Financing Constraints and Workplace Safety

• Author: JONATHAN B. COHN,MALCOLM I. WARDLAW
• Published date: 01 October 2016
• Date: 01 October 2016
• DOI: http://doi.org/10.1111/jofi.12430

THE JOURNAL OF FINANCE •VOL. LXXI, NO. 5 •OCTOBER 2016

Financing Constraints and Workplace Safety

JONATHAN B. COHN and MALCOLM I. WARDLAW∗

ABSTRACT

We present evidence that financing frictions adversely impact investment in work-

place safety,with implications for worker welfare and firm value. Using several iden-

tification strategies, we find that injury rates increase with leverage and negative

cash flow shocks, and decrease with positive cash flow shocks. We show that firm

value decreases substantially with injury rates. Our findings suggest that invest-

ment in worker safety is an economically important margin on which firms respond

to financing constraints.

OVER 3.5 MILLION WORKPLACE INJURIES and illnesses occur in the United States

each year. The estimated annual cost of these injuries is $250 billion, more

than the cost of all forms of cancer combined (Leigh (2001)). While workplace

safety has been studied extensively in fields as diverse as industrial relations,

operations management, and industrial-organizational psychology, its connec-

tions with finance remain largely unexplored. This paper studies how financing

constraints impact workplace safety, which has implications for firm value and

employee welfare.

Firms invest resources in improving workplace safety just as they invest

in research and development, property, plant, and equipment, and organiza-

tional capital. As with other forms of investment, spending on safety must

be financed out of either internal cash flow or externally raised capital. In

a world with financing frictions, a firm’s investment may be sensitive to the

financial resources available to finance that investment. Thus, the safety of

a firm’s workplaces could depend on the financial resources at its disposal.

∗Jonathan B. Cohn is with the University of Texas at Austin. Malcolm I. Wardlaw is with the

University of Texas at Dallas. Wewould like to thank Michael Roberts (Editor), the Associate Edi-

tor,two anonymous referees, Manuel Adelino, Ashwini Agrawal, Andres Almazan, Heitor Almeida,

Aydo ˘

gan Altı, Sugato Bhattacharyya, Andres Donangelo, Jay Hartzell, Marcin Kacperczyk, David

Matsa, Marco Pagano, Gordon Phillips, Giovanni Pica, Uday Rajan, Adriano Rampini, Avri Ravid,

Nancy Rose, Bill Schwert, Amit Seru, Denis Sosyura, Sheridan Titman, Yongxiang Wang, Toni

Whited, and seminar participants at the University of British Columbia, University of Illinois at

Urbana–Champaign, University of Michigan, New York University, University of Oklahoma, Uni-

versity of Rochester, University of Southern California, University of Texas at Austin, University

of Texas at Dallas, Bureau of Labor Statistics, 2013 CSEF Conference on Finance and Labor, and

2014 AFA for their comments. We would also like to thank Nicole Nestoriak from the Bureau

of Labor Statistics for her assistance with the injury data. This paper benefited greatly from a

UT–Austin McCombs Research Excellence Grant. The authors have read the Journal of Finance’s

disclosure policy and have no conflicts of interest to disclose.

DOI: 10.1111/jofi.12430

2017

2018 The Journal of Finance R

Investment in safety may be especially vulnerable to cuts in the face of financing

constraints, as its payoffs accrue slowly over time and are difficult to evaluate.

In this paper, we explore the impact of financing constraints on workplace

safety by examining the sensitivity of workplace injury rates to the financial

resources available to a firm using establishment-level injury data from the

Bureau of Labor Statistics’ (BLS) annual Survey of Occupational Injuries and

Illnesses (SOII). As we lack exogenous variation in financial resources with

which to completely isolate the effect of financing on injuries, we employ sev-

eral empirical strategies. Each approach produces evidence pointing toward

increased financial resource availability leading to fewer injuries, suggesting

that financing constraints impair investment in safety. While any one piece of

evidence is open to alternative interpretations, the evidence taken together is

difficult to reconcile with any specific alternative.

We begin by examining the empirical relationship between injury rates and

well-established drivers of a firm’s capacity to finance investment, including

cash flow, cash balances, and financial leverage. Cash balances and cash flow

are sources of internal financing. Debt reduces cash flow through interest pay-

ments, and existing debt claims can make it difficult to raise additional external

capital (Myers (1977)). Prior research shows that investment in general tends

to increase with available cash (e.g., Fazzari, Hubbard, and Petersen (1988),

Lamont (1997), Rauh (2006)) and decrease with leverage (e.g., Denis and Denis

(1993), Lang, Ofek, and Stulz (1996)). If investment in safety is sensitive to a

firm’s financial resources, then injury rates should decrease with cash flow and

cash balances and increase with leverage.

We find a robust positive relation between injury rates and leverage, con-

trolling for establishment and firm characteristics as well as establishment,

industry-year, and state-year fixed effects. A one-standard-deviation increase

in a firm’s debt-to-assets ratio is associated with a 5.6 percentage point in-

crease in total workplace injuries the following year, relative to the sample

mean injury rate. It is further associated with a 6.5 percentage point increase

in injuries serious enough that the injured employee misses at least one full

day of work, suggesting that it is not just minor injuries that are sensitive

to leverage. These estimates are larger than existing estimates of the impact

of penalty-imposing Occupational Safety and Health Administration (OSHA)

inspections or plant unionization on workplace injury rates (Mendelhoff and

Gray (2005)). Injury rates are sensitive to at least the first two annual lags of

leverage but are unrelated to contemporaneous and future leverage, partially

alleviating concerns about reverse causality.

Injury rates show strong negative relations with cash flow and cash balances

in the cross-section. However, these relations disappear when we control for

establishment fixed effects, suggesting that they may be driven by unobserved

firm- or establishment-level heterogeneity. We also find some evidence that

injury rates are negatively related to dividend payout and firm size. As these

characteristics are often seen as inverse proxies for the severity of financing

constraints, this evidence provides added support for the role of financing con-

straints in limiting investment in safety.Overall, the results from this analysis

Financing Constraints and Workplace Safety 2019

provide some support for the hypothesis that injury rates decrease with finan-

cial resources, though the evidence here is far from conclusive.

Tobetter isolate the effect of financial resources on injury rates, we next study

three quasi-natural experiments involving cash flow shocks. These include a

repatriation tax holiday in 2004, the onset of the financial crisis in 2007 and

2008, and large oil price fluctuations during the 2000s. Variants of each have

been used in prior papers to study the effect of cash flow on capital investment.1

The cash flow shocks involved are large, plausibly exogenous with respect to

injury rates, uncorrelated with each other, and affect some firms but not others.

We exploit this last feature to conduct difference-in-differences analysis using

matched samples, which mitigates though does not eliminate concerns about

unobserved counterfactual changes in injury rates absent a shock. One useful

feature of the set of experiments we study is that exposure is almost completely

uncorrelated across the three, suggesting that they can be treated as three

independent tests.

The results of these tests broadly support a negative (positive) response of

injury rates to a positive (negative) cash flow shock, especially in firms with

higher leverage. The estimates imply that, on average, a firm’s injury rate

would fall by 8.4 to 11.9 percentage points in response to a one-standard-

deviation increase in its cash flow. We verify that the difference-in-differences

estimates are unlikely to be driven by ex ante differences in treated and control

establishments or by differential preexisting trends in injury rates. While we

cannot rule out the possibility that unobserved factors correlated with future

trends in injury risk impact assignment, they would have to do so in ways that

coincidentally produce consistent results across three separate experiments to

explain the results. We also explicitly consider specific omitted characteristics,

such as managerial skill and production technology, and conclude that it would

be difficult for any single characteristic to explain all of the results in the paper.

The weight of the evidence thus appears to support financing constraints as

the most likely mechanism driving the results.

Stakeholder theory argues that a firm indirectly bears costs, in expectation,

that its financial policies impose on nonfinancial stakeholders such as employ-

ees ex post (Titman (1984)). The natural channel in the case of costs due to

higher workplace injury risk is a compensating wage differential that employ-

ees require to bear this risk. Firms may also bear costs directly in the form of

decreased productivity resulting from increased downtime and poor employee

morale.2In the last part of our analysis, we find a substantial negative rela-

tion between firm value and injury rates, with firm value decreasing by 6.1%

for each one-standard-deviation increase in injury rate. While we cannot rule

out the possibility of alternative explanations for this relation, the estimates

1Dharmapala, Foley,and Forbes (2011) and Faulkender and Petersen (2012) examine the effect

of the repatriation tax holiday, Almeida et al. (2012) study the effect of the onset of the financial

crisis, and Lamont (1997) studies the effect of an oil price shock in 1985.

2Danna and Griffin (1999) argue that these costs are likely to be greater than those due to

compensating wage differentials.