It’s Not Optimism When You Know You’re Right: Optimism, Attribution And Corporate Investment Policy

Richard Walton

Arizona State University (ASU) – Finance Department

September 15, 2015


CEOs whose observed personal option-holding patterns are not consistent with theoretical predictions are variously described as overconfident or optimistic. Existing literature demonstrates that the investment and financing decisions of such CEOs differ from those of CEOs who do not exhibit such behavior and interprets the investment and financing decisions by overconfident or optimistic CEOs as inferior. This paper argues that it may be rational to exhibit behavior interpreted as optimistic and that the determinants of a CEO’s perceived optimism are important. Further, this paper shows that CEOs whose apparent optimism results from above average industry-adjusted CEO performance in prior years make investment and financing decisions which are actually similar, and sometimes superior to, those of unbiased CEOs.

It’s Not Optimism When You Know You’re Right: Optimism, Attribution And Corporate Investment Policy – Introduction

Since the articulation of the principal-agent problem it has been widely accepted that firms are not necessarily run in the best interests of their owners. The general acknowledgment of the disproportionate impact of the Chief Executive Officer (CEO) on both the behavior and the performance of the firm has prompted research in corporate finance to better understand why CEOs behave as they do. Agency problems occur when the interests of the CEO and the shareholders diverge and the CEO can intentionally follow policies that further his own interests at the expense of those of the shareholders. A different line of research considers how a CEO’s personal bias may cause him to unintentionally act against the best interests of the shareholders even in the absence of any deliberate attempt to further his own advantage. Hirshleifer (2001) presents an overview of biases which can affect investment. Two such biases in the literature are optimism about likely outcomes and a CEO’s overconfidence in the precision of his information. The literature is however not always consistent in its use of the terms “overconfidence” and “optimism”. This paper follows the traditional approach of Hackbarth (2008) which defines optimistic agents as predicting “that favorable future events are more likely than they actually are” and overconfident agents as believing “that they have more precise knowledge about future events than they actually have”. In short, optimists overestimate expected values, and overconfident agents underestimate risk.

There is a broad and growing literature in which the personal option-holding patterns of CEOs are used to identify CEOs with an “upward bias in the assessment of future outcomes” (Malmendier & Tate (2005a)). The bias is labelled overconfidence, by authors who wish to distinguish a CEO’s tendency to overestimate his personal attributes and outcomes from a tendency to overestimate exogenous outcomes, and as optimism, by authors who wish to distinguish it from the tendency to underestimate risk. The bias is generally claimed to be present when a CEO fails to exercise exercisable options until the final year of the options or when a CEO fails to exercise exercisable options exceeding a threshold in-the-moneyness (ITM). Such behavior has been shown to be associated with a variety of negative behaviors. Malmendier & Tate (2005a, 2008) find that a biased CEO’s reluctance to raise external financing makes his investment decisions more sensitive to the availability of internal funds. In addition, CEOs subject to this bias engage in more mergers and acquisitions and that the market reacts less favorably to the announcement of their acquisitions. Deshmukh et al. (2013) finds that a biased CEO’s preference for internal financing causes them to pay lower dividends but the dividend policy deviates less when their firms have growth opportunities. However, there are theoretical and empirical findings demonstrating positive consequences of CEO optimism. Hirshleifer et al. (2012) shows biased CEOs in innovate industries are more successful at investing in innovation. Campbell et al. (2011) claims there is an optimal level of CEO optimism and find that “CEOs with relatively low or high optimism face a higher probability of forced turnover than moderately optimistic CEOs face.” Whilst these behavioral distortions have been identified and well documented, it is less well known how CEOs become subject to the optimism bias and whether inter-temporal variation in the exhibition of this bias can provide information about their causes and the resultant quality of a CEO’s decisions.

This paper considers whether seemingly optimistic behaviors might sometimes be an unbiased response to short-term conditions and therefore looks at the annual variation in CEO measured optimism rather than optimism as a permanent characteristic. Information on CEOs’ optionholdings found in Compustat’s Execucomp database is used to estimate the average in-themoneyness (ITM) of unexercised exercisable options as a measure of a CEO’s optimism. Concentrating on short term variation in optimism allows the identification of the determinants of changes in optimism and the classification of optimistic CEOs into distinct groups. The existing literature suggests that, as a whole, CEOs who exhibit optimistic option-holding behaviors are likely to implement sub-optimal and value-destroying investment and financing decisions. Van den Steen (2004) shows how “choice-driven overoptimism” can result from rational behavior under uncertainty but the CEO’s choice is still ex post incorrect, in that it was the result of inaccurate expectations. Optimistic option-holding behaviors essentially identify CEOs whose expectations of the future value of the firm are more positive than those of the market. This can be the result of bias or a response to superior information concerning the prospects of the firm or the CEO’s abilities. This paper finds that CEOs exhibiting both optimistic option-holding behaviors and superior prior industry-adjusted performance actually implement investment and financing decisions which are comparable to or better than those of non-optimistic CEOs, supporting the hypothesis that, for a significant proportion of CEOs, such behavior is most likely a rational response to temporary conditions rather than the unfortunate consequence of bias.



See full PDF below.