The measurement of bank output, a difficult and contentious issue, has become even more important in the aftermath of the devastating financial crisis of recent years. In this paper, we argue that models of banks as processors of information and transactions imply a quantity measure of bank service output based on transaction counts instead of balances of loans and deposits. Compiling new and comparable output measures for the United States and a range of European countries, we show that our counts-based output series exhibit significantly different growth patterns from those of our balances-based output series over the years 1997 to 2009. Since the U.S. official statistics rely on counts while European statistics rely on balances, this implies a potentially considerable bias in the estimate of bank output growth in Europe vis-à-vis that in the United States.
JEL Classifications: E01, E44, O47