Berrone and Gomez-Mejia (2009)

Berrone, Pascal and Luis R. Gomez-Mejia. "Environmental Performance and Executive Compensation: An Integrated Agency-Institutional Perspective." Academy of Management Journal, 2009.

Explores the relationship between environmental performance and executive pay for 469 U.S. firms (2,008 firm-level observations) over the 1997-2003 time period, focusing on companies in polluting industries. Citing prior research showing potential financial and valuation impacts to poor environmental performance, the authors argue that "[i]t makes sense that firms should reward their executives for environmental actions that confer greater legitimacy and may directly or indirectly improve firm performance."

The study tests the following hypotheses:

Environmental performance has a positive effect on CEO total pay.
Evidence of pollution prevention strategies has a greater impact on CEO total pay than evidence of end-of-pipe pollution control strategies.
(a) The positive effect of environmental performance on CEO's pay is higher for firms that have environmental governance mechanisms in place. (b) Firms with environmental governance mechanisms in place put greater emphasis on pollution prevention than on end-of-pipe pollution control as a criterion for CEO pay.
(a) Long-term pay has a positive effect on subsequent environmental performance. This positive effect is greater on pollution prevention than end-of-pipe results. (b) Long-term pay has a greater effect on subsequent environmental performance as a firm's presence in polluting industries increases. This positive moderating effect is greater for pollution prevention than end-of-pipe result.

The empirical analysis consists of a group of cross-sectional time series regressions.

For hypotheses 1 and 2, the dependent variable is CEO total pay, "the sum of salary, bonus, and all long-term components of CEO pay (stock options, restricted stock, and other long-term compensation). Rather than use an option pricing formula, the authors value stock option grants at 25% of their exercise value - the authors report that "separate analyses using Black-Scholes values yielded results almost identical to those reported here." The key independent variable is environmental performance, which is measured using Toxic Release Inventory data from the Environmental Protection Agency, weighted by 'human toxicity potential factor', an adjusted method introduced by Hertwich (2001). Many control variables are used, notably firm size, ROE, and Tobin's Q. This analysis shows that environmental performance has "a positive and significant effect on CEO total pay." The authors comment that "although the increase in variance was modest, the impact of environmental performance on CEO compensation was not due to random chance."

For hypothesis 3, the dependent variable is again CEO total pay, but the environmental measure used is 'end-of-pipe pollution control', defined as "a ratio in which the numerator was the sum of chemicals recycled, treated on-site, and transferred to other locations for further treatment, and the denominator was the total waste generated by a firm." Control variables were that same as those used for hypotheses 1 and 2. Results of this analysis were inconclusive, and the authors comment that "as a whole these results fail to support both Hypothesis 3a and Hypothesis 3b."

For hypothesis 4, the dependent variable is environmental performance, with long-term pay (total compensation excluding salary and bonus) as the as the primary independent variable and controls as in prior regressions. This analysis appears to support Hypothesis 4a and Hypothesis 4b.

The authors comment that "good environmental behaviors are important in achieving social legitimacy; firms should support environmental strategies by using environmental performance criteria to reward their CEOs. Such incentives can help shareholders, managers, and the general public benefit from good environmental performance."

The paper includes a very strong literature review.