Does Corporate Governance Causally Predict Firm Performance? Panel Data and Instrumental Variables Evidence

Abstract

This study investigates the impact of overall level as well as of separate elements of corporate governance on enterprise performance for open joint-stock companies (OJSC) in transitional country, Ukraine. We use unique data on corporate governance choices for above 5 thousand firms (about a universe of OJSCs in Ukraine) for three years from 2000 to 2002. We construct overall index of corporate governance (UCGI) and sub-indices of corporate governance describing such aspects of corporate governance as shareholder rights, transparency/information disclosure, board independence and chairman independence. We use two different sets of instrumental variables to deal with econometrics problems. First, we use regional variations in social trust factors to instrument for corporate governance choices in cross-sectional framework. Our instruments include political diversity, religion (number of Christian churches) and ethnic diversity across 24 regions in Ukraine. We thoroughly discuss the plausibility of our instruments in a separate section. The hypothesis that our instruments are weak is rejected based on critical values of Stock-Yogo test and thus we use two-stage least squares (2SLS) estimator, and two-stage generalized method of moments (2SGMM) estimator to account for arbitrary heteroskedasticity. Additionally we use time-invariant variables plus time-varying religion variable as instruments after first-differencing transformation. Since in this case we are not able to reject the hypothesis that our instruments are weak we employ also limited-information maximum likelihood estimator (LIML) and testing procedures that are robust to the presence of weak instruments. We find strong evidence that corporate governance predicts firm performance in the transition context. We could not reject the null that UCGI is exogenous in various cross-sectional specifications but we document presence of endogeneity when first differencing estimator employed. Our results predict that one-point-increase in our UCGI index would result in around 0.4-1.9% increase in performance; and based on HOLS results worst to best change in UCGI predicts about 40% increase in company's performance. Additionally we document nonlinear effect of corporate governance suggesting that companies with lower then average level of governance incur losses from an incremental improvement in their governance while those with higher then average level of governance enjoy excessive improvement from that in their performance. We document statistically and economically strong effects of shareholder rights, transparency and board independence on performance. We also find a negative effect of the independence of the board chairman on performance.