Over the past two decades, CEO compensation has interested researchers from various disciplines such as economics, finance and strategic management (Deschenes et al., 2014; Hou, Priem & Goranova, 2014). According to Barkema & Gomez-Mejia (1998), a total of more than 300 studies have accumulated examining the relationship between CEO compensation and various organizational outcomes including firm performance, innovation, corporate strategy and more recently CEO risk-taking (Balkin, Markman & Gomez-Mejia, 2000; Finkelstein & Boyd, 1998; Sanders, 2001; Wright et al., 2007). The vast majority of research in this area of study has been grounded on agency theory prescriptions originally developed by Jensen & Meckling (1976). Despite the large number of studies conducted, the results of this stream of research have best been described as mixed and inconsistent (Barkema & Gomez-Mejia, 1998; Gomez-Mejia & Wiseman, 1997; Pass, 2003). To move this stream of research forward, several suggestions have been advanced to examine the effects of various internal and external contextual factors that might interact with these relationships (Barkema & Gomez-Mejia, 1998; Hambrick, 2007).
Similarly, Hambrick & Mason (1984) call for the use of TMT demographic characteristics to study the content and process of firm strategies has led to an impressive body of empirical investigations in the last twenty years. This body of literature can be broadly classified into two distinct streams. One stream focuses on how the characteristics of the CEO and/or the Top Management Team (TMT) relate to various organizational outcomes (Prasad & Junni, 2017; Laufs, Bembom & Schwens, 2016; Nguyem, Rahman & Zhao, 2018; Herri, Handika & Yulihasric, 2017; Meeks, 2015; Henderson, Miller & Hambrick, 2006; Simsek, 2007; Wei & Ling, 2015). The second stream investigates the dynamics of the relationships between CEO/TMT demographics and their behaviors (Ahn, Minshall & Mortara, 2017; Saeed & Ziaulhaq, 2018; Wei, Ouyang & Chen, 2018; Smith et al., 1994; Simons, Pelled & Smith, 1999). The emergence of the second stream, to a great extent, was a response to the criticisms about the "black-box" of organizational demography (Lawrence, 1997). That is, the upper echelons approach has been often faulted for its use of CEO and/or TMT demographics as proxies of executive's cognitive frames, limiting its accuracy, validity and completeness of appreciation for the real psychological and social dynamics that drive executive behaviors (Hambrick, 2007). Despite this criticism, research using the organizational demography approach has continued to flourish, primarily because of its ability to predict organizational phenomena. As Hambrick points out in his retrospective, there is no sign that interest in upper echelons theory is waning or that the theory has been tapped out.
This study is a response to Hambrick (2007) insightful call for theory and research integrating the combined effects of executive compensation system and CEO characteristics. There has been a growing interest in recent years in CEO risk-taking and/or firm risk-taking (Wright et al., 2007; Simsek, 2007). Our effort is to explain the variance in CEOs' risk-taking behaviors by simultaneously examining both CEO compensation arrangements and CEO characteristics. For example, what are the effects of an aggressive long-term incentive plan on the behavior of a 45-year-old CEO compared to a 65-year-old CEO?
Similarly, what are the effects of an aggressive long-term incentive plan on the behavior of a CEO in his first year compared to a CEO who has been in the saddle for ten years?
These are the two questions this paper is trying to answer.
In order to accomplish the above objectives, the remainder of the paper is organized as follows: The next section reviews relevant literatures and develops our model and hypotheses. We then outline the methodology, data and sample, operationalization of variables and inference methods. Section four presents the findings from the statistical analysis. We conclude by discussing the findings and their implications.
THEORY AND HYPOTHESES
According to agency theory, one important area where the interests of principals and agents diverge is their attitudes and preferences toward risk. While shareholders regard their investment in a firm as a single investment within a well-diversified portfolio, thus diversifying away their risk relatively easily and inexpensively, managers are normally less able to diversify their risks in the forms of employment (Amihud & Lev, 1981), human capital (Wang & Barney, 2006) and ownership stakes in the firm. As a result of this, interests of managers and owners tend to diverge on this issue. While shareholders, who are well-diversified, are in favor of risk-seeking strategies, managers, who are constrained by their firm-specific investments, tend to adopt strategies that reduce risk.
Different mechanisms have been suggested to align the interests of managers and owners and help ensure that managers act on behalf of shareholders, thereby mitigating potential agency problems and costs. These include incentive alignment, monitoring and bonding (Jensen & Meckling, 1976). One important incentive alignment mechanism is executive compensation, which has captured the interest of researchers since the beginning of the past century (Taussig & Barker, 1925). In subsequent decades, a huge body of research has accumulated that examines both the antecedents and outcomes of CEO and/or TMT compensation (Barkema & Gomez-Mejia, 1998). Agency theory holds that the design of executive compensation provides a powerful mechanism to align the traditionally divergent interests of managers and owners, especially since firm outcomes such as performance tend to interact non- uniformly with the various elements of executive compensation (Lewellen, Loderer & Martin, 1987; Murphy, 1985). For example, Lewellen, Loderer & Martin (1987) point out that a component designed to control for one problem (i.e., time horizon) may tend to intensify another problem (i.e., risk exposure).
Agency theorists argue that an appropriately designed executive compensation system should reduce managerial opportunism, promote positive risk-taking attitudes on the part of executives and induce wealth maximizing investment strategies, decisions and behaviors that presumably would enhance firm performance (Jensen & Murphy, 1990). For example, the incentive alignment argument posits that executives who are paid large amounts of long-term incentives such as stock options in their compensation should be reluctant to engage in investments that do not increase shareholders wealth (Amihud & Lev, 1981; Sanders, 2001). However, so far, according to Finkelstein & Hambrick (1996), only limited effort has been made to explore the effects of executive compensation structure on CEO behaviors, especially in regards to strategic decisions and risk-taking behavior (Bloom & Milkovich, 1998; Sanders, 2001; Gilley et al., 2004; Wright et al., 2007). Therefore, according to Bloom & Milkovich (1998), underestimating the important role of risk attitudes and behaviors in this area of research may result in telling only part of the story about whether and when incentive pay leads to positive organizational outcomes.
To what extent can the design of CEO/TMT compensation promote congruence among shareholders and managers and thus induce greater risk-seeking on the part of executives? Despite the agency theory prediction that the introduction of long-term incentives, such as stock options, in CEO compensation can help reduce managerial opportunism and induce shareholder wealth maximizing corporate investment decisions and behaviors (Jensen & Meckling, 1976; Jensen & Murphy, 1990), the actual empirical results seem to be somewhat mixed and more nuanced. For example, DeFusco, Johnson & Zorn (1990) reported a positive relationship between the adoption of stock option plans and managerial risk-taking behavior, measured as variance in both stock price and stock returns. However, later studies provided less consistent evidence. While Hoskisson, Hitt & Hill (1993); Balkin, Markman & Gomez-Mejia (2000) found no significant relationships between executive long-term incentives and managerial risk-taking, a negative relationship was reported by Gray & Cannella (1997).
Sanders (2001) examined the influence of executive incentives on corporate decisions and strategies that entail risk-taking by testing the effects of executive's stock option pay on firm's acquisition and divestiture strategy. He reported that executive option pay had positive relationships with both acquisition and divestiture activities. The author's interpretation that these results lend support to the incentive alignment argument, however, may be problematic in light of prior research on diversification. The general consensus emerging from diversification literature is that agents have the motive to pursue excessive diversification that goes beyond the optimal level for shareholders because diversification is one of the means through which they can reduce their employment risk (Amihud & Lev, 1981). In addition, firm size, according to previous work in this area, is known to be a strong predictor of executive compensation (Hambrick & Finkelstein, 1995; Tosi et al., 2000). From a shareholder's perspective, diversification and acquisition are not beneficial due to the low cost of portfolio diversification in the external capital market (Amihud & Lev, 1981). Therefore, more acquisition activity associated with stock option pay found in this study is more an indicator of interest misalignment than an evidence of alignment. In a similar vein, Wright et al. (2007) attempted to examine the influences of various components of TMT compensation (i.e. salary and bonus, stock options) on subsequent firm...