Does Branch Religiosity Influence Bank Risk‐Taking?

• Author: Michele Fabrizi,Elisabetta Ipino,Justin Chircop,Antonio Parbonetti
• DOI: http://doi.org/10.1111/jbfa.12227
• Published date: 01 January 2017
• Date: 01 January 2017

Journal of Business Finance & Accounting

Journal of Business Finance & Accounting, 44(1) & (2), 271–294, January/February 2017, 0306-686X

doi: 10.1111/jbfa.12227

Does Branch Religiosity Influence Bank

Risk-Taking?

Justin Chircop, Michele Fabrizi, Elisabetta Ipino

and Antonio Parbonetti∗

Abstract: Using branch-level data on public and private US banking institutions, we investigate

the importance of branch religiosity in shaping bank risk-taking behavior. Our results show

robust evidence that branch religiosity is negatively related to bank risk-taking. This effect

persists after controlling for several bank-level and county-level variables that might correlate

with religiosity. Moreover, this result is robust to controlling for headquarter religiosity,

suggesting that the effect of branch religiosity is additive and not washed out by headquarter

religiosity. Overall, our findings document that headquarter religiosity does not capture the full

effect of religiosity on bank behavior, as claimed by previous research, but that the religiosity of

the geographic area in which the bank operates significantly influences bank behavior.

Keywords: religion, risk-taking, risk aversion

1. INTRODUCTION

This paper tests the importance of branch religiosity in shaping bank risk-taking. A

rich literature across several disciplines provides robust evidence of a strong relation

between headquarter religiosity and risk-taking in non-financial (Hilary and Hui,

2009; Grullon et al., 2010; Kumar et al., 2011; Dyreng et al., 2012; and McGuire

et al., 2012) and financial firms (Kanagaretnam et al., 2013; 2015; and Adhikari and

Agrawal, 2016).

A common research design choice of prior literature is its exclusive focus on the

religiosity of the geographical area of the corporate headquarters. In most cases such

a choice is implicitly motivated by the location of important firm stakeholders (Hilary

and Hui, 2009). Specifically, given that the corporate headquarters is the venue in

which firm strategic decisions are made, it is reasonable to assume that firm senior

employees reside in the vicinity of the corporate headquarters. Similarly, given that

the corporate headquarters is the center of information exchange between the firm

∗The first author is at the Lancaster University Management School, Lancaster, UK. The second and fourth

authors are at the University of Padova, Padova, Italy. The third author is at the Concordia University,

Montreal, Canada. The authors also gratefully acknowledge helpful comments and suggestions on previous

versions of this paper from the editor Andrew Stark, an anonymous reviewer, Andreas Charitou, Luo He,

Michel Magnan, Patricia O’Brien and Stephen Penman. (Paper received February 2015, revised revision

accepted October 2016).

Address for correspondence: Antonio Parbonetti, Department of Economics and Management, University

of Padova, Via del Santo, 33 35123 Padova, Italy.

e-mail: antonio.parbonetti@unipd.it

C

2016 John Wiley & Sons Ltd 271

272 CHIRCOP,FABRIZI, IPINO AND PARBONETTI

and investors, and the documented investor ‘home bias’ (Coval and Moskowitz, 1999),

it is also reasonable to assume that firm shareholders would reside in the vicinity of

corporate headquarters. Given this, prior literature finds that headquarter religiosity

is inversely related to bank risk-taking (Adhikari and Agrawal, 2016).

In solely focusing on headquarter religiosity, the prior literature assumes that all

firm decisions are made at the corporate headquarters and that thus only these

decisions, which are influenced by the religiosity of the headquarter geographical

location, give rise to bank risk. However, this might not be the case for banks, which

tend to have geographically dispersed operations with decentralized decision making.

There are three reasons why we expect branch-level operations to influence bank

risk-taking. First, bank customers’ religiosity is likely to influence bank risk-taking.

Studies in economic literature find that religious people are risk-averse (Miller and

Hoffmann, 1995; Diaz, 2000; and Miller, 2000) and, thus, banks with more religious

customers are likely to have less risky borrowers. Second, building on organizational

legitimacy theory, we argue that when banks operate in very religious, and thus

risk-averse, geographic areas, the level of risk of each branch has to be adjusted

to the expectations of stakeholders to avoid a misalignment between the two value

systems. Finally, religious norms of the local population in which the firm operates

will influence branch employees irrespective of whether they are themselves religious

(Dyreng et al., 2012). This prediction is ultimately grounded in social norm theory,

which posits that individuals will act in ways that conform to the behavioral norms

of the groups with which they associate. Moreover, such influence is amplified by the

need for the bank to maintain organizational legitimacy.

We conjecture that the effect of branch religiosity on bank risk-taking is above and

beyond the previously documented effect of headquarters’ religiosity on bank risk-

taking. Specifically, if the geographic area in which the company operates influences

bank risk-taking and if we view religious individuals as more risk-averse (similar to

Miller and Hoffmann, 1995; Diaz, 2000; and Miller, 2000), then we should expect

banks operating in highly religious areas to take on lower risk. However, based on

previous research, whether the geographical location of bank branches influences

bank risk-taking is ultimately an empirical question.

To test whether branch religiosity influences risk-taking, we focus our study on each

branch of 1,578 public and private US banking institutions over the period 2000–2010.

We measure branch religiosity using data collected from the American Religion Data

Archive (ARDA) and proxy for bank risk using the Z-score, the amount of charge-

offs, the amount of non-performing loans and the probability of bank failure over the

sample period. Further, in an attempt to minimize measurement error included in

these proxies, we use principal component analysis (PCA) to compute a composite

measure of bank risk. In line with our expectations and the conjecture that branch

religiosity influences the bank’s risk profile, we find a significant negative relation

between our measure of branch religiosity and bank risk-taking. Importantly, our

results are robust to controlling for headquarter religiosity, thus suggesting that the

effect of branch religiosity is additive and not washed out by the effect of headquarter

religiosity.

In further analyses, we take advantage of the large number of bank mergers and

acquisitions (M&A) over the sample period to tease out the branch versus headquarter

effect. Specifically, we find that branch religiosity of the target firm is significantly

negatively related to the change in bank risk after the M&A deal. This result provides

C

2016 John Wiley & Sons Ltd