A Nonparametric and Aggregation Theoretic Approach for Measuring Productivity of US Banks during 2006-2016

Abstract

Existing literature related to efficiency measurement and productivity analysis of banks is swarmed with the input-output classification of banks based on using accounting conventions. This usage varies from paper to paper. No two research papers are in consensus as to which classification should be used . This present work, however, uses the input-output classification of banks based on Barnett’s generalized model of production for financial intermediaries originally proposed in Barnett (1987) [1] . This model is based on economic theory definitions of inputs and outputs of a bank. Using this classification, the paper applies Data Envelopment Analysis to US banks during 2006-2016. This new methodology seeks to resolve and fix the issue of lack of consensus regarding which inputs and outputs to use for productivity analysis of banks. Furthermore, a standardized way of measuring productivity across banks is developed which can be used all over the world. This is accomplished by using the Malmquist Index of Productivity which is a tool used under Data Envelopment Analysis. The paper further establishes the connection of this tool with Barnett’s generalized model of production for financial intermediaries. Results indicate very high efficiency levels for US banks even post financial crisis. The reason for this performance is the cleansing of the financial system as unhealthy banks either left the scene or were merged. Better risk management, cost management and efficiency of structure of funding are some other reasons for high efficiency.