0.12) (0.14) (0.24) (0.85)
DIRECTOR 0.38 0.81 0.88 0.82 0.64
(0.50) (1.07) (1.20) (1.07) (0.80)
Adjusted R2 0.26 0.28 0.32 0.27 0.19
2017. Models 1, 2, 3, 4 and 5, respectively, include BRANCH, ATM, CREDIT, DEPOSIT
and ACCOUNTS as financial inclusion measures. ROA, return on assets; SR, stock return;
BRANCH, number of bank branches, ATM, number of ATMs; CREDIT, amount of credit
provided by each bank; DEPOSIT, amount of deposit received by each bank; ACCOUNTS,
number of accounts; SIZE, firm size; LEVERAGE, level of debt in capital structure; ATO,
asset turnover; PM, profit margin; DIRECTOR, proportion of independent directors. All
variables are explained in Table 1. ***, ** and * indicate statistical significance at the 1, 5 and
10 percent level, respectively.
5.2.2. CSR expenditure, financial inclusion and subsequent firm performance
Thus far, we have examined CSR expenditure and financial inclusion’s

contemporaneous association with banking firm performance. This analysis
reveals that current CSR expenditure and financial inclusion are unlikely to
drive banks’ current financial performance. However, the studies of Navarro,
(1998), Greening and Turban (2000), Fombrun (2005) and Margolis et al.
(2009), among others, report that current CSR expenditure may lead to
improved future financial performance of firms due to improved legitimacy and
improved acceptability of the firms to their stakeholders. In this subsection, we
empirically verify this conjecture by explaining variation in firm performance
using lagged CSR expenditure and financial inclusion variables. In particular,
we estimate the following regression models:
X
Performancet ¼ aa þ b1a CSRt1 þ li;a Control variablesi;t1 þ e ð4Þ
X
Performancet ¼ ab þ b2a FIt1 þ li;b Control variablesi;t1 þ e ð5Þ
Performancet ¼ ac þ b1b CSRt1 þ b2b FIt1

X ð6Þ
þ li;c Control variablesi;t1 þ e

Table 8
C 0.01*** 0.01*** 0.02* 0.01 0.01

(3.62) (5.15) (1.92) (1.12) (6.30)
CSR 0.00 0.00 0.00** 0.00 0.00
(0.54) (0.53) (2.78) (0.77) (0.41)
BRANCH 0.00
(0.76)
ATM 0.00
(0.75)
CREDIT 0.00**
(2.79)
DEPOSIT 0.00
(0.74)
ACCOUNTS 0.00
(0.82)
Control variables
SIZE 0.00 0.00 0.00*** 0.00 0.00
(0.07) (0.35) (3.09) (1.33) (0.54)
LEVERAGE 0.02*** 0.02*** 0.02*** 0.02*** 0.02***
(8.19) (8.53) (9.43) (8.70) (8.60)
ATO 0.13*** 0.13*** 0.13*** 0.13*** 0.12***
(7.66) (7.64) (8.07) (7.74) (7.36)
PM 0.09*** 0.09*** 0.09*** 0.09*** 0.09***
(42.51) (43.41) (42.78) (42.66) (43.03)
DIRECTOR 0.00 0.00 0.00 0.00 0.00
(0.19) (0.34) (0.14) (0.27) (0.35)
Adjusted R2 0.97 0.97 0.97 0.97 0.97
C 0.31 1.47 2.43 2.84 0.55

(0.14) (1.04) (0.31) (0.81) (0.50)
CSR 0.43*** 0.37 0.34 0.55*** 0.46***
(3.76) (3.19) (1.19) (3.74) 5.61)
BRANCH 0.16
(0.49)
ATM 0.04
(0.18)
CREDIT 0.10
(0.15)
DEPOSIT 0.36
(1.22)
ACCOUNTS 0.09
(1.25)
(continued)

Table 8 (continued)
Control variables
SIZE 0.41 0.23 0.18 0.67* 0.46**
(1.25) (0.90) (0.27) (1.84) (2.10)
LEVERAGE 1.73 1.43 1.38 2.05 1.82
(1.09) (0.93) (0.83) (1.31) (1.20)
ATO 10.43 10.94 10.86 8.48 7.82
(0.99) (1.04) (1.03) (0.80) (0.73)
PM 0.80 0.64 0.61 0.94 0.93
(0.61) (0.49) (0.46) (0.72) (0.71)
DIRECTOR 0.89 0.84 0.84 0.79 0.80
(1.21) (1.15) (1.15) (1.09) (1.10)
Adjusted R2 0.32 0.32 0.32 0.33 0.33
2017. Models 1, 2, 3, 4 and 5, respectively, include BRANCH, ATM, CREDIT, DEPOSIT
and ACCOUNTS as financial inclusion measures. CSR, actual amount spent on CSR
activities during the year; ROA, return on assets; SR, stock return; BRANCH, number of
bank branches, ATM, number of ATMs; CREDIT, amount of credit provided by each bank;
DEPOSIT, amount of deposit received by each bank; ACCOUNTS, number of accounts;
SIZE, firm size; LEVERAGE, level of debt in capital structure; ATO, asset turnover; PM,
profit margin; DIRECTOR, proportion of independent directors. All variables are explained
in Table 1. ***, ** and * indicate statistical significance at the 1, 5 and 10 percent level,
respectively.
Relevant results are reported in Table 9. Panels A and B show the results
from a model with dependent variables of ROA and SR, respectively. We find
that although lagged CSR expenditure does not have a statistically significant
association with firm performance when performance is measured as ROA
(Panel A), it has a significant negative association with market-based firm
performance (SR) (Panel B). This result implies that mandatory CSR
expenditure is not associated with firms’ one-year ahead performance, but is
associated with lower subsequent performance (when stock return is a measure
of firm performance). This is consistent with the results from the contempo-
raneous regressions. Overall, this result supports the idea that financial
inclusion programs are driven more by social and political motives than
economic ones. As such, financial inclusion activities are negatively perceived
by shareholders. Nonetheless, we find evidence of a moderating effect of CSR
expenditure and financial inclusion on firm performance. Neither CSR
expenditure nor financial inclusion has a statistically significant association
with future firm performance when they are combined in a model (column 3 in
Panels A and B).

Table 9
Variable Model 1 Model 2 Model 3
Panel A: Dependent variable ROA
C 0.01 0.00 0.01

(0.59) (0.44) (0.99)
CSR 0.00 0.00
(0.07) (1.22)
FI 0.00 0.00
(0.32) (1.26)
SIZE 0.00 0.00 0.01
(0.95) (0.95) (1.49)
LEVERAGE 0.06*** 0.06*** 0.05***
(5.12) (4.98) (3.57)
ATO 0.09 0.07 0.08
(0.75) (0.55) (0.67)
PM 0.06*** 0.06*** 0.07***
(5.68) (5.86) (5.71)
DIRECTOR 0.00 0.00 0.00
(0.58) (0.55) (0.64)
Adjusted R2 0.67 0.67 0.68
Year fixed effect Included Included Included
Panel B: Dependent variable SR
C 4.06 2.85 3.73

(1.77) (1.34) (2.46)
CSR 0.40*** 0.26
(3.56) (0.37)
FI 0.72*** 0.27
(3.50) (0.67)
SIZE 0.69** 0.12 0.39
(2.87) (0.41) (0.79)
LEVERAGE 0.35 2.17 1.02
(0.14) (0.85) (3.02)
ATO 18.52 17.49 19.17
(0.75) (0.71) (24.90)
PM 1.95 3.23 2.36
(0.88) (1.52) (2.46)
DIRECTOR 0.25 0.17 0.22
(0.23 (0.16) (1.08)
Adjusted R2 0.21 0.21 0.21
Dependent variable: Firm performance measured by ROAt+1 and SRt+1. Entries are the
regression coefficients and t-statistics (standard) within parentheses. Models 1, 2 and 3 are,
respectively, estimated based on Equations (4), (5), and (6) using the OLS method. The
sample includes 149 banking firm-years over the period 2015–2017. CSR, actual amount
spent on CSR activities during the year; FI, financial inclusion; ROA, return on assets; SR,
stock return; SIZE, firm size; LEVERAGE, level of debt in capital structure; ATO, asset
turnover; PM, profit margin; DIRECTOR, proportion of independent directors. All variables
are explained in Table 1. ***, ** and * indicate statistical significance at the 1, 5 and 10
percent level, respectively.

5.2.3. Alternative model estimation
Since we have used two measures of firm performance, an accounting-based

measure and a market-based measure, our results may not be free from
endogeneity concerns. As noted in the previous subsections, our results do not
indicate an association between CSR expenditure or financial inclusion
activities and the accounting performance measure, but they show negative
associations between stock return and both CSR expenditure and financial
inclusion activities. It is unlikely that poor (good) market performance is a
determinant of high (low) CSR spending, because it is a manifestation of
investors’ reaction to CSR expenditure and it is forward looking. However,
Masulis and Reza (2014) and Bose et al. (2017) among others argue that
accounting-based firm performance may be a determinant of firms’ voluntary
expenditure on social activities. So, in this subsection, we check the sensitivity
of our results (presented in Section 5.1) to alternative model estimations to
consider possible reverse causality or endogeneity. More specifically, while the
regression models (Eqns 1, 2 and 3) were previously estimated using OLS, we
now estimate the models using two-stage least squares (2SLS) and generalised
method of moments (GMM) approaches.
We use the instrumental variable (IV) approach to estimate the 2SLS
regression. To implement the IV approach, we select firm age as a variable that
is correlated with CSR expenditure and financial inclusion but not with ROA.
This follows the prior literature, such as Jo and Harjoto (2011, 2012).
Theoretically, older firms have a higher capacity to be engaged in CSR and
financial inclusion activities. However, firm age is unlikely to be associated with
better firm performance. We use the Durbin–Wu–Hausman test to check if
CSR expenditure and financial inclusion are endogenous. We find that the
coefficient for CSR expenditure is marginally significant, implying that the
relation between CSR expenditure and ROA might be endogenous, but there is
no indication of endogeneity between financial inclusion and ROA.
The 2SLS regression is estimated in two steps. In the first step, the
endogenous variable (for example, CSR expenditure) is regressed on the
instrumental variable (for example, firm age). In the second stage regression,
the residual derived from the first stage is used as a proxy for CSR expenditure
and included as an independent variable in estimating the regression models.
This process addresses the potential endogeneity bias as residuals are
determined exogenously rather than by firm-specific variables (Gul et al.,
2011). A two-step system GMM estimator is also used as an alternative model
estimation. In particular, Wintoki et al.’s (2012) approach is used to derive
system GMM estimates. As indicated earlier, GMM estimation addresses
biases associated with dynamic endogeneity, unobserved heterogeneity and
simultaneity by incorporating instruments in the estimation process (Wintoki
et al., 2012). Moreover, estimation bias arising from persistence is controlled by
the system GMM approach (Sufian and Habibullah, 2010). The relevant results

Table 10
2SLS GMM
C 0.01*** 0.01**
(5.16) (3.26)
CSR 0.00 0.00
(0.48) (0.43)
FI 0.00 0.00
(0.11) (0.07)
SIZE 0.00 0.00
(0.38) (0.23)
LEVERAGE 0.02*** 0.02**
(6.62) (2.82)
ATO 0.13*** 0.13***
(5.14) (3.21)
PM 0.09*** 0.09***
(39.32) (18.62)
DIRECTOR 0.00 0.00
(0.64) (0.55)
C 3.50** 3.50**
(2.41) (2.23)
CSR 0.50** 0.50*
(2.54) (2.27)
FI 0.07 0.06
(0.18) (0.20)
SIZE 0.50 0.50
(1.11) (1.24)
LEVERAGE 0.66 0.66
(0.30) (0.37)
ATO 9.90 9.89
(0.57) (0.70)
PM 0.89 0.89
(0.61) (0.81)
DIRECTOR 0.39 0.39
(0.49) (0.46)
coefficients and t-statistics (standard) within parentheses. The models are alternatively
estimated using 2SLS and GMM approaches. The sample includes 132 banking firm-years
over the period 2015–2017 (44 banks each year). CSR, actual amount spent on CSR activities
during the year; FI, financial inclusion; ROA, return on assets; SR, stock return; SIZE, firm
size; LEVERAGE, level of debt in capital structure; ATO, asset turnover; PM, profit margin;
DIRECTOR, proportion of independent directors. All variables are explained in Table 1. ***,

are presented in Table 10. Panels A and B display results derived from the
models with ROA and SR as alternative dependent variables.14
We find that results from both 2SLS and GMM approaches are consistent in
terms of direction and strength of the relation of CSR expenditure and financial
inclusion with firm performance. ROA is found to be invariant to the level of
CSR expenditure and financial inclusion while SR responds negatively to CSR
expenditure. The sign, magnitude and statistical significance of the regression
coefficients are in line with the results derived from the OLS estimation
presented in Section 5.1. Overall, this result indicates that our key findings
remain robust to the use of these alternative model estimation approaches.
6. Conclusion
In this paper, we have examined the effects of both mandatory CSR

expenditure and bank-level financial inclusion on bank performance in India,
using both accounting-based and stock market-based measures of financial
performance. While the existing literature only considers voluntary CSR
expenditure as proxied by CSR index scores, the novelty of our paper is the use
of actual CSR expenditure data. We find that when they are assessed
individually, neither CSR expenditure nor financial inclusion is associated with
banks’ operating performance, but both have a negative association with
banks’ stock return. When both are assessed together, neither CSR expenditure
nor financial inclusion has any relation with operating performance and only
CSR expenditure has a negative relation to stock return. This result remains
robust to different model specifications. To some extent, CSR expenditure has a
moderating effect on the relation between financial inclusion and stock return.
Overall, our findings clearly show that mandated CSR expenditure and
financial inclusion are not associated with better financial performance by
banks; rather they are perceived negatively by investors. Although the
legislation is designed to improve social equality and economic development
in India, its success will depend on banks’ willingness to engage with this goal in
the expectation of improved financial performance and/or improved stock
return. Because engagement with the goal and compliance with the legislation
can be avoided simply by providing an explanation for non-compliance, it is
unlikely to occur.
Our findings contribute to the wider policy debate on voluntary versus
mandatory CSR expenditure. Our primary contribution is to present clear
evidence of (i) an insignificant association of mandated CSR spending and
financial inclusion with accounting-based performance, and (ii) a significant
negative association when performance is measured by stock market return.
Additionally, our findings add to the literature on (i) the association of CSR
expenditure and performance, (ii) financial inclusion’s association with banks’
14
These results are based on a common sample of 44 banking firms each year.

performance at the firm level, (iii) the function of firm-specific variables in

determining performance, (iv) the economic consequences of mandated CSR
expenditure, and (v) industry-specific CSR.
He et al. (2012) report that voluntary CSR spending leads to positive
economic values in countries with strong law enforcement mechanisms.
However, this relation has not been examined in the context of a weak law
enforcement country using mandated CSR expenditure. Our study provides
empirical evidence that mandated CSR spending is perceived negatively by
investors in India, a country that is generally considered to be a weak law
enforcement country. Moreover, van Beurden and Gossling (2008) called for
industry-specific investigations of organisational social activities to improve the
effectiveness of CSR research. We add to the industry-specific CSR literature
by investigating a critical social aspect of banking firms’ operations: financial
inclusion activities.
Therefore, our study will be particularly valuable to government authorities of
developing and weak law enforcement countries that are considering mandated
CSR expenditure and activities. Additionally, our findings are critical to the
financial institutions that are spending on financial inclusion activities and
considering the incorporation of financial inclusion activities in their operational
activities. Furthermore, our findings are valuable to country-level regulators and
various international organisations that promote financial inclusion activities.
Our study is limited by the relatively small number of banks listed on the
Indian stock exchanges. Our sample period is also constrained from 2015 to
2017. Compliance with Section 135 of the Indian Companies Act 2013 was
made mandatory in 2014. Therefore, CSR expenditure data is mostly available
from 2015. We used banking firms that are listed on the stock exchanges.
Future research could include both listed and non-listed banks and financial
institutions to examine the relations between financial inclusion, CSR
expenditure and performance, as they are equally important to both types of
organisations. Beyond economic performance, financial inclusion and manda-
tory CSR expenditure’s impact on firm productivity and firm export perfor-
mance could be an interesting research area to explore.
References
Abdmoulah, W., and R. B. Jelili, 2013, Access to finance thresholds and the finance-
growth nexus, Economic Papers 32, 522–534.
Ahmed, A., and S. Ali, 2017, Boardroom gender diversity and stock liquidity: evidence
from Australia, Journal of Contemporary Accounting and Economics 13, 148–165.
Ahmed, A., R. Monem, D. Delaney, and C. Ng, 2017, Gender diversity in corporate
boards and continuous disclosure: evidence from Australia, Journal of Contemporary
Accounting and Economics 13, 89–107.
Alhaj-Ismail, A., S. Adwan, and J. Stittle, 2019, Share-option based compensation
expense, shareholder returns and financial crisis, Journal of Contemporary Accounting
and Economics 15, 20–35.

Ardic, O. P., M. Heimann, and N. Mylenko, 2011, Access to financial services and the
financial inclusion agenda around the world: a cross-country analysis with a new data
set, World Bank Policy Research Working Paper No. 5537. Available at: https://ssrn.c
om/abstract=1747440
Association of Chartered Certified Accountants, 2008, Going Concern? A Sustainability
Agenda for Action (ACCA, Glasgow).
Baker, M., and J. Wurgler, 2006, Investor sentiment and the cross-section of stock
returns, The Journal of Finance 61, 1645–1680.
Barnett, M. L., 2007, Stakeholder influence capacity and the variability of financial returns
to corporate social responsibility, The Academy of Management Review 32, 794–816.
Baron, D. P., 2001, Private politics, corporate social responsibility, and integrated
strategy, Journal of Economics and Management Strategy 10, 7–45.
Beck, T., A. Demirg€ ußc-Kunt, and V. Maksimovic, 2005, Financial and legal constraints
to growth: does firm size matter?, Journal of Finance 60, 137–177.
Beck, T., A. Demirg€ußc-Kunt, and M. S. Martinez Peria, 2007, Reaching out: access to and
use of banking services across countries, Journal of Financial Economics 85, 234–266.
Black, B. S., and V. S. Khanna, 2007, Can corporate governance reforms increase firm
market values? Event study evidence from India, Journal of Empirical Legal Studies 4,
749–796.
Black, B., and W. Kim, 2012, The effect of board structure on firm value: a multiple
identification strategies approach using Korean data, Journal of Financial Economics
104, 203–226.
Bose, S., A. Bhattacharyya, and S. Islam, 2016, Dynamics of firm-level financial
inclusion: empirical evidence from an emerging economy, Journal of Banking and
Finance Law and Practice 27, 47–68.
Bose, S., J. Podder, and K. Biswas, 2017, Philanthropic giving, market-based
performance and institutional ownership: evidence from an emerging economy, The
British Accounting Review 49, 429–444.
Brown, N., and C. Deegan, 1998, The public disclosure of environmental performance
information – a dual test of media agenda setting theory and legitimacy theory,
Accounting and Business Research 29, 21–41.
Burgess, R., and R. Pande, 2005, Do rural banks matter? Evidence from the Indian
social banking experiment, American Economic Review 95, 780–795.
Chauvet, L., and L. Jacolin, 2017, Financial inclusion, bank concentration, and firm
performance, World Development 97, 1–13.
Cheng, S., K. Z. Lin, and W. Wong, 2015, Corporate social responsibility reporting and
firm performance: evidence from China, Journal of Management and Governance 20,
503–523.
Coase, R. H., 1937, The nature of the firm, Economica 4, 386–405.
Cochran, P. L., and R. A. Wood, 1984, Corporate social responsibility and financial
performance, Academy of Management Journal 27, 42–56.
Dhaliwal, D., Z. Li, A. Tsang, and G. Yang, 2011, Voluntary non-financial disclosure
and the cost of equity capital: the case of corporate social responsibility reporting, The
Accounting Review 87, 723–759.
Di Giuli, A., and L. Kostovetsky, 2014, Are red or blue companies more likely to go
green? Politics and corporate social responsibility, Journal of Financial Economics 111,
158–180.
Dupas, P., and J. Robinson, 2009, Savings constraints and microenterprise development:
evidence from a field experiment in Kenya, NBER Working Paper No. 14693 (National
Bureau of Economic Research).
Fombrun, C. J., 2005, Building corporate reputation through CSR initiatives: evolving
standards, Corporate Reputation Review 8, 7–12.

Freeman, R. E., 2010, Strategic Management: A Stakeholder Approach (Cambridge

University Press, Cambridge).
Friedman, M., 1970, The social responsibility of business is to increase its profits, The
New York Times Magazine, September 13, 122–126.
Frooman, J., 1999, Stakeholder influence strategies, Academy of Management Review 24,
191–205.
Frost, G. R., 2007, The introduction of mandatory environmental reporting guidelines:
Australian evidence, Abacus 43, 190–216.
Godfrey, P. C., 2005, The relationship between corporate philanthropy and shareholder
wealth: a risk management perspective, Academy of Management Review 30, 777–798.
Gopalan, R., V. Nanda, and A. Seru, 2007, Affiliated firms and financial support:
evidence from Indian business groups, Journal of Financial Economics 86, 759–795.
Greening, D. W., and D. B. Turban, 2000, Corporate social performance as a competitive
advantage in attracting a quality workforce, Business and Society 39, 254–280.
Gul, F. A., B. Srinidhi, and A. C. Ng, 2011, Does board gender diversity improve the
informativeness of stock prices?, Journal of Accounting and Economics 51, 314–338.
Gupta, N., 2005, Partial privatization and firm performance, The Journal of Finance 60,
987–1015.
Hawkins, D. E., 2006, Corporate Social Responsibility: Balancing Tomorrow’s Sustain-
ability and Today’s Profitability (Palgrave Macmillan, Basingstoke).
He, X., T. J. Wong, and D. Young, 2012, Challenges for implementation of fair value
accounting in emerging markets: evidence from IFRS adoption in China, Contem-
porary Accounting Research 29, 538–562.
Hillman, A., and G. Keim, 2001, Shareholder value, stakeholder management, and
social issues: what’s the bottom line?, Strategic Management Journal 22, 125–139.
Himmelberg, C. P., R. G. Hubbard, and D. Palia, 1999, Understanding the
determinants of managerial ownership and the link between ownership and
performance, Journal of Financial Economics 53, 353–384.
International Auditing and Assurance Standards Board (IAASB), 2007, Assurance
Engagements on Carbon Emissions Information (IAASB, New York).
Jamali, D., 2008, A stakeholder approach to corporate social responsibility: a fresh
perspective into theory and practice, Journal of Business Ethics 82, 213–231.
Jensen, M. C., and W. H. Meckling, 1976, Theory of the firm: managerial behavior,
agency costs and ownership structure, Journal of Financial Economics 3, 305–360.
Jo, H., and M. A. Harjoto, 2011, Corporate governance and firm value: the impact of
corporate social responsibility, Journal of Business Ethics 103, 351–383.
Jo, H., and M. A. Harjoto, 2012, The causal effect of corporate governance on corporate
social responsibility, Journal of Business Ethics 106, 53–72.
Jo, H., H. Kim, and K. Park, 2015, Corporate environmental responsibility and firm
performance in the financial services sector, Journal of Business Ethics 131, 257–284.
Karnani, A., 2003, India makes CSR mandatory: a really bad idea, The European
Financial Review, October 29. Available at: https://www.europeanfinancialreview.c
om/india-makes-csr-mandatory-a-really-bad-idea/
Kempson, E., A. Atkinson, and O. Pilley, 2004, Policy level response to financial
exclusion in developed economies: lessons for developing countries, Report of Personal
Finance Research Centre (University of Bristol).
Lys, T., J. P. Naughton, and C. Wang, 2015, Signaling through corporate accountability
reporting, Journal of Accounting and Economics 60, 56–72.
Manchiraju, H., and S. Rajgopal, 2017, Does corporate social responsibility create
shareholder value? Evidence from the Indian Companies Act 2013, Journal of
Accounting Research 55, 1257–1300.

Margolis, J. D., H. A. Elfenbein, and J. P. Walsh, 2009, Does it pay to be good . . . and
does it matter? A meta-analysis of the relationship between corporate social and
financial performance. Available at: http://ssrn.com/abstract=1866371
Masulis, R. W., and S. W. Reza, 2014, Agency problems of corporate philanthropy, The
Review of Financial Studies 28, 592–636.
McWilliams, A., and D. S. Siegel, 2001, Corporate social responsibility: a theory of the
firm perspective, Academy of Management Review 26, 117–127.
Mukherjee, A., R. Bird, and G. Duppati, 2018, Mandatory Corporate Social
Responsibility: The Indian Experience, Journal of Contemporary Accounting &
Economics 14, 254–265.
Navarro, P., 1998, Why do corporations give to charity?, The Journal of Business 61, 65–
93.
Orlitzky, M., and J. D. Benjamin, 2001, Corporate social performance and firm risk: a
meta-analytic review, Business and Society 40, 369–396.
Park, C.-Y., and R. Mercado Jr., 2015, Financial inclusion, poverty, and income
inequality in developing Asia, Asian Development Bank Economics Working Paper
Series 426.
Peloza, J., 2006, Using corporate social responsibility as insurance for financial
performance, California Management Review 48, 52–72.
Pfeffer, J., and G. R. Salancik, 1978, The External Control of Organizations: A Resource
Dependence Perspective (Stanford Business Books, Stanford, CA).
Ramsey, J. B., 1969, Tests for specification errors in classical linear least-squares
regression analysis, Journal of the Royal Statistical Society: Series B (Methodological)
31, 350–371.
Richardson, G., B. Wang, and X. Zhang, 2016, Ownership structure and corporate tax
avoidance: evidence from publicly listed private firms in China, Journal of Contem-
porary Accounting and Economics 12, 141–158.
Safiullah, M., and A. Shamsuddin, 2018, Risk-adjusted efficiency and corporate
governance: evidence from Islamic and conventional banks, Journal of Corporate
Finance 55, 105–140.
Sarma, M., and J. Pais, 2011, Financial inclusion and development, Journal of
International Development 23, 613–628.
Servaes, H., and A. Tamayo, 2013, The impact of corporate social responsibility on firm
value: the role of customer awareness, Management Science 59, 1045–1061.
Shane, P. B., and B. H. Spicer, 1983, Market response to environmental information
produced outside the firm, The Accounting Review 58, 521–538.
Sufian, F., and M. S. Habibullah, 2010, Does economic freedom fosters banks’
performance? Panel evidence from Malaysia, Journal of Contemporary Accounting and
Economics 6, 77–91.
United Nations Conference on Trade and Development (UNCTAD), 2014, Bangladesh
Country Paper: Impact of Access to Financial Services. Available at: http://unctad.org/
meetings/en/Presentation/ciem6_2014_Bangladesh_en.pdf
van Beurden, P., and T. Gossling, 2008, The worth of values – a literature review on the
relation between corporate social and financial performance, Journal of Business
Ethics 82, 407–424.
Vig, V., 2013, Access to collateral and corporate debt structure: evidence from a natural
experiment, The Journal of Finance 68, 881–928.
Wang, H., and C. Qian, 2011, Corporate philanthropy and corporate financial
performance: the roles of stakeholder response and political access, Academy of
Management Journal 54, 1159–1181.
Wintoki, M. B., J. S. Linck, and J. M. Netter, 2012, Endogeneity and the dynamics of
internal corporate governance, Journal of Financial Economics 105, 581–606.

Wooldridge, J. M., 2015, Introductory Econometrics: A Modern Approach, 6th edn

(Cengage Learning, Boston, MA).
World Bank, 2014, Financial Inclusion: Global Financial Development Report.
Available at: http://siteresources.worldbank.org/EXTGLOBALFINREPORT/Re
sources/8816096-1361888425203/9062080-1364927957721/GFDR-2014_Complete_
Report.pdf

Appendix
Table A1
C 0.01*** 0.01 0.01

(4.35) (4.17***) (3.81***)
CSR 0.00 0.00
(0.18) 0.16)
FI 0.00 0.00
(0.21) (0.21)
SIZE 0.00 0.00 0.00
(1.21) (0.61) (0.26)
LEVERAGE 0.02*** 0.02*** 0.02***
(3.12) (3.19) (3.20)
ATO 0.13*** 0.13*** 0.13***
(4.46) (4.39) (4.38)
PM 0.09*** 0.09*** 0.09***
(19.48) (19.30) (18.91)
DIRECTOR 0.00 0.00 0.00
(0.27) (0.28) (0.27)
Adjusted R2 0.97 0.97 0.97
C 1.28 0.55 1.75

(1.24) (0.55) (1.31)
CSR 0.38*** 0.49***
(6.56) (2.81)
FI 0.56*** 0.20
(5.45) (0.71)
SIZE 0.27** 0.48*** 0.53
(2.15) (2.71) (1.47)
LEVERAGE 1.48 0.00 1.88
(1.17) (0.00) (1.27)
ATO 10.78 18.81 9.29
(1.07) (1.56) (0.96)
PM 0.66 0.39 0.88
(0.79) (0.46) 0.94)
DIRECTOR 0.83 0.75 0.83
(1.06) (0.93) (1.05)
(continued)

Table A1 (continued)

2
Adjusted R 0.33 0.28 0.33
coefficients and t-statistics (Huber-White) within parentheses derived from estimating the
models using the OLS method. The sample includes 149 banking firm-years over the period
2015–2017. Models 1, 2 and 3 are, respectively, estimated based on Equations (1), (2) and (3).
CSR, actual amount spent on CSR activities during the year; FI, financial inclusion; ROA,
return on assets; SR, stock return; SIZE, firm size; LEVERAGE, level of debt in capital
directors. All variables are explained in Table 1. ***, ** and * indicate statistical significance
at the 1, 5 and 10 percent level, respectively.
Table A2
CSR expenditure, financial inclusion and firm performance (common sample)
C 0.01*** 0.01*** 0.01***

(5.18) (5.19) (5.16)
CSR 0.00 0.00
(1.52) (0.48)
FI 0.00 0.00
(1.44) (0.11)
SIZE 0.00 0.00 0.00
(1.20) (0.07) (0.38)
LEVERAGE 0.02*** 0.02*** 0.02***
(8.65) (7.87) (6.62)
ATO 0.13*** 0.14*** 0.13***
(5.60) (5.37) (5.14)
PM 0.09*** 0.09*** 0.09***
(40.85) (42.51) (39.32)
DIRECTOR 0.00 0.00 0.00
(0.65) (0.63) (0.65)
Adjusted R2 0.97 0.97 0.97
C 3.51** 2.70* 3.50**

(2.42) (1.86) (2.41)
CSR 0.47*** 0.50**
(6.16) (2.54)
FI 0.80*** 0.07
(5.47) (0.18)
(continued)

SIZE 0.42** 0.49** 0.50

(2.32) (2.16) (1.11)
LEVERAGE 0.41 2.64 0.66
(0.24) (1.45) (0.30)
ATO 11.18 18.97 9.90
(0.71) (1.09) (0.57)
PM 0.83 0.46 0.89
(0.58) (0.33) (0.61)
DIRECTOR 0.39 0.33 0.39
(0.49) (0.41) (0.49)
Adjusted R2 0.34 0.30 0.33
coefficients and t-statistics within parentheses derived from estimating the models using the
OLS method. The results are based on a common sample of 44 firms over a period 2015–2017
(132 firm-year observations). Models 1, 2 and 3 are, respectively, estimated based on
Equations (1), (2) and (3). CSR, actual amount spent on CSR activities during the year; FI,
financial inclusion; ROA, return on assets; SR, stock return; SIZE, firm size; LEVERAGE,
level of debt in capital structure; ATO, asset turnover; PM, profit margin; DIRECTOR,
proportion of independent directors. All variables are explained in Table 1. ***, ** and *
indicate statistical significance at the 1, 5 and 10 percent level, respectively.
Table A3
Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
C 0.01*** 0.01*** 0.01*** 0.01*** 0.01*** 0.01***

(3.47) (3.68) (6.94) (7.44) (5.70) (6.01)
CSR 0.00 0.00 0.00 0.00
(0.63) (0.23) (0.18) (0.31)
FI 0.00*** 0.00 0.00 0.00
(3.48) (0.33) (0.30) (0.32)
SIZE 0.00 0.00 0.00 0.00
(1.53) (1.00) (0.40) (0.85)
LEVERAGE 0.02*** 0.02*** 0.02*** 0.02***
(8.89) (9.00) (8.33) (8.52)
ATO 0.13*** 0.13*** 0.13*** 0.12***
(7.74) (7.91) (7.55) (6.82)
PM 0.09*** 0.09*** 0.09*** 0.09***
(continued)

Model 1 Model 2 Model 3 Model 4 Model 5 Model 6
(43.56) (44.46) (42.54) (41.11)

DIRECTOR 0.00 0.00 0.00 0.00
(0.31) (0.32) (0.31) (0.11)
Adjusted R2 0.01 0.07 0.97 0.97 0.97 0.96
Year fixed effect Included Included Included Included Included Not included
C 2.37*** 1.16*** 1.28 0.55 1.75 1.26

(7.62) (4.20) (1.38) (0.62 (1.59) (1.12)
CSR 0.37*** 0.38*** 0.49*** 0.42**
(7.28) (6.64) (3.38) (2.79)
FI 0.47*** 0.56*** 0.20 0.04
(6.49) (5.56) (0.81) (0.17)
SIZE 0.27* 0.48** 0.53 0.35
(1.71) (2.33) (1.47) (0.94)
LEVERAGE 1.48 0.00 1.88 1.15
(0.99) (0.00) (1.19) (0.71)
ATO 10.78 18.81* 9.29 4.62
(1.03) (1.77) (0.87) (0.42)
PM 0.66 0.39 0.88 0.17
(0.51) (0.30) (0.66) (0.13)
DIRECTOR 0.83 0.75 0.83 1.17
(1.15) (1.00) (1.14) (1.56)
Adjusted R2 0.30 0.25 0.33 0.28 0.33 0.27
Year fixed effect Included Included Included Included Included Not included
coefficients and t-statistics (standard) within parentheses derived from estimating the models
using the OLS method. The sample includes 149 banking firm-years over the period 2015–
2017. Models 1 and 2, respectively, include CSR expenditure and financial inclusion as
explanatory variables; models 3 and 4, respectively, include CSR expenditure and financial
inclusion along with the control variables; and models 5 and 6, respectively, include and
exclude year fixed effects incorporating CSR expenditure, financial inclusion and control
variables. CSR, actual amount spent on CSR activities during the year; FI, financial
inclusion; ROA, return on assets; SR, stock return; SIZE, firm size; LEVERAGE, level of
debt in capital structure; ATO, asset turnover; PM, profit margin; DIRECTOR, proportion
of independent directors. All variables are explained in Table 1. ***, ** and * indicate
statistical significance at the 1, 5 and 10 percent level, respectively.

Is Better Banking Performance Associated With Financial Inclusion and Mandated CSR Expenditure in A Developing Country?

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Is Better Banking Performance Associated With Financial Inclusion and Mandated CSR Expenditure in A Developing Country?

Uploaded by

Copyright:

Available Formats

Accounting & Finance

Is better banking performance associated with ﬁnancial

Asit Bhattacharyyaa , Sue Wrightb , Md Lutfur Rahmana

Motivated by legislation mandating CSR expenditure to improve social

Key words: Mandated CSR spending; Financial inclusion; Performance; Bank

JEL classification: M14, M40, M41, M48

Please address correspondence to Sue Wright via email: sue.wright@uts.edu.au

© 2019 Accounting and Finance Association of Australia and New Zealand

In emerging countries, poverty, poor working conditions and abuse of

© 2019 Accounting and Finance Association of Australia and New Zealand

the global vision of the United Nation’s Millennium Development Goals

© 2019 Accounting and Finance Association of Australia and New Zealand

World Bank). We demonstrate that performance improvements and gains in

2. Theory, literature and hypotheses

According to Freeman’s stakeholder theory, a ﬁrm should consider the

© 2019 Accounting and Finance Association of Australia and New Zealand

© 2019 Accounting and Finance Association of Australia and New Zealand

© 2019 Accounting and Finance Association of Australia and New Zealand

There is substantial evidence on the beneﬁts of greater ﬁnancial inclusion at

© 2019 Accounting and Finance Association of Australia and New Zealand

We examine whether the levels of ﬁnancial inclusion and CSR expenditure

Figure 1 Summary of hypotheses tested.

© 2019 Accounting and Finance Association of Australia and New Zealand

© 2019 Accounting and Finance Association of Australia and New Zealand

4. Data and descriptive statistics

© 2019 Accounting and Finance Association of Australia and New Zealand

Panel A: Firm performance measures

Panel B: CSR expenditure variable

CSR Natural logarithm of actual CSR expenditure

Panel C: Financial inclusion variables

BRANCH Natural logarithm of number of branches

Panel D: Control variables

information company.3 The studies of Gupta (2005), Gopalan et al. (2007),

© 2019 Accounting and Finance Association of Australia and New Zealand

4.2. Descriptive statistics and correlation matrix

Descriptive statistics are presented in Table 2. Panels A, B, C and D,

© 2019 Accounting and Finance Association of Australia and New Zealand

Mean Median Standard deviation 25th percentile 75th percentile

Panel A: Firm performance measures

ROA (%) 0.87 0.73 1.59 0.14 1.40

Panel B: CSR expenditure variable

CSR 192 27 449 5 133

Panel C: Financial inclusion variables

BRANCH 2,258 1,457 2,734 505 3,210

Panel D: Control variables

SIZE 2,440,283 1,315,684 3,734,358 353,184 2,744,731

© 2019 Accounting and Finance Association of Australia and New Zealand

Measures of firm performance

© 2019 Accounting and Finance Association of Australia and New Zealand

Table 3 presents bivariate correlations between the variables considered in

5.1. CSR expenditure and firm performance

We ﬁrst examine the association between ﬁrm performance and mandated

© 2019 Accounting and Finance Association of Australia and New Zealand

Next, we investigate the relation between ﬁnancial inclusion and ﬁrm

© 2019 Accounting and Finance Association of Australia and New Zealand

indicating that very little information is lost in the aggregating process.9

© 2019 Accounting and Finance Association of Australia and New Zealand

The associations of control variables with ﬁrm performance in the presence

© 2019 Accounting and Finance Association of Australia and New Zealand

© 2019 Accounting and Finance Association of Australia and New Zealand

form of misspeciﬁcation, we conduct the regression speciﬁcation error test

5.2. Robustness checks

5.2.1. Individual financial inclusion variables

C 0.01* 0.01* 0.01 0.01* 0.01

C 6.08* 3.77* 11.34* 8.20* 1.39

C 0.01* 0.01* 0.02* 0.01 0.01

C 0.01* 0.01* 0.01***

SIZE 0.42 0.49 0.50

C 0.01* 0.01* 0.01* 0.01* 0.01* 0.01*

C 2.37* 1.16* 1.28 0.55 1.75 1.26