What Doesn't Kill You Will Only Make You More Risk‐Loving: Early‐Life Disasters and CEO Behavior

• Author: VINEET BHAGWAT,P. RAGHAVENDRA RAU,GENNARO BERNILE
• DOI: http://doi.org/10.1111/jofi.12432
• Date: 01 February 2017
• Published date: 01 February 2017

THE JOURNAL OF FINANCE •VOL. LXXII, NO. 1 •FEBRUARY 2017

What Doesn’t Kill You Will Only Make You More

Risk-Loving: Early-Life Disasters and CEO

Behavior

GENNARO BERNILE, VINEET BHAGWAT, and P. RAGHAVENDRA RAU∗

ABSTRACT

The literature on managerial style posits a linear relation between a chief executive

officer’s (CEOs) past experiences and firm risk. We show that there is a nonmono-

tonic relation between the intensity of CEOs’ early-life exposure to fatal disasters

and corporate risk-taking. CEOs who experience fatal disasters without extremely

negative consequences lead firms that behave more aggressively, whereas CEOs

who witness the extreme downside of disasters behave more conservatively. These

patterns manifest across various corporate policies including leverage, cash hold-

ings, and acquisition activity.Ultimately, the link between CEOs’ disaster experience

and corporate policies has real economic consequences on firm riskiness and cost of

capital.

“I know of no one who has achieved something significant without also

in their own lives experiencing their share of hardship, frustration, and

regret . . . if you’re like me and you occasionally want to swing for the

fences, you can’t count on a predictable life.”

—Tim Cook, CEO of Apple Inc., Auburn University Spring 2010 Com-

mencement

∗Gennaro Bernile is with Lee Kong Chian School of Business, Singapore Management Univer-

sity.Vineet Bhagwat is with Lundquist College of Business, University of Oregon. P. Raghavendra

Rau is with Judge Business School, University of Cambridge. We are grateful to Michael Roberts

(Editor), two anonymous referees, and Michael Anderson, Michelle Baddeley, Yosef Bonaparte,

David de Cremer, Henrik Cronqvist, Peter Cziraski, Larry Dann, David Feldman, Ben Hardy,

Jianfeng Hu, Mark Humphery-Jenner, Ivalina Kalcheva, Shimon Kogan, Steve McKeon, Maur-

izio Montone, Vikram Nanda, Lalitha Naveen, Micah Officer, Chris Parsons, David Reeb, Adam

Reed, Andreas Richter, VidyaSharma, Avanidhar Subrahmanyam, Jo-Ann Suchard, Scott Yonker,

Bohui Zhang, Weina Zhang, and seminar participants at the Ig Nobel MIT Lecture 2015, West-

ern Finance Association 2015 meetings, ISB CAF conference 2015, IFABS conference 2015, Asian

Finance Association 2015 meetings, China International Finance Conference 2015, European Fi-

nance Association 2014 meetings, the UBC summer finance conference 2014, the University of

Alberta 2014 Frontiers of Finance conference, the Erasmus University Research in Behavioural

Finance Conference 2014, Arizona State University, Frankfurt School of Finance & Management,

London Business School, National University of Singapore, Norwich Business School, Singapore

Management University,University of Adelaide, University of Cambridge, University of New South

Wales, University of Sussex, and the University of Oregon for comments and suggestions. Wealso

thank Carina Cuculiza, Daniel Dodson, Eric Higgins, and Dee Zaster of the University of Miami

for valuable research assistance. We have no conflicts of interest to disclose.

DOI: 10.1111/jofi.12432

167

168 The Journal of Finance R

PRIOR RESEARCH SHOWS THAT CHIEF executive officers’ (CEO) managerial styles

explain a large part of the variation in firm capital structure, investment, com-

pensation, and disclosure policies (e.g., Bertrand and Schoar (2003), Bamber,

Jiang, and Wang (2010), Graham, Li, and Qiu (2012)). Prior work further shows

that at least part of the heterogeneity in CEOs’ managerial styles reflects vari-

ation in individual life and career experiences (e.g., Graham and Narasimhan

(2005), Malmendier and Tate (2005), Malmendier, Tate, and Yan (2011), Schoar

and Zuo (2011), Benmelech and Frydman (2014), Lin et al. (2014), Dittmar and

Duchin (2016)).

A common thread underlying this line of research is the existence of a mono-

tonic relation between treatment and effect. Specifically, existing studies find

that exposure to a particular macroeconomic, personal, or career-specific event

has a unidirectional effect on CEO risk-taking and consequently on corporate

policies. In this study, we test whether the intensity of early-life experiences

has a nonmonotonic impact on CEOs’ attitudes toward risk and thus on the

corporate policies that they influence. In medical terms, the question we ad-

dress is whether, in addition to a treatment being administered, the strength

of the dosage affects the treatment outcome. This hypothesis, while a stan-

dard prediction in the psychiatry literature (e.g., Yerkes and Dodson (1908)), is

relatively unexamined in the finance and economics literature.

Early-life exposure to natural disasters may affect a CEO’s risk-taking in

several ways. CEOs with exposure to fatalities from natural disasters may

be more sensitized to the consequences of risk and in turn be more wary of

decisions that increase firm risk. However, it is also possible that childhood

exposure to natural disasters may increase CEOs’ ability to deal with risky

situations as well as their confidence when making decisions involving firm

risk. Hence, the effect of exposure to natural disasters on subsequent behavior

may be nonmonotonic, as Castillo and Carter (2011) find with respect to trust

and reciprocity between individuals. CEOs with disaster experience that does

not involve significant fatalities may develop a higher risk tolerance, whereas

those with exposure to major fatal disasters may behave more conservatively.1

To test this conjecture, we examine the relation between CEO early-life ex-

posure to natural disasters and firms’ subsequent corporate financial and in-

vestment policies. We begin by identifying the name, date, and place of birth of

1,508 U.S.-born CEOs in a sample of S&P1500 firms from 1992 to 2012. Next,

we assemble a unique database of U.S. county-level natural disaster events

over the period from 1900 to 2010, including earthquakes, volcanic eruptions,

tsunamis, hurricanes, tornadoes, severe storms, floods, landslides, and wild-

fires. We then combine the two databases to infer CEOs’ likely exposure to

the consequences of natural disasters during their formative years, which we

define to be the ages of 5 to 15 (Nelson (1993)).

1Apple Inc. CEOs Tim Cook and Steve Jobs are examples of CEOs with widely differing child-

hood disaster experiences when growing up: Tim Cook, born in Mobile, Alabama in 1960, witnessed

1.15 deaths across 57 natural disaster events between the ages of 5 and 15, while Steve Jobs, born

in San Francisco, California in 1955, witnessed 31.6 deaths across 39 natural disasters during the

early years of his life.

What Doesn’t Kill You Will Only Make You More Risk-Loving 169

In our baseline tests, we group CEOs based on the number of disaster-related

fatalities in their birth county scaled by the population of their birth county

during the formative years of their childhood. We then examine the relation

between CEO early-life disaster experience and several firm decisions and out-

comes pertaining to capital structure, acquisition activity,and return volatility.

Our results provide a consistent picture across the firm decisions and out-

comes on which CEOs typically have a large influence (e.g., Graham, Harvey,

and Puri (2012,2013)). Specifically, we find robust evidence of an inverse U-

shaped relation between a CEO’s early-life exposure to fatal disasters and

corporate risk-taking. For example, all else equal, firms whose CEOs experi-

enced a moderate level of fatalities from natural disasters have a 3.4% higher

leverage ratio than firms whose CEOs experienced no fatal disasters, a mag-

nitude comparable to the Depression babies effect documented in Malmendier

and Nagel (2011). In contrast, firms whose CEOs experienced extreme levels of

fatalities from natural disasters have a 3.7% lower leverage ratio than firms

whose CEOs have no fatal disaster experience (and hence an overall 7.1% lower

leverage ratio than firms whose CEOs have a moderate level of fatal disaster ex-

perience). These results are due to active financing choices made by the CEOs.

Specifically, CEOs with moderate levels of fatal disaster experience are signif-

icantly more likely to meet net financing deficits with debt rather than equity,

to pay higher interest expenses and loan spreads on their debt, and to have

lower credit ratings than CEOs with no fatal disaster experience. Their firms

are also significantly more likely to go through bankruptcy. In contrast, CEOs

with extreme levels of fatal disaster experience display the opposite patterns.

Similarly, CEOs with moderate (extreme) levels of fatal disaster experience

hold significantly less (more) cash, announce more (fewer) acquisitions, are less

(more) likely to pay for acquisitions with stock, and are more (less) likely to

announce unrelated acquisitions than CEOs with no fatal disaster experience.

Consistent with excessive risk-taking, acquisitions by CEOs with moderate

fatal disaster experience earn worse announcement returns. Firms managed by

moderate disaster CEOs also display higher volatility, especially idiosyncratic

volatility, than CEOs with no disaster experience, while CEOs with extreme

disaster experience again display the opposite patterns.

Overall, our results support the view that experiencing fatal disasters with-

out extreme consequences desensitizes CEOs to the negative consequences of

risk. In contrast, CEOs who experienced extreme fatal disasters and witnessed

the downside potential of risky situations appear to be more cautious in ap-

proaching risk when at the helm of a firm.

All of our empirical specifications include time, state of birth, year of birth,

and firm fixed effects, which effectively purge cohort effects from our analysis.

Moreover, all of our tests control for the historical incidence of disaster-related

fatalities in the CEO’s county of birth over the period 1900 to 2010 to help us

separate the effect of a CEO growing up in a “high-risk” county from actually

having lived through a fatal disaster during his formative years. For exam-

ple, a CEO who did not experience any major fatalities despite growing up in

the “tornado belt” (states like Kansas, for example) may underestimate the