Bank branches efficiency under management and regulatory constraints/Eficiencia en sucursales bancarias bajo restricciones gerenciales y regulatorias/Eficiencia nas agencias bancarias sob restricoes gerenciais e regulatorias.
| Autor | Ferro, Gustavo |
| Cargo | Research article |
-
Introduction
The main purpose of this paper is to assess the cost efficiency of Banco Ciudad de Buenos Aires's (BCBA) bank branches using a Stochastic Frontier Analysis (SFA). Likewise, we use efficiency measures to gauge new branches. We estimate cost efficiency and then compare the cost efficiency of the worst situated branches with the best performers according to the cost efficiency criteria. This is one of several possible uses of the tool. Many other possible managerial recommendations can be developed.
The BCBA is the official bank of Buenos Aires City. It was founded in 1878 and ranks as the eighth largest provider of bank loans in Argentina and the second local one in the mortgages market. It has 66 branches and some 3,200 employees. Nowadays, one of the bank's goals is to move beyond the boundaries of its historical area of influence (Buenos Aires City and suburbs). As a public bank, this goal faces some management constraints (for instance, it is not a pure profit maximizer, and employees enjoy stability as public servants), plus regulations of the Central Bank of Argentina restrict decisions to all banks in the country concerning free branches' creation or closing.
At management level, the BCBA currently uses two methods to analyze branch performance. The first is a system of quarterly and annual targets, determined by business volume and the complexity of each branch. Targets are set for "strategic products and services", such as loans, new current account and savings account openings, the number of desk and cashier operations, insurance policies sold, and so on. Targets are summarized in an output index by assigning weights to each individual goal and by dividing the weighted output average by its maximum. The second method is a partial productivity ratio set, computed over a set of performance indicators such as loans per employee, loans per branch, or customers per employee.
More comprehensive and exhaustive measures to assess efficiency and to compare relative performance are frontier studies. Two strands of the latter are mathematical programming models (mainly using Data Envelopment Analysis - DEA) and econometric models (mainly using Stochastic Frontier Analysis - SFA). This study employs the SFA technique. Each method has comparative advantages and disadvantages against the other; SFA is a preferred method when estimating cost efficiency (our goal), and DEA is customarily used when the purpose is to estimate technical efficiency (Coelli, Rao, O'Donnell & Battese, 2005).
We use SFA for estimating cost efficiency, not technical efficiency. In the latter case the multiple output criteria cannot be applied (for instance, when estimating a production frontier y = f([x.sub.1] [x.sub.2], ..., xn) where [x.sub.i] from 1 to n are inputs and y is the (only) output (1). Nevertheless, we are estimating a cost frontier C=([y.sub.1] [y.sub.2],..., [y.sub.n], [w.sub.1] [w.sub.2], ..., [w.sub.n-1]) where C is used for denoting costs, yi from 1 to n are the different outputs and wi from 1 to n-1 are the relative prices of the different inputs relative to the n-th input (the numeraire). Thus, the cost function allows to estimate cost efficiency in a multiple-output environment.
With respect to strengths and weaknesses of SFA against DEA, there are relative advantages of the second method to address multi-output technical efficiency computing, the possibility of working with small samples, the sensitivity of the method to outliers -which allows the investigator to discover errors in the database-, and the absence of constraining to a particular and probably arbitrary functional form. On the other hand, SFA allows statistical significance tests, permits multi-output estimates -in the case of cost functions-, SFA yields a function which describes the behavior of the bank, under some economic objective (such as cost minimization or profit maximization), and SFA allows to separate inefficiency from statistical noise (at the cost of making some hypothesis on the statistical distribution of the inefficiency component).
In frontier models, a numeric value is assigned to the efficiency of each branch which makes it possible to identify both the over-utilization of inputs--or the under-production of outputs--and best practices. SFA models can supplement more commonly used managerial indicators (mainly accountancy ratios) to determine inefficiency and help to provide remedies.
Identifying a branch's relative performance is a starting point in a comprehensive evaluation of some branch network efficiency. The efficiency scores allow building a performance ranking by identifying the worst and best performers, detecting for remedial action, for incentive design, or for the reallocation of resources. In the same vein, high-performing branches may serve as instructive role models for low-performing ones (Pastor, Knox & Tulkens, 2003). Or, at least, given binding constraints on management, or regulations that prevent the exploitation of all the results potential (2), new branches could be designed with the attributes of the best performers, which is the criterion followed in this exploratory study.
The structure of the paper is the following: after this introduction, section 2 presents the literature review. Section 3 summarizes the method. Section 4 describes the database. Section 5 presents the results and the managerial implications, and section 6 discusses them. Lastly, section 7 concludes.
-
Literature review
According to Hughes and Mester (2008), there are two broad approaches to measure banking efficiency and thus to explain nonstructural and structural performance. The former uses a variety of financial ratios or the market value of the firm, capturing aspects of performance to compare banks. The structural approach relies on the theoretical models of the bank's technology and on the assumption of some objective-function optimization goals. The studies assume that the banks choose a production plan that minimizes costs given their output mix and input prices or that they maximize profits given input prices and outputs.
It is difficult to estimate banking efficiency because of the variety of services commercial banks offer. Following Stavarek (2005), three main approaches in the literature define the input-output relationship in financial institutions.
First, the production approach considers banks as producers of accounts and loans, defining output as the number of such accounts and loans. This method defines inputs as the number of employees and capital expenses in fixed assets. The approach centers on operative costs and ignores interests paid. Second, the intermediation approach originates in the traditional role of banks in transferring financial resources from savings suppliers to savings demanders. Operative and interest costs are the most important inputs contemplated here, while the main outputs are interest income for credits and investments and charges for services. Third, the assets approach highlights the role of financial entities as credit originators. This view is a variant of the intermediation approach, differing in the outputs than the latter approach.
None of the three approaches can fully capture the dual role of financial entities as providers of transaction services and conveyors of savings from suppliers to demanders. Nevertheless, there is a reasonable consensus in the literature about the banks' inputs and outputs. Loans and other relevant assets should be considered outputs according to intermediation and asset approaches. Some controversy exists over the role of deposits. There is an empirical test to determine whether deposits act as an input or output. Let us call variable costs (VC) and deposit levels (x). If deposits are an input, then [partial derivative]VC/[partial derivative]x 0. Thus, output can be increased only if input expenditures are increased (Hughes & Mester, 2008).
Inputs and outputs are flows. When data on input flows are not available, stocks are used as proxies. A third input, beyond labor and capital, are the flows of financial services, which are hard to measure with the data that are normally available. Thus, they are approximated by stocks, such as loans and deposits (Lopez, Appennini & Rossi, 2002).
Berger and Humphrey (1997) encourage the efficiency research at branch level after presenting a survey of empirical literature of branch and bank system efficiency. They observe that the literature at bank branch level was, at the time of their survey, still limited by comparison to bank system efficiency measurement.
Firstly, we describe the literature on branch efficiency after the seminal survey of Berger and Humphrey (1997), and secondly, we summarize the findings and their importance for this study. The literature related to banking system efficiency, by opposition to bank branches efficiency, continues to be more extensive. For a very recent and exhaustive discussion of the former literature, see Asimakopoulos, Chortareas and Xanthopoulos (2018).
Athanassopoulos, Sotiriou and Zenios (1997) analyze the efficiency of bank branches networks in different countries (the United Kingdom, Greece and Cyprus), suggesting guidelines for branch efficiency improvement.
In turn, Berger, Leusner and Mingo (1997), measure the efficiency of a branch network over a large American bank. They find evidence of severe "over-branching". The X-inefficiencies they find exceeded 20% of operating expenditures.
Camanho and Dyson (1999) assess the performance of Portuguese bank branches. The analysis focused on the relation between branch size and performance.
Zenios, Zenios, Agathocleous and Soteriou (1999) develop a study commissioned by the Bank of Cyprus, with the objective of set benchmarks of performance and to address the effects of a cyclical component of the demand due to tourism. The analysis revealed branch resource underutilization during the low season.
...
Para continuar leyendo
Solicita tu pruebaCOPYRIGHT GALE, Cengage Learning. All rights reserved.
