Standard principal-agent theory predicts that the pay-performance sensitivity (PPS) decreases in the risk of the firm. An alternative literature argues that entrenched executives as in weakly governed firms use compensation contract to extract the rent, which renders risk irrelevant in determining PPS. This paper uses event study approach to test both principal-agent model and CEO power theory by studying changes in executives’ compensation contract around litigation events. Consistent with principalagent model prediction, we find that, after the initiation of litigation, PPS drops, compensation shifts from performance-sensitive component (equity) to performanceinvariant component (cash). In addition, all the changes reverse themselves after litigation settlements. To test CEO power theory, we further partition the event firms into firms with good and bad corporate governance. We find that the PPS in firms with bad corporate governance increases after lawsuit and decreases after the settlement. This suggests that litigation brings the bad compensation practice of poorly governed firms to the limelight and forces firms to discipline their CEOs temporarily during the litigation period (so called “limelight effect”), which lends indirect support to CEO power theory. Our results are robust to a battery of sensitivity tests including two endogeneity tests.