臺大管理論叢 NTU Management Review VOL.28 NO.3

Does CEO Career Horizon Lead to Corporate Misconduct? Evidence of Taiwanese Semiconductor Firms 162 As Table 2 shows, we first enter the control variables in the baseline model (Model 1), and then enter CEO career horizon and its squared term in Model 2. The results for Model 2 reveal that the linear term and squared term of CEO career horizon separately have significant positive ( p < 0.05) and negative ( p < 0.05) effects on the likelihood of corporate misconduct. Hypothesis 1 is thus supported by our findings. Hypothesis 2 predicts that board size can negatively moderate the relationship between CEO career horizon and corporate misconduct. According to the existing literature, there is no standard method to test the moderating effect on the nonlinear relationship. This study thus used Kim, Kim, and Lee’s (2008) approach to estimate the moderating effect on an inverted U-shaped relationship. Firstly, we estimated the direct effect of board size on corporate misconduct in Model 3. The results reveal that the direct effect of board size is negatively, but not significantly, correlated with corporate misconduct. Such a finding shows that large board size may not always imply better monitoring (Amar, Boujenoui, and Francoeur, 2011). In Model 4, we further examined the moderating effect of board size on the inverted U-shaped relationship between CEO career horizon and corporate misconduct. Although the coefficient on the interaction term is negative and consistent with Hypothesis 2, it fails to reach statistical significance. Hypothesis 2 is thus not supported. As reported in Model 5, the significant negative effect of board functional diversity ( p < 0.01) echoes the early studies, and reveals that the functional diversity of board members can fundamentally shape effective board governance (Bear, Rahman, and Post, 2010). The result shown in Model 6 indicates that the interaction of CEO career horizon and board functional diversity is negatively significant ( p < 0.05), suggesting that the inverted-U effect of CEO career horizon on corporate misconduct is more moderate. Hypothesis 3 is thus supported. To illustrate the complex interaction effect, drawing upon the recommendations of Richard, Barnett, Dwyer, and Chadwick (2004), we plotted the moderating effect of board functional diversity on the relationship between CEO career horizon and corporate misconduct, as shown in Figure 1. Specifically, we divided the entire sample two subgroups (high board functional diversity vs. low board functional diversity) by the median of the board functional diversity and reran Model 2 for each of the two groups; the nuanced differences between the groups can demonstrate the effects of the moderator. Figure 1 shows that, in the firms with low board functional diversity, there is a strong inverted U-shaped relationship between CEO career horizon and corporate misconduct. In the firms with high board functional diversity, the inverted U-shaped curve