41 Pages Posted: 3 Sep 2014
Date Written: May 10, 2013
We propose generalized frontier analysis (GFA), a simple and computationally efficient method for estimating economic frontiers. While previous research assumes that the production function under which firms operate is either parametric or deterministic in nature, we apply the theory of asymmetric loss functions in combination with a nonparametric and stochastic estimation method to relax these potentially restrictive assumptions. We use GFA to estimate shareholder value efficiency of US banks on a large sample of 118,164 bank-year observations for the period 1994-2010. Using a broad set of criteria, we find that GFA provides valid efficiency scores, which are economically and statistically more meaningful in explaining value creation of US banks than both managerial ability and conventional efficiency scores. We also demonstrate the generality of GFA by investigating other economic frontiers such as cost efficiency. Overall, our analysis indicates that GFA is an attractive method for modeling economic measures of efficiency.
Keywords: Financial Institutions; Efficiency; Artificial Neural Networks; Managerial Ability
JEL Classification: C45, G21, G28, E58
Suggested Citation: Suggested Citation
Andreou, Panayiotis C. and Philip, Dennis and Robejsek, Peter Paul, Shareholder Value Efficiency: Methods and Evidence from the US Banking Industry (May 10, 2013). Available at SSRN: https://ssrn.com/abstract=2489014 or http://dx.doi.org/10.2139/ssrn.2489014