Use of data envelopment analysis and regression for establishing manpower requirements in a bank

Abstract

We describe an approach towards forecasting the manpower requirements in each of the branches of a bank, based on regression models fitted to the sets of efficient branches. DEA is employed to identify the efficient branches within a category, using the numbers of employees in the different grades at each branch as input variables, and the average volumes of different types of work performed by them during a month as output variables. Forecasts of future volumes of work are obtained by fitting a model which takes into account branch and seasonal effects, as well as separate trend effects for each of the branches. The models have been tested on data from a large bank, with very encouraging results. The approach holds great promise for use towards a decision support system for managing the bank's total branch manpower requirements.