The Impact of Ambiguity on Managerial Investment and Cash Holdings

• Date: 01 September 2014
• Author: Monica Neamtiu,Christopher D. Williams,Hal D. White,Nemit Shroff
• DOI: http://doi.org/10.1111/jbfa.12079
• Published date: 01 September 2014

Journal of Business Finance & Accounting

Journal of Business Finance & Accounting, 41(7) & (8), 1071–1099, September/October 2014, 0306-686X

doi: 10.1111/jbfa.12079

The Impact of Ambiguity on Managerial

Investment and Cash Holdings

MONICA NEAMTIU,NEMIT SHROFF,HAL D. WHITE

AND CHRISTOPHER D. WILLIAMS∗

Abstract: Standard finance theory suggests that managers invest in projects that, in expecta-

tion, produce returns that justify the use of capital. An underlying assumption is that managers

have the information necessary to understand the distributional properties of the pay-offs

underlying the decision. This paper examines firm investment behavior when managers are

likely to find it more challenging to develop expectations of pay-offs, namely during periods of

increased macroeconomic ambiguity. In particular, we examine how macroeconomic ambiguity

– proxied by the variance premium (Drechsler, 2010) and the dispersion in forecasts of

corporate profits from the Survey of Professional Forecasters (Anderson et al., 2009) – impacts

managerial capital investment and cash holdings. Consistent with ambiguity theory, we find

that macroeconomic ambiguity is negatively associated with capital investment and positively

associated with cash holdings. These results are robust to alternative explanations related to risk,

investor sentiment and economic conditions. Moreover, consistent with recent theoretical real

options literature, we find that ambiguity reduces the value of investment opportunities, while

risk increases the value of such opportunities. Overall, these findings provide initial empirical

evidence on the economic distinction between ambiguity and risk with respect to managerial

investment and cash holdings.

Keywords: ambiguity, risk, investment, cash holdings

1. INTRODUCTION

Within a firm, the allocation of capital to its highest value use is one of the most

important roles of a manager. The central determinant of successful capital allocation

is the precision with which a manager is able to identify cash flow prospects and

apply appropriate discount rates for the respective investment decisions. Typically,

the investment literature assumes that managers know, or behave as if they know,

∗The first author is an Associate Professor of Accounting at the University of Arizona, Tucson, AZ. The

second author is an Assistant Professor of Accounting at Massachusetts Institute of Technology, Boston,

MA. The third and fourth authors are Assistant Professors of Accounting at the University of Michigan,

Ann Arbor, MI. The authors would like to thank Andrew Stark (Editor), an anonymous referee, Ilia Dichev,

Raffi Indjejikian, Feng Li, Russell Lundholm, Greg Miller, Mike Minnis, Marlene Plumlee and workshop

participants at the 2010 University of Michigan Harvey Kapnick Conference and the Seoul National

University for useful comments and suggestions.

Address for correspondence: Christopher D. Williams, Ross School of Business, University of Michigan, 701

Tappan Street, Ann Arbor, MI 48109, USA.

e-mail: williacd@umich.edu

C

2014 John Wiley & Sons Ltd 1071

1072 NEAMTIU, SHROFF, WHITE AND WILLIAMS

the probability distributions of the cash flows and discount rates related to potential

projects. In other words, although project pay-offs are ex-ante unknown, managers feel

confident in their assessment of investment pay-off probabilities.

Recent theoretical work by Nishimura and Ozaki (2007), however, argues that

managers facing a decision regarding irreversible investment may not always have

complete confidence in their perceived pay-off probability measures. As Nishimura

and Ozaki (2007) state, managers “might think other probability measures perturbed

from the original one are also possible.” This uncertainty regarding investment pay-

off – characterized by a set of probability measures rather than a single probability

measure – is defined as ambiguity, or Knightian uncertainty.1The intuition developed

in Nishimura and Ozaki (2007) draws from the large theoretical and experimental

literature on ambiguity (e.g., Ellsberg, 1961; Gilboa and Schmeidler, 1989; Bewley,

2002), which builds on Knight (1921).2

The purpose of this paper is to examine the impact of ambiguity on managerial

investment and cash holding decisions. The ambiguity literature suggests that, in the

presence of ambiguity, ambiguity-averse individuals behave as if they assume “worst-

case’ outcome scenarios – i.e., they incorporate the worst possible pay-off probability

measure in their consideration set into their decision. With respect to managerial

investment, ambiguity-averse managers place a lower reservation price on potential

projects than would be predicted using a rational expectations framework (Nishimura

and Ozaki, 2007). As a result, their investment level declines, as anticipated project

pay-offs do not meet the lower reservation prices and they shift to holding riskless

assets instead (Dow and Werlang, 1992; Epstein and Schneider, 2010). Following this

intuition, we predict that when managers find it more difficult to develop expectations

of future investment pay-offs (i.e., when they are faced with ambiguity), they are likely

to reduce their capital expenditures and increase their cash holdings.3

To capture the impact of ambiguity, we develop our models of investment and cash

holdings based on prior literature. In particular, we model a firm’sinvestment decision

using Euler equations (e.g., Whited, 1992; Hubbard et al., 1995; Love, 2003), and we

model a firm’s cash holdings following Opler et al. (1999).4Our proxies for ambiguity

are the variance premium (Drechsler, 2010) and dispersion in forecasts of corporate

profits by the Survey of Professional Forecasters (Anderson et al., 2009).

1 The term “Knightian uncertainty” is based on the seminal work of Knight (1921), which points out that

ambiguity is a fundamentally different concept than risk. Specifically, risk relates to known, or objective,

uncertainty, as in the roll of a die, where the outcome probabilities are known, or the probabilities can

be estimated with confidence. In contrast, ambiguity relates to subjective uncertainty, where outcome

probabilities are unknown, and the decisionmaker fears the estimated model is incorrect. Note that,

although both risk and ambiguity ultimately relate to uncertainty regarding project pay-offs, ambiguity

relates to uncertainty as to the probability measure governing those outcomes as well. In this paper, we

use the term ‘uncertainty’ to mean either risk or ambiguity or both. That is, uncertainty is comprised of

both risk and ambiguity. Weuse the terms ambiguity and “Knightian uncertainty” interchangeably.

2 See Epstein and Schneider (2010) for a review of the ambiguity literature.

3 Note that a manager’s ability to develop expectations regarding future investment pay-offs may also

be affected by the manager’s overconfidence. However, in the context of investment, we believe that

overconfidence is likely to work in the opposite direction as ambiguity. Specifically, we expect ambiguity

to decrease the manager’s precision of the pay-off distribution and thus lead to less investment. In contrast,

as stated in Gervais et al. (2011, p.1,738), “overconfidence captures the idea that individuals overestimate

the precision of their information or their ability to interpret that information when they make economic

decisions.” If overconfident managers overestimate the precision of their investment pay-off distribution, we

would expect overconfidence to lead to greater investment, biasing against our hypothesized results.

4 We also model cash holdings following Ozkan and Ozkan (2004) as an alternative analysis.

C

2014 John Wiley & Sons Ltd