Todd Henderson’s paper finds that compensation arrangements of solvent and insolvent firms are similar to each other. The empirical strategy involves the assumption that firms in bankruptcy are a useful control group for testing agency theory explanations of executive compensation because those costs are significantly lower for insolvent firms. I don’t know enough about bankruptcy to know how “clean” this control group is, but Henderson’s strategy is a simple, creative, and powerful empirical method for resolving competing theories about executive compensation. What would be really nice (from an econometric perspective I mean) to see is a test of how bankrupt and non-bankrupt firms respond to some additional, exogenous shock impacting executive compensation levels. HT: Larry Ribstein.