O.R. Applications
Personnel allocation among bank branches using a two-stage multi-criterial approach

https://doi.org/10.1016/S0377-2217(02)00350-8Get rights and content

Abstract

This paper presents a normative two-stage approach for personnel allocation among branches of a large-scale commercial bank. It invokes and generalizes upon a decision rule named “Condorcet practical rule” first introduced by Marquis de Condorcet (1743–1794) and used by Geneva’s parliament during two years around 1792. It lay dormant for more than two centuries. The approach presented is based on a multi-criteria ordinal ranking model, taking into account a threshold of insensitivity within branch performance evaluations. An underlying homogeneity assumption is relaxed by introducing the concept of homogeneous branch subsets, using a new clustering algorithm. These subsets are based on workload patterns within the branches. Lastly the method is applied to real-life data from the largest private commercial bank (Yapı Kredi Bank) in Turkey.

Introduction

Commercial bank operations differ because of varying organizational formats, modes of competition, and size or level of development. Despite the developments in communication and computer technologies branch banking is still a necessary reality. Hence, a bank’s success in attaining its objectives is very much dependent on the operational performance of its branches. Bank branches remain labor and technical resource intensive operations (Belasco, 1997). To achieve the highest efficiency and effectiveness levels, it is necessary to understand the work process dynamics so that human resources (HR) managers can maintain and control optimal staff size throughout the network by quickly responding to changing conditions while maintaining a proper balance between human and technical resources.

The literature has recorded various judgmental and analytic methods for estimating both supply and demand of HR. These typically include both quantitative and qualitative variables among the work measurement statistics (Correa and Craft, 1999). The analytic methods generally require significant amounts of reliable data on past experience. In this paper, we propose a normative two-stage multi-criteria approach to personnel allocation requiring little past information. Ordinal rules are used in ranking of branches. These are based on data available in the bank and require very low inputs from managers. The class of rules used herein are based on a generalization of the so-called “Condorcet practical rule” introduced by Marquis de Condorcet (1743–1794). This rule was used for decision making in Geneva’s parliament during two years around 1792 and lay dormant for more than two centuries. The time is now ripe for operations researchers to note these rules.

HRs need for a mechanism to allocate available personnel among branches under increasing demand is discussed in Section 2. Section 3, discusses the proposed two-stage method of solution through an illustrative example. In this section, it is assumed that the branches are homogenous, i.e. the workload pattern is similar for these branches. In Section 4, the suggested method is modified to fit the case of heterogeneous branches. In Section 5, an empirical case study is explained by using actual data from Yapı Kredi Bank (YKB) a major Turkish private commercial bank, The paper ends with the concluding remarks of Section 6.

Section snippets

A normative framework for solving the HR department’s personnel allocation problem

One potential HR problem is as follows. HR has a pool of available employees coming from the Training department. Each branch provides HR its current personnel requirements. HR evaluates the eligibility of these demands and allocates available personnel in line with bank objectives. It is assumed that the supply of personnel is less than the combined branch demand.

Thus, it is necessary to rank branches according to their contribution to the bank objectives, i.e. there is a need to measure the

Homogeneous treatment of branches: The two-stage personnel allocation method

In this Section all branches of a bank are considered as homogeneous, i.e., they perform in the same environment, with the same personnel categories, say, tellers, etc. The homogeneity of branches enables the management to compare the performances of all branches.

Heterogeneous treatment of branches: The three-stage personnel allocation method

Clearly, in real life the branches are not homogeneous. Therefore, for banks with numerous branches this homogeneity assumption turns out to be a drastic simplification. The reason for this is threefold. First, the geographical distribution of the banking activity may not be homogeneous. Therefore, local markets may have different patterns of demand for banking products and services. Second, a bank can also be considered as a working team consisting of branches, headquarters, computer centers,

Application

The method above was applied to data obtained from Yapı Kredi Bank (YKB) from Turkey. As indicated, YKB is one of the largest private commercial bank in Turkey and employs more than 9000 personnel in 370 different size and scope branches. First, to enhance the applicability of the proposed method to real-life data, the bank branches are clustered according to their workload pattern. Then, two-stage allocation procedure is applied to homogeneous subsets of branches.

Concluding remarks

This paper presents a normative methodology for personnel allocation among bank branches to reduce any inefficiencies resulting from the gap between supply and branch demand for HR. The approach takes into consideration the bank managers’ objectives, and evaluation respective branches. Its informational requirements are moderate as it uses data collected by almost any bank.

It is worth pointing out that the method as presented does not tackle the problem of removing resources from branches with

Acknowledgements

We are strongly indebted to Prof. Arnold Reisman for many useful comments and help, and Mr. Cengiz Gŭndeş for discussions during the application of the method in YKB. We thank the comments of two anonymous referees as well.

References (10)

There are more references available in the full text version of this article.

Cited by (7)

View all citing articles on Scopus
View full text