*3.5. Descriptive Statistics*

Our final sample contains 1625 firms and 11,718 firm-year observations. Table 1 provides the summary statistics of hedging behavior, pension incentive, equity incentive and control variables for this final sample. We find that in general hedging appears in 52% of the firm-years and the notional amount of hedging position is on average USD 19.51 million, an equivalent of about 8.2% of total assets for the full sample of firm-year observations, and rises to USD 260.7 million. These results are consistent with the literature. For example, Campello et al. (2011) report that 50% of their sample firms use foreign currency derivative to hedge and the positions of derivatives account for about 7.5% of firm assets, respectively. As to the pension incentive variables, on average, CEOs are paid with USD 2.63 million of pension benefits. Such liability-based compensation constitutes 36% of CEOs' total compensation or 54% of equity-based compensation.

The variable of CEO pension relative leverage has a mean (median) value of 4.72 (0.30) in our final sample. Less than half (39%) of the firm-year observations have this variable greater than one. For equity incentive, the mean values of Delta incentive and Vega incentive are 0.14 and 0.01, respectively. We also note that the ratio of Vega/Delta has an average value of 0.24, suggesting that risk-taking incentive tied to stock volatility represents about a quarter of the value-increasing incentives driven by stock price. Our sample firms on average have a moderate leverage ratio of 0.18 and a healthy interest coverage ratio of 24.61. On average market-to-book ratio is 1.66 and the tangible assets ratio is 76%. The correlation between cash flow and firm investments has a mean (median) value of 0.41 (0.48). In addition, the sample firms hold 18% assets in cash and 3% assets in convertible bond contracts. We also find that 56% sample firm-year observations have a positive tax credit and the average tax convexity is USD 2.13 million, both suggesting that the tax benefits associated with smoothing incomes are considerably attractive. Finally, about 71% of the board members are from out of the firms to serve as independent directors and 33% firm stakes are held by the institutional blockholders.


**Table 1.** Summary statistics.

This table shows summary statistics of pension incentive variables, hedging variables, equity incentive variables, and firm-related variables for the sample of 11,718 firm-year observations from 2006 through 2015 with winsorization at the 1th and 99th percentiles. All variable definitions are reported in Appendix A.

Moreover, we present the Pearson correlation matrix of hedging activity and incentive variables in Table 2, which indicates a positive relation between hedging activity and CEO pension incentives. For example, *CEO Pension Relative Leverage* is significantly positive correlated with both of hedging measures, including hedging propensity (0.36) and hedging intensity (0.35). These relations suggest that greater pension incentives may lead to more hedging activities. In addition, Delta incentive and Vega incentive—these two equity-based incentives—are found to have a different correlation with hedging activity. Note that Delta incentive has an insignificant correlation with hedging, while Vega incentive has a significantly negative correlation with hedging, which is consistent with the literature of equity compensation that granting equity, particularly stock options to managers should, ceteris paribus, generate incentives to take more risks, or consequently hedge less.

**Table 2.** Correlations matrix between hedging and pension incentive.


This table presents the Pearson correlation coefficients between hedging variables and CEO pension incentive for the final sample from 2006 through 2015. \* denotes statistical significance under 5% level. All variable definitions are reported in Appendix A.
