Jurnul Ekonotni Malaysicr 48(2) 2014 Ill - 125

Technological Innovation in Services and the Efficiency of Malaysian. Commercial Banks

(Inovasi kknologi dalam Sektor Perkhidmatan dan Kecekapan Bank Komersil di Malaysia)

Schoor or Man"i*:;Tlr:i#.:l Sains Maraysia

Sok-Gee. ChanFaculty of Business and Accountancy, University of Malaya

ABSTMCT

This paper revisits relative fficiency of commercial banks in Malaysia using Data Envelopment Analysis (or.a) overthe 2004-2012 periods. We take into account technological innovation in bankservices, a new dimension thatwerenot investigated by any previous literature. Our mainfindings are: (l) Commercial banks in Malaysian have higherpure technical fficient as compare to scale fficient. (2) Foreign banks are relatively inefficient but the gap has beengradually closing after the end of the capital control by the Malaysian government in 2005. (3) Excess numbers ofBranches, Cash Deposit Machine and Cheque Deposit Machine appeared to be the major weakness of the infficiencybanl<s. Our result implies technological innovation in bank services is one important dimension in addressing the relativefficiency of commercial bqnl$. Banlcs that provide better technological services do acquire competitive advantageagainst their peers.

Kewords: Bank; technical fficiency; scale fficiency; technological innovation; DEl

ABSTMK

Kajian ini melihat kembali kecekapan relative bank komersil di Malaysia menggunakan Analisis Penyampulan Data(ot'e) bagi tahun 2004-2012. Dengan mengambilkira inovasi teknologi dalam perkhidmatan bank satu dimensi baru yangbelum pernah dikaji disajikan. Dapatan utama kajian adalah: (l) Bank Komersil di Malaysia mempunyai kecekapantehtikal asli yang lebih tinggi berbanding kecekapan skala. (2) Bank asing secara relatifuya kurang cekap tetapiiurangnya semakin mengecil selepas dasar kawalan modal yang dilaksanakan kerajaan Malaysia pada tahun 2005 (3)Terlebih bilangan cawangan Mesin Deposit Tunai dan Mesin Deposit Cek muncul sebagaifactor utama ketidakcekapanbank. Keputusan kaiian menunjukkan bahawa inovasi teknologi dalam perkhidmatan perbankan adalah satu dimensipenting bagi kecekapan relative bank komersil. Kesimpulannya banleyang memberikan perkhidmatan berteknologilebih baik akan mencapai kelebihan saingan berbanding yang lain.

Kata kunci: Bank; kecekapqn teknikal; kecekapan skala; inovasi teknologi; DEA

INTRODUCTION

The traditional bank delivery system has been thebranch network. Modern banks. however. do notrely on the number of its branch network anymoreas customers have demanding for more advance andconvenient delivery systems. Success or failure ofcommercial banking today is depending very much onthe capabilities of bank to anticipate and react to fastchanges in the marketplace. For example, the use ofelectronic banking recently has change the nature ofbanking activities and provides a lot ofother advantagesover traditional banking delivery channels, and thistrend is well coped by cornrnercial banks in Malaysia.Over the last 2 decades, the quality of bank servicesin Ma lays ia has a l so been enhanced by L r t i l i za t i on

of technological innovation in the delivery system.Technological innovation is expected to add value tothe bank. It happen either through increasing revenueat marginal cost or through reducing costs at marginalchanges in revenue. Howeveq existing l iterature onbanking efficiency focus on tangible value creation suchas the amount of profit gained or costs reduction. Otheraspect such as operation process and system efficiencyin the delivery system received little attention. Thisis the knowledge gap that the present paper tryingto gauged. The dimensions of new delivery channelsinclude increasing customer base, cost savings, massprodr.rct customization and innovation, rnarketing andcommunications, developrnent of non-cole busrnessesand the offering of services regardless of geographicalea and t inre (Giannakoudi 1999).

112

Since the 1980s, Malaysian commercial banks haveconsistently harnessed state-of-the-art technology in itseffort to upgrade new delivery channels. TheAutomatedTeller Machine (eru) in the 1980s to electronic bankingand electronic cards in the 1990s, were all innovationsof the information and computer technology (tcr).Computerization and self-service terminals are essentialin answering the pressure of cost reduction and improvedservice qualities. Migration to electronic payments andreduction in the quantity of over-the-counter serviceshave becomes part of the bank's agenda to increasethe efficiency of the payment systems. Installing moreself-service terminals not only can achieve cost savingobjective but also deliver better services to customersin terms of accessibility and convenience. Today, ATMdoes not only provide serves to withdraw cash, but italso offers multi-purposes functions which include billpayments and credit upload. In this regards, Malaysiancommercial banks have expand quite tremendously.In this paper, we explore whether the technologicalinnovation in banking delivery system affects therelative efficiency of commercial banks in Malaysia.To be more specific, we considered the role of allkinds of banking self-service terminals, which includeATM that offers more type of services as compared to10 years ago, in determining the relative efficiency ofMalaysian banks. We shall call these terminals as bank'stechnological services henceforth.

ln the context of technical efficiency, we yet to findany research conducted on the role of technologicalservices offered by banking. Existing literature on banktechnical efficiency whether on developed market, or onemerging markets including Malaysia, were all build onconventional inputs and outputs variables such as staffnumbers, branches and financial indicators or ratio. Ourstudy offers a different perspective in efficiency study interms of modeling and data. We follow the productionapproach in defining our variables, but we introducednew variables to gauge for technological innovation inbank services, i.e. the new service facilities offered bymodern banks, covering ATM machines, cash depositmachine, cheque deposit machine, cheque scan machine,and passbook update machine. The fast reformation inthe banking sectors over the last 2 decades implies thatthese physical facil i t ies deserved a special focus rnaddressing the issue ofbanking efficiency.

The rest of this paper is organized as follows:Section 2 portrayed the overall picture of the trendand acceptance level of technological innovat ionin Malaysian banking delivery system. Sections 3discussed the current accepted measures of bankefficiency, and some literature review on Malaysiabanking ef f ic iency. Sect ion 4 in t roduced the DEArnethodology in brief. Section 5 reports the results frornDEA and it is divided into 4 sub sections to facil i tate thedifferent stages ofour nea analysis. The final section isour concluding lemark.

Jurnal Ekononi Malaysia 48(2)

ACCEPTANCE OF TECHNOLOGICAL. INNOVATION IN MALAYSIAN BANKING

DELIVERY SYSTEM

Technological innovation in Malaysian banking industrybegan in the 1980s, with the introduction of credit cards,and Automatic Teller Machine (.qtu) networks. Thiswas followed by telephone banking, electronic bankingand electronic cards in the late 1990s. Technologicalinnovation allows customer to perform many bankingtransactions that would traditionally be served over thecounter. Due to its convenience, the acceptance ratefor these new technologies is growing rapidly and theterritory of some old system ofpaymenthas been invaded.ln the US, the rise of electronic mode of payments hascaused the value ofchecks paid to drop significantly fromabout $49 billion in 1995 to about $42 billion in2002(Gerdes and Walton 2002).

In Malaysia, such trend is not observed significantly.As reported in Table 1, over the last half decade thevalue of per capita cash and cheque transactions arestill in a rise, partly reflecting the stable improvementof income, and partly reflecting the attach to cash andcheque payment are still solid. However, the values ofthe uses of various modes of e-payments are mostlyin a significant rise. This is especially obvious whencome to ATM internet banking and mobile banking.For example the per capita transaction value throughRTM is only nul46.5 in 2005, but the value hasjumped to nrral,387.3 in2012, nearly 10 times higherin about 7 years time. For internet banking, the percapita value has increased from RM668 in 2005 toRM64,259.5 in 2012, a 100 times jumped in value.For mobile banking which was first introduced in2005, the value has also jumped significantly from20 cent per capi ta to RMl24. l per capi ta in2012.Thetotal value of E-payments in 2005, i.e. RM 3,517 .9 percapita, has increase significantly to nu5 13,846.8 percapita. This could be due to the direct link betweenReal-time Electronic Transfer of Funds and SecuritiesSystem (neNrns) for the settlement of ringgit andthe uso cHATS system (Clearing Housed AutomatedTransfer System) in Hong Kong (Hong Kong MonetaryAuthority) for the settlement of us dollar, which enablethe simultaneous settlement of ringgit in Malaysia andus dollar in Hong Kong during Malaysian businesshours. The link is aims to eliminate foreign exchangesettlement risk for ringgit and us dollar transactions.RENTAS is a large value payment system that enablesthe transfer and settlement of high value interbankfunds and scripless securit ies transactions.

Table 2 takes a closer look at the payment statisticsfor the ATM, internet banking and mobile banking. Thefigures reveal that the volume of transactions involving,qTM has jumped more than l2 tirnes from 3.6 milliontransactions in 2005 to 46.4 million transactions in2012.rvlri le in ten.ns of values it also.junrped rnore than l0 times

Technological Innottation in Sentices and the Efficiencv of Malaysian Contntercictl Bcmks

TABLE l. General Economic Information and Basic Pavment Statistics

I r3

2008200720062005 2009 2010 20tl 2012

Population (million)

GDP (RM million)

Cash in circulation(RM million)

26.t 26.6522,445 574,44130,177.6 33,5t9.4

27.2 27.7639,776 738,677

36,247.1 40,424.6

28 .1 29 .1712,851 797 ,32'743,439.2 47,68s.2

29.1 29.5884,456 941,23757,395.6 63,264.1

Transactions per capita (nu)

Cash in circulation

Cheque

E-payments:

Credit card

Charge card

Debit card

E-money

ATMI

Internet banking2

Mobile banking3

Others

1,155.0 1,258.25t,964.t 54,t27.63,5t7.9 22,n4.3I ,565.3 I ,785.1

82.8 87.810.0 24.437.7 47.8

146.5 80.6668.0 940.3

0.2 0.41,007.5 19,147 .9

1,333.963,078.3

) ) ) 1n) )

2,068.689.041.560.3

858.0t,602.5

0.8217,98r.s

1,457.963,528.884,445.02,354.6

I10 .470.674.5

692.01,630.8

2.6279,s09.5

1,546.9s9,669.2

7q5 175 R

2,468.5135 .998.878.4

704.r23,264.s

4.7268,62r

I,668.064,437.9

340,33r.42,79r .7

164.0t65.694.'l

t.094.347,430.5

4.2288,586.3

1840.568,125.3

453,014.53,055.4

185 .42 t6 .6n9.7

1,282.970,355.6

24.4377,773.6

1944.468,880.7

513,846.83,187.4

2 1 8 . 8293.4143.8

1,387.364,259.5

r24.1444,229.9

Soarce: Bank Negara Malaysia 2012Note: ... denotes Negligible; n.a. denotes Not available

l. Comprise bill payments, payments for electronic share application, interbank funds transfer, reloading of MEPS Cash, Touch 'n Go andmobile prepaid value

2. Exclude non-financial transactions, credit card and IBG transactions performed online3- Others include Interbank GIRO, Direct Debit, Financial Process Exchange (FPX) and Real Time Electronics Transfer of Funds and

Securities (RENTAS) - Third party transactions which include Government, custom duty and third party payments. Third party paymentrefers Interbank Funds Transfer System transaction with a minimum amount of RM10,000, where the beneficiary or ordering party is anon-RENTAS member.

from nu3.8 million in 2005 to RM4l million in 2012. Forintemet banking and mobile banking, the expansion rateis equally dramatic.

Whilst we can understand why internet and mobilebanking undergo a fast acceptance rates from thepopulation, such pattern of increasing usage of ATMmight be puzzling. This could be possibly explainedby the fact that the function of ATM nowadays is muchmore sophisticated than those in the 1990s. Besides cash

withdraw and checking account balances, customers nowcan also pay routine bills; paying Touch 'n Go reload,purchase pre-paid cell phone credit; transferring funds,and etc. Other technological innovations of in bank self-service terminals include cash deposit machine, chequedeposit machine, cheque scan machine and passbookupdate machine. All these terminals facilitate custom€rsto do banking activities without constraint to officehours. The costs incurred are much more competitive

TABLE 2. Basic Payment Statistics

200s 2006 2007 2008 2009 2010 20n 2012

Volume of Transactions(million)

ATMI

Intemet banking2

Mobile banking

3.621.60.4

1 . 833.50.9

20.662.6

T ,4

2t.s84.91 . 6

24.7 38.8 ',42.r

46.4n0.2 153.3 203.4 232.6

2 .s 2 .3 2 .2 7 .1Value ofTransactions(RM billion)

AIMI

Intemet bankingr

Mobile banking

3.8259.1

2.1 23.3 19.2334.8 417.8 624.4

19 .8 3 1 .3 37 .3 4 l . 0702.0 1.415.4 2137.0 3078.1140.9 137.9 852.0 4237.04.s 10.5 21.2 7 1.5

Solare: Bank Negara Malaysia.2012Note: .-. Negligible: n.a. Not available

l. Comprise bill payrnents, payments for electronic share application. funds transfer, reloading of MEPS Cash. Touch n'Go and rnobileprepaid r,alue, Data on funds transfer is only available fr-orn 2007.

2. lnclude non-financial transactions. antl lransaclior.rs bv conrorate subscribers liorn 2(X)5.

u4

as compared to ernploying a bank teller or clerk overthe counter.

ln short, the technological innovation iri Malaysianbanking landscape in neatly on track. Migrating frompaper-based payments to electronic and terminal baseddelivery system has provided an opportunity for thebanks to improve productivity levels and lower the costof doing business, and thus it is expected to improve theoverall efficiency of the banking sector, as well as to theMalaysian economy as a whole.

DEA EFFICIENCY STUDIES ON MALAYSIANBANKS

Basically the efficiency analysis in the banking sector canbe measured in terms ofratio analysis, technical efficiency,cost efficiency, and profit efficiency. In this paper, wefocus on technical efficiency. Technical efficiency canbe furthered decomposed into pure technical efficiencyand scale efficiency. Pure technical efficiency deals withthe management ability in maximizing it production ofoutputs at a given set of input level or minimizes theuse of inputs given the output vectors while the scaleefficiency deals measure the performance of the bankswhen the scale ofoperations increases.

Data Envelopment Analysis (oen) is one of themost commonly applied method in capturing technicalefficiency. DEA is a non-parametric frontier efficiencytechnique, which is essentially a relative benchmarkingmethod, providing an overall, objectively determined,numerical efficiency value (X-efficiency) and rankingof the production units that is not otherwise available(Berger and Humphrey 1997). The concept of frontierefficiency techniques can be dated back to the hallmarkpaper ofFarrell (1957), but Sherman and Gold (1985) isperhaps the earliest study to apply ore in banking studies.An extensive discussion on some recent applicationof nee, in banking studies can be found in Berger andHumphrey (1997), and Thanassoulis (1999).

In the Malaysia context, Katib and Mathews (2000)had studied the characteristics ofthe technical efficiencyof twenty domestic commercial banks using data fromI 989 to I 995. Their DEA results imply that economies ofscale exist in small banks; diseconomies of scale existin large-sized banks. They further conclude that there isa deterioration of operational efficiency of Malaysianbanks in the 1990s. Another pioneer study that is morerelevant with our paper is Salleh et al. (2001). Salleh etal. (2001) covered Malaysian local commercial banks,as well as foreign banks and utilized physical input suchas ArM, staffnumber and bank branches, and their oEarelative efficiency is measured based on 3-year analysisover the Asian financial crisis horizon ( I 997- I 999). Thepaper docurrented considerable relative inefficiencyin Malaysian local banks, and they are rnainly due tooversupply of inputs. especially ATIv!. They also lbund the

Jurnal Ekonomi Malay"sict 48(2)

inefficient banks suffer a great disadvantage in terms ofprofit. The paper indicates that foreign bank, especiallyCitibank, despite having tough regulatory restriction, stillperform superbly. The study concludes that the relativeefficiency of the commercial banks in Malaysia is notchanging very much across the crisis horizon, despitehaving to face the Asian financial crisis and the bankingmerger program initiated by the Malaysian governmentin 1999.

As a result of the banking merger program inMalaysia, many scholars have investigated the impactof the merger program. Krishnasamy et al. (2004) forexample, investigated productivity changes of Malaysianlocal banks in the post-merger year in 2000-2001. Thepaper applies only monetary variables, where the inputsused were overhead expenses on labor, and the total assets(excluding loans, advances and financing) while loansand advances and total deposits are defined as outputs.The paper documented that Malaysian banks achieveda total factor productivity growth of 5.lYo as the resultof the merger. However, they find that the merger hasnot resulted in better scale efficiency. Another study bySufian (2004) also focuses on the effect ofmergers andacquisitions towards the technical efficiency of the localcommercial banks over 1998-2003 periods. The inputsused were labor expenses; fixed assets, retail and otherfinancial institutions deposits, while the output coveredare total loans, and investment and dealing securities,which are all monetary measures. The paper found thatcommercial banks experienced an improvement in overalltechnical efficiency scores after the merger process, butit was the small and medium size banks that benefitedthe most from the merger program. Large banks seem toface reduction in scale efficiency after the merger. Thestudy highlighted that most of the banks in Malaysiahave achieved 100% in terms ofpure technical efficiencyeven though the average pure technical efficiency scoresof all commercial banks is relatively low as comparedto scale efficiency.

In a more general study Mohd. Azmi et al. (2006)follows up to study I I local Malaysian commercialbanks for the period 2000-2004 using loans andadvances, capital market investments, and money marketinvestments as outputs, and total deposits, personnelexpenses, and capital expenses as inputs. This studyalso reported an increase of total factor productivityin the industry as a whole and concludes that scaleefficiency is relatively important than pure efficiency.They conclude that bank size does matter in improvingbank efficiency and more importantly, and the negativegrowth oftechnical efficiency signal the needs ofhigherutilization of technology among the Malaysian localbanks. The result is occurred by Tahir et al. (2009a,2009b) who expanded the sample to include foreignbanksover the sarnple of2000-2006 periods. They find that thesource of inefficiency of the locally-owned comrnercialbanks in Malaysia is due to pure technical inef f ic iency.

Technological Innot,utiotr in Services cmd the Efficienct,ol Mukn;siun Commercial Bonks

But the domestic banks are relatively rnore technicalefficient as compared to foreign banks. The main sourceof foreign banks inefficiency is due to scale inefficiency.Tahir et al. (2009a,2009b) however only covered threemonetary variables; i.e. total deposits and total overheadexpenses as their inputs and total eaming assets as theiroutput. Similar results are also documented in Sufian(2006, 2007) on the non-commercial bank financialintermediaries whereby the scale inefficiency dominatespure technical inefficiency. A recent study on Islamicbanking by Muhammad-Rus et al. (2011) comparedefficiencies of Islamic foreign banks and conventionalbanks with Islamic bank subsidiaries during financialliberalisation. The study employs the Stochastic FrontierAnalysis (sre) and report that domestic Malaysian banksare more efficient. Their results show that the latterhave cost-saving technology as compared to the Islamicforeign banks. The conventional banks however do notshow any improvement in efficiencies over time althoughthe government have set to improve their efficiency underthe financial liberalization era.

As a whole. we see that the above studies have notincorporated the role bank delivery channels, or evenphysical inputs to study Malaysian banking performance.Unlike orl banking studies in developed countries, theinput and output used in their oea analysis for Malaysianbanks are mainly relies on the accounting financial items.This is due to the problem ofbank data unavailabilityin emerging countries as highlighted by Kwan (2003).The only study that involved physical input that relate totechnological innovation which come close to our settingis Salleh et al. (2001), which applied ena as an input.However, the importance and functionality of aru todayis far advance than before. We would like to see how theimprovement in technological innovation in bank servicesis able to gauge the relative efficiency of commercialbanks in Malaysia.

METHODOLOGY

onn provides multiple inputs and multiple outputs wayof measuring efficiency, and it is widely been used toassess the operating efficiency of public sector andnon-profit organizations. The optimization procedurein DEA ensures that a particular commercial bank beingevaluated is given highest score possible by maximizingits relative efficiency ratio, at the same time maintainingequity for all other bank. ose basically established arelative scoring system lead by the benchmark e{frciencyscore ofunity that no individual unit's score can exceed.This is a mathematical approach based on the concept ofPareto efficiency in developing the production frontier inestimating the relative efficiency of the decision-makingunits (orraus) involved in the analysis. ln other case, thernost efficient banks are located on the production fi'ontierrvith a scole of one while othels inefficient banks nriqht

score between zero and one and lie below the productionfrontier. Therefore, DEA rneasures the relative efficiencyofthe luus orbanks as compared to the benchmark onausor banks. Hence, all the banks that lie on the frontier areknown as the reference peers. The main advantage of oe,A.approach is that it does not require a prespecified functionas compared to the econometric methods. However, theshortcoming of pgR is that the unspecified technologylevel of all individual banks is assumed to be the same.

The relative efficiency score E" for k-banks, or so-called decision making units (otr,tu), is given as follows:

$ , . . .2Y i t r i r

E x = # ( l )

Er,,N,o

where Irn andXn denote theT{h t-output and l-th k-inputrespectively for k-bankss (the ft{h DMU). V1* and U6 arethe weight placed on the 7-th k-output and i-th /r-input

. i . . , m

respecuvery &r le L/ , r : Lqo fo. a l l k (k : 1,2, . . . ,9) .

This paper applied the basic constant variable-to-scaleDEe model. This setting is realistic as the change ininputs and outputs vary in different proportion. Theestimation of oeA based on variable return to scale iscarried out because it allows a further decompositionof the technical efficiency score into pure technical andscale efficiency. Transforming the model into a linearfractional programming formula, the focus is to solvethe normalized, E* that is e*:

Maximizing "*:fin,rr,r e)

Subject to the constraints of:

l ,u,rx,r: I

3gl ivi*Y1*- Liiui[ ik<0

V j k > 0 , j : 1 , 2 , . . . , n

U , * 2 0 , i = 1 , 2 , . . . , mn m

2v,r:2.u,*l = t t = l

A very important concern in applying DEA inmeasuring bank efficiency is to define inputs and outputsof the model. Traditionally there are two approaches to befollowed; the production approach and the intermediationapproach. The production approach interpret bank asproducer of services for account holders, by inputtingoperating resources. Some of the halhlark works relatedto this approach are Sherman and Gold (1985), Ferrierand Lovell ( 1990), Fried et al. ( 1993) and Wheelock andWilson ( 1995a). ln contrast, the intermediation approachtreated banks as an intermediary, transferring financialassets from the surplns units to the deficit units. Amongthe studies that followed this approach includes Chan-reset a l . (1990). Berger and Humphrey (199l ) . Wheelock

u5

116

and Wilson (1995b), Mil ler and Noulas (1996), Haslemand Scheraga (1999) and Saha and Ravisankar (2000),to name a few. Latly, there is another modefn approachwhich mix up the

To gauge for technological innovation, the inputvectors employed in this study covered the number ofbranches, number of staff, number of atu machines,number of cash deposit machine, number of chequedeposit machine, number of cheque scan machine, andnumber of passbook update machine. For the outputswe use total loans and advances and total deposits.With our setting, it is believed that the advancementin technological employment enables better and moreconvenient services to the banks customers which willattract more customers to the banks and improve theamount of loans and deposits customers. Basically whatwe assumed here is that banks employed various factorsof production in generating the banks financial goodsand services, namely the loans and deposits ofthe banks.Thus, we are following the production approach for theestimation of technical efficiency scores of technologyinnovation. The reason is all our inputs are physicalfacilities that provide services for the account holderswhich are highly consistent with the definition of theproduction approach. Our model does not fir to explainthe intermediate role of banks in the fund transferringprocess. The two outputs, i.e. loans and advances anddeposits employed are expressed in rul million.

We cover 14 out of 24 (before 200'7 there were 25)commercial banks operate in Malaysia for the year 2004,2006,and2008. The reasonwe skip annual data is that theinformation for the service terminals are only availableonce in every two years, provided by The Association ofBanks in Malaysia (enu). The list of the banks covered isreported in Appendix A. We exclude some banks due toincomplete data or insufficient data variation, especiallythose do not have any self-service terminals, includingABN AMRo Bank Berhad, Bangkok Bank Berhad, BankofAmerica Malaysia Berhad, Bank of China (Malaysia)Berhad, Bank of Tokyo-Mitsubishi (Malaysia) Berhad,Deutsche Bank (Malaysia) Berhad, J.P Morgan ChaseBank Berhad, The Bank of Nova Scotia Berhad and TheRoyal Bank of Scotland Berhad, which are all foreignbanks.

RESULTS AND DISCUSSION

DESCRIPTIVE STATISTICS

The descriptive statistics of all the inputs and outputsinvolved are reported in Table 3. ln general, there issubstantial variation in the distribution of each of theinputs and outputs across space and time. For example,the mean unit of Branch in 2004 is 124 units, with amedium of 93 branches, but a range of 331 branches(maxirrum is 334 branches rnir.rus minimunr 3 branches).

Jurnal Ekonomi Malaysiu 48(2)

The standard deviation is 98 branches with a coefficientof variation of 0.79. This wide dispersion is due to therestrictive policy of Bank Negara Malaysia (aNu) inallowing foreign banks to expand. Relatively foreignbanks have very small number of branches in Malaysia,with United Overseas Bank (Malaysia) Bhd having themost number of branches, i.e. 36 branches in 2004.Many foreign banks only have I branch (all of theseforeign banks are not covered in the present study). Forour sample, Citibank is the one with the least numberof branches - only 3 branches in 2004 (increases to 1 Ibranches in2012). The number ofbranches also increasesover years, where in 2008 the mean is 141 but in2012the mean is 156. Of course, the expansion is due mainlyto local banks. Similar pattern of huge variation areobserved for other inputs and outputs more or less forthe same reasons mentioned above.

In fact, the variation in the new facilities, i.e. the4 machines that facilitates payment system, are evenbigger. Note that not all of the sample banks offer afull set of these modern facilities. In 2004 for example,Public bank still do not have any unit of Cheque DepositMachine, while Alliance and Eon Bank do not providethis modern payment facility in the whole sample. RHBand Citibank do not have Cheque Scan Machine in oursample. Hong Leong Bank has 2 units in2004, but thisfacility was not offered after 2005, while Affin and clrr.lsonly offer this payment facility in 2005. Only 3 banksoffer Passbook Update Machine, they are Maybank, ocncbank and Standard Charted Bank Malaysia. It seems thatthe leading local banks have offered a huge quantify ofthese new payment facilities. For example Maybankhave 359 units of Cash Deposit Machines in 2004, andthe number continue to rise to 986 units of Cash DepositMachines in20l2. Other local banks however, averagelyonly have 50 units in 2004 but also have increases to 434units in 2012. However, if we were to look at quantityper branch, foreign banks actually top the list. By branchlevel, Citibank, HSBC and ocec offer actually offeringmore of these facilities relative to Maybank.

TECHNICAL AND SCALE EFFICIENCY FROM DEA

The results ofore scores are tabulated in Table 4. Thetable shows the technical efficiency of the commercialbanks over year 2004-20012 and the decompositionof pure technical efficiency and scale efficiency. Onaverage, in the context of technological innovation inbanking delivery system, the average technical efficientof commercial banks in Malaysia are relatively close withaverage efficiency score of 81 .77% in 2004,92.98Vo tn2006 ̂ 89.66% in 2008. 9 5 .1 5% in 20 1 0 and 81 .7 2 rn 2012.This means that the inefficient banks (those scored lessthan 100%) would have to reduce less than 20o/o of theirfactor of production in the delivery system in order toproduce the same level offinancial output as their efficientcounterpaft. ln addition. it can be seen that these banks

Technological Innovation in Services and the Elficiency of Maktysian Commercial Banl<s

TABLE 3. Descriptive Statistics of DEA Inputs and Outputs

t 1 7

Inputs Outputs

Branch staffCash Cheque

Deposit DepositMachine Machine

Cheque PassbookScan Update

Machine Machine

Loans &DeDOSlt

Advances .: . .(mil l ion)

(mll l lon)

2004MeanMedianMaximumMinimumStd. Dev.C.V.

12493

J J +

J

980.79

5r2432 t220764146250930.99

2901 5 5I 128

3t71.09

5526359

A

90t.64

5428290079t.47

207

t02029t.46

t 70

203054

3 .14

23673.07 29873.r416101 .50 18012 .5086718.00 l l 1046.007221.00 6680.002060r.33 2697r.48

0.87 0.902006MeanMedianMaximumMinimriniStd. Dev.c.v.

1 3 5)25353.

J

l 0 l0.75

64604714

2081320to49460.77

35068.9323074.00t27848.0012901.00303s6.92

0.87

41t9t.6428670.501s3175.0016255.0037472.47

0.91

J J J

t691 5 1 8324ll1 .23

8952509l 81261.41

9243

z l 5

094t .o2

342 l187050t.46

t 70

218058

3.382008MeanMedianMaximumMinimumStd. Dev.c.v.

l4ln4372

1170.83

6977533622465201258420.84

462178

268038719L56

18810583739226r.20

86t z

3360

l l lt .28

90563470

l l 6t .29

9096026

2.86

41495.1426309.50t36224.0013019.0035108.28

0.85

54066.s034328.00149576.0017787.0043246.50

0.802010MeanMedianMaximumMinimumStd. Dev.C.V.

7,4905,15922,8561 ,8186,0720 .81

575208

2,831J J

8381.46

t43120384

1170.82

26013494428303t . 1 7

121394420

146t .2 l

93524830

140l . 5 l

60460l 3

2 . 1 7

47 ,084 66,91828,75t 42,539144,432 t93,575t6,278 21,81538,964 54,0510.83 0.81

2012MeanMedianMaximumMinimumStd. Dev.c.v.

9,61'l4,43442,6932,08410,821l . l 3

707231

2,8663 89031.28

340136

I ,104483561.05

15698392l l

0.84

244'l l0 '759'35 '244'L00.

202507

5 . 4 |

85,616 t02,59755,193 65,607214,852 243,97019,727 26,95972,913 83,2220.85 0.81

Std. Del'. Denote standard deviation and C.V. is Coefficient of Variation; " The Association of Banks in Malaysia has merged the statistics of ChequeDeposit Machine and Cheque Scan Machine in year 2012 and onwards.

TABLE 4. Technical and Scale Efficiencv Scores

Bank (1) Technical Efficiency (2) Pure Technical efficiency Scale Efficiency = [(l)/(2)] Type

2004Affin

Alliance

AmBank

CIMB

Eon

Hong Leong

Maybank

PublicRHB

II0 .71I0.593l0.402II

I

I

0 . 7 lI

0.593I0.402II

IRS

IRS

DRS

Technological Innovation in Services and the Elficiency of Malaysian Commercial Banks

TABLE 3. Descriptive Statistics of DEA Inputs and Outputs

1 1 7

lnputs Outputs

Branch Staff ATMCash Cheque

Deposit DepositMachine Machine

Cheque PassbookScan Update

Machine Machine

Loans &Advances .':ryt'l

. (mrl lron)(mrllton)

2004MeanMedianMaximumMinimumStd. Dev.C.V.

12493

J J +

J

980.79

5t24321220764r46250930.99

2901 5 5n28

3t71.09

55263s9490t .64

54282900791.47

207

t020291.46

r70

203054

3 . t 4

236"t3.07 29873.1416101 .50 18012 .5086718.00 111046.00722t.00 6680.002060t33 2697r.48

0.87 0.902006MeanMedianMaximumMinimumStd. Dev.c.v.

646047 t420813201049460.77

35068.9323074.00t27848.0012901.0030356.92

0.87

4n91.6428670.s0153 I 75.0016255.0037472.47

0.91

135125353

J

1010.75

J J J

t691 5 1 8

4ll1 .23

8952509l 8t261.41

92+5

273094t.02

342 l1870501.46

1 70

218058

3.382008MeanMedianMaximumMinimumStd. Dev.C.V.

t4ltl4372

1170.83

697753362246520t2s8420.84

4621782680387191 .56

188105837392261.20

86323360

l l l1 .28

90563470

l l 61 .29

9096026

2.86

41495.1426309.50136224.0013019.0035108.28

0.85

54066.5034328.00149576.00t7787.0043246.s0

0.802010MeanMedianMaximumMinimumStd. Dev.c.v.

7,4905 ,159

22,8561 , 8 1 86,0720.81

57s208

2,831J J

838t.46

1431203847

1170.82

260t3494428303t . t 7

121394420

146r .2 l

9352

4830

140l .5 t

60

460

I J

2 .17

47,084 66,91828,757 42,539t44,432 193,57516,278 2 l ,81538,964 54,0510.83 0.81

2012MeanMedianMaximumMinimumStd. Dev.C.V.

9,6174,43442,6932,08410,821l . 1 3

70723r

2,866389031.28

340t36

1 ,104483561.05

15698392l l1320.84

244"I l0'7 59.J )

244.1.00 '

20250

3.47

85,616 102,59755,193 65,607214,852 243,97019,727 26,9s972913 83,2220.85 0.81

Std. Dev. Denote standard deviation and C.V. is Coefficient of Variation; 'TheAssociation

of Banks in Malaysia has merged the statistics of ChequeDeposit Machine and Cheque Scan Machine in year 2012 and onwards.

TABLE 4. Technical and Scale Efficiencv Scores

Bank ( I ) Technical Efficiency (2) Pure Technical efficiency Scale Efficiency : t(1)i(2)l Type

2004Affin

Alliance

AmBank

CIMB

Eon

Hong Leong

Maybank

PublicRHB

II0 .7 |I0.593I0.402II

II0 .7 |I0.593I0.402II

IRS

IRS

DRS

1 1 8

SouthemCitibankHSBC

OCBC

United Overseas

0.902I0.5030.6270.711

II0.640 .715I

0.90210.7850.8780.711

Jurnal Ekonomi Malaysia 48(2)

IRS

DRSIRSIRS

Average 0.8t77 0.9539 0.85582006Affin

Alliance

AmBank

CIMB

Eon

Hong Leong

Maybank

Public

RHB

Southem

CitibankHSBC

OCBC

United Overseas

0.9980.894I0 .918I0.9980.918I0.7670.668I0.856II

0.9980.894I0 .918I0.9980 .918I0.7670.66810.856II

IRS.IRS

IRS

IRS

IRS

IRS

IRS

Average 0.9298 1.0000 0.92982008Affin

Alliance

AmBank

CIMB

Eon

Hong Leong

Maybank

Public

RHB

Standard Charted

CitibankHSBC

OCBC

United Overseas

0.6580.5811I0.8940.980.58IIII0 .859II

0.6580.581II0.8940.980.58IIII0 .85911

IRSIRS

IRSIRSDRS

IRS

Average 0.8966 1.0000 0.89662010Affin

Alliance

AmBank

CIMB

Eon

Hong Leong

Maybank

Public

RHB

Standard Charted

CitibankHSBC

OCBC

United Overseas

I0.7681II0.9290.624IIIIIII

I

0.76811I0.9290.624IIIIIII

IRSDRS

IRS

L0000Average 0 .9515 0.95 l52012Affin

All iance

0 . 6 1 30.396

0 .61 30.396

IRS

IRS

Technological Innowfion irt Services and the Et'ficienc.v ol Malatsian Commercial Bcmks t 19

AmBank

CIMB

Eon (delisted)

Hong Leong

Maybank

Public

RHB

Standard Charted

Citibank

HSBC

OCBC

United Overseas

0.869I

0.746I1I0.955I0.8241I

I

I

0 .869

l

0.146

I

I

I

0.9s5I

0.824

I

I

IRS

IRS

IRS

IRS

Average 0.8772 1.0000 0.8772

Note: CRS : Constant Retum to Scale; VRS: Variable Retum to Scale; IRS/DRS - Increasing/Decreasing Retum to Scale .

are relatively pure technical efficient as compare to scaleefficient, indicating banks management is able to decideon their mixed of inputs effectively in the production offinancial goods and services. This result is very consistentwith Sufian (2004), Mohd. Azmi et al. (2006), and Tahiret al. (2009a,2009b).

It is obvious that most of the foreign banks arerelatively inefficient compared to their local counterpartsin2004. This result is consistent with Tahir et al. (2009a,2009b). Both HSBC and ocgc banks have scored a verylow relative technical inefficient in year 2004 mainlydue to relatively low pure technical efficiency scores.This indicates that HSBC and ocsc banks are relativelyinferior in dealing with the mix of factors of productionin order to produce the financial output. Nevertheless,as time goes by, these foreign banks have gained morecompetencies when Malaysia government continuesto liberalize the banking sector, especially after the(gradual) lift of the capital control which started inSeptember 1998, and the return to the flexible exchangerate regime in mid of 2005. In addition, Maybank isfound to experience decreasing return to scale for allthe years under observation and this had affected theiroperational efficiency in year 2004,2008 and 2010 withthe reported technical efficiency score of 40.2%,58%and 62.4Yo, respectively. This indicates that the bankshad expanded beyond the optimal size and resulted indiseconomies of scale. Despise generally inefficient,one ofthe foreign banks appeared to be very outstandingrelative to Malaysian local banks - Citibank manages toachieve 100% efficiency across the 5 sample periods. lnfact, Citibank and Public Bank are the only two banksthat successfully remain in the perfect efficient group (fullscores across the three periods) but as to be discussed inthe next section, Citibank is the most referred bank in thenEa benchmarking process.

ANALYSES ON T}IE EFFICIENT BANKS

Note that in the above efficiency score tabulation, theordel of the etficient banks does not iruply efEciency

ranking. The efficient banks with 100% scoring areequally efficient. In the nE,A. scoring process, the rankingfor each bank is based on their benchmark to the 100%scored bank, or so called the efficient banks. Furtheranalysis is needed to tell which one of these efficientbanks is actually the most refereed efficient bank in thebenchmarking process. We proceed to analyze the DEAbenchmarking process by studying the reference peer(s)for each bank across all the 3 sample periods. The resultsare tabulated in Table 5.

The reference peers refers to the selected efficientbank(s) that each bank is benchmarking with incalculating the individual IEA score. Ifabank is efficient,it will not have to refer to any other bank as referencepeer, and the inefficient banks might have more than onereference peer in the benchmarking process. In Table 5,we tabulate not only the reference peers'identity but alsothe weight distribution if the bank has more than onereference peer. We also tabulate the total frequency a bankis been selected as reference peer for other inefficientbank in the last column.

From Table 5, relative to the others, both Publicbank and Citibank have appeared to be the overall mostfrequently refereed efficient banks. This is in termsof their percentage weight in the inefficient banks'benchmarking process, as well as the total count numberthat they sewed as reference peer for the inefficient banks.A further analysis shows that Citibank is actually moreefficient relative to Public Bank as Citibank has higherweight as well as total count in every benchmarkingprocess and in the two cases where both ofthem appearedas the reference peers for the inefficient bank, Citibankactually dominates the weighting process. For example,in 2004, both Citibank and Pr.rblic Bank have appearedtogether as the reference peer for the sarne inefficientbank twice. The first is for Eon Bank where Citibank has aweight of 52.4% while Public Bank only has2.9o/o weight.The second is for Hssc where Citibank's weight is93.2%while Public Bank is only 4.8Vo. For 2006 and 2008, and2010, Citibank shows similar dominant, but oCBC appearsto be tlre dominant ref-erence neel in vear 20 12.

120 Jurnal Ekonomi Malaysia 48(2)

TABLE 5. Reference Peer in the DEA Benchmarkins process

Reference Peer in DEA Total

2004 SelfLocalBank

ForeignBank

Count asReference

I2J

4

AffinAllianceAmBankCIMB

YesYesYesYes

No

YesYesYesYesYesYes

No

No

No

Public(2.e%)

0t20

05 Eon

6 Hong Leong7 Maybank8 Public9 RHB10 Southem11 Citibank

12 HSBC

13 OcBc

14 United Overseas

Alliance(323%)

poiri"(4.8%)

AmBank(2.2%)

AmBank(r0.3w

Southern(r2.4%)

Maybank(2%)

Citibank(s2.4%)

Citibank(e3.2%)Citibank(e7.8%)Citibank(89.7o/o)

0I

20I+

0

0

02006

I2J

AffinAllianceAmBank

YesYesYes

No

YesYesYesYes

No

00I

U

000J

0

0

J

0

0J

No

Yes

No

4 CIMB

5 Eon

6 Hong Leong

7 Maybank

8 Public

9 RHB

l0 Southem

11 Citibank

12 HSBC

13 OcBc14 United Overseas

Public(60.6yo)

PuUti"(?o 10/^\

PuOfi. emSunt(4.e%) (3.2%)

unit"aoon:

_^ ' -

Citibank(60.60/0)

Citibank Unitedo(e2.e%) (7.1%)

Citibank UnitedO(68.so/o) (23A%)

YesYes

2008I2J

45

AffinAllianceAmBankCIMBEon

I

0000

0

0I

I

0

z

0

YesYesYesYesYes

No6 Hong Leong

7 Maybank8 Public9 RHBl0 Standard ChartedI I Citibank

I2 T{SBC

YesYesYesYesYes

No

RHB(48.e%)

Public(8 .2%\

Affin(3 .5? ; )

Citibank(51 .1%)

Cit ibank(88.3ol0)

Technological Innovation in Services and the Elficiency of Malaysian Conrmercial Banks 121

I3 OCBC

14 United Overseas

YesYes

00

2010I Affin

2 Alliance

3 AmBank

4 CIMB

5 Eon

6 HongLeong

7 Maybank

8 Public

9 RHB

l0 Standard Charted

11 Citibank

12 HSBC

I3 OCBC

14 United Overseas

0000I0

0

000I000

YesYesYesYesYesYes

No

YesYesYesYesYesYesYes

Eon(78.6%)

Citibank(2r.4%)

2012I Affin

2 Alliance

3 AmBank

4 CIMB

5 Eon

6 HongLeong

7 Maybank

8 Public

9 RHB

10 Standard Charted

11 Citibank

12 HSBC

l3 OcBc

14 United Overseas

StandardL luDanK

Lna' ( ) 1o/^\(s.2%)

OCBC Standard

(6\ 10/^\ Lna(19.9%',)

OCBC(2.8%)

Yes

No

No

Eon(100%)

CIMB(r4.4%)

Eon(37.6%)

1

0

OCBC(73.8o/o)

YesYes

I

2

0

002I0J

0

No

No

YesYesYesYesYesYesYes

Affin(se.6%)

ANALYSES ON THE INEFFICIENT BANKS

This section proceeds to analyze the inefficient banksto investigate the reason why they are relatively inferiorin terms of performance. This analysis is particularlyimportant for the inefficient banks to identify the area ofweakness for improvement on every aspect (both inputand output). This allows us to tell the ideal improvementeach inefficient bank need to target on. These are basicallythe slack values calculated in oER linear prograrnming.To conserve space, we tabulate the values of the targetimprovement for each sample year in Table 6. The detailsof the target improvement values for each bank areavailable upon request.

The value repolted for inputs are the excess units ofresources an inefficient bank should cut down in order

to become equally efficient as its reference peer. Forexample, given the outputs of Eon Bank relative to itsgroup of peers as shown in Table 4 (Citibank AllianceBank, Southern Bank, and Public Bank), Eon Bankneeds to close off47 branches, laid off 1029 workers,and reduces many of its new payment facilities so that itcan be performing equally efficient as its reference peers.The value reported for the output on the contrary is theamount of output Eon Bank need to increase in order tobe qualify as equal efficient as its reference peers, giventhe current level ofinputs they have.

The last panel in Table 6 reports the total frequency(count) of each input and output appeared in the slackcolumn. We interpret the frequency of each variable beingreported with slack value as a signal showing the variablervorth rnore attention in order to improve the efhciency

122 Jurnol Ekonotni Malaysia 48(2)

TABLE 6. Input and Output Slack of the Inefficient Banks

Input Output

2004 Branches staff ATMCash Cheque

Deposit DepositMachine Machine

Cheoue Passbook

;tafll" **{ ^',ff.:, Deposi'lEon

HSBC

OCBC

United Overseas

47I1 324

1029

501

40

1 8

26

l 5

1 4o2230

89I8

31 5

4643.7 2087.33435.3207 | 367 5.1

3206.3 4859.1

2006CIMB

RHB

SouthemHSBC

70102n68

826I 802622I 083

64735412846

1 1 5A <

764

29

;

9694668

- 11542- n42l

2562.7 10025- 176.7

2008Hong Leong

Citibank

931 l

2 l50 47

'741

2128967

6456

s722.410955

2010Maybank 127.5 430.4 192 70 88.3 6667.2

2012Alliance

AmBank

Hong Leong

Maybank

87101 .4201.s8.3

10252441.22902.7'7 63.5

135492.5t96.261.9

5644.7132.361.4

) 6

32.194"97.2.14.3"

629.8 3092.7- 12094.7

162s07 558.3

Total Count20042006200820102012

n

4

2

I

4

z

42I4

2A

2IA

442IA

+

+

2I4

43I0A

20000

I

I

I

I

3

J

420z

il1 0t 2l 5l 5l 3l 3l 5Total' The inputs ofCheque Deposit Machine and Cheque Scan Machine are merged in year 2012.

ranking. ln terms of input, Branch, Cash Deposit Machineand Cheque Deposit Machine appeared to be the majorproblem for the inefficient banks (which are mainlyforeign banks), with a frequency of 15 times occurredwith slack values, while Deposit is the major problemin terms of output with 11 counts in total. Branch hasbecome a key point differentiating banking efficiency inMalaysia. This is easily explainable given the fact thatCitibank, with 3 branches (and 11 branches in 2012), hasserved as the most important reference peer among all.

The excess supply of Cash Deposit Machine andCheque Deposit Machine is one of the major sourcesof inefficiency to other foreign banks. These foreignbanks have to cope with nNrvr restriction and operatewith lirrited number of branches. They have also takenthe alternative to offer rnany self-service tenlinals inevery branch. However. as cotnpared to the output level

of Citibank and many other local Malaysian banks, theseforeign banks have not produce the required scale ofoutput level as their counterparts. ln terms of output, itappears that these inefficient banks need to worry moreabout their deposit generating ability in order to lift uptheir performance to the level oftheir reference peer(s)is they want to keep the current level of physical inputs.Anyway, we can conclude that optimum technologicalinnovation does matter in determining relative bankingefficiency in Malaysia.

CONCLUSION

Comrnercial banks in Malaysia have been adopting newtechnology and services since the mid 1990s. However,there is no attempt in empirical studies to quantify the

Technological Innovation in Senices and the Efficiency o/ Malaltsian Commercial Banks 123

contribution of technological innovation in bank servtcesin accessing their performance and efficiency. Existingliterature on bank technical efficiency were all build onconventional inputs and outputs variables such as staffnumbers. branches. and financial indicators or ratio.Our study offers a different perspective of efficiency interms of modeling and data. Basically we introducednew variables to gauge for technological innovation,which covered all the new service facilities offered bycommercial banks in the last decade or so, includingcash deposit machine, cheque deposit machine, chequescan machine, and passbook update machine. Thesetechnology platforms not only are cost saving to thebanks but they also improve customer satisfaction withmore easy, speedy and reliable transaction processing.

Based on data over 2004-2012 periods, our newmodeling that incorporated technological services showsthat Malaysianbanks have betterpure technical efficient ascompared to scale efficient, indicating banks managementis able to decide on their input mixed effectively in theproduction of financial goods and services. This resultis consistent with the previous literature, includingSufian (2004), Mohd. Azmi et al. (2006), and Tahiret al. (2009a, 2009b). With technological services asinputs, we documented that foreign banks in general arerelatively inefficient compared to local Malaysian bankssomewhat consistent with Tahir et al. (2009a,2009b), butafter the Malaysia govemment ended its capital controlin 2005, the gap has been gradually closing. The resultis of no surprises given that the Malaysian governmenthas embark on liberalization in the banking and financialsectors, which surely allow for more competition betweenlocal and foreign banks. Sooner or later we can expect theforeign banks to become a dominant force in Malaysianshould the local banks is not well prepared for the openingup of the banking sectors in the near future. Despisegenerally inefficient, one foreign bank - Citibank appearto be very outstanding relative to Malaysian local banks.Our efficient ranking basically is consistent with Sallehet al. (2001) although their study uses conventional setoffinancial input in the efficiency benchmarking processusing the data in the late 1990s.

In short, our study implies that technologicalinnovation in banking services does contribute todifferentiating the relative efficiency of banks. In thecontext of policy recommendation, our novel resultindicates banks that provide better technological servicesdo acquire competitive advantage against their peersand so local bankers should keep abreast with the latesttechnology needs oftheir customers in order to stay aheadof the stiff competition in this industry.

REFERENCES

Berger , A .N. & Humphrey , D.B. 1991. The Dominance o fine f f i c ienc ies over sca le and produc t mix econorn iesi n b a n k i n - e . J o t t r n u l o . f M o n e l u n , E t o n o m i t s 2 8 .| 7 1 4 8 .

Bergeg A.N. & Hurnphrey, D.B. 1997 . Efficiency of financial'

institutions: Intemational survey and directions for futureresearch. European Journal of Operational Research 98:r75-212.

Chames, A., Coopers, W.W., Huang, Z.M. & Sun, D.B. 1990.Polyhedral cone-ratio DEA models with an illustrativeapplication to large commercial banks. Journal oJ'Econometric 46 73-91.

F a r r e l l , M . J . 1 9 5 7 . T h e m e a s u r e m e n t o f p r o d u c t i v e

efficiency. Journal of the Royal Statistical Society A

120:253-281.Fenier, G.D. & Lovell, C.A.K. 1990. Measuring cost efficiency

in banking: Econometric and l inear programming

evidence. Journal ofEconometric 46:229 245.Fried, H.O., Lovell, C.A.K. & Eeckaut, P.V. 1993. Evaluating

the performance ofU.S. credit unions. -/ournal of Bankingand Finance 17 (2-3\: 251-265 .

Gerdes. G.R. & Walton" II J.K. 2002. The use of checks andother noncash payment instruments in the United States.Federal Reserve Bulletin 88 360-3'1 4.

Giannakoudi, S. 1999. Intemet banking: The digital voyageof banking and money in cyberspace. Information andCommunications Technologt Law 8(3): 205-244.

Haslem, J.A. & Scheraga,C.A.1999. DEA efficiencyprofiles ofU.S. banks operating intemati onally. International ReviewofEconomics and Finance 8: 165-183.

Katib, M.N., Mathews, K. 2000. Anon-parametric approach toefficiency measurement in the Malaysian banking sector.The Singapore Economic Review 44l-89-114.

Krishnasamy, G., Ridzwa, A. F. & Vignesan, P. 2004. Maiaysianpost-merger banks' productivity: Application of malmquistproductivity tndex. Managerial Finance 30 : 63-7 4.

Kwan, S.H. 2003. Operating performance ofbanks amongAsianeconomies: An international and time series comparison.Journal of Banking and Finance 27:471489.

Miller, S.M. & Noulas, A.G. 1996. The technical efficiency oflarge bank production. Journal of Banking and Finance20:495-509.

Mohd. Azmi, O., Abdul Rahim, A.R., Rosylin, M.Y., Shabri,A.M. & Mohd. Eskandar Shah, M.R. 2006. Efficiencyof commercial banks in Malaysia. Asian Academic o/Management Journal of Accounting and Finance 2(2):1942.

Muhammad-Rus, A.D., Abdul-Majid M., and Mohd-Nor,A.H.S. 2011. Financial liberalization and cost efficiencyin islamic and conventional banking: A stochastic frontierapproach. Journal Ekonomi Malaysia 45: 89-100.

Saha, A. & Ravisankar, T.S. 2000. Rating of Indian commercialbanks: A DEA approach. European Journal ofOperationalResearch 124l-187-203.

Salleh, Y., Hooy, C.W, Goh, W.K. 2001. Using DEA to ratecommercial banks. Banker b Journal Malaysia I I 8: 5-1 8.

Sherman, H.D. & Gold, F. 1985. Bank branch operatingefficiency: Evaluation with data envelope analysis. Jotu"nalo/ 'Banking und Finance 9:297 316.

Sufian, F. 2004. The efficiency effects of bank mergers andacquisit ions in developing economy: Evidence fromMalaysia. International Journal of Applied Econonen'itscttrd Quantitatit,e Studies l(4):53 74.

Sufian, F. 2006. The efficiency ofnon-bank financial institutions:Ernpirical evidence fror-r-r Malaysia. International Journulol Firrunc'e ancl Econoruit ,r 6: 49 65.

124 Jurnal Ekonomi Malaysict 48(2)

Sufian, F. 2007. What explains differences in the efficiency of Chee-Wooi, Hooy*non-commercial bankfinancial intermediaries: Empirical School of Managementevidence from Malaysia. Asian Academic of Management Universiti Sains MalaysiaJournal of Accounting and Finance 3(l): 37-57 . 1 1 800 Penang

Tahir I.M., Nor Mazlina, A.B. & Haron, S. 2009a. Estimating MALAYSIAtechnical and scale efficiency of Malaysian commercialbanks:Anon-parametricapproach. International Review Sok-Gee,Chan**ofBusiness Research Papers 5(l): 113-123. Departrnent ofFinance and Banking

Tahir I.M., Nor Mazlina, A.B. & Haron, S. 2009b. Evaluating Faculty of Business and Accountancy Buildingefficiency of Malaysian banks using data envelopment University of Malayaanalysis. International Journal of Business and 50603 Kuala LumpurManagement 4(8): 96-106. MALAYSIA

Thanassoulis, E. 1999. Data envelopment analysis and its usein banking. Intedaces 29(3): l-13. *cwhooy@usm.my (corresponding author)

Wheelock, D.C. & Wilson, P.W. 1995a. Evaluating the **sokgee@um.edu.my

efficiency of commercial banks: Does our view of whatbanks do matter? Review, Federal Reserve Bank of St.Louis 77;39-52.

Wheelock, D.C. & Wilson, P.W. 1995b. Explaining bankfailures: Deposit insurance, regulation, and efficiency. TheReview of Economics and Statistics 77 : 689-700.