How do Debt Markets React to Mandatory CSR? Evidence from the

How do Debt Markets React to Mandatory CSR?

Evidence from the Indian Companies Act 2013

Jitendra Aswani

N. K. Chidambaran

Iftekhar Hasan

∗

October 2019

Abstract

This paper examines how debt markets price ﬁrms’ Corporate Social Responsibility (CSR) activi-

ties, in a setting where it is not a strategic choice but rather mandated by regulation. Using both

a diﬀerence-in-diﬀerence and a regression discontinuity empirical speciﬁcation, the paper reports

that binding CSR rule increase yield-spreads on bonds by 22 basis points. Firms with a business

group aﬃliation show a reduction in yield spreads and state-owned companies an increase in yield-

spreads. The increase in yield-spreads is mitigated by good governance with well governed ﬁrms

having lower yield-spreads. These ﬁndings add to the contribution of Manchiraju and Rajgopal

(2017), which reports that such mandatory rules are detrimental to stockholders wealth.

JEL: G30

Keywords: Corporate Social Responsibility, Yield Spread, Regression Discontinuity Design (RDD),

Diﬀ-in-Diﬀ, Indian Company Act 2013

.

∗

Aswani ([email protected]), Chidambaran (chidam[email protected]) and Hasan ([email protected])

are at Fordham University, New York.

We thank Vikas Agarwal, Kose John, Nagpurnanand R. Prabhala Thomas Noe, Prasanna Tantri, Pradeep Yadav

and participants at the NSE-NYU conference on Indian Financial Markets and seminar presentations at Fordham for

their feedback and suggestions. We are grateful for a ﬁnancial grant from the 2018 NSE-NYU Stern Initiative on the

Study of Indian Financial Markets.

How do Debt Markets React to Mandatory CSR?

Evidence from the Indian Companies Act 2013

Abstract

This paper examines how debt markets price ﬁrms Corporate Social Responsibility (CSR) activi-

ties, in a setting where it is not a strategic choice but rather mandated by regulation. Using both

a diﬀerence-in-diﬀerence and a regression discontinuity empirical speciﬁcation, the paper reports

that binding CSR rule increase yield-spreads on bonds by 22 basis points. Firms with a business

group aﬃliation show a reduction in yield spreads and state-owned companies an increase in yield-

spreads. The increase in yield-spreads is mitigated by good governance with well governed ﬁrms

having lower yield-spreads. These ﬁndings add to the contribution of Manchiraju and Rajgopal

(2017), which reports that such mandatory rules are detrimental to stockholders wealth.

Keywords: Corporate Social Responsibility, Yield Spread, Regression Discontinuity Design (RDD),

Diﬀ-in-Diﬀ, Indian Company Act 2013

I Introduction

Advocates of Corporate Social Responsibility (CSR) have long argued for a socially responsible

strategy that includes broader participants such as employees, community, environment and lenders.

(e.g. Freeman 1984, Kim, Park, and Wier 2012, Lins, Servaes, and Tamayo 2017, Kitzmueller and

Shimshack 2012, Benabou and Tirole 2010). While CSR diverts cash ﬂows from bondholders and

stockholders, the argument is that ﬁrms should view CSR as a value enhancing corporate strategy

(see e.g. Freeman (1984)) as CSR engenders beneﬁts by way of increased goodwill that enhances

productivity and ﬁrm performance. Whether CSR is value enhancing or is merely costly social

externality, therefore, depends on the beneﬁts relative to the costs. A rich literature has examined

the net impact of CSR on stockholders. In this paper, we add to the empirical evidence by examining

the impact of exogenously imposed CSR on the bond market.

We focus our analysis on the impact of CSR on the bond markets for multiple reasons. We

expect bond markets to be more sensitive to the positive impact of CSR. Simple logic implies

that spending on CSR activities reduces resources available to meet obligations to bond holders.

However, this may be oﬀset by a rise in cash ﬂow due to the increased goodwill from the ﬁrm’s

involvement in the community. Any such positive impact will ﬁrst the bond markets, debt being

the ﬁrst claimant on cash ﬂows. CSR may also impact bondholders diﬀerently than stockholders.

It is plausible that CSR may serve to reduce the volatility of future cash ﬂows resulting in a positive

impact on bondholders but negatively impact shareholders. Finally, we add to the literature on

the impact of CSR on bond markets in a setting when CSR is exogenously mandated.

That CSR is an important issue can be judged by the emphasis on CSR by corporate leaders

and the resources devoted on social investments. The PwC Global CEO Survey 2016 reveals that

64% of CEOs believe CSR is core to their business rather than being a stand-alone program. U.S.

and European markets had over $8.72 trillion and $11.4 trillion in certiﬁed socially responsible

assets in 2015 (Social Investment Forum, 2016). Researchers have previously studied the impact of

CSR on the bond markets. Oikonomou, Brooks, and Pavelin (2014) and Cooper and Uzun (2015)

examine the impact of CSR on credit ratings and the cost of debt. They ﬁnd that ﬁrms engaging in

CSR activities see an increase in credit rating and their cost of debt decreases. Goss and Roberts

(2011) similarly ﬁnd that CSR ﬁrms have a lower cost of bank loans as compared to non-CSR

ﬁrms. Several issues cloud the interpretation of these results as they primarily examine ﬁrms that

1

voluntarily engage in CSR activities. Firms that engage in CSR activities voluntarily are those

that expect to beneﬁt from such activities or are proﬁtable enough to have the resources for CSR

spending.

To resolve the endogeniety issues in analyzing the impact of CSR on debt markets, we use

a unique setting in India. The Indian government incorporated a clause mandating minimum

amounts of CSR spending for proﬁtable ﬁrms as part of the 2013 Company Act. Provisions in the

2013 Act (henceforth the CSR rule) imposed a mandatory requirement that ﬁrms meeting speciﬁc

cutoﬀs with respect to Net Worth, Sales, and Net Proﬁt spend at least 2% of their proﬁt on CSE

related activities. Discussions on the CSR provisions began in 2009 and clauses speciﬁc to requiring

CSR by proﬁtable ﬁrms, were passed by the Lok Sabha in 2012 and was included as Clause 135 of

the 2013 Company Act. We therefore designate 2013 as the CSR YEAR, i.e. the year after which

CSR was required for Indian ﬁrms.

1

The CSR Rule speciﬁes the following cut oﬀs to identify ﬁrms

subject to minimum CSR spending: the ﬁrm should have either (1) a net worth of Indian Rupees

(INR) 5 billion (about U.S. $83 million) or more; (2) sales of INR 10 billion (about U.S. $167

million) or more; or (3) a net proﬁt of INR 50 million (about U.S. $0.83 million) or more. Firms

meeting this criteria are required to spend 2% of their average net proﬁt, calculated over a three

year period, on CSR related activities.

2

The exogenously imposed CSR Rule presents a natural

setting for examining the causal impact of CSR on bondholders.

We obtain data on bond issues by Indian ﬁrms in the eight year period from 2009 to 2017 from

the SDC Platinum Fixed-Income Issues database. We ignore all preferred stock issues and bonds

with contingent features such as step-up and convertible bonds. We augment the bond issue data

with company data from CMIE’s ProwessDx database. The ProwessDx database is widely used

in studies on Indian markets, e.g. Bertrand, Mehta, and Mullainathan (2002), Gopalan, Nanda,

and Seru (2007), Khanna and Palepu (2000) and Manchiraju and Rajgopal (2017) for conducting

research on large sample of Indian ﬁrms. As there is no common identiﬁer in between SDC and

1

Appendix A gives more details on the time line of the passage of the Indian Companies Act. Manchiraju and

Rajgopal (2017) evaluate the market reaction for announcements at the various dates indicated in the time line and

ﬁnd that 2012 was associated with a signiﬁcant negative reaction to the passage of the CSR provision in 2012.

2

Approved CSR activities comprises of: (i) eradicating extreme hunger and poverty; (ii) promotion of education;

(iii) promoting gender equality and empowering women; (iv) reducing child mortality and improving maternal health;

(v) combating HIV, AIDS, malaria and other diseases; (vi) ensuring environmental sustainability; (vii) employment-

enhancing vocational skills; (viii) social business projects; (ix) contribution to the Prime Ministers National Relief

Fund or any other fund set up by the Central Government or the state governments for socioeconomic development,

and relief and funds for the welfare of the scheduled castes, the scheduled tribes, other backward classes, minorities

and women; and (x) such other matters as may be prescribed.

2

prowess database, we hand match the two datasets using the name of the ﬁrm. We are able to

match data for 236 ﬁrms with 3,466 bond issues over the nine year period from 2009 to 2017.

We next apply the ﬁlters speciﬁed by the 2013 companies act to determine whether a ﬁrm is

aﬀected by the mandatory CSR rules. The dummy variable AFFECTED is set equal to one if a

bond is issued by a ﬁrm that is aﬀected by the CSR Rule and zero otherwise. Of the three criteria,

proﬁt and net worth are the primary determinants of whether a ﬁrm is subject to mandatory CSR

spending. We ﬁnd that there are 3,357 bonds issued by ﬁrms aﬀected by the mandatory CSR

Rule and 109 bonds issued by ﬁrms not aﬀected by the rule. We use AFFECTED to isolate the

eﬀects of CSR on bond yield-spreads and implement two empirical speciﬁcations. First, we use a

Regression Discontinuity Design to capture the diﬀerential eﬀects of the CSR Rule on ﬁrms that

just meet the CSR cutoﬀ to those that just miss the CSR cutoﬀ. Any CSR eﬀects is likely to be

the sharpest between these two sets of ﬁrms. We ﬁnd that the yield spreads on bonds issued by

ﬁrms that just meet the criteria is higher than the yield spreads on bonds that just miss the CSR

criteria. The impact of mandatory CSR on yield-spreads is positive and signiﬁcant. Second, we use

a Diﬀ-in-Diﬀ approach to examine the impact of the passage of the CSR rule in 2013. Our focus is

on the interaction between the variable AFFECTED and the time dummy POSTCSR that is equal

to one for bonds issued int he period 2013-2016 and zero for bonds issued in the period 2008-2012.

In running our regression, we also control for industry ﬁxed eﬀects. A large percentage of bonds

issued in India are by banks and industry ﬁxed eﬀects controls for unobservable characteristics

across industries. We also control for bond characteristics, such as bond ratings and maturity,

and for ﬁrm characteristics. We ﬁnd that the coeﬃcient on the interaction term AFFECTED and

POSTCSR is positive and signiﬁcant. Both the Regression Discontinuity Design and Diﬀ-in-Diﬀ

approaches thus show that yields on bonds are higher in the POSTCSR period for ﬁrms aﬀected

by the CSR Rule. To explore this further, we individually examine the impact of each of the three

criteria used to determine whether a ﬁrm is subject to the CSR mandate. We ﬁnd that bonds

issued by ﬁrms that are subject to mandatory CSR spending based on the individual criteria also

have higher spreads.

Our tests on bonds issued by ﬁrms subject to the individual criteria rather than the collective

criteria also address another important methodological issue. Diﬀ-in-diﬀ and RDD tests assume

that the CSR Rule treatment eﬀects, i.e. the the sample of treated ﬁrms subject to the CSR Rule

and the sample of non-treated ﬁrms, are truly exogenous. While manager’s have discretion on

3

reported income, it is less likely that manager’s are able to simultaneously manipulate the total

revenue of the ﬁrm, its net worth and its total proﬁt. The treatment eﬀect is therefore truly

exogenous in our speciﬁcation.

Our ﬁndings indicate that that mandatory CSR has a detrimental eﬀect on the bond market

and the negative eﬀects of CSR therefore dominate. Mandatory CSR reduces the ﬂexibility ﬁrms

have in using their cash ﬂows to meet debt obligations. Such lack of ﬂexibility has a negative impact

on the market’s perceptions of bond value, leading to a higher cost of capital for ﬁrms. Moreover,

allowed types of CSR activity beneﬁts society without beneﬁtting the ﬁrm. As noted by John,

Nair, and Senbet (2005), socially conscious investments that generate non-monetizeable beneﬁts

to society but are negative NPV for the ﬁrm should be rejected by the ﬁrm. If such projects are

mandated, shareholders and bondholders can lose value. Mandatory CSR may also exacerbate the

moral hazard between insider managers and shareholders. In a mandatory CSR environment, ﬁrms

have to pick from an approved list of acceptable CSR activities to show compliance. Several of the

approved avenues for CSR could allow for private beneﬁts to an insiders.

3

It is plausible that ﬁrms that are not subject to the CSR Rule voluntarily engage in CSR

activity because CSR spending is optimal. Given the lack of data on CSR spending, we cannot

identify and exclude ﬁrms that voluntarily engage in CSR activity or examine them separately in

our tests. This raises two issues. One, some ﬁrms in our treatment sample would have engaged

in CSR voluntarily and as such should not experience the negative impact of reduced ﬂexibility.

That is, while CSR activity is mandated for treatment ﬁrms, the CSR activity matches what would

have been optimal anyways. This eﬀect biases against ﬁnding signiﬁcant eﬀects from the tests as

designed. Our results are however signiﬁcant in spite of such confounding issues. Two, the lower

credit spreads for some of the bonds issued by ﬁrms in the untreated sample could be because

of their voluntary CSR activities. Our basic point however continues to hold – requiring ﬁrms to

engage in CSR increases bond yield spreads compared to the yield spreads on bonds issued by ﬁrms

that are not required to engage in CSR activity.

We extend our analysis of yield-spreads by examining the diﬀerential impact of the CSR

Rule by analyzing carefully constructed subsamples based on shareholder ownership and corporate

governance. In addition to comparing aﬀected and unaﬀected ﬁrms, we also analyze the cross-

3

Reports in the popular press, e.g. The Economic Times on Oct 21, 2015; The Guardian on Apr 05, 2016; and

The Wire on Dec 22, 2018; have alleged that some ﬁrms have misused the CSR funds for private and political gain

through Trusts that are lightly supervised.

4

section of aﬀected ﬁrms in these tests. In our ﬁrst set of extensions, we contrast bonds issued by

ﬁrms that have concentrated shareholding, bonds issued by ﬁrms aﬃliated with business groups),

and bonds issued by government owned ﬁrms, the government owned. Promoter holdings in India

represent the stake in the ﬁrm held by the original founder/promoter of the ﬁrm and is many

ﬁrms have a higher fraction of shares held by the original founder. CONC HLDG is equal to

one if the level of promoter holding is greater than 52% the median value for the sample, and

zero otherwise. We ﬁnd that CONC HLDG does not aﬀect yield spreads for aﬀected ﬁrms in

the POSTCSR period. The promoters shareholding therefore does not contribute, or ameliorate,

the adverse impact of CSR. We use a dummy variable BG, which is equal to one for ﬁrms with

business group aﬃliations, to examine the impact of group aﬃliation. Bonds issued by ﬁrms that

are aﬃliated to a business group have lower yields post CSR. Membership in business groups can

result in best practices with respect to CSR and also allow the group ﬁrms to collaborate on an

eﬀective CSR strategy. Finally, we identify bonds issued by government owned ﬁrms. We deﬁne a

dummy variable GOVT OWNED, which is equal to one if the ﬁrm is governed owned. We ﬁnd that

yield spreads are higher in the POSTCSR period for government owned ﬁrms. Bonds issued by

state owned ﬁrms and stand alone ﬁrms unaﬃliated with a business group, therefore, are negatively

aﬀected by the mandatory CSR Rule.

In our second set of extensions, we contrast ﬁrms with good governance and poor governance.

One measure of good governance is BI, a dummy variable that is equal to one if the bond issuer has

above median board independence and zero if the level of board independence is below the median.

Firms that have a larger fraction of their directors who are independent are considered to be

better governed, with the fraction of independent directors on the board used as a measure of good

governance. Another measure of good governance we use is BIG4, which is a dummy variable equal

to one for ﬁrms that are audited by aﬃliates of leading multinational auditing ﬁrms. Aﬃliates of

large multinational accounting ﬁrms have reputational concerns that ensures high quality audits of

their clients. Having an aﬃliate of a multinational ﬁrm can, therefore, improve external monitoring

and is a measure of good governance. We ﬁnd that yield-spreads are lower for bonds issued by

ﬁrms that have a value one for these metrics, i.e. for better governed ﬁrms. These results suggest

that better governed companies may be better able to target their CSR spending and maximize

the strategic beneﬁts of CSR activity.

5

In summary, our ﬁndings are consistent with CSR reducing the cash available to meet obliga-

tions and increases the perceived costs of ﬁnancial distress. Good governance, and group aﬃliation

mitigates these eﬀects, and suggests that well governed ﬁrms may be better able to beneﬁt from

the positive eﬀects of CSR activity. One size, thus, does not ﬁt all. Our result augments the

results of Manchiraju and Rajgopal (2017) who show that mandatory CSR activity reduces wealth

of stockholders.

The rest of the paper is organized as follows. The next section provides a brief review of

the literature. Section 3 describes our data and methodology. Section 4 presents our results and

Section 5 concludes.

II Relevant Literature

The impact of CSR on value to insiders of the ﬁrm has been a subject of much debate in the

literature. As CSR diverts the cash ﬂows of the ﬁrm to other stakeholders it is a negative externality

for shareholders and bondholders. Adam Smith and Milton Friedman ﬁnd little use for CSR and

have argued that ﬁrms should only focus on proﬁt maximization. In their view, even CSR that

seeks to mitigate the negative eﬀects of business, e.g. pollution, should be left to the government

and other institutions. Executives engaging in social activities, therefore, do so at the expense of

shareholders.

An alternate stream of thought sees CSR as a strategic activity that can add to ﬁrm value.

Freeman (1984) and Kitzmueller and Shimshack (2012) argue that CSR engenders beneﬁts for the

shareholders and bondholders indirectly, by way of increased goodwill that enhances productivity

and ﬁrm performance. If executives engage in CSR activities, they will Do well by doing good.

Benabou and Tirole (2010) conclude that CSR, if strategically used, can be a win-win situation for

both society and the ﬁrm.

Whether CSR is solely a negative externality or if it indeed can result in value increases for

shareholders and bondholders is an empirical issue. Researchers have tried to explain the relation

between CSR and ﬁrm performance in diﬀerent settings and through diﬀerent channels. Lys,

Naughton, and Wang (2015) checks three hypotheses: CSR as charity, CSR as investment, and

CSR as signal for ﬁrms future performance. Using the sample of Russell 1000 ﬁrms over the

sample period from 2002 to 2010, they conclude that ﬁrms undertake CSR expenditures when they

6

anticipate stronger future ﬁnancial performance i.e. CSR is used as a signaling mechanism for

ﬁrms future performance. Martin and Moser (2016) explain the relation between CSR and ﬁrm

performance through experimental economics. They conduct an experiment with 90 participants

from which they randomly chose groups of 5 participants. Of this group of 5 people, one act as

manager, another act as current shareholder, and remaining three act as potential investors. The

reaction of investors is noted when the manager discloses a report on green investments and details

on the level of carbon emission and other pollutants. They ﬁnd that potential investors respond

more positively to voluntary disclosures of a green investment. They also ﬁnd that managers and

shareholders are willing to bear the marginal cost of the project in order to provide societal beneﬁts.

While these studies have intuitive results, it is diﬃculty to infer causality. The level of CSR

activity is an endogenous choice of the ﬁrm and it is likely that it is the well performing ﬁrms who

engage in CSR activity. It is therefore hard to disentangle the eﬀect of CSR from the strategic

investment behavior of the management. Several researchers have tried to resolve the endogeniety

issue and the associated reverse causality problems. Flammer (2015) use data on shareholder spon-

sored CSR proposals that pass or fail by a small margin of votes at annual meetings as an exogenous

shock. They calculate the cumulative abnormal returns surrounding the shareholder proposal vote

for these proposals to infer the impact of exogenous shocks to CSR on ﬁrm performance. Using

a diﬀerence-in-diﬀerences and RDD techniques, they ﬁnds that CSR signiﬁcantly increases ﬁrm

performance. Similarly, Lins, Servaes, and Tamayo (2017) show that ﬁrms with CSR activities

provided better returns than others at the time of crisis. Studies have also examined the impact

of CSR on Analyst recommendations.Albuquerque, Durnev, and Koskinen (2013)) show that ﬁrms

with CSR activities have more positive sell-side analysts recommendations. Bushee and Noe (2000)

show that CSR ﬁrms have higher abnormal returns and Deng, koo Kang, and Low (2013) show

that they have higher long-term post acquisition returns.

Researchers have also use mandates on the disclosure of CSR to resolve endogeniety concerns.

Chen, Hung, and Wang (2017) examines how mandatory disclosure of CSR impacts ﬁrm perfor-

mance in China. China made CSR disclosure mandatory for a subset of ﬁrms from 2008. Using

this an exogenous shock, the authors examine how CSR activities aﬀects the ﬁrm proﬁtability and

social externalities for ﬁrms listed on Shanghai Stock Exchange (SSE) and Shenzhen Stock Ex-

change (SZSE) between 2006-2011. They ﬁnd that the treatment ﬁrms experience a decrease in

return on assets (ROA) and return on equity (ROE), sales revenue and increase in operating costs

7

and impairment charges. On the other hand, Industrial waste water discharge and the level of SO2

emission reduced after the CSR disclosure mandate.

Manchiraju and Rajgopal (2017) use the unique setting of mandatory disclosure in India to

examine the impact on shareholder value. Clause 135 of Indian Companies Act 2013 requires ﬁrms

that cross any of the three individual thresholds – (1) a net worth of Indian Rupees (INR) 5 billion

(about U.S. $83 million) or more; (2) sales of INR 10 billion (about U.S. $167 million) or more; or

(3) a net proﬁt of INR 50 million (about U.S. $0.83 million) or more – have to spend a minimum

of 2% of their net proﬁt averaged over three years, on CSR activities. The clause was ratiﬁed and

passed by the Lok Sabha of the Indian Parliament in 2012 and Manchiraju and Rajgopal (2017)

ﬁnd a signiﬁcant negative abnormal return associated with the passage of the CSR rule in 2012.

While the average impact of CSR on ﬁrms that would have and would not have engaged in CSR

voluntarily, companies that advertise their CSR activity do not have a negative abnormal return.

They conclude that involuntary CSR has a negative impact on shareholder value.

All the articles cited above examine the impact on shareholders, but is useful to examine the

eﬀect on bondholders as well. Oikonomou, Brooks, and Pavelin (2014) andCooper and Uzun (2015)

examine the impact of CSR on credit ratings and the cost of debt. They ﬁnd that if ﬁrm engages in

CSR activities then their credit rating increases and their cost of debt decreases. Goss and Roberts

(2011) similarly ﬁnd that CSR ﬁrms have a lower cost of bank loans as compare to non-CSR ﬁrms.

As in the case of studies that examine the impact of CSR on shareholder value, these studies are

subject to the endogeniety problem and reverse causality.

Our paper extends these studies by examining the cost of debt for Indian ﬁrms before and after

the passage of the CSR provisions in 2012 and included in the 2013 Companies Act. We focus on the

bond markets for several reasons. If ﬁrms spend 2% of proﬁt on CSR, it reduces the funds available

to pay shareholders and debtholders, thereby making the debt riskier. Any positive impact of CSR

would necessarily ﬁrst impact bondholders when debt is risky. We therefore expect bond markets

to be more sensitive to the positive impact of CSR. CSR may also impact bondholders diﬀerently

than stockholders. It is plausible that CSR may serve to reduce the volatility of future cash ﬂows

resulting in a positive impact on bondholders but negatively impact shareholders. Finally, we add

to the literature on the impact of CSR on bond markets in a setting when CSR is exogenously

mandated.

8

Our sample consists of bonds issued by Indian ﬁrms for which ﬁrm level data is available on

Global Compustat. We examine the impact of CSR on bonds issued in the ﬁve-year period from

2009 to 2013, the PRECSR period, and the four-year period from 2014-2017, the POSTCSR period.

We use a Regression Discontinuity Design that examines ﬁrms that are just above and just below

the cutoﬀs for mandatory CSR. We also use a Diﬀ-in-Diﬀ approach to compare the yield on bond

issues before and after 2012 for ﬁrms aﬀected by the CSR rule and those that are not aﬀected by

the CSR rule.

III Data

The bond issues data for our study are obtained from the SDC Platinum database. SDC Platinum

reports the issue of ﬁxed-income securities by Indian ﬁrms in the nine-year period from 2009 to

2017, across several categories. The sample mostly comprises debentures, ﬁxed/straight bonds,

secure bond/debentures, subordinate bond/debentures, and zero-coupon bond/debentures. We

exclude all issues that are preferred stock issues, bond issues that have contingent features, and

when the yield-to-maturity is not reported.

We obtain data on ﬁrm fundamentals from the Center for Monitoring Indian Economy (CMIE)

ProwessDx database. As there is no common identiﬁer in between SDC and prowess database, we

hand match the two datasets using the name of the ﬁrm. We are able to match data for 236 ﬁrms

and 3,466 observations. All the ﬁrm fundamentals are from consolidated ﬁnancial statements data

of ProwessDx. Table 1, Panel A, shows the yearly distribution of these observations. Year 2011

saw the largest number of bond issues. There is no monotonic pattern in the number of bond issues

each year.

Our focus is on the impact of the CSR provision passed in the Lok Sabha of the Indian

Parliament mandating a minimum level of CSR spending. We therefore classify ﬁrms as being

aﬀected by the CSR Rule in each year. Speciﬁcally, we label a ﬁrm as AFFECTED by the CSR

rule if during any ﬁscal year a ﬁrm has either (i) a net proﬁt of INR 50 million (about U.S. $0.83

million) or more; (ii) a net worth of Indian Rupees (INR) 5 billion (about $83 million) or more; or

(iii) sales of INR 10 billion (about U.S. $167 million) or more.

We construct three variables, R1, R2, and R3, that can be used to determine whether the ﬁrm

is subject to CSR requirements using the three measures individually. The three variables R1, R2,

9

and R3 are centered around the cutoﬀ threshold for proﬁt, net worth, and sales, respectively, and

are expressed as the percentage diﬀerence from the cutoﬀ threshold as follows:

R1 =

P RET AX INCOM E − 50

50

(1)

R2 =

NET W ORT H − 5, 000

5, 000

R3 =

T OT AL REV ENUE − 10, 000

10, 000

with all values speciﬁed in INR millions.

To capture the requirements under the 2013 Company Act, we create a measure M that is

equal to the minimum positive value of R1, R2, or R3, if at least one of the three variables is

positive and is the maximum value if all three individual variables R1, R2, and R3 are negative.

We calculate M as follows:

M =



















































































min(R1, R2, R3) if R1 ≥ 0&R2 ≥ 0&R3 ≥ 0

min(R1, R2) if R1 ≥ 0&R2 ≥ 0&R3 < 0

min(R2, R3) if R1 < 0&R2 ≥ 0&R3 ≥ 0

min(R1, R3) if R1 ≥ 0&R2 < 0&R3 ≥ 0

R1 if R1 ≥ 0&R2 < 0&R3 < 0

R2 if R1 < 0&R2 ≥ 0&R3 < 0

R3 if R1 < 0&R2 < 0&R3 ≥ 0

max(R1, R2, R3) if R1 < 0&R2 < 0&R3 < 0

(2)

All four variables, M , R1, R2, and R3, are continuous variables. The critical cutoﬀ value for all four

binding score measures, M, R1, R2, and R3, is equal to zero. We therefore develop the following 4

variables to designate ﬁrms aﬀected by the 2013 Company Act. AFFECTED that takes a value of

1 if M > 0, AFFECTED R1 that takes a value of 1 if R1 > 0, AFFECTED R2 that takes a value

of 1 if R2 > 0, and AFFECTED

R3 that takes a value of 1 if R3 > 0.

M and AFFECTED are our primary measure to reﬂect the requirements of the 2013 Act. For

robustness, we run our empirical tests using the collective criterion M and the component speciﬁc

criteria R1, R2, and R3. Diﬀ-in-diﬀ and RDD tests require that the CSR Rule treatment eﬀects are

truly exogenous. While manager’s have discretion on reported income, it is ulikely that manager’s

10

are able to simultaneously manipulate the total revenue of the ﬁrm, its net worth and its total

proﬁt.

Table 1, Panel B, shows the distribution of bonds issued by Aﬀected/Unaﬀected ﬁrms based

on the criteria described above. There are 3,357 bonds issued by AFFECTED ﬁrms and 109 bonds

issued by ﬁrms not aﬀected by the mandatory CSR spending requirement. We examine the three

individual criteria as well. There are 3,306 bonds issued aﬀected by the Net Proﬁt criteria, 2,576

bonds issued by ﬁrms aﬀected by the Net Worth criteria, and 492 ﬁrms aﬀected by the sales criteria.

It is clear, therefore, that the proﬁt and net worth criteria are the primary criteria for determining

whether a ﬁrm is subject to Mandatory CSR spending.

We also note that the CSR Rule was passed in the year December 2012 and subsequently

included in the 2013 Indian Company Act. We therefore label the period from 2009 to 2013 as

the PRECSR period and the period from 2014 to 2017 as the POSTCSR period. A total of 2,035

bonds were issued in the PRECSR period, of which 1,352 bonds were issued by AFFECTED ﬁrms.

A total of 1,431 bonds were issued in the POSTCSR period, of which 2,005 bonds were issued by

AFFECTED ﬁrms

The dependent variables for this study is the yield spread, deﬁned as the oﬀer yield to maturity

minus a reference treasury bill rate. In determining the appropriate benchmark treasury rate, we

match the maturity of the bond to the maturity of the treasury bonds Speciﬁcally, if the bond’s

maturity is 1 year or less, we use 1-year T-bill rate, if securitys maturity is 5 years but greater than

1 year then we use 5-year treasury rate

4

. Table 2 presents data on the yield-spread and other bond

and issuer characteristics for the bonds and issuers in our sample. As shown in the table, the mean

(standard deviation) of yield spread is 2.17% (1.985%).

The control variables used for this study are SIZE, LEVERAGE, TOBIN Q, CREDIT RANK,

and MATURITY. Firm size, leverage, and Tobin Q are calculated as log of total assets, ratio of

long term debt to total assets, and ratio of market value to book value and are for the lagged ﬁscal

year. All three variables are winsorized at the 1% level. CREDIT RANK, is a rank for securities

rating grade assigned based on ratings given by diﬀerent rating agencies. In India, mainly three

rating agencies, CARE, ICRA, and CRISIL dominates the market and provides the ratings to

various corporate securities. IProwessDx provides a rating grade, that ranks the bond in terms of

4

We include perpetual bonds in our sample. As perpetual bonds do not have a maturity, we set the maturity of

the bond to be 100 years and the corresponding treasury rate to be that of the longest maturity treasury security.

11

safety, in eight steps: Highest Safety, High Safety, Moderate Safety, Adequate safety, Inadequate

Safety, Substantial Risk, High Risk, and Default. We convert these rating grades into ranks from

8 (Highest Safety) to 2 (High Risk) and develop a CREDIT

RANK measure.

5

As seen in Table

2, CREDIT ANK is available for 3,037 ﬁrms. The average credit rank is 7.479% with a standard

deviation of 0.641%. Bond maturity is calculated as diﬀerence between issued year and maturity

year. The mean(median) maturity is 7.879 (3.000) years, suggesting that Indian ﬁrms largely issue

short-term bonds, with some ﬁrms issuing long-term bonds.

IV Research Design

To examine the eﬀect of CSR rule on ﬁrms cost of debt, we compare the yield to maturity of

AFFECTED ﬁrms to those of UNAFFECTED ﬁrms in the PRECSR and POSTCSR period. We

use two approaches to make the comparison, a Diﬀ-inDiﬀ approach and a Regression Discontinuity

Design (RDD).

A Diﬀ-in-Diﬀ Design

We use a diﬀ-in-diﬀ speciﬁcation around the CSR Year, year 2013. We capture the impact of the

passage of the law using the following regression speciﬁcation:

Y ieldSpread = α + β

1

AF F ECT ED + β

2

P OST CSR + β

3

AF F ECT ED X P OST CSR +(3)

γX + IN DU ST RY F IXED EF F ECT S + 

Of interest is β

3

, the coeﬃcient on the interaction term AF F ECT EDXP OST CSR, that

captures the impact of the exogenous imposition of a minimum level of CSR expenditure. X

represents the set of ﬁrm and bond level controls in the regression.

For robustness, we also use component speciﬁc measures AFFECTED R1, AFFECTED R2,

and AFFECTED R3 and interaction variables to capture their interaction with POSTCSR.

5

We We drop one ﬁrm from the sample, issued by Bhushan Steel, from our sample.

12

B Regression Discontinuity Design

We also use a regression discontinuity design (RDD) to document the eﬀect of CSR rule on yield-

spreads The RDD technique has been used in various studies in accounting and ﬁnance (see e.g.

Flammer (2015), Manchiraju and Rajgopal (2017) and Iliev (2010)) for resolving the diﬀerent

endogeniety concerns. RDD is applicable in our case as it is diﬃcult for ﬁrms to manipulate

three diﬀerent thresholds used to determine whether they fall under the minimum CSR spending

requirement.

The inference drawn under an RDD approach are considered to be credible because the as-

signment of individuals in treatment and control groups is “as good as randomized” given that

individuals cannot precisely control the assignment variable near the exogenously determined cut-

oﬀs (Lee and Lemieux 2010).

Our research setting diﬀers from the basic RDD applications listed above, in that the manda-

tory CSR rule relies on more than one rating score to determine treatment status. We, therefore,

implement multivariate RDD (MRDD). To estimate the treatment eﬀects under MRDD, we follow

the binding-score regression discontinuity method of Reardon and Robinson (2012). The Reardon

and Robinson (2012) methodology is intuitively simple and easy to use and reframes the multi-

dimensional vector of rating scores into a single dimension for determining treatment status, and

hence ensures minimal loss of data in estimation.

We ﬁt the following model using zero as the critical value of the binding score.

Y ieldSpread = α + f (M ) + γX +  (4)

We use the overall criteria M as the binding score in our regressions. We also use the individual

components R1, R2, and R3 as alternate metrics for robustness. Firms to the right of the cutoﬀ

score of zero are AFFECTED by the mandatory CSR rule whereas ﬁrms to the left of zero are

UNAFFECTED. A discontinuity in YIELD SPREAD at the cutoﬀ captures the impact of the

exogenously imposed mandate for minimum CSR spending.

To implement MRDD we use RDPLOT and RDROBUST command in STATA Calonico,

Cattaneo, Farrell, and Titiunik (2017) using M as the binding score variable. The RDPLOT

command generates the plots and RDROBUST tabulates the results. We run MRDD regressions

for full sample and for sample before and after CSR rule, controlling for other covariates, with

13

Epanechnikov weighting scheme, and a bandwidth of 10 in both the left and right side of the

cut-oﬀ point, M = 0. Our results are robust to other choices in these parameter values.

V Empirical Results

In this section we discuss the results of our diﬀ-in-diﬀ regressions and our RDD regressions.

A Base Results

Table 3 shows the results for diﬀ-in-diﬀ regressions. Eight models are presented. Columns 1-4

presents results when controlling only for whether ﬁrms are aﬀected by the CSR Rule, the post-

CSR rule period from 2013-2017, and the interaction term between whether ﬁrms are aﬀected and

the post CSR rule period. Column 1 uses the overall metric AFFECTED that uses the three

criteria depending on net worth, and sales, as the metric for determining whether the ﬁrm is

aﬀected by the CSR Rule. Columns 2-4 present the case when the individual measures alone are

used to determine whether ﬁrms are aﬀected by the CSR rule. In these regressions we replace the

variable AFFECTED with AFFECTED R1, AFFECTED R2, and AFFECTED R3, respectively.

POSTCSR is set equal to one for the post CSR period. All the regressions are run with industry

ﬁxed eﬀects based on the Fama-French 30 industry classiﬁcations. Models 5 through 8, augment

the regressions in Models 1-4 with control variables.

We ﬁnd that the coeﬃcient on POSTCSR is negative in all models, but is signiﬁcant in only

three of the eight models. This suggests that yield spreads are somewhat lower for bonds issued

after 2013, i.e. in the period from 2014-2017. The coeﬃcient on the AFFECTED measures are also

negative but are signiﬁcant only models 1, 3, and 4, i.e. in the regressions without controls for bond

and ﬁrm characteristics. As ﬁrms that meet the criteria for the CSR Rule are large and proﬁtable,

it is not surprising that the yields are somewhat lower for AFFECTED ﬁrms, but the variation is

explained by bond and ﬁrm characteristics. The interaction variables between AFFECTED and

POSTCSR, our variable of interest that captures the impact of exogenously imposed CSR activity,

is positive and statistically signiﬁcant in seven of the eight models. Thus the impact of the CSR

Rule is robust to using the overall measure M or the individual components R1, R2, or R3, in

determining whether ﬁrms are aﬀected by the CSR Rule and with and without controls. The

magnitude of the coeﬃcient vary and reﬂect the variations in the sample of ﬁrms that would be

14

aﬀected by the CSR Rule using these alternate speciﬁcations. Using the criteria for mandatory

CSR as speciﬁed by the 2013 Company Act, i.e. ﬁrms are subject to mandatory CSR if at least

one of the measures is positive, we ﬁnd that yield-spreads increase by 0.22% or 22 basis points for

ﬁrms aﬀected by the CSR Rule in the POSTCSR period. The result is consistent with mandating

CSR expenditure reduces future cash ﬂows and higher perceived costs of ﬁnancial distress.

The table also shows the sign and signiﬁcance of the control variables. Bonds issued by larger

ﬁrms have lower credit spreads as large ﬁrms are likely to have larger future cash ﬂows and have

have more assets that can serve as collateral. Interestingly bonds issued by ﬁrms that have higher

leverage have lower spreads. This is perhaps because ﬁrms with higher leverage have greater debt

capacity. The coeﬃcient on CREDITRANK is negative and signiﬁcant. As expected, higher rated

ﬁrms have lower spreads.

The economic signiﬁcance of the CSR Rule can be examined by adding the coeﬃcients on

POSTCSR (-0.825) and the interaction term (1.046), which is equal to 0.221. While there is a

general decline post 2013, perhaps reﬂecting robust economic growth in India, the oﬀsetting eﬀects

of mandatory CSR spending increases the cost of capital by 22 basis points.

The 22 basis point increase in yield spreads indicates the signiﬁcant causal economic eﬀect of

mandated CSR. Earlier studies have shown that voluntary CSR can have positive eﬀects on bond

markets, but such studies are subject to endogeniety and reverse causality issues. Our empirical

design resolves these econometrics issues and allow us to examine the value impact on the bond

markets of CSR spending.

Our ﬁndings indicate that that mandatory CSR has a detrimental eﬀect on the bond market.

While the mandate for CSR spending targets ﬁrms that are large and proﬁtable, the lack of ﬂexi-

bility has a negative impact on the market’s perceptions of the ﬁrms ability to meet its obligations

to debt holders.

Our ﬁndings indicate that that mandatory CSR has a detrimental eﬀect on the bond market.

The mandate for CSR spending targets ﬁrms that are large and proﬁtable, but the lack of ﬂexibility

reduces the cash ﬂows available to the ﬁrm to meet debt obligations. Such lack of ﬂexibility

has a negative impact on the market’s perceptions of bond value, leading to a higher cost of

capital for ﬁrms. Mandatory CSR also exacerbates the moral hazard between insider managers and

shareholders. Firms have to pick from an approved list of acceptable CSR activities, which could

allow for private beneﬁts to the insider manager. It is also plausible that all allowed types of CSR

15

activity beneﬁts society without beneﬁtting the ﬁrm. As noted by John, Nair, and Senbet (2005),

socially conscious investments that generate non-monetizeable beneﬁts to society but are negative

NPV for the ﬁrm should be rejected by the ﬁrm. If such projects are mandated, shareholders and

bondholders can lose value.

We next examine the importance of the 2013 Company Act by using a regression discontinuity

framework. Given that in the post CSR period, there is an exogenously imposed criteria for

mandating a minimum CSR expenditure, we expect to see a discontinuity around the scale measures

centered around zero. That is indeed what we ﬁnd. Table 4 and Figure 1 present results and plots

for the binding score MRDD. As Table 4 shows, the coeﬃcient on the RDD variable, POSTCSR,

is positive and highly statistically signiﬁcant. Figure 1 shows the RDD plots. As Panel A of the

ﬁgure shows, there is discontinuity at M = 0 in the overall sample. Yield-spreads for bond issued by

ﬁrms that just meet the CSR criteria are higher than the yield-spreads for ﬁrms that just miss the

criteria. Panel B and Panel C show similar RDD plots for bond issues in the PRECSR period and

POSTCSR period respectively. As shown in these panels, the increased spreads and discontinuity

at M = 0 only exists in the POSTCSR period when the CSR mandate kicks in and not in the

PRECSR period. Our RDD tests thus also conﬁrm that yield-spreads increased after the passage

of the CSR rule.

B Extensions

We augment our base speciﬁcation with subsample analysis to more fully capture the impact of

the CSR Rule. We extend our base case analysis to consider whether ownership structure of bond

issuers impact on bond yield-spreads. We also examine the impact of corporate governance features

on yield-spreads. Both ownership structure and good governance can impact on the strategic use

of CSR spending to maximize potential beneﬁts.

The extensions we examine also serve another purpose. As Table 1, Panel B shows, the sample

of bonds sold by ﬁrms that are not aﬀected by the CSR Rule is relatively small, only 109 of the total

3,466 bond issues are sold by ﬁrms not aﬀected by the CSR Rule. We therefore compare subsamples

of bonds sold by AFFECTED ﬁrms classiﬁed by exogenous variation in order to conﬁrm our results

and examine the cross-sectional characteristics of AFFECTED ﬁrms that impact on bond yield-

spreads.

16

B.1 Ownership Structure

We develop three measures to capture the diﬀerences in ownership structure across ﬁrms. One,

we deﬁne a dummy variable CONC HLDG, which is equal to one if the shareholding of the ﬁrm’s

promoters

6

is greater than the median promoter holdings in the sample.

Two, we develop a dummy variable GOVT OWNED, which is equal to one if either the central

Indian government or the governments of individual states have an equity stake in the ﬁrm.

Three, we develop a dummy variable BG, which is equal to one if the ﬁrm is aﬃliated with a

business group. Group ownership can bring professional management and best practices all their

aﬃliate, improving corporate governance at member ﬁrms. Group aﬃliation can also impose costs

on ﬁrms because of potential agency problems between shareholders representing the group and

non-group shareholders. Such intra-shareholder agency concerns are, however, less relevant to the

impact of CSR on debt markets. We argue therefore that group membership should allow ﬁrms

to better deal with the externalities imposed by the CSR Rule. To examine the impact of group

membership, we therefore use a dummy, BG, that is equal to 1 if the ﬁrm is aﬃliated with a business

group, and otherwise zero.

Our base results show that mandatory CSR spending increases yield-spreads as the direct

reduction in cash ﬂows is greater than perceived beneﬁts from strategically targeting CSR spending.

However, ﬁrms can diﬀer in the ability to utilize their CSR spending strategically. We explore such

cross-sectional diﬀerences by examining the impact of CONC HLDG, and GOVT OWNED, and

BG variables.

We run two sets of models. In our ﬁrst set, we use all data and examine the coeﬃcient on a triple

interaction terms between AFFECTED, POST CSR, and CONC HLDG/Govt Hldg/BG. In our

second set of models, we use only bonds issued by AFFECTED ﬁrms and examine the interaction

terms between AFFECTED and CONC HLDG/GOVT OWNED/BG. Table 5 presents the results

of our tests. Columns 1-3 present models for the full sample and Column 4-6 present results

when only using AFFECTED ﬁrms. As before, we use several control variables and industry ﬁxed

eﬀects. We ﬁnd that the coeﬃcient in the triple interaction terms diﬀer for the diﬀerent ownership

measures. The coeﬃcient on the triple interaction term when using CONC HLDG is insigniﬁcant

suggesting the promoter holdings are not an important determinant on how debt markets react to

6

Founders of Indian ﬁrms are designated as promoters of the ﬁrm. Promoters of Indian ﬁrms tend to hold a large

fraction of shares, even after the ﬁrm goes public.

17

the mandatory CSR Rule. The coeﬃcient on the triple interaction term with GOVT OWNED is

positive and signiﬁcant and suggest that state owned ﬁrms are not as eﬃcient in strategically using

their mandatory CSR spending. The coeﬃcient on the triple interaction term with BG is negative

and signiﬁcant. Firms belonging to business group can coordinate their CSR spending with other

ﬁrms in the group maximizing potential beneﬁts. Group aﬃliation also increases the resources and

expertise fo better manage the CSR spending of the ﬁrm. The table also shows that the coeﬃcients

on the dummies CONC HLDG/GOVT OWNED/BG and the coeﬃcients on the interaction term

between POSTCSR and CONC HLDG/GOVT OWNED/BG in Models 4-6 are similar in sign and

signiﬁcance as the corresponding variables in Models 1-3 and conﬁrm the ownership eﬀects within

the set of ﬁrms subject to the mandatory CSR Rule.

The coeﬃcients on the interaction term AFFECTEDxPOSTCSR in models 1-3 captures the

impact of the CSR Rule on aﬀected ﬁrms compared to unaﬀected ﬁrms as before. The coeﬃcients on

the treatment dummy and treatment period is in the range 0.959%-1.29%, indicating a signiﬁcant

increase in yield-spreads in ﬁrms aﬀected by the CSR Rule. Models 4-6 only consider AFFECTED

ﬁrms and, therefore, the coeﬃcient on POSTCSR captures both the trend in the interest rates

and the impact of the CSR Rule. The coeﬃcient on POSTCSR in models 4-6 and the sum of

the coeﬃcients on POSTCSR and the interaction term AFFECTEDxPOSTCSR in models 1-3 are

similar in magnitude. The results in models 4-6 also indicate that ﬁrms aﬀected by the CSR Rule

see and increase in yield-spreads.

The results for the other control variables are similar to our base case results.

B.2 Corporate Governance

We next examine the impact of corporate governance on the role of the CSR Rule on yield-spreads.

Firms with good corporate governance can be expected to more strategically utilize the publicity

value of CSR spending, even mandated CSR spending. We therefore expect that ﬁrms that better

governed ﬁrms mitigate the overall negative impact of the mandatory CSR Rule.

To examine this hypothesis, we consider two measures of good corporate governance. The ﬁrst

measure relates to the degree of board independence. A board composed largely of independent

directors is considered to be an indicator of better corporate governance. We therefore create a

dummy BI that is equal to 1 if the fraction of the board that is independent, is above the median

for the sample. Next we examine the impact of the quality of the ﬁrm’s auditors. We develop

18

a dummy variable, BIG4, that is equal to one if the auditing ﬁrm is an aﬃliate of multinational

auditing ﬁrms Deloitte & Touche, PWC, E & Y or KPMG and is otherwise zero. Foreign ﬁrms

are not allowed to conduct auditing business in India due to the norms of the 1949 Chartered

Accountants Act and these ﬁrms therefore conduct business through their aﬃliates. We determine

the aﬃliations based on disclosure on company websites. BIG4 is one for bonds issued by ﬁrms

audited by the following accounting companies:

• Aﬃliates of Deloitte & Touche: C. C. Chokshi & Co., S. B. Billimoria Co, A. F. Ferguson &

Co, Fraser & Ross, MCA & Co P. C. Hansotia and Deloitte Haskins & Sells.

• Aﬃliates of KPMG: Bharat S. Raut & Co, SRBC & CO., SRB & ASSOCIATES.

• Aﬃliates of PWC: Price Waterhouse & Co, Lovelock & Lewes, Dalal & Shah.

• Aﬃliates of Ernst & Young: S.R. Batliboi & Co, S.R. Batliboi & Associates

Table 6 presents the results our our tests. Columns 1 and 2 present models that use the en-

tire sample of bond issues and Columns 3 and 4 present models that only examine bonds sold by

AFFECTED ﬁrms. Columns 1 and 3 use BI as the proxy for good governance and Columns 2 and

4 use BIG4 as the proxy for good governance, respectively. All regressions are run with control

variables and industry ﬁxed eﬀects. As shown in the table, the coeﬃcient in the triple interaction

term between AFFECTED, POSTCSR and BI/BIG4 is -0.785%/-0.456% and are statistically sig-

niﬁcant. Similarly, the coeﬃcients on POSTCSR and BI/BIG4 in Models 4-6 are -0.769%/-0.458%

and are also statistically signiﬁcant. These results are consistent with the hypothesis that good

governance ameliorates the negative impact of mandated CSR expenditure. These results are con-

sistent with the hypothesis that good governance ameliorates the negative impact of mandated

CSR expenditure.

As in Table 5, the coeﬃcient on the interaction term AFFECTEDxPOSTCSR captures the

impact of the CSR Rule on aﬀected ﬁrms. The coeﬃcient is positive and statistically signiﬁcant,

consistent with our base case results. The sum of the coeﬃcients on POSTCSR and the interac-

tion term AFFECTEDxPOSTCSR in columns 1-2 is similar in magnitude and signiﬁcance as the

coeﬃcient on POSTCSR in models 3 and 4, conﬁrming the increase in yield-spreads in regressions

using only AFFECTED ﬁrms. The results for the other control variables are similar to our base

case results.

19

In summary, our results indicate that the cost of debt capital is higher for ﬁrms aﬀected by

mandatory CSR rule and complements Manchiraju and Rajgopal (2017) ﬁndings focused on the

equity market.

VI Conclusions

The debate on the rationale for corporation actions classiﬁed as corporate social responsibility

activities has become a focal point among corporations, policy makers, and academics. In this

paper, we examine the impact of CSR on yield-spreads and our analysis contributes to better

understand the broader impact of CSR. Our empirical tests use bond data on Indian ﬁrms. The

Indian Parliament passed the 2013 Company Act which mandated a minimum 2% CSR spending

for ﬁrms that were considered to be proﬁtable, or have large levels of sales or have high net worth.

CSR spending by Indian ﬁrms post 2013, therefore, is not a strategic choice of management, but

is mandated by regulation. Thus unique setting allows us to examine the causal impact of CSR

spending without running into endogeniety or reverse causality issues.

We use diﬀ-in-diﬀ analysis and regression discontinuity models to examine the impact of CSR

spending on bond yield-spreads. Our diﬀ-in-diﬀ analysis shows that the interaction variable between

a dummy variable that indicates whether a ﬁrm meets the criteria for mandatory CSR spending

and the time period over which the rule is in eﬀect, has a positive and signiﬁcant coeﬃcient. Our

regression discontinuity analysis shows a positive and signiﬁcant jump in yield spreads from ﬁrms

that just meet the criteria for mandatory spending compared to ﬁrms that just miss the criteria.

Our ﬁndings indicate that that bond yield-spreads increased after 2013, by about 22 basis

points, for ﬁrms subject to the CSR rule. The negative eﬀects of CSR on bond markets can arise

for several reasons. The mandate for CSR spending targets ﬁrms that are large and proﬁtable, but

the lack of ﬂexibility reduces the cash ﬂows available to the ﬁrm to meet debt obligations. Such

lack of ﬂexibility has a negative impact on the market’s perceptions of bond value, leading to higher

yield spreads. Mandatory CSR also exacerbates the moral hazard between insider managers and

shareholders. Firms have to pick from an approved list of acceptable CSR activities, several of

which allows for private beneﬁts to the insider manager. It is also plausible that all allowed types

of CSR activity beneﬁts society without beneﬁtting the ﬁrm. Socially conscious investments that

have social beneﬁts that cannot be monetized can be negative NPV and reduce ﬁrm value.

20

Our results also reveal that good governance mitigates the negative eﬀects of mandatory CSR

requirements as well-governed ﬁrms can target CSR spending more strategically such that the posi-

tive eﬀects of CSR ameliorate the increased costs arising from a reduction in the cash ﬂows available

to pay oﬀ debt. Our study implies that any mandatory approach to community commitment by

ﬁrms is costly and any regulations designed to promote CSR should oﬀer ample ﬂexibilities in

implementations.

21

References

Albuquerque, R. A., A. Durnev, and Y. Koskinen (2013): “Corporate social responsibility

and ﬁrm risk: Theory and empirical evidence,” Working Paper.

Benabou, R., and J. Tirole (2010): “Individual and corporate social responsibility,” Economica,

77, 1–19.

Bertrand, M., P. Mehta, and S. Mullainathan (2002): “Ferreting out Tunneling: An Ap-

plication to Indian Business Groups,” Quarterly Journal of Economics, 117, 121–148.

Bushee, B. J., and C. F. Noe (2000): “Corporate disclosure practices, institutional investors,

and stock return volatility,” Journal of Accounting Research, 36, 171–202.

Calonico, S., M. Cattaneo, M. Farrell, and R. Titiunik (2017): “rdrobust: Software for

Regression Discontinuity Designs,” Stata Journal, 17, 372–404.

Chen, Y.-C., M. Hung, and Y. Wang (2017): “The eﬀect of mandatory CSR disclosure on

ﬁrm proﬁtability and social externalities: Evidence from China,” Journal of Accounting and

Economics.

Cooper, E. W., and H. Uzun (2015): “Corporate social responsibility and the cost of debt,”

Journal of Accounting and Finance, 15, 11.

Deng, X., J. koo Kang, and B. S. Low (2013): “Corporate social responsibility and stakeholder

value maximization: Evidence from mergers,” Journal of Financial Economics, 110, 87–109.

Flammer, C. (2015): “Does corporate social responsibility lead to superior ﬁnancial performance?

A regression discontinuity approach,” Management Science, pp. 2549–2568.

Freeman, R. E. (1984): Strategic Management: A Stakeholder Approach. Pitman, Boston, MA.

Gopalan, R., V. Nanda, and A. Seru (2007): “Aﬃliated Firms and Financial Support: Evi-

dence from Indian Business Groups,” Journal of Financial Economics, 86, 759–795.

Goss, A., and G. S. Roberts (2011): “The impact of corporate social responsibility on the cost

of bank loans,” Journal of Banking and Finance, 35, 1794–1810.

22

Iliev, P. (2010): “The Eﬀect of SOX Section 404: Costs, Earnings Quality, and Stock Prices,”

The Journal of Finance, 65, 1163–1196.

John, K., V. B. Nair, and L. W. Senbet (2005): “Law, Organizational Form and Taxes: A

Stakeholder Perspective,” https://ssrn.com/abstract=676987.

Khanna, T., and K. Palepu (2000): “Is Group Aﬃliation Proﬁtable In Emerging Markets? An

analysis of Diversiﬁed Indian Business Groups,” The Journal of Finance, 55, 867–891.

Kim, Y., M. S. Park, and B. Wier (2012): “Is Earnings Quality Associated with Corporate

Social Responsibility?,” The Accounting Review, 87, 761–796.

Kitzmueller, M., and J. Shimshack (2012): “Economic perspectives on corporate social re-

sponsibility,” Journal of Economic Literature, 50, 51–84.

Lee, D. S., and T. Lemieux (2010): “Regression Discontinuity Designs in Economics,” Journal

of Economic Literature, 48, 281–355.

Lins, K. V., H. Servaes, and A. Tamayo (2017): “Social capital, trust, and ﬁrm performance:

The value of corporate social responsibility during the ﬁnancial crisis,” The Journal of Finance,

72, 1785–1824.

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 (CSR) create

shareholder value? Evidence from the Indian Companies Act 2013,” Journal of Accounting

Research, 55, 1257–1300.

Martin, P. R., and D. V. Moser (2016): “Managers green investment disclosures and investors

reaction,” Journal of Accounting and Economics, 61, 239–254.

Oikonomou, I., C. Brooks, and S. Pavelin (2014): “The eﬀects of corporate social performance

on the cost of corporate debt and credit ratings,” Financial Review, 49, 49–75.

Reardon, S. F., and J. P. Robinson (2012): “Regression Discontinuity Designs with Multiple

Rating Score Variables,” Journal of Research on Educational Eﬀectiveness, 5, 83–104.

23

Appendix A: Timeline of events related to CSR rule (Manchiraju and Rajgopal (2017))

IMPACT OF CSR ON SHAREHOLDER VALUE 1267

Introduction

of the

Company’s

Bill 2009 in

the Lok

Sabha

Event # 1

Aug 3, 2009

Event # 2

Aug 31, 2010

Event # 3

Feb 28, 2011

Event # 4

Dec 14, 2011

Event # 5

Jun 26, 2012

Event # 6

Dec 18, 2012

Event # 7

Aug 8, 2013

Event # 1

Aug 29, 2013

For the first

time,

Standing

Committee

on Finance

inserts the

notion of

Mandatory

CSR in its

report

Bowing to

corporate

pressure, the

Ministry of

Corporate

Affairs

announces

that it is

considering

making CSR

voluntary

and not

mandatory

Government

eventually

goes ahead

with the

mandatory

CSR and

introduces a

revised

Company’s

Bill in the

Lok Sabha

Standing

Committee

on Finance

endorses the

bill with no

new changes

relating to

mandatory

CSR

The

Companies

Bill 2012,

containing

mandatory

CSR clause,

is passed in

the Lok

Sabha

The

Companies

Bill 2012 is

passed in

the Rajya

Sabha

The

President

signs the

bill and it

is enacted

FIG. 1.—Timeline of important events related to the passage of the mandatory CSR rule.

August 31, 2010. The Finance Committee introduced the notion of

mandatory CSR for the ﬁrst time in its report dated August 31, 2010.

Anecdotal evidence and reports in the popular press suggest that the

Finance Committee, perhaps anticipating popular backlash that might

result from a ver y progressive pro-business bill, inserted several new

clauses to make the bill slightly more pro-development. The proposal

mandating CSR spending was among these clauses and noted that:

“every company having [(net worth of rupees 500 crore or more, or

turnover of rupees 1000 crore or more)] or [a net proﬁt of rupees 5 crore

or more during a year] shall be required to formulate a CSR Policy to ensure

that every year at least 2% of its average net proﬁts during the three im-

mediately preceding ﬁnancial years shall be spent on CSR activities as may

be approved and speciﬁed by the company.”

10

!

Event # 3: On account of protests from industry, the Ministry of Cor-

porate Affairs announced on February 28, 2011 that it is considering

making only the disclosure, as opposed to the actual spending, of CSR

mandatory. The proposal for mandatory CSR attracted a lot of criti-

cism from companies who argued that what they spend on community

welfare, education, health, development, and environmental activism

is for them to decide. Azim Premji, Chairman of Wipro Ltd., opposed

mandatory CSR and argued that “my worry is the stipulation should

not become a tax at a later stage ... Spending two per cent on CSR is a

lot, especially for companies that are trying to scale up in these difﬁcult

times. It must not be imposed.” He also felt that a distinction should

be made between personal philanthropy and CSR, which is a company

activity.

11

Murli Deora, the Minister of Corporate Affairs (the ministry

10

From a research design standpoint, as discussed later, it is important to note that these

numerical cutoffs introduced on August 31, 2010, remained unchanged through various iter-

ations of the bill until the bill became law.

11

http://businesstoday.intoday.in/story/azim-premji-aima-convention-corporate-social-

responsibility/1/198960.html

• Event 1: The Companies Act introduced in the Lok Sabha (the lower house of the Indian parliament)

on August 3, 2009 as the Companies Bill, 2009. The Bill in its original form had no CSR clause.

• Event 2: The 2009 version of the Bill referred to the Parliamentary Standing Committee on Finance,

which submitted its report on August 31, 2010. The Finance Committee introduced the notion of

mandatory CSR in its report. Anecdotal evidence and reports in the popular press suggest that

the Finance Committee, perhaps anticipating popular backlash that might result from a progressive

pro-business bill, inserted several new clauses to make the bill more pro-development.

• Event 3: On account of protests from industry, the Ministry of Corporate Aﬀairs announced on

February 28, 2011 that it is considering making only the disclosure, as opposed to the actual spending,

of CSR mandatory.

• Event 4: The Ministry of Corporate Aﬀairs resisted pressure from corporate houses and went ahead

with the mandatory CSR rule. Keeping in view the recommendations made by the Finance Committee,

a revised Bill was prepared and the original Bill was withdrawn. The new Bill was introduced in the

Lok Sabha on December 14, 2011. The Bill was again referred to the Finance Committee as certain

new provisions, which had not been referred earlier to the committee, were included in this new Bill.

• Event 5: The Finance Committee submitted its report on June 26, 2012.

• Event 6: The Lok Sabha subsequently approved the Bill on December 8, 2012 and labeled it as the

Companies Bill, 2012.

• Event 7: The Companies Bill, 2012 was then considered and approved by the Rajya Sabha (the upper

house of the Indian parliament) on August 8, 2013 as The Companies Bill, 2013.

• Event 8: Following the Presidents assent, The Companies Bill, 2013 become law on August 29, 2013.

24

Appendix B: Variable Deﬁnition and Description

Variables Deﬁnition

YIELD The bond’s yield to maturity as reported by SDC Platinum

YIELD SPREAD YIELD - Maturity matched Indian Treasury Interest Rate

MATURITY Bond maturity

CSR Year 2013, The year of the passage of the Indian Company Act

PRECSR Dummy variable, equal to 1 if year < 2014

POSTCSR Dummy variable, equal to 1 if year

is greater than or equal to 2013, otherwise 0.

R1 Is the percentage diﬀerence between the ﬁrm’s

PRETAX INCOME and INR 50 million

R2 Is the percentage diﬀerence between the ﬁrm’s

Total Equity value and INR 5 billion

R3 Is the percentage diﬀerence between the ﬁrm’s

Total Revenue and INR 10 billion

M The minimum positive value of R1, R2 and R3

if at least one of the is positive or the maximum value

if R1, R3, R3, are all negative

AFFECTED R1 Dummy variable, equal to 1 if R1 > 0, otherwise 0

AFFECTED R2 Dummy variable, equal to 1 if R2 > 0, otherwise 0

AFFECTED R3 Dummy variable, equal to 1 if R3 > 0, otherwise 0

AFFECTED Dummy variable, equal to 1 if M < 0, otherwise 0

SIZE Logarithm of total assets

LEVERAGE Ratio of long term debt to total assets

TOBINQ Mkt Value of Equity / (Mkt Value of Equity + Book Value of LT Debt)

BI Dummy variable, equal to 1 if the fraction of independent

directors is higher than the sample median

BIG4 Dummy variable, equal to 1 if the auditor of the ﬁrm

aligned with Deloitte, E &Y, PWC, or KPMG

BG Dummy variable, equal to 1 if the ﬁrm

is aligned with a business group, otherwise 0

CONC HLDG Dummy Variable, equal to 1 if promoter holding is greater

median holding in the sample, otherwise 0

GOVT OWNED Dummy variable, equal to one if the ﬁrm is government owned, otherwise 0

25

(Continued)

Variables Deﬁnition

AFFECTED x POSTCSR Dummy variable, equal to 1 if AFFECTED is 1

and POSTCSR is 1, otherwise 0.

AFFECTED R1x POSTCSR Dummy variable, equal to 1 if AFFECTED R1 is 1

and POSTCSR is 1, otherwise 0.

AFFECTED R2x POSTCSR Dummy variable, equal to 1 if AFFECTED R2 is 1

and POSTCSR is 1, otherwise 0.

AFFECTED R3x POSTCSR Dummy variable, equal to 1 if AFFECTED R3 is 1

and POSTCSR is 1, otherwise 0.

AFFECTEDxPOSTCSRxBI Dummy variable, equal to 1 if AFFECTED is 1

POSTCSR is 1 and BI is 1, otherwise 0.

AFFECTEDxPOSTCSRxBG Dummy variable, equal to 1 if AFFECTED is 1

POSTCSR is 1 and BIG4 is 1, otherwise 0.

AFFECTEDxPOSTCSRxGOVT OWNED Dummy variable, equal to 1 if AFFECTED is 1

POSTCSR is 1 and GOVT OWNED is 1, otherwise 0.

AFFECTEDxPOSTCSRxCONC HLDG Dummy variable, equal to 1 if AFFECTED is 1

POSTCSR is 1 and CONC HLDG is 1, otherwise 0.

POSTCSRxBI Dummy variable, equal to 1 POSTCSR and BI is 1,

otherwise 0; for AFFECTED ﬁrms only

POSTCSRxBG Dummy variable, equal to 1 if POSTCSR is 1

and BIG4 is 1, otherwise 0; for AFFECTED ﬁrms only

POSTCSRxGOVT OWNED Dummy variable, equal to 1 if POSTCSR is 1

and GOVT OWNED is 1, otherwise 0; for AFFECTED ﬁrms only

POSTCSRxCONC HLDG Dummy variable, equal to 1 if POSTCSR is 1

and CONC HLDG is 1, otherwise 0; for AFFECTED ﬁrms only

26

Table 1: Distribution of Bond Issuers by Year and Aﬀected by CSR Mandate

This table presents data on the distribution of bond issues over the period from 2009 to 2017 in

our sample. Pane A presents the distribution of bond issues by year. Panel B presents the number

of bonds issued by ﬁrms that met the cut-oﬀ criteria for mandatory CSR in the period before the

law went into eﬀect and in the period after.

Panel A: Bond Issues by Year

Year #Bonds Percent Cumulative

2009 289 8.34% 8.34%

2010 367 10.59% 18.93%

2011 460 13.27% 32.20%

2012 615 17.74% 49.94%

2013 304 8.77% 58.71%

2014 326 9.41% 68.12%

2015 307 8.86% 76.98%

2016 421 12.15% 89.12%

2017 377 10.88% 100.00%

TOTAL 3,466 100%

Panel B: Unaﬀected/Aﬀected Distribution

AFFECTED PERIOD TOTAL

PRE POST

0 30 79 109

1 2,005 1,352 3,357

TOTAL 2,035 1,431 3,466

27

Table 2: Descriptive Statistics

This table presents gives descriptive statistics of the Indian bond ﬁrms that issued the 3,446 bond

issues. All variables are winsorized at the 1% level.

Variable #Obs Mean Median Std. Dev. Min Max

YIELD SPREAD 3,466 2.174 1.485 1.985 0.111 8.200

SALES 3,466 2121.056 1,103.833 2443.321 0.364 8656.865

PROFIT 3,466 264.200 176.602 264.961 -109.483 774.837

NET WORTH 3,466 1835.733 1190.933 1736.816 -71.045 5188.324

TOTAL ASSETS 3,466 8.774 8.874 1.484 3.330 11.685

TOBIN Q 3,466 2.303 2.004 1.382 0.409 6.018

LEVERAGE 3,428 0.509 0.554 0.252 0.022 0.899

MATURITY 3,466 7.879 3.000 16.706 0.000 100.000

CREDIT RANK 3,037 7.479 7.615 0.641 2.000 8.000

28

Table 3: Bond Yield-Spreads and the CSR Rule: Diﬀerence-in-Diﬀerence Regressions

This table shows results of a Diﬀerence-in-Diﬀerence speciﬁcation on the determinants of the yield-spread for

bonds issue by ﬁrms aﬀected/un-aﬀected by the mandatory CSR rule. Columns 1-4 present results with ﬁrm

controls for ﬁrm characteristics and Columns 5-8 present results including the full set of control variables.

Columns 1 and 5 present results when using all three possible criteria, i.e. proﬁt, net worth, and sales, to

determine whether a ﬁrm is subject to mandatory CSR spending. The remaining columns present results

when using the three criteria individually. All regressions are run with industry ﬁxed eﬀects using Fama-

French 30 Industry classiﬁcation. Standard errors in parentheses. The superscripts,

∗∗∗

,**, and * represent

coeﬃcients that are signiﬁcant at the 1%, 5%, and 10% respectively.

(1) (2) (3) (4) (5) (6) (7) (8)

VARIABLES Yield Yield Yield Yield Yield Yield Yield Yield

Spread Spread Spread Spread Spread Spread Spread Spread

POSTCSR -1.592

∗∗∗

0.293 -0.249* 0.032 -0.825 -0.676 -0.405** 0.030

(0.307) (0.350) (0.150) (0.102) (0.565) (0.433) (0.164) (0.108)

AFFECTED -0.449* -0.652

(0.267) (0.498)

AFFECTED R1 -0.450 -0.578

(0.303) (0.359)

AFFECTED R2 -0.511

∗∗∗

-0.009

(0.098) (0.126)

AFFECTED R3 -0.486

∗∗∗

-0.044

(0.089) (0.127)

AFFECTED 0.635** 1.046*

xPOSTCSR (0.311) (0.570)

AFFECTED R1 -0.207 0.904**

xPOSTCSR (0.357) (0.440)

AFFECTED R2 0.524

∗∗∗

0.733

∗∗∗

xPOSTCSR (0.168) (0.182)

AFFECTED R3 0.248* 0.301**

xPOSTCSR (0.137) (0.143)

SIZE -0.111** -0.107** -0.168

∗∗∗

-0.124**

(0.046) (0.044) (0.050) (0.049)

TOBINQ 0.00213 -0.000 0.005 0.009

(0.027) (0.027) (0.027) (0.028)

LEVERAGE -0.378* -0.389** -0.401** -0.302

(0.195) (0.192) (0.190) (0.203)

MATURITY 0.003 0.004 0.004 0.003

(0.002) (0.002) (0.002) (0.002)

CREDITRANK -0.399

∗∗∗

-0.412

∗∗∗

-0.391

∗∗∗

-0.408

∗∗∗

(0.0658) (0.0662) (0.066) (0.0665)

CONSTANT 10.05

∗∗∗

2.536

∗∗∗

2.458

∗∗∗

2.331

∗∗∗

6.760

∗∗∗

6.755

∗∗∗

6.567

∗∗∗

6.260

∗∗∗

(0.265) (0.299) (0.0813) (0.0608) (0.717) (0.672) (0.579) (0.589)

Observations 3407 3407 3407 3407 2978 2978 2978 2978

R-squared 0.102 0.005 0.009 0.011 0.026 0.026 0.031 0.026

29

Table 4: Bond Yield-Spreads and the CSR Rule: Regression Discontinuity Results

This table presents results for the Regression Discontinuity design. All variables are winsorized at

the 1% level. Superscripts

∗∗∗

, ** and * respectively denote statistical signiﬁcant at the 1%, 5%,

and 10% levels.

Sample Period Method Coef. Std. Err. z P >z [95% Conf. Interval]

Full Sample Conventional 1.172

∗∗

0.555 2.111 0.035 0.084 2.260

PRECSR Conventional -4.219

∗∗

1.974 -2.137 0.033 -8.089 -.3495

POSTCSR Conventional 1.60

∗∗∗

0.543 2.948 0.003 0.536 2.665

30

Figure 1: Regression Discontinuity Plots

This ﬁgure shows the regression discontinuity plots for the yield on bonds in the PRECSR

period 2009-2013 and the POSTCSR period from 2014-2017. Firms to the right of the cut-

oﬀ value 0 subject to mandatory CSR spending and ﬁrms to the left of the cutoﬀ value 0

are not subject to the mandate. Panel A presents the plot for the full sample period, Panel

B presents the plot for the PRECSR period and Panel C presents the plot for the POSTCSR period.

Panel A: Full Sample

0 5 10 15 20

-10 0 10 20 30

Sample average within bin Polynomial ﬁt of order 1

Regression function ﬁt

Panel B: Pre CSR Period

-60 -40 -20 0 20

-10 0 10 20 30

Sample average within bin Polynomial ﬁt of order 1

Regression function ﬁt

Panel B: Post CSR Period

0 5 10 15 20 25

-10 0 10 20 30

Sample average within bin Polynomial ﬁt of order 1

Regression function ﬁt

31

Table 5: Eﬀect of Ownership Structure

This table shows results of a Diﬀerence-in-Diﬀerence speciﬁcation on the determinants of the yield-spread

for bonds issue by ﬁrms aﬀected/un-aﬀected by the mandatory CSR rule and classiﬁed by the safety rank

of the bond. Columns 1 and 3 present results for bonds that are the safest and have the highest rating by

ProwessDx. Columns 2 and 4 present results for bonds where the ﬁrms is owned by the government. Column

1 and 2 present results when using both Aﬀected and Un-aﬀected ﬁrms. Column 3 and 4 present results

when using only the Aﬀected ﬁrms. All regressions are run with industry ﬁxed eﬀects using Fama-French 30

Industry classiﬁcation. Standard errors in parentheses. The superscripts,

∗∗∗

,**, and * represent coeﬃcients

that are signiﬁcant at the 1%, 5%, and 10% respectively.

VARIABLES Yield Yield Yield Yield Yield Yield

Spread Spread Spread Spread Spread Spread

POSTCSR -0.829 -0.846 -0.844 0.311*** 0.109 0.450***

(0.565) (0.562) (0.564) (0.115) (0.0796) (0.153)

AFFECTED -0.645 -0.674 -0.737

(0.498) (0.495) (0.497)

AFFECTED 1.149** 0.959* 1.291**

xPOSTCSR (0.578) (0.567) (0.580)

Conc Holdg 0.00879 -0.00915

(0.0996) (0.101)

GOVT OWNED 0.292** 0.297**

(0.133) (0.133)

BG -0.237* -0.254**

(0.124) (0.126)

AFFECTED -0.171

POSTCSRxCONC HLDG (0.146)

AFFECTEDx 0.689***

POSTCSRxGOVT OWNED (0.186)

AFFECTED -0.302*

xPOSTCSRxBG (0.173)

POSTCSRxCONC HLDG -0.158

(0.147)

POSTCSRxGOVT OWNED 0.699***

(0.186)

POSTCSRxBG -0.310*

(0.174)

SIZE -0.115** -0.143*** -0.107** -0.108** -0.126** -0.0880*

(0.0458) (0.0459) (0.0459) (0.0515) (0.0514) (0.0515)

TOBINQ -0.00206 0.0842*** 0.0466 -0.00209 0.0855*** 0.0490*

(0.0273) (0.0301) (0.0291) (0.0275) (0.0303) (0.0294)

LEVERAGE -0.391** -0.114 -0.178 -0.349 -0.0271 -0.0855

(0.197) (0.198) (0.200) (0.214) (0.215) (0.218)

MATURITY 0.00349 0.00231 0.00250 0.00390 0.00268 0.00282

(0.00237) (0.00235) (0.00236) (0.00241) (0.00239) (0.00240)

CreditRank -0.393*** -0.463*** -0.422*** -0.381*** -0.463*** -0.426***

(0.0663) (0.0661) (0.0661) (0.0724) (0.0723) (0.0723)

Constant 6.752*** 7.162*** 6.962*** 5.941*** 6.282*** 6.045***

(0.718) (0.715) (0.723) (0.606) (0.604) (0.611)

Observations 2978 2978 2978 2922 2922 2922

R-squared 0.027 0.039 0.032 0.023 0.036 0.029

Sample Full Full Full Aﬀected Aﬀected Aﬀected

32

Table 6: Eﬀect of Corporate Governance

This table shows results of a Diﬀ-in-Diﬀ speciﬁcation of the yield-spread for bonds issue by ﬁrms aﬀected/un-

aﬀected by the mandatory CSR rule and classiﬁed by the issuers governance characteristics. We use two

governance dummies, BI that is equal to one if the ﬁrm has above median Board Independence or otherwise

zero and BIG4 that is equal to one is the ﬁrm is audited by one of the four multinational auditing ﬁrm or

otherwise zero. Columns 1 and 3 present results when using BI, Columns 2 and 4 present results when using

BIG4, respectively. Columns 1 & 2 present results when using the full sample. Columns 3 & 4 present results

when using only AFFECTED ﬁrms. All regressions are run with industry ﬁxed eﬀects using Fama-French 30

Industry classiﬁcation. Standard errors in parentheses. The superscripts,

∗∗∗

,**, and * represent coeﬃcients

that are signiﬁcant at the 1%, 5%, and 10% respectively.

VARIABLES Yield Yield Yield Yield

Spread Spread Spread Spread

POSTCSR -0.811 -0.934* 0.698*** 0.371***

(0.561) (0.564) (0.107) (0.0901)

AFFECTED -0.651 -0.702

(0.494) (0.496)

AFFECTED 1.516*** 1.306**

xPOSTCSR (0.570) (0.570)

BI 0.00279 -0.0142

(0.0934) (0.0952)

BIG4 -0.115 -0.110

(0.0978) (0.0992)

AFFECTED -0.785***

xPOSTCSRxBI (0.146)

AFFECTED -0.456***

xPOSTCSRxBIG4 (0.149)

POSTCSR -0.769***

xBI (0.147)

POSTCSR -0.458***

xBIG4 (0.150)

SIZE -0.128*** -0.109** -0.119** -0.0968*

(0.0455) (0.0458) (0.0510) (0.0513)

TOBINQ 0.0593** 0.0259 0.0595** 0.0257

(0.0282) (0.0275) (0.0284) (0.0278)

LEVERAGE -0.366* -0.403** -0.312 -0.342

(0.194) (0.195) (0.210) (0.211)

MATURITY 0.00207 0.00223 0.00246 0.00257

(0.00235) (0.00235) (0.00238) (0.00239)

CreditRank -0.432*** -0.449*** -0.425*** -0.445***

(0.0655) (0.0662) (0.0715) (0.0725)

CONSTANT 7.018*** 7.175*** 6.211*** 6.300***

(0.716) (0.718) (0.605) (0.609)

Observations 2978 2978 2922 2922

R-squared 0.041 0.034 0.038 0.031

Sample Full Full Aﬀected Aﬀected

33