CEO overconfidence and corporate risk taking: Evidence from pension policy

• DOI: http://doi.org/10.1002/jcaf.22470
• Author: Joohyung Ha,Carol M. Graham,Cathy S. Goldberg
• Date: 01 October 2020
• Published date: 01 October 2020

BLIND PEER REVIEW

CEO overconfidence and corporate risk taking: Evidence

from pension policy

Cathy S. Goldberg | Carol M. Graham | Joohyung Ha

School of Management, University of San

Francisco, San Francisco, California

Correspondence

Cathy S. Goldberg, School of

Management, University of San Francisco,

San Francisco, CA.

Email: goldberg@usfca.edu

Abstract

We examine the relationship between managerial overconfidence and corporate

risk taking. More specifically, we investigate how overconfidence affects the risks

that CEOs take in managing their firm's pension plans. Using both options-based

and firm-based measures to proxy for CEO overconfidence, we find that over-

confident CEOs are more likely to take on greater risk in managing their pension

plans by increasing the amount of pension assets invested in equities and the vola-

tility of the pension fund. Our results are significant in the post-global financial cri-

sis of 2008 during a period when data indicates average equity allocation in pension

assets decreased thus highlighting that overconfident managers behaved contrary to

the overall market. This study adds to the literature on both managerial over-

confidence and its consequences, and on the drivers of corporate pension policy.

KEYWORDS

CEO overconfidence, pension plans, risk taking

1|INTRODUCTION

Research shows that overconfidence leads corporate

managers to take risk.

1

This dynamic of overconfidence

leading to risk manifests itself in a host of corporate poli-

cies where CEOs have discretion. This article examines

overconfidence in the context of pension policy, building

upon similar studies of managerial overconfidence in

other areas of finance, such as merger activity, dividend

payout, earnings management, and earnings forecasts.

CEOs tend to be overconfident; they take more risk and

are more optimistic than the general population. Further-

more, these personal traits affect corporate financial policies

(Graham et al., 2013). Levels of optimism have differential

effects on how managers approach investing (Campbell

et al., 2011). Because overconfident managers underesti-

mate downside risk and overestimate expected cash flows

from investments, they are more likely to increase corpo-

rate risk relative to other CEOs (Banerjee et al., 2015).

Roll (1986) first introduced overconfidence (or the

“hubris hypothesis”) in finance as a possible explanation

for some of the observed differences between the theory

and practice of managers with respect to takeover activ-

ity. Since Roll's study, the firm-level consequences of

managerial overconfidence have been the focus of a large

number of studies in accounting and finance. Previous

research has found that managerial overconfidence

affects key investment and financing decisions.

Malmendier and Tate (2008) examine the impact of over-

confidence on merger activity. Since overconfident man-

agers overestimate their ability to generate returns, they

note that managers pay too much for target firms. Their

results indicate that overconfident managers make riskier

investment decisions that consequently destroy value for

shareholders. This is especially true when they have

available internal financing and do not need to resort to

external financing. In a complementary study, Campbell

et al. (2011) show that low, moderate and high levels of

optimism have differential effects on how managers

approach investing. With regards to financing activities,

Malmendier and Tate (2005) find that overconfident man-

agers are less likely to seek external financing, and equity

Received: 21 December 2019 Revised: 3 August 2020 Accepted: 18 August 2020

DOI: 10.1002/jcaf.22470

J Corp Acct Fin. 2020;31:135–153. wileyonlinelibrary.com/journal/jcaf © 2020 Wiley Periodicals LLC 135

financing in particular, because they believe that these

sources of financing are overpriced. Overconfident man-

agers are likely to use more debt (in particular short-term

debt) and have a lower likelihood of paying dividends.

Sunder et al. (2010) further investigate the impact of over-

confidence on debt contracts and find that for merger-

related activities, overconfident CEOs are more likely to

experience restrictions on their investing activities as

reflected in debt covenant design.

Aside from key investment and financing decisions,

research also shows that managerial overconfidence has

an impact on financial reporting and forecasting. Ahmed

and Duellman (2012) examine whether overconfidence is

associated with accounting conservatism and find a nega-

tive and significant relationship even after controlling for

known determinants of conservatism. Prior studies also

show that not only is there a positive relationship

between managerial overconfidence and the decision to

issue voluntary forecasts, but that overconfidence affects

the accuracy of the forecasts and leads managers to inject

more of an upward bias into their forecasts. (Libby and

Rennenkemp (2011), and Hribar and Yang (2016)). Fur-

thermore, this optimistic bias appears to have a cascade

effect that manifests itself in earnings management

behavior (Schrand and Zechman (2012)).

More than a decade ago, Ben-David et al. (2007) argued

that “theeffectofoverconfidenceshouldbeexplicitly

modeled when analyzing corporate decision-making”.From

the research summarized above, it is clear that the impor-

tance of individual heterogeneity in corporate finance and

corporate governance has indeed become a primary focus in

behavioral finance (Graham et al., 2013) and can potentially

help explain a wide variety of corporate decisions.

Given that prior research has consistently docu-

mented the impact of managerial overconfidence on key

investment and financing decisions, we anticipate that

managerial overconfidence will also have an impact on

the risks that CEOs take with regard to the pension plans

that they manage.

The paper is structured as follows: Section 2 provides

our hypotheses development. Section 3 outlines our

research design and sample selection. Descriptive statistics

are included in Section 4. Section 5 provides a discussion

of our empirical results and Section 6 concludes the paper.

2|HYPOTHESES DEVELOPMENT

2.1 |The context of pension plans

Our study focuses on the impact of managerial over-

confidence and risk taking in the context of pension plan

investment risk. It is widely understood that managers

can exercise discretion with pension-related assumptions

and pension funding in both the private and public sec-

tors. Comprix and Muller III (2011) argue that due to the

lack of specific guidance in the Accounting Standards

Codification (ASC) 715: Compensation-Retirement Bene-

fit (Financial Accounting Standards Board, 2009) that

governs pension-related accounting issues, private com-

panies have considerable discretion over the choice of

pension-related assumptions. Bonsall et al. (2019) argue

that, at the state level, gatekeepers such as auditors or

actuaries do not play an important role in limiting the

opportunistic use of pension plan assumptions due to

their limited market power. Furthermore, the extant liter-

ature shows that managerial incentives affect pension

strategy. For example, Anantharaman and Lee (2014)

find that the structure of executive compensation affects

risk shifting through pension underfunding and the allo-

cation of pension assets to risky securities.

Given that managers can exercise considerable discre-

tion over pension plan assets and the assumptions under-

lying pension-related liabilities, the factors that influence

the risk taken by managers in this context is arguably an

important area of interest. We investigate whether CEO

overconfidence can explain risk taking as reflected in the

asset allocation decisions of managers and subsequent

volatility of fund portfolios. Our results indicate that

overconfidence is directly and significantly, related to

pension investment risk.

By examining overconfidence and risk-taking in the

context of pension accounting, our paper adds to the lit-

erature on both the effects of managerial overconfidence,

and on pension policy drivers. As far as we are aware,

our study is the first to investigate these issues in tandem.

2.2 |Hypotheses development

Previous research on defined-benefit pensions has exam-

ined factors that affect variations in pension fund invest-

ment strategies such as tax provisions (Black, 1980;

Frank 2002; Tepper 1981), compensation structures

(Anantharaman & Lee, 2014), economic factors and

sponsor firm characteristics (Bartram, 2018). Evidence

from this stream of research remains mixed, however,

and calls for further research in the area of pension fund

policies and investment strategies persist (Rauh, 2009;

Bartram, 2018). What previous defined-benefit pension

studies have neglected to focus on, are the behavioral

characteristics of those who manage, and exercise discre-

tion over the fund.

Malmendier and Tate (2008) show that overconfident

managers overestimate their abilities to generate returns

and thus overpay for merger targets. Along these same

136 GOLDBERG ET AL.