Productivity and computers in Canadian banking

Canadian banks have invested millions in computer systems in the last two decades. Yet the banks and outside observers have been uncertain that these investments have had net benefits. In this paper, unique data collected directly from a bank is used to investigate the impact of these investments on bank output, input and productivity. Using data from 1974–1987, a translog cost model is estimated. Both capital and labor are divided into information and noninformation inputs. The results are generally consistent with economic theory. The attempt to separate technical change from possible scale effects is very sensitive to alternative specification. Overall there has been some productivity growth associated with the changing computer technology. However, many of the benefits seem to have accrued to the customer and have not directly lead to gains for the bank.