Relative Performance Benchmarks: Do Boards Get it Right?

Abstract

Standard principal-agent models suggest that boards design incentive contracts that filter out common shocks in performance to motivate costly effort from the CEO ― a process entailing the judicious selection of benchmarks for relative performance evaluation (RPE). We evaluate the efficacy of firms' chosen RPE benchmarks and document that, relative to a normative benchmark, index-based benchmarks perform 14% worse in their time-series return-regression R2 and 16% worse in measurement error variance; firms choosing specific peers only modestly under-perform. Structural estimates suggest that, absent frictions, the underperformance of index-based benchmarks imply a performance penalty of 106-277 basis points in annual returns. Consistent with these estimates, firms choosing index-based benchmarks exhibit lower annual returns and ROA. Finally, reduced-form analyses suggest that the inefficient benchmarking is associated with governance-related frictions. Collectively, these findings provide new evidence on the explicit practice of RPE and its implications for corporate governance and firm performance.