资源描述
ContentslistsavailableatScienceDirect Pacific-BasinFinanceJournal journal homepage: Dofinancialtechnologyfirmsinfluencebankperformance? DinhHoangBachPhan a ,PareshKumarNarayan b, ,R.EkiRahman c , AkhisR.Hutabarat c a DepartmentofEconomicsandFinance,LaTrobeBusinessSchool,LaTrobeUniversity,Australia b CentreforFinancialEconometrics,DeakinBusinessSchool,DeakinUniversity,Melbourne,Australia c BankIndonesiaInstitute,BankIndonesia,Indonesia ARTICLEINFO Keywords: Financialtechnology Bankperformance Predictability Estimator ABSTRACT Wedevelopahypothesisthatthegrowthoffinancialtechnology(FinTech)negativelyinfluences bankperformance.WestudytheIndonesiamarket,whereFinTechgrowthhasbeenimpressive. Usingasampleof41banksanddataonFinTechfirms,weshowthatthegrowthofFinTechfirms negativelyinfluencesbankperformance.Wetestourhypothesisthroughmultipleadditionaltests and robustness tests, such as sensitivity to bank characteristics, effects of the Global Financial Crisis,andtheuseofalternativeestimators.OurmainconclusionthatFinTechnegativelypredicts bankperformanceholds. 1. Introduction The lastdecade or so has seen strong growth in digital innovation, especially in financialtechnology (FinTech). However, the traditional players (financial institutions) in the financial sector have only slowly begun to participate in new technological in- novations(BrandlandHornuf,2017).AlthoughtherehavebeenacquisitionsofFinTechfirmsbybanksrecently,mostFinTechstart- upsareindependentofbanksandareopentoinvestmentinterests.Becausemanybanks,apartfromthewell-knownbigbanks,still offerold-fashioned,costly,andcumbersomefinancialservices(BrandlandHornuf,2017),FinTechfirmshavetheopportunitytotake overseveralkeyfunctionsoftraditionalbanks(Lietal.,2017).Putdifferently,FinTechfirmsarelikelytotriggerasubstitutioneffect, wherebybanksarelikelytocedesomebusinessactivity.TowhatextentbankswillbeaffectedandhowmuchFinTechfirmswill replacetheactivitiescurrentlycontrolledbybanksisanempiricalissue. TheeffectofFinTechfirmsonbankscanbeexplainedbytheconsumertheory(AakerandKeller,1990)anddisruptiveinnovation theory(Christensen,1997).Theconsumertheorysuggeststhatnewservices(suchasthoseprovidedbyFinTechfirms)bymeetingthe sameconsumerdemandcanreplacetheoldservices(suchasthoseprovidedbytraditionalbanks).Basedonthedisruptiveinnovation theory,newentrantswhoapplyinnovativetechnologytoprovidemoreaccessibleandcost-effectivegoodsandservicescancreate competitioninthemarket.TheremitsofthesetheoriesarerelevanttoourstorywherenewentrantsareFinTechfirmsandestablished incumbentsaretraditionalbanksplementingthislineofthoughtistheworkofJunandYeo(2016),whoprovideamodelofa two-sided market with vertical constraints, emphasising on firm entry. Their model focusses on end-to-end and front-end service providersadistinctionthatwedonotmakepetitioninourstoryisgeneratedbynewentrantsregardlessofwhotheyare.Akey featureofFinTechfirmsisthattheyapplyinnovativetechnology toperformtaskspreviouslyreservedforbanks,suchaslending, payments,orinvestments(ChishtiandBarberis,2016;BrandlandHornuf,2017;Puschmann,2017).Recently,FinTechfirmshave doi/10.1016/j.pacfin.2019.101210 Received9November2018;Receivedinrevisedform1August2019;Accepted23September2019 Corresponding author at: Centre for Financial Econometrics, Deakin Business School, Deakin University, 221 Burwood Highway, Burwood, Victoria3125,Australia. E-mailaddress:paresh.narayandeakin.edu.au(P.K.Narayan). Pacific-Basin Finance Journal 62 (2020) 101210 Available online 05 November 2019 0927-538X/ 2019 Elsevier B.V. All rights reserved. Tbeendevelopingpracticalapplicationstoimproveefficiencyinfinancialservicesacrossarangeofservices,including(butnotlimited to):contactlessandinstantpayments;assetmanagementservices;investmentandfinancialserviceadvice;andinformationanddata management/storage(VilleroydeGalhau,2016).Inthisvein,JagtianiandLemieux(2018)arguethatnon-banklenderscansecure softinformationrelatingtocreditworthiness.Thisserviceisconsideredvaluableforconsumersandsmallbusinessalike,particularly thosethatarecharacterizedbyweakcredithistory.Onthecontrary,banksoperateonoldinformationtechnologysystemandare perceived to be slow in adopting new technology (Hannan and McDowell, 1984; Laven and Bruggink, 2016; Brandl and Hornuf, 2017). The main conclusion, therefore, is that eventuallyFinTech firms can substitute the traditional banks by providing less ex- pensiveandmoreefficientservices.Ourhypothesis,therefore,isthatFinTechgrowthwillnegativelyinfluencebankperformance. Despitetheemergenceofdigitalinnovationanditsperceivedeffectonthefinancialindustry,theeffectofdigitalinnovationand FinTech growth on the financial system are less understood. Exceptions include: (a) Cumming and Schwienbacher (2016), who investigatethepatternofventurecapitalinvestmentinFinTechusingaglobalsampleoffirms;(b)HaddadandHornuf(2018),who test the determinants of the globalFinTech market; (c) Brandl and Hornuf (2017), who trace the transformation of the financial industryafterdigitalization;and(d)Lietal.(2017),whofocusonhowretailbankssharepricesreacttoFinTechstart-ups. Wetestourhypothesisusingbank-leveldatafromIndonesia.WeconsiderIndonesiabecause,amongemergingmarkets,itsFinTech growthhasbeenphenomenal,asshowninFig.1.ThistrendinthegrowthofFinTechfirmsmakesIndonesiaaninterestingcasestudyto analysehowFinTechinfluencesbankperformanceinanemergingmarketcontext.Ingeneral,weunderstandlittleabouthowFinTech impactsthebankingsector.Usingdatafrom41banks,ourpanelmodelsofthedeterminantsofbankingsectorperformancesuggestthat FinTechfirmshaveanegativeeffectonIndonesianbankperformance.FinTech,weshow,alsonegativelypredictsbankperformance. Specifically, we summarize our key findings as follows. First, we find thatFinTech reduces net interest income to total assets (NIM),net income tototal equities(ROE), netincome to totalassets (ROA),and yieldon earningassets (YEA) by0.38%, 7.30%, 1.73%,and0.38%oftheirsamplemeanvalues(reportedinTable1),respectively. Second, FinTech predicts bank performance. With every new FinTech firm introduced into the market, we find that FinTech negativelypredictsNIM,ROE,ROA,andYEAby0.53%,9.32%,2.07%,and0.48%oftheirsamplemeans,respectively.Third,wetest whetherbankcharacteristics,suchasmarketvalue(MV)andfirmage(FA)influencethewayFinTechinfluencesbankperformance. Wefindthattheydo.Specifically,theeffectofFinTechisstrongeron(a)largebankscomparedtosmallbanks,and(b)maturebanks comparedtoyounger (new) banks.We concludeour analysis bytesting whetherFinTechaffects bank performancedifferentlyfor state-ownedversusprivatebanks.WeshowthatFinTechhasabiggereffectonstate-ownedbanks. Weconfirmourresultsthroughmultiplerobustnesstests.Usingfourmeasuresofbankperformance,wetestthesensitivityofthe relationbetweenFinTechandbankperformancetomeasuresofperformance.Wefindnoevidencethatmeasuresofbankperformance mattertotherelationbetweenFinTechandperformance.WeexploretheeffectsofFinTechonbankperformancebyaskingwhether the wayFinTech affects performance is dependent on specific bank characteristics. By and large, we find thatFinTech negatively influencesperformanceregardlessofbanksizeandage,andwhilewedouncoversomepositiveeffectofFinTechforyoungerbanks, thereisnoevidencethatFinTechpredictsperformanceoftheseyoungerbanks.WeexplainthispositiveeffectbydrawingonGiunta andTrivieri(2007)andHallerandSiedschlag(2011).Theseauthorsfindthatyoungerfirmsadoptandusetechnologicalinnovations 0 20 40 60 80 100 120 140 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 Accummulated number of fintech firms Number of fintech firms established in a year Fig. 1.FinTechfirmsinIndonesiain1998-2017. ThisfigureplotsthenumberandaccumulatednumberofFinTechfirmsestablishedineachyearinIndonesiain19982017.Dataareobtainedfrom theFintechIndonesiaAssociation. D.H.B.Phan,etal. Pacific-Basin Finance Journal 62 (2020) 101210 2muchmoresuccessfully.Inaddition,intestingtheeffectsofFinTech,weutilizeawiderangeofcontrolvariablesconsistentwiththe bankingperformancedeterminantsliterature.TheroleofFinTechininfluencingperformancesurvivesthesetests.Wealsocheckfor thesensitivity of our results by (a)controlling for 2017 Global Financial Crisis (GFC) effects and (b) using a differentpanel data estimator.WeconcludethatthenegativeeffectofFinTechonbankperformanceholdsacrossalltheseadditionaltests. OurpapersmaincontributionistoshowhowFinTechinfluencesbankperformance.Therearenostudiesonthissubjectatpresent.Our paper,therefore,representsthefirstempiricalstudyexploringthehypothesisthatFinTechnegativelyinfluencesbankperformance.Using bank-leveldatafromIndonesia, 1 weshowthatFinTechnegativelyinfluencesbankperformanceandthatthisrelationisrobust. Thispaperisorganizedintothreeadditionalsections.WediscussthedataandtheempiricalframeworkinSectionII.Adiscussion oftheresultsappearsinSectionIII.Finally,SectionIVsetsforthourconcludingremarks. 2. Data and empirical framework Thissectionhastwoobjectives.First,wediscussthedata.Then,wepresenttheempiricalframeworkfortestingourhypothesis thatFinTechhasanegativeeffectonbankperformance. 2.1. Data Wecollectdatafrommultiplesources.ThedataonFinTechfirmsareobtainedfromFinTechIndonesiaAssociation. 2 Wecollect theannualnumberofFinTechfirmsregisteredtotheFinTechIndonesiaAssociation.TheFinTechfirmsarethosenewsupplyfirms andsettlementprocessesrelatedtothebankingsector,suchaslending,payments,personalfinancemanagement,crowdfunding,and cryptocurrencies.InIndonesia,thebulkoftheFinTechactivitiesarecenteredonlending(45%)followedbypayments(38%).Bank- level dataNIM,ROA,ROE,YEA, total assets (SIZE), ratio of equity to total assets (CAP), cost to income ratio (CTI), loan loss provision(LLP),annualgrowthofdeposits(DG),interestincomeshare(IIS),andfundingcost(FC)areobtainedfromDatastream. Ofthesedata,NIM,ROA,ROE,andYEAareproxiesforbankperformanceourdependentvariableinregressionmodel(1).Variables SIZE, CAP, CTI, LLP, DG, IIS, and FC are firm-specific control variables. We also use macroeconomic variablesgross domestic product(GDP)growthrateandinflation(INF)rateasadditionalcontrols.ThesedataareobtainedfromtheGlobalFinancialDa- tabase.Alldataareannualovertheperiod1998to2017.Specificdetails,includingvariabledefinitions,areprovidedinTable1. AdescriptionofourdatasetappearsinTable2.Selectedbasicstatisticsarereportedtoobtaininsightsonthedata.Thesestatistics arefortheentiresampleaswellasforbanksatthe25thand75thpercentiles.ThenumberofnewFinTechfirmsaverages7perannum overthe1998to2017period.Thebankperformancestatistics(foroursampleof41banks)revealthefollowing.AverageNIMis 4.94%perannumwhileROEis7.99%perannum.Bycomparison,ROAstandsat0.40%perannum.Moreover,YEAisvaluedatover 10%perannum.AnnualaverageCAP,ameasureofmarketcapitalization,isaround12%.Theseperformancestatistics,asexpected, arehigheratthe75thpercentilecomparedtothe25thpercentile.Amongthecontrolvariables,interestincomeis91.2%oftotal income,withaCTIofaround56%perannumforoursample.Growthofdepositsis16.32%perannum. Table 1 Variabledescription. Variable Definition Source Expectedsign FinTech Numberoffinancialtechnology(FinTech)companiesfounded FintechIndonesiaAssociation NIM Ratioofnetinterestincometototalassets Datastream ROA Ratioofnetincometototalassets Datastream ROE Ratioofnetincometototalequities Datastream YEA Yieldonearningassets Datastream SIZE Logoftotalasset($USmillion) Datastream +/ CAP Capitalratioequalsequityovertotalassets Datastream +/ CTI Cost-to-incomeratioequalstotalexpensesovertotalgeneratedrevenues Datastream LLP Loanlossprovisionsequalsloanlossprovisionsovertotalloans Datastream DG Annualgrowthofdeposits Datastream +/ IIS Interestincomeshareequalstotalinterestincomeovertotalincome Datastream FC Fundingcostequalsinterestexpensesoveraveragetotaldeposits Datastream GDP IndonesiaannualGDPgrowthrate GlobalFinancialDatabase + INF Indonesiaannualinflationrate GlobalFinancialDatabase +/ Thistablecontainsdescriptionsandsourcesofvariables. 1 TheliteratureonIndonesianbanksisrich.SeveralstudiesexaminedtheIndonesianbankperformance(Avilianietal.,2015;Wuetal.,2016; Ekananda, 2017a, 2017b; Irawan and Kacaribu, 2017; Ekananda, 2017a, 2017b; Shaban and James, 2018a, 2018b; Ibrahim, 2019), efficiency (Widiartietal.,2015;Anwar,2016;,PurwonoandYasin,2019),risk(Agusmanetal.,2008;Hidayat,Kakinaka,andMiyamoto2012;Agusman etal.,2014),stability(Mulyaningsihetal.,2016;Karimetal.,2016;Dienillahetal.,2018),andIslamicbanking(Pepinsky,2013;Gustianietal., 2010;Hidayatietal.,2017a,2017b;AnwarandAli,2018). 2 fintech.id. This data is not available to public. We obtained data from Bank Indonesia which was sourced from Asosiasi FinTech Indonesia(Aftech). D.H.B.Phan,etal. Pacific-Basin Finance Journal 62 (2020) 101210 32.2. Empiricalframework Our empirical model is motivated by the literature that estimates the determinants of bank performance (Dietrich and Wanzenried,2011,2014;Trujillo-Ponce,2013;KsterandPelster,2017;ShabanandJames,2018a,b).Weaugmentthisconventional modelofperformancedeterminantswiththeFinTechvariable.Ourregressionmodelis: = + + + + + + + + + + + + PER FinTech PER CAP SIZE CTI LLP DG IIS FC GDP INF i t t i t i t i t i t i t i t i t i t t t i t , 1 2 , 1 3 , 4 , 5 , 6 , 7 , 8 , 9 , 10 11 , WecollectdataforallIndonesianbanksfromDatastream.Dataavailabilityleadstoasampleof41banks.Oursampleofbanks excludesunlistedbankssince theyarelikelyto introducepotentialestimationbias. Indonesianbanksarerequired toreveal their performancethroughannualreports submittedtothecentralbanktheBankIndonesia.However,therearedifferencesbetween listedandunlistedIndonesianbanksinthelevelofriskdisclosurethatisconveyedintheirannualreports.Adheringtocapitalmarket regulation, listed firms commit to extensive public disclosures and transparency in showing their performance in order to attract investors for external funds. Unlisted firms, with fewer stakeholders, however, have lack of incentives and the absence of trans- parencywhenrevealingtheirperformanceinannualreports(GoktanandMuslu,2018). Ourdatasamplespans1998,whenthefirstFinTechfirmwasestablished,to2017.Atwo-stepgeneralizedmethodofmoments (GMM)systemdynamicpanelestimatorisemployedtotestthenullhypothesisthatFinTechnegativelyinfluencesbankperformance inIndonesia. SpecificdefinitionsandexpectedsignsoneachofthevariablesaresetforthinthelastcolumnofTable1.Webrieflydiscussthese relationshere.ThefirstcontrolvariableisCAP,measuredasequityscaledbytotalassets.Previousstudiesthattestthecapitalbank performancenexusfailtofindconclusiveevidenceonhowthisrelationunfolds.Somestudiesdocumentapositiveeffectofcapitalon bank performance (Berger, 1995; Holmstrom and Tirole, 1997; Jacques and Nigro, 1997; Rime, 2001; Iannotta et al., 2007a,b; MehranandThakor,2011;NaceurandOmran,2011;BergerandBouwman,2013),whileothersfindtheopposite(Altunbasetal., 2007; Lee and Hsieh, 2013) or mixed results (Dietrich and Wanzenried, 2014). Berger (1995) draws on the bankruptcy cost hy- pothesis to explain the relation between capital and bank profits. This hypothesis suggests that banks with a higher capital ratio increase their
展开阅读全文