NBER WORKING PAPER SERIES
CEO OVERCONFIDENCE AND CORPORATE INVESTMENT
Ulrike Malmendier
Geoffrey Tate
Working Paper 10807
http://www.nber.org/papers/w10807
NATIONAL BUREAU OF ECONOMIC RESEARCH
1050 Massachusetts Avenue
Cambridge, MA 02138
September 2004
We are indebted to Brian Hall and David Yermack for providing us with the data. We are very grateful to
Jeremy Stein for his invaluable support and comments. We also would like to thank Philippe Aghion, George
Baker, Stefano DellaVigna, Edward Glaeser, Rick Green, Brian Hall, Oliver Hart, Caroline Hoxby, Dirk
Jenter, Larry Katz, Tom Knox, David Laibson, Andrei Shleifer and various participants in seminars at
Harvard University, MIT, University of Chicago, Northwestern University, UC Berkeley, Stanford
University, UCLA, CalTech, Yale University, Univeristy of Michigan, Duke University, NYU, Columbia
University, Wharton, LSE, CREST, CEMFI, LMU Munich, and at the EER, the Russell Sage Summer
Institute for Behavioral Economics and at the SITE for helpful comments. Mike Cho provided excellent
research assistance. Malmendier acknowledges financial support from Harvard University (Dively
Foundation) and the German Academic Exchange Service (DAAD). The views expressed herein are those
of the author(s) and not necessarily those of the National Bureau of Economic Research.
©2004 by Ulrike Malmendier and Geoffrey Tate. All rights reserved. Short sections of text, not to exceed
two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is
given to the source.
CEO Overconfidence and Corporate Investment
Ulrike Malmendier and Geoffrey Tate
NBER Working Paper No. 10807
September 2004
JEL No. G31, G32, D21, D23, D82
ABSTRACT
We argue that managerial overconfidence can account for corporate investment distortions.
Overconfident managers overestimate the returns to their investment projects and view external
funds as unduly costly. Thus, they overinvest when they have abundant internal funds, but curtail
investment when they require external financing. We test the overconfidence hypothesis, using panel
data on personal portfolio and corporate investment decisions of Forbes 500 CEOs. We classify
CEOs as overconfident if they persistently fail to reduce their personal exposure to company-specific
risk. We find that investment of overconfident CEOs is significantly more responsive to cash flow,
particularly in equity-dependent firms.
Ulrike Malmendier
Graduate School of Business
518 Memorial Way
Stanford University
Stanford, CA 94305-5015
and NBER
ulrikem@stanford.edu
Geoffrey Tate
Wharton School
University of Pennsylvania
tate@wharton.upenn.edu
In this paper, we argue that personal charact eristics of CEOs in large corporations lead to
distortions in corporate investmen t policies. In particular, we study the investment deci-
sions of CEOs who overestimate th e future returns of their companies, measured by failure
to divest company-specific risk on their personal accounts. We find that overconfident CEOs
have a heightened sensitivity of corporate investment to cash flow, particularly among equity-
dependent firms.
The two traditional explanations for investment distortions are the misalignment of managerial
and shareholders int erests (Jensen and Meckling (1976), Jensen (1986)) and asymmetric infor-
mation between corporate insiders and the capital market (Myers and Majluf (1984)). Both
cause investment to be sensitive to the amount of cash in the firm. Under the agency view,
managers overinvest to reap private benefits such as “perks,” large empires, and en trenchment.
Since the ex ternal capital market limits the extent to which managers can pursue self-interested
investment, an influx of cash flow enables the manager to invest more and increases investment
distortions. Under asymmetric information, the managers themselves (who act in the interest
of shareholders) restrict external financing in order to avoid diluting the (undervalued) shares
of their company. In this case, cash flow increases investment, but reduces the distortion.
The empirical literature, starting with Fazzari, Hubbard and Petersen (1988), confirms the
existence and robustness of investment-cash flow sensitivity after controlling for investment
opportunities. While most of the literature relates investment-cash flow sensitivity to imper-
fections in the capital market, this interpretation remains controversial (Kaplan and Zingales,
(1997), (2000); Fazzari, Hubbard and Petersen, (2000)).
We propose an alternative explanation for investment-cash flow sensitivity and suboptimal
1
investment behavior. Rather than focusing on firm-level characteristics, w e relate corporate
investment decisions to personal characteristics of the top decision-maker inside the firm. Build-
ing on Roll (1986) and Heaton (2002), we argue that one important link between investment
levels and cash flow is the tension between the beliefs of the CEO and the market about the
value of the firm. Overconfident CEOs systematically overestimate the return to their invest-
ment projects. If they have sufficient internal funds for investment and are not disciplined by
the capital market or corporate governance mechanisms, they overinvest relative to the first
best. If they do not ha ve sufficient internal funds, however, they are reluctant to issue new
equity because they perceive the stock of their company to be undervalued by the market. As
a result, they curb their investment. Additional cash flow provides an opportunity to invest
closer to their desired level.
Our overconfidence story builds upon a prominent stylized fact from the social psychology
literature, the “better than average” effect. When individuals assess their relative skill, they
tend to overstate their acumen relative to the average (Larwood and Whittaker, (1977); Sven-
son, (1981); Alic ke, (1985)). This effect extends to economic decision-making in experiments
(Camerer and Lovallo (1999)). It also affects the attribution of causality. Because individuals
expect their behavior to produce success, they are more likely to attribute good outcomes to
their actions, but bad outcomes to (bad) luck (Miller and Ross (1975)). Executives appear to be
particularly prone to display overconfidence, both in terms of the “better-than-average effect”
and in terms of “narrow confidence intervals” (Larwood and Whittak er, (1977); Kidd, (1970);
Moore, (1977)).
1
This finding is attributed to three main factors, which trigger overconfidence:
the illusion of control, a high degree of commitment to good outcomes, and abstract reference
2
poin ts that make it hard to compare performance across individuals (Weinstein (1980); Alicke
et al. (1995)). All three factors are pertinent in the context of corporate investment. A CEO
who hand-picks an investment project is likely to believe he can control its outcome and to
underestimate the likelihood of failure (March and Shapira (1987); Langer, (1975)). The typi-
cal CEO is also highly committed to good company performance since his personal wealth and
the value of his human capital fluctuate with the company’s stock price. And finally, assessing
relative managerial skill or, specifically, the abilit y to pick profitable investment projects is
difficult — ev en ex post — due to other factors that influence overall firm performance.
Heaton (2002) first showed that common distortions in corporate in vestment may be the result
of managers overestimating the returns to their investment. We expand on Heaton’s insight in
two ways. First, we model how the pre-existing capital structure affects the role of overconfi-
dence. Second, we empirically test the predictions of the model.
To construct measures of overconfidence, we exploit the over-exposure of typical CEOs to their
firms’ idiosyncratic risk. CEOs receive large grants of stock and options as compensation. They
cannot trade their options or hedge the risk by short-selling company stock, and the value of
their human capital is intimately linked to the firm’s performance. Because of this under-
diversification, risk-averse CEOs should exercise their options early given a sufficiently high
stock price (Lambert, Larcker and Verrechia (1991); Meulbroek, (2001); Hall and Murphy,
(2000), (2002)). We take two main approaches to translate this logic into overconfidence
measures. First, we identify a benchmark for the minimum percentage in the money at which
CEOs should exercise their options for a given y ear immediately following the vesting period.
If a CEO persisten tly exercises options later than suggested by the benc hmark, we infer that he
3
