Purpose The purpose of this study is to analyze the CEO compensation sensitivity to firm performance, termed as the pay-for-performance sensitivity (PPS) in the Tunisian context and to test the robustness of this relationship when corporate governance (CG) mechanisms are considered. Design/methodology/approach The consideration of past executive pay as one of the explanatory variables makes this estimation model a dynamic one. Furthermore, to avoid the problem of endogeneity, this study uses the system-GMM estimator developed by Blundell and Bond (1998). For robustness check, this study aims to use a simultaneous equation approach (three-stage least squares [3SLS]) to estimate the link between performance and CEO pay with a set of CG mechanisms to control for possible simultaneous interdependencies. Findings Using a sample of 336 firm-years from Tunisia over the 2009–2015 periods, this study finds strong evidence that the pay-performance relationship is insignificant and negative, and it becomes more negative or remains insignificant after introducing CG mechanisms consistently with the managerial power approach. The findings are robust to the use of alternative performance measures. This study provides new empirical evidence that CEOs of Tunisian firms abuse extracting rents independently of firm performance. Originality/value This study contributes to the unexamined research on PPS in a frontier market. This study also shows the ineffectiveness of the Tunisian CG structure and thus recommends for the legislator to impose a mandatory CG guide.