What is the Relation (If Any) between a Firm's Corporate Governance Arrangements and its Financial Performance?

We present a model to test the null hypothesis that firms organize their corporate governance arrangements optimally given the constraints they face. Following the literature, the model rejects the null if the conditional correlation between governance and performance is significantly different from zero. Our model provides a clean test of this hypothesis by controlling for measurement errors in all observed variables and avoiding simultaneous equation biases by casting our model as a reduced-form bivariate equation. We model governance, performance and the constraints on the firm's investment decisions as latent variables. Our estimate of the conditional correlation between our measures of governance and performance is statistically speaking equal to zero, which therefore provides empirical support for the in-equilibrium view proposed by Demsetz (1983), of corporate governance arrangements