A Replication of “Are Competitive Banking Systems More Stable?” (Journal of Money, Credit, and Banking, 2009)

This study replicates Schaeck, Čihák, and Wolfe (2009), henceforth SCW, and performs a variety of robustness checks. Using a cross-country, time series sample of 45 countries from 1980-2005, SCW investigate the relationship between competition and concentration in the banking system, and the occurrence of country-level systemic crises. Their primary measure of competition in the banking industry is Panzar and Rosse’s H-statistic. Concentration is measured using a concentration ratio of the three largest banks. They conclude that (i) competition and concentration measure two separate dimensions of the banking sector, and (ii) greater competition is associated with fewer systemic crises. Using data and code provided by the authors, we are able to exactly reproduce the original results of SCW. However, we find that their results are not robust. In none of our many robustness checks do we find that H-statistic is significant at the 5-percent level. Further, the estimated effect sizes are small and economically inconsequential. Unfortunately, our results with respect to concentration are inconsistent, depending on the estimation procedure, variable specification, and measure of financial stability that we employ. Thus, while we are confident in concluding that competition, as measured by H-statistic, plays an inconsequential role with respect to systemic crises, further research needs to be done to better understand the importance of concentration in the banking sector.