Sharing Risk with the Government: How Taxes Affect Corporate Risk Taking

• Author: LIANDONG ZHANG,ALEXANDER LJUNGQVIST,LUO ZUO
• Published date: 01 June 2017
• Date: 01 June 2017
• DOI: http://doi.org/10.1111/1475-679X.12157

DOI: 10.1111/1475-679X.12157

Journal of Accounting Research

Vol. 55 No. 3 June 2017

Printed in U.S.A.

Sharing Risk with the Government:

How Taxes Affect Corporate Risk

Taking

ALEXANDER LJUNGQVIST,

∗LIANDONG ZHANG,

†

AND LUO ZUO

‡

Received 19 December 2015; accepted 18 October 2016

ABSTRACT

Using 113 staggered changes in corporate income tax rates across U.S. states,

we provide evidence on how taxes affect corporate risk-taking decisions.

Higher taxes reduce expected profits more for risky projects than for safe

ones, as the government shares in a firm’s upside but not in its downside.

Consistent with this prediction, we find that risk taking is sensitive to taxes,

albeit asymmetrically: the average firm reduces risk in response to a tax in-

crease (primarily by changing its operating cycle and reducing R&D risk) but

does not respond to a tax cut. We trace the asymmetry back to constraints on

risk taking imposed by creditors. Finally, tax loss-offset rules moderate firms’

∗Stern School of Business, New York University,and NBER; †College of Business, City Uni-

versity of Hong Kong; ‡Johnson Graduate School of Management, Cornell University.

Accepted by Christian Leuz. We gratefully acknowledge helpful comments from two anony-

mous reviewers; Morten Bennedsen, Nathan Goldman, Abhiroop Mukherjee, and Stefan

Zeume (our discussants); Eli Bartov,Sanjeev Bhojraj, Robert Bloomfield, Agnes Cheng, Robert

Engle, Michelle Hanlon, Shane Heitzman, Sudarshan Jayaraman, Andrew Karolyi, Anne

Marie Knott, Clive Lennox, Ji-Chai Lin, Kenneth Merkley, Jeffrey Ng, Joseph Piotroski, Mark

Soliman, K.R. Subramanyam, Ross Watts, John Wei, Haibin Wu, Eric Yeung, Paul Zarowin,

and Zilong Zhang; as well as participants at various seminars and conferences. We thank

Charles Choi and Chuchu Liang for competent research assistance. Zuo gratefully acknowl-

edges generous financial support from the Institute for the Social Sciences at Cornell Univer-

sity. An online appendix to this paper can be downloaded at http://research.chicagobooth.

edu/arc/journal-of-accounting-research/online-supplements.

669

Copyright C, University of Chicago on behalf of the Accounting Research Center,2016

670 A.LJUNGQVIST,L.ZHANG,AND L.ZUO

sensitivity to taxes by allowing firms to partly share downside risk with the

government.

JEL codes: G32; H32

Keywords: risk taking; corporate taxes

1. Introduction

Taxation is one of the most important tools governments use to influence

the economy. Taxes affect many aspects of economic activity, from individu-

als’ labor supply, consumption, and savings decisions to companies’ hiring,

location, and capital investment choices. In this paper, we ask how taxes on

corporate income affect corporate risk taking. As Solow [1956] notes, risk

taking is essential for both firms and economies to grow in the long run.

Income taxes affect corporate risk taking because they induce an asym-

metry in a firm’s payoffs. This basic insight can be traced back to early

work on individuals’ risk-taking choices in response to personal income taxes

(Domar and Musgrave [1944], Feldstein [1969], Stiglitz [1969]) and to sub-

sequent applications to the corporate setting (Green and Talmor [1985]).

A simple numerical example serves to illustrate the intuition. Suppose

there are two projects (A and B) and two equally likely outcomes (“good”

and “bad”). Project A yields a profit of $40 in both scenarios; project B

yields a profit of $100 in the good scenario but a loss of $20 in the bad

scenario. Project risk is idiosyncratic and hence diversifiable. Absent taxes,

the expected profit of each project is $40 and so a risk-neutral firm will be

indifferent between the projects.

Now suppose the tax rate increases from zero to 30%. This reduces the

expected after-tax profit of both projects, but risky project B is more af-

fected than safe project A: B’s expected profit falls to $25 while A’s falls

to only $28.1The greater reduction (of $3) in project B’s expected profit

stems from the fact that the government shares in the profit but not in

the loss. Given this asymmetry, a risk-neutral firm will now prefer the safe

project to the risky project.2Generalizing from the example, we predict

that firms should respond to a tax increase by choosing safer projects and

thereby reducing the risks they take.

As Domar and Musgrave [1944] argue, introducing loss-offsets into the

tax code can modify this prediction. Consider the extreme case in which

losses can be completely written off against past or future profits. In this

case, the pretax and posttax ordering of the two projects is identical be-

cause both the upside and the downside are reduced at the same tax rate.3

In practice, the tax code permits at most a partial offset of losses, in which

1For project A, $40 ×(1 – 0.3) =$28; for project B, 0.5 ×[(1 – 0.3) ×$100 – $20] =$25.

2Firms are commonly modeled as being risk-neutral, but this is not crucial to the argument.

3In our numerical example, project B’s expected profit with full loss offsets is $28, the same

as A’s.

HOW TAXES AFFECT CORPORATE RISK TAKING 671

case the upside is reduced by more than the downside. A tax increase will

then reduce the expected profit of the risky project by more than that of

the safe project and firms should respond by reducing risk.

Absent other frictions, these arguments apply symmetrically to tax in-

creases and tax cuts, so firms should respond to cuts by increasing risk.

In practice, there is reason to expect asymmetry. As the literature on risk-

shifting emphasizes, higher risk reduces the value of claims held by cred-

itors. Whether a firm can respond to a tax cut by increasing risk then de-

pends on the extent to which creditors constrain its behavior, for example,

by means of debt covenants. In the presence of such constraints, the effect

of a tax cut on risk taking is likely attenuated for many firms.

A key challenge when testing how taxes affect corporate policies is that

a firm’s tax status is often endogenous to its policies. For example, a firm’s

choice of investment projects will affect its future marginal tax rate. The

literature confronts this identification challenge in various ways. One ap-

proach is to exploit changes in federal income tax rates. Unfortunately,

federal tax changes suffer from two shortcomings: they are few and far be-

tween, and they affect virtually all firms in the economy at the same time,

making it difficult to find control firms with which to establish a plausible

counterfactual. A second approach is to exploit cross-country differences

in tax policies. This typically results in a larger number of tax “shocks” than

in studies using federal tax changes, but often requires implausible assump-

tions about treated firms and their controls being comparable despite op-

erating in different countries.

We adopt a third approach, pioneered by Asker, Farre-Mensa, and

Ljungqvist [2015], Heider and Ljungqvist [2015], and Farre-Mensa and

Ljungqvist [2016]. The approach exploits the fact that U.S. companies pay

not only federal income tax but also taxes in the various states in which

they operate. As Heider and Ljungqvist note, state taxes are a meaningful

part of U.S. firms’ overall tax burden, accounting for about 21% of total

income taxes paid in Compustat. Changes in state corporate income tax

rates are numerous (we count 113 between 1990 and 2011) and, because

they are staggered across states and time, lend themselves to a difference-

in-differences research design. As long as the usual identifying assumptions

plausibly hold, such a design can disentangle the effects of tax changes

from other macroeconomic shocks that affect firms’ risk taking by estab-

lishing a counterfactual—what level of risk would firms have chosen absent

the tax change?—using as controls firms that experience similar economic

conditions (in time, space, and industry) but are not themselves subject to

a tax change.

By way of preview, we have four main results. First, we find that firms re-

spond to tax increases by reducing their earnings volatility, as expected. To

illustrate, a treated firm reduces its earnings volatility by 2.6% in response

to the average tax increase of 136 basis points, relative to other firms in

the same industry that are not subject to a tax change in their headquarter