Are Saudi Bank Efficient? Evidence uing Data Envelopment Analyi (DEA) Khalid AlKhathlan (Correponding Author) Economic Department, College of Buine Adminitration, King Saud Univerity, P.O.Box 2459, Riyadh 11451, Saudi Arabia Phone No. 00966-555-227217, Email: kaa8161@hotmail.com Syed Abdul Malik Finance Department, College of Buine Adminitration in Al kharj, King Saud Univerity, P.O.Box 2459, Riyadh 11451, Saudi Arabia
ABSTRACT Saudi Arabia ha a bank-centric and divere financial ytem compared with other countrie in the region. Thi paper ue baic DEA model i.e. CCR and BCR to evaluate the relative efficiency of Saudi Bank uing annual data from 2003 through 2008. The reult how that, on a relative cale, Saudi bank were efficient in the management of their financial reource. In addition, the reult would provide crucial information about Saudi bank financial condition and management performance for the benefit of bank regulator, manager and bank tock invetor. Keyword: Saudi Banking Efficiency, Saudi Financial Sytem, data Envelopment Analyi, Technical Efficiency, Scale Efficiency
1. INTRODUCTION In it Forty-Fourth Annual Report the Saudi Arabian Monetary Agency-SAMA (2008, pg. 51) ummarized the performance of Saudi bank a follow: Commercial bank recorded good growth rate in their financial poition during 2007. Thi wa due to efficient management by commercial bank of their financial reource. Thi paper give the quantitative proof of the above ummarized view of SAMA and invetigate the relative efficiency of Saudi bank during the period of 2003-2008 by uing the Data Envelopment Analyi (DEA) technique. Over the pat two decade, DEA ha become a popular methodology for evaluating the relative efficiencie of deciion making unit (DMU). A DMU i an entity that produce output and ue up input, in thi tudy, each bank contitute a DMU. DEA i a linear programming model that meaure the efficiency of DMU in multiple-input, multiple-output etting. Typically, each of the DMU in a given population ue the ame multiple input in varying quantitie to produce varying quantitie of the ame multiple output. Uing the actual oberved value for the input and output for each DMU, DEA contruct a piecewie linear production urface, which in economic term repreent the revealed bet-practice production frontier the maximum output empirically obtainable for any DMU in the oberved population, given it level of input. By projecting each unit onto the frontier, it i poible to determine the level of inefficiency by comparion to a ingle reference unit or a convex combination of other reference unit. The projection refer to a hypothetical DMU which i a convex combination of one or more efficient DMU and not an actual DMU. Efficient DMU typically utilize the ame level of input and produce the ame level or higher of output. Although DEA model jointly handle the multiple input and multiple output characteritic of financial ervice, they have everal limitation. Firt, with the baic DEA model there are uually a large number of zero weight in input and output variable. Second, the incluion of a large number of input and output reduce the degree of freedom of thi program and, a a reult, the number of efficient unit increae. To the reearcher bet knowledge, thi i the firt time thi technique i being ued to analyze both the technical and cale efficiencie of Saudi bank uing two baic DEA model. The reult would provide u explicit indication a to whether the Saudi bank are efficient a claimed by SAMA even in the wake of global melt down. In addition, the reult would provide crucial information about Saudi bank financial condition and management performance for the benefit of bank regulator, manager and bank tock invetor.
The paper i divided into five part. Following thi, Section two review briefly the previou tudie on bank frontier efficiency in Arab World. Section three proceed with the methodology and data ued to carry out the efficiency analyi. Section four examine the empirical finding and ection five conclude the paper. 2. LITERATURE REVIEW Empirical evidence on performance evaluation and efficiency of the banking indutry i much reearched globally. However, there i dearth of reearch in the Arab world. The earliet technique, ued to meaure performance change wa ratio analyi which examine the financial tatement of individual firm and comparing them with a benchmark. However, thi technique failed to take into account the fact that bank produce multiple output from multiple input and conitent aggregation wa not poible (See, for example, Barne, 1987; Smith, 1990). The hort coming of uch a decriptive and tatic analyi of the data are overcome by later reearcher with the ue of parametric and non-parametric technique. The parametric and non-parametric technique differ mainly in how they handle random error and their aumption regarding the hape of the efficient frontier. The three main parametric (or econometric) methodologie ued by reearche to examine financial intitution include the tochatic frontier analyi, the thick frontier analyi, and the ditribution free analyi. In general, the parametric approache pecify a functional form for the cot, profit, or production relationhip among input, output and environmental factor, and allow for random error. The two non-parametric (or mathematical programming) technique ued in the banking ector efficiency literature include Data Envelopment Analyi (DEA) and Malmquit Productivity Indice (MPI). In general, a nonparametric technique doe not require the pecification of an a priori functional form and therefore i the mot favored approach. The literature examining the efficiency of financial intitution with parametric and/or nonparametric frontier technique ha expanded rapidly in recent time. While, a large body of literature panning a half-century exit on banking efficiency in the United State (ee urvey in Berger et al., 1993; Berger and Humphrey, 1997; and Berger, 2007), more recent tudie examine everal other countrie uch a India (Ataullah and Le, 2006), Hong Kong (Drake et al., 2006), Singapore (Sufian, 2007), Croatia (Jemric and Vujcic, 2007), Turkey (Iik, 2008), Ukraine (Kyj and Iik, 2008), and Thailand (Supachet Chanarn, 2008). What follow i a brief expoition on empirical reearch on bank frontier efficiency in Arab world. Al-Faraj et al. (1993) evaluated the relative efficiency of 15 bank branche of one of the larget commercial bank in the eatern province of Saudi Arabia by mean of DEA. Uing one year data he found out that 12 branche were efficient baed on eight input and even output identified.
Al-Shammari and Salimi (1998) evaluated the comparative operating efficiency of bank in Jordan uing a modified verion of DEA. The reult ugget that the majority of bank invetigated are fairly inefficient over the period 1991 1994. Haan et al. (2004) employing a panel of 31 bank for the year 1998 and 2000 invetigated relative efficiency of the banking indutry in Bahrain. The reult indicate that all bank have improved their efficiency level and experienced ome gain in productivity. Al-Faraj et al. (2006) invetigated the performance of the Saudi commercial banking indutry uing DEA to evaluate the technical efficiency of Saudi bank for the year 2002 and compared with world mean efficiency core. Their tudy revealed that the mean efficiency core of Saudi commercial bank compare very well with the world mean efficiency core. They recommend that Saudi bank hould continue their effort of adapting new technologie and providing more ervice in order to utain competitive advantage a Saudi Arabia continue to deregulate the banking indutry. Saeed Al-Muharrami (2007) examined productivity change in the countrie of the Gulf Cooperation Council (GCC) bank uing Malmquit DEA. The reult indicated a negative change in efficiency during the period from 1993 to 2002 for the 52 GCC bank conidered. Further, they found that the decreaed efficiency wa the due to catching up effect. Motafa, M. M. (2007) invetigated the efficiency of top 85 Arab bank uing DEA and Neural network for the year 2005. He found that, eight bank a per the CCR Score and four bank a per BCC Score were poitioned on the efficient frontier. He uggeted that future tudie hould tet the exitence of poitive rank-order correlation between efficiency core obtained from DEA analyi and traditional efficiency meaure uch a financial ratio. Hi reult further demontrate that, Al-Rajhi Bank and National Commercial Bank were placed among the top ten Arab bank with a relative ranking of eight and ten repectively. Emrouznejad, A. and Anouze, A.L. (2009) found only ix bank poitioned on the efficient frontier a per the CCR Score uing the ame et of bank and the ame et of input output variable a in Motafa (2007). Their finding indicate that the efficiency of Arab bank reported in Motafa (2007) i incorrect. In ummary, their tudy overcome with ome data and methodology iue in meauring efficiency of Arab bank and highlight the importance of encouraging increaed efficiency throughout the banking indutry in the Arab world uing the new reult. Interetingly, their reult alo demontrate that, Al-Rajhi Bank and National Commercial Bank were placed among the top ten Arab bank but with a new relative ranking of ix and even repectively. The above review revealed that there i hardly any comprehenive tudy on meauring the efficiency of Saudi commercial bank. Thu, thi tudy add to the exiting literature and evade previou empirical tet limitation in
the following way. Firt, the fact that depite the ubtantial tructural change and importance of the Saudi Arabian banking ector, the ector ha remained under reearched compared to tudie in other countrie. The preent tudy thu addree an important gap in the literature. Second, compared to earlier paper, thi tudy ha the following merit. Firtly, unlike Saeed Al-Muharrami (2007), Motafa, M. M. (2007), and Emrouznejad, A. & Anouze, A.L. (2009) invetigated GCC or top Arab bank efficiency where a we invetigate the efficiency of Saudi commercial bank. Although, Al-Faraj et al. (2006) invetigated technical efficiency of Saudi bank uing 2002 data, our tudy invetigate both technical and cale efficiency of Saudi commercial bank during the period 2003-08. Latly, the tudy ha important public policy implication to achieve a more competitive and efficient financial ytem. The tudy could help the regulatory authoritie in determining the future coure of action to be purued to further trengthen the Saudi Arabian banking ector in particular the dometic incorporated bank. 3. DATA & METHODOLOGY 3.1 Data We ue conolidated annual data compiled mainly from balance heet and income tatement of bank, their webite, and related web page of Saudi tock exchange (Tadawul) and Saudi Arabian Monetary Agency (SAMA) on the internet. We cover only ten out of twelve commercial bank operating in Saudi Arabia which are a follow: Arab National Bank Al-Rajhi Bank Bank Al-Jazira Banque-Saudi Frani National Commercial Bank Riyadh Bank Saudi Britih Bank Samba Financial Group Saudi-Hollandi Bank Saudi Invetment Bank ANB ARB BAJ BSF NCB RYB SABB SAMBA SHB SIB 3.2 Methodology We apply two baic Data Envelopment Analyi (DEA) model i.e. Charne Cooper Rhode (CCR) model and Banker Charne Cooper (BCC) model to evaluate the relative efficiency of Saudi Bank uing annual data from 2003 through 2008.
3.2.1 CCR Model Charne, Cooper and Rhode (1978) have coined the term data envelopment analyi (DEA). They extended Farrell (1957) piecewie-linear convex hull approach to frontier etimation by expanding multiple input and ingle output to multiple input and multiple output and utilized linear combination to convert it to ingle virtual input and output. Their model aumed contant return to cale (CRS) to meaure the relative efficiency of each DMU which i between 0 and 1 and can determine whether a DMU i in contant, increaing or decreaing return to cale. Following, Emrouznejad, A. and Anouze, A.L. (2009), the linear programming formulation i a follow: Minimize θ Subject to: λ ϑ 0 input x I x I kx λ 0 output y O x λ 0 O kx The value of θ obtained will be the efficiency core for the i-th DMU. 3.2.2 BCC Model Banker, Charne and Cooper (1984) widened the CCR model to account for variable return to cale (VRS). The CRS linear programming problem can be eaily modified to account for VRS by adding the following contraint to above model: λ = 1 Thi approach form a convex hull of interecting plane which envelope the data point more tightly than the CRS conical hull and thu provide technical efficiency core which are greater than or equal to thoe obtained uing the CRS model. Baed on the literature review we found that there i no uniform opinion concerning what contitute input and output for bank in the context of a DEA tudy. We employed the intermediation approach which view bank to intermediate aving to productive invetment through the upply of credit to buinee and conumer. Baed on thi approach, the input variable ued in thi tudy are operating expene, equity capital, and depoit where a loan and advance (net) i the only output variable conidered. We ran the DEA model eparately for each year uing input-orientation. By running thee program with the ame data under contant return-to-cale (CRS) and variable return-to-cale (VRS) aumption, meaure of overall
technical efficiency (E) and pure technical efficiency (PT) are obtained. In order to obtain a meaure of cale efficiency (S), we divide overall technical efficiency (E) by pure technical efficiency. The following ection dicue the reult obtained 4 RESULTS DEA efficiency core baed on contant return to cale (CCR Model) are hown in Table 1. Average technical efficiency in the Saudi bank during the tudy period range from 0.81913 (2003) to 0.86784 (2008). ARB and BSF were the only two bank with efficiency core of 1.0000 each year, implying that they are on the efficiency frontier and were peer (or bench marked) during the tudy period. In 2007, five bank emerged on the efficient frontier, indicating efficient management by 50% of Saudi bank of their financial reource. Although, BAJ, NCB, and SIB are having efficiency core le than 0.75 indicating that 30 per cent of Saudi bank are inefficient and require further probing. Table 1 DMU No. DMU Input-Oriented CRS Efficiency Name 2003 2004 2005 2006 2007 2008 1 ANB 0.68215 0.93401 0.94741 1.00000 1.00000 0.96705 2 ARB 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 3 BAJ 0.71441 0.62352 0.58609 0.48448 0.55842 0.60983 4 BSF 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 5 NCB 0.77912 0.75986 0.76404 0.64942 0.61297 0.64264 6 RYB 0.63118 0.73445 0.79837 0.81111 0.92106 0.89811 7 SABB 0.92876 0.88302 0.87337 0.77927 1.00000 1.00000 8 SAMBA 0.63316 0.86666 0.80830 0.79927 0.81303 0.82318 9 SHB 0.85749 0.90388 1.00000 1.00000 1.00000 1.00000 10 SIB 0.96499 0.79477 0.89504 0.78542 0.71127 0.73761 Mean Efficiency 0.81913 0.85002 0.86726 0.83090 0.86167 0.86784 Table 2 how DEA efficiency core baed on variable return to cale (BCC Model) for each year. Mean efficiency in the Saudi bank during the tudy period range from 0.87879 (2003) to 0.95336 (2004). ARB, BAJ, and BSF were the only three bank with efficiency core of 1.0000 each year, implying that they are on the efficiency frontier and were peer (or bench marked) during the tudy period. In 2006 and 2007, ix bank emerged on the efficient frontier, indicating efficient management by majority (60%) of Saudi bank of their
financial reource. Although, NCB with efficiency core of 0.62319 wa till inefficient during the year 2007. figure 1 how the ame reult. Table 2 DMU No. DMU Input-Oriented VRS Efficiency Name 2003 2004 2005 2006 2007 2008 1 ANB 0.72516 0.96918 0.98959 1.00000 1.00000 0.97664 2 ARB 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 3 BAJ 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 4 BSF 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 5 NCB 0.91388 0.93688 0.89230 0.68357 0.62319 0.66517 6 RYB 0.63593 0.77378 0.84512 0.81273 0.92109 0.90154 7 SABB 0.94250 0.91375 0.88533 0.77936 1.00000 1.00000 8 SAMBA 0.63466 0.93996 0.91573 0.86765 0.91070 0.88773 9 SHB 0.93576 1.00000 1.00000 1.00000 1.00000 1.00000 10 SIB 1.00000 1.00000 1.00000 1.00000 0.94207 0.84560 Mean Efficiency 0.87879 0.95336 0.95281 0.91433 0.93970 0.92767 Table 3 how the mean efficiency each year by decompoing technical efficiency into pure technical efficiency and cale efficiency. Decompoing technical efficiency into pure technical efficiency and cale efficiency allow u to gain inight into the main ource of inefficiencie. The average index of technical efficiency during the tudy period varie in between 81.91% to 86.78%, of pure technical efficiency varying at 87.88% to 95.34%, and of cale efficiency varying at 89.16% to 93.55%. Table 3 Year Mean Technical Efficiency (CRS) Input-Oriented Mean Technical Efficiency (VRS) Mean Scale Efficiency 2003 0.8191 0.8788 0.9321 2004 0.8500 0.9534 0.8916 2005 0.8673 0.9528 0.9102 2006 0.8309 0.9143 0.9087 2007 0.8617 0.9397 0.9170 2008 0.8678 0.9277 0.9355
5 CONCLUSION Thi paper ue the two baic Data Envelopment Analyi (DEA) model i.e. CCR and BCR to invetigate and provide the quantitative proof to the claim of Saudi Arabian Monetary Agency about the efficient management of financial reource by Saudi bank. The empirical reult do confirm that majority of Saudi bank efficiently managed their financial reource and the mean efficiency during the year 2007 wa 86.17% and 93.97% a per CCR and BCR approach repectively. In 2007, we found that five bank a per the CCR Score and ix bank a per BCC Score were poitioned on the efficient frontier. The empirical reult indicate that ARB and BSF hould be benchmarked or peer to other Saudi bank a they were the only bank found to be on the efficient frontier uing both CCR and BCR model. NCB being the only bank found to be le efficient compared to the other bank in term of CCR and BCR model. REFERENCES Al-Faraj, T., Alidi, A., & Bu-Bhait, K. (1993). Evaluation of bank branche by mean of data envelopment analyi. International Journal of Operation & Production Management, 13, 45 52. Al-Faraj, T., Bu-Bhait, K. & Al-Muhammad, W. (2006). Evaluating the efficiency of Saudi commercial bank uing data envelopment analyi. International Journal of Financial Service Management, 1(4), 466-477. Al-Shammari, M., & Salimi, A. (1998). Modeling the operating efficiency of bank: A nonparametric methodology. Logitic Information Management, 11, 5 12. Al-Sharka A, Haan MK, Lawrence S. (2008). The impact of merger and acquiition on the efficiency of the U.S. banking indutry: further evidence. Journal of Buine, Finance and Accounting, 35: 50 70. Ataullah A, Le H. (2006). Economic reform and bank efficiency in developing countrie: the cae of the Indian banking indutry. Applied Financial Economic, 16: 653 663. Banker, R., Charne, A., & Cooper, W. (1984). Some model for etimating technical and cale inefficiencie in data envelopment analyi. Management Science, 30, 1078 1092. Banxia Frontier Analyt Uer Guide (2001). Frontier analyt profeional (Verion 3.0). Glagow, UK: Banxia Holding Ltd.
Berger AN, Humphrey DB. (1997). Efficiency of financial intitution: international urvey and direction for future reearch. European Journal of Operational Reearch 98: 175 212. Berger AN. (2007). International comparion of banking efficiency. Financial Market, Intitution and Intrument 16: 119 144. Charne, A., Cooper, W., & Rhode, E. (1978). Meauring the efficiency of deciion-making unit. European Journal of Operational Reearch, 2, 429 444. Drake L, Hall MJB, Simper R. (2006). The impact of macroeconomic and regulatory factor on bank efficiency: a non-parametric analyi of Hong Kong banking ytem. Journal of Banking and Finance, 30(5): 1443 1466. Emrouznejad, Ali and Anouze, A.L. (2009). Expert Sytem with Application, 36, 5741 5744. Iik, I., & Haan, M. (2002). Technical, cale and allocative efficiencie of Turkih banking indutry. Journal of Banking and Finance, 26, 719 766. Iik I. (2008). Productivity, technology and efficiency of de novo bank: a counter evidence from Turkey. Journal of Multinational Financial Management 18(5): 427 442. Jemric I, & Vujcic B. (2007). Efficiency of bank in Croatia: a DEA approach. Comparative Economic Studie 44: 169 193. Motafa, M. M. (2007). Modeling the efficiency of top Arab bank: A DEA neural network approach. Expert Sytem with Application 36: 309-320. Sufian F. (2007). Trend in the efficiency of Singapore commercial banking group: a non-tochatic frontier DEA window analyi approach. International Journal of Productivity and Performance Management 56: 99 136. 44 th Annual Report, Saudi Arabian Monetary Agency, 2008.