A major prediction of agency theory is that there is a trade-off between risk and incentive compensation. Aggarwal and Samwick (1999) [Aggarwal, R., Samwick, A., 1999. The other side of the trade-off: the impact of risk on executive compensation. Journal of Political Economy, 107, 65–105.] directly test and find results consistent with agency theory—pay-performance sensitivity is decreasing in risk. However, Prendergast (2002, 2000) [Prendergast, C. 2002. The tenuous trade-off between risk and incentives. Journal of Political Economy 110 (5), 1071–1102; Prendergast, C. 2000. What trade-off risk and incentives? The American Economic Review 90 (2), 421–425.] offers a number of reasons why the sensitivity of pay to performance can be higher in risky environments. We use data from a sample of Internet firms for 1997–1999 to provide empirical evidence on these competing arguments regarding the relation between risk and CEO compensation. Consistent with Aggarwal and Samwick (1999), our results show that pay–performance sensitivity declines with increases in variance in a base model. After controlling for size, we find that pay–performance sensitivity is positively related to risk, consistent with the theoretical predictions in Prendergast (2002, 2000). However, sensitivity tests in later periods show that the Aggarwal and Samwick (1999) results are more robust to changes in the economic environment.