The Technical Efficiency of Nigerian Banks: The Data Envelopment Analysis Model

The Technical Efficiency of Nigerian Banks: The Data Envelopment Analysis ModelFour periods namely 1984/1985, 1994/1995, 1999/2000 and 2003/2004, 1984/1985 was chosen to examine the technical efficiency of banks. This is to enable us examine the technical efficiency of individual banks in Nigeria during periods of pre and post liberalization and democratization. This is intended to give us a broader understanding of the efficiency of banks bearing in mind different policies and styles of government in Nigeria over-time. It will also shed light on the evolution of the degree of technical inefficiency over-time. Furthermore, considering that it is a study in retrospect, we chose years before the Soludo and Sanusi consolidation to enable further assessment of the banks and the actions taken.
A cross section of both commercial and merchant banks was used for each period. In selecting the banks, all the banks that have complete data for the four periods under review were selected. This is done in order to see the changes in efficiency of banks in the four periods. In order to select other banks, we used randomization. According to Ndiyo randomization gives “a more reliable and valid estimate of the population being studied than a sample which is composed by selection regardless of whether such selected sample is random or not”.
In obtaining the data, we used the Annual Reports of the banks in the Nigeria Banking, Finance and Commerce Books compiled by Research and Data Services Limited (Redasel), Lagos Nigeria for the periods of the study.
Following Nyong, we adopt the Grigorian approach – a variant of the modern approach of bank production in choosing inputs and outputs. Thus, we assume that the labour (personnel management, X1), fixed assets (computer hardware and premises) and also captured extensive branch network, X2 and interest expenses (leverage funds) X3. The outputs are revenues (emphasizes profit maximization, Y1), loans and advances (service provision Y2) and liquid assets including securities investments (liquidity services, Y3). The model is adapted from Ali and Seiford and Nyong.
Two types of envelopment surfaces are used in DEA. These are Constant Returns to Scale (CRS) and Variable Returns to Scale (VRS) (Ali and Seiford, 1993). The VRS model will be utilized here. This is because it gives technical efficiency of DMUs under investigation without scale effect. In this model all the points (Xj, Yj) lie on or beneath the hyper plane and the hyper plane passes through at least one of the points (Obsersteiner, 1999; Ali and Seiford, 1993, 1994). further
Let U1 = virtual multiplier associated with output 1 for DMUj.
Let U2 = virtual multiplier associated with output 2 for DMUj.
Let U3 = virtual multiplier associated with output 3 for DMUj.
Let Us = virtual multiplier associated with output s for DMUj
w = virtual multiplier associated with VRS.
The objective function indicates the distance of the DMU from the hyper plane. Maximization of the objective function selects a hyper plane, which minimizes this distance.
The DEA analysis requires solving a linear programming problems; one for each DMU. While ‘xj and ‘yj are the observed values for the DMUs and are constants, u, v, w are the variables. The latter gives the feasible solution.
The values u, v, w have been interpreted as virtual multipliers. Thus, the linear programming problem, VRSm has been referred to as the multiplier side.
The technical efficiency of the banks is shown in table 2 below.

Table-1: Selected Studies on the Impact of Financial Liberalization on the Efficiency of the Banking Industry

Author Country Findings
Bhattacharyya, Bhattacharyya, and Kumbahakar Eastern & Central European Countries including China Increased efficiency
Berg, Forsund and Jansen Norway Efficiency declined and then rose
Zaim Turkish Banks Increased efficiency
Gilbert and Wilson Korea Increased efficiency
Humphrey and Pulley, Grifell-Tatje and Lovell US and Spain Decline in efficiency
Denizer et al Turkey Did not improve efficiency
Ziorklui Ghana Improved efficiency
Hardy and Patti Pakistan Did not improve efficiency
Barajas, Steiner and Salazar Columbia Enhanced efficiency
Ikhide and Alawode Nigeria Bank’s health deteriorated
Asogwa Nigeria Did not change the level of competition
Oyaromade Nigeria Had positive impact on financial savings mobilization
Adeoye and Adewuyi Nigeria No growth in savings, no improvement in the level of financial dependence
Koeva Indian Commercial Banks Increased competition, decline in bank spreads, reduction in cost of intermediary
Barajas, Steiner and Salazer Colombian Banks Increased competition, lower intermediation cost
Galindo A. et al 12 developing countries Improved efficiency
Demirguc-Kuit and Detragiache Panel of 53 developed and developing countries Banking crises more likely to occur in a liberalized system
Nyong Nigeria No improvement in efficiency

Table-2: Efficiency Scores of the Individual Banks, 1985-2003

S/N Banks 1985 1995 2000 2003
1 Afribank 1.000 1.000 0.854 0.435
2 Bank of the North 0.521 0.383 1.000
3 ACB 0.418 1.000 1.000
4 Citi Bank 1.000 1.000 1.000 1.000