Management estimation in banking

• Published in: European journal of operational research Vol. 284; no. 1; pp. 355 - 372
• Main Authors: Delis, Manthos D., Iosifidi, Maria, Tsionas, Mike
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.07.2020
• Subjects: Bayesian methods; Cost function; Management practices; Monte Carlo simulation; OR in banking; Cost function; OR in banking; Bayesian methods; Management practices; Monte Carlo simulation
• ISSN: 0377-2217; 1872-6860
• DOI: 10.1016/j.ejor.2019.12.023

•We propose a new model for measuring management practices of banks.•We model management practices as a latent input in a cost function.•We estimate our model using US bank-quarter data and Bayesian techniques.•We validate our approach using Monte Carlo methods and against existing techniques. We present and estimate a model of management practices as an unobserved (latent) input in a standard production function of banks. We derive estimates of management at the bank-quarter level, using data for all U.S. banks from 1984 to 2016 and Bayesian techniques. We show that management practices are an important missing input in banks’ operational processes. We validate our approach using repeated random sampling (Monte Carlo simulation) within a rather unfavorable environment. We also show that management is highly correlated with bank profitability (positively) and the probability of default (negatively). Our new measure compares favorably with existing measures of management derived from the residual of the regression of frontier efficiency on bank characteristics that are beyond management's capacity to change. Our model can be considered a benchmark to estimate good management practices, to analyze their sources, and to establish good management as an important determinant of efficient and sound banking.