Using a new data set from the Office of the Superintendent of Financial Institutions, I conduct an in-depth study on cost efficiency and returns to scale (RTS) in Canadian banking. I estimate a transcendental log cost function for the six largest Canadian commercial banks which account for approximately 90 percent of chartered bank assets over the 1996-2011 sample period. The minimal amount of firm entry and exit simplifies many difficulties in the analysis, and the panel dynamic ordinary least squares estimator (PDOLS) provides less biased results than the fixed-effect OLS. Departing from previous studies in banking, I calculate whether the estimated cost function satisfies the microeconomic properties of a monotonicity and price concavity. To my knowledge, this is the first paper to find evidence of constant RTS among the Canadian banks. The result is robust to a number of different asset and price specifications. Furthermore, there is little evidence to suggest cost inefficiencies among the large Canadian banks. This is true whether the Greene (2005) true fixed effects ML estimator is estimated or a distribution-free approach is measured. Combining these two results, the large Canadian banks managed costs efficiently and minimized costs from 1996 to 2011.
QED Working Paper Number