PoliciesandSustainableEconomicDevelop ment |1 TheImpactofNon-PerformingLoansonBankProfitability andLendingBehavior: EvidencefromVietnam NGUYENTHIHONGVINH BankingUniversityofHochiminhCi
Trang 1PoliciesandSustainableEconomicDevelop ment |1
TheImpactofNon-PerformingLoansonBankProfitability
andLendingBehavior:
EvidencefromVietnam
NGUYENTHIHONGVINH
BankingUniversityofHochiminhCity-vinhnth@buh.edu.vn
Abstract
Theaimofthisstudyistoinvestigatetheimpactofnon-performingloansonprofitabilityandlendingbehavior,usinganempiricalframeworktha tincorporateswhetheranincreaseof NPLscanleadbankstoreducetheirprofitabilityandl endingactivity.To accountforprofit and lendingpersistence, thepaperappliestheGeneralizedMethodofMomentstechniquefordynamicpanelsusin
gbank-leveldatafor34Vietnamesecommercialbanksovertheperiod2005to2015.Theextantli
teraturepresentnon-performingloansasoneofthemostimportantfactorseffectingonprofitabilityandlendin
gbehavior.Throughoutthewholesample,wefoundsomeevidencesthatnon-performingloanshasstatisticallysignificantnegativeeffectonVietnamesecommerci albanksprofitabilityandlendingbehavior.Thesefindingsshowthatinordertoimprovet h
e performanceoftheVietnamcommercialbanks,bankmanagersandgovernorshaveto dealwiththenon-performingloanproblem.
Keywords:Vietnam;non-performingloan;profitability;lendingbehavior;GMMmodel
Trang 21 Introduction
Theissueofnon-performingloans(NPLs)hasrecentlybecomeacauseforconcerninVietnam,e s
p e c i a l l y
asthelevelofnon-performingloansmayeffectonbankprofitabilityandlendingbehavior.Theratioof NPLsinVietnamsharplyincreasedintheyearof2012.SBVreportsthattheratioofno
was4.3%bythethirdquarterof2012.IMFandWorldBank1(2014)estimatetheratioo fNPLsforVietnambankingsectorwas12%bytheendof2012.Meanwhile,Moody
2(2014)showstheratioofNPLstototalassetsinVietnamwas15%bytheFebruaryo f2014.Althoughthei m p a c t o f NPLso n bankbehaviori s i m p o r t a n t i n Vietna
m,therearefewstudiesaddressedonimpactofnon-performingloansinVietnam.Besides,studiesforVietnamesebanksm a i n l y us esstaticpaneldatamethodssuchastheRandomEffectsModelandtheFixedEffects Model.Thestaticpaneldatamethodsmayleadtobiasinresultsbecausetheyhave notdealwithendogenousi s s u e Thepaperthusappliesthedynamicpaneldata toexaminetherelationbetweenNPLsandprofitabilityandloangrowth.Therese
whethermattersf o r banks’profitabilityandloangrowthinVietnamesecommerci albanks.Theresearchresultsallowsthebank’smanagementtofocus
onissuesthatwillletthemenhancethebank’soverallprofitabilityandlendingacti vityinthefuture.Thisalsohelpspolicymakersfindsuitablebankingpoliciestodeal
w i t h thenon-performingloanproblemforcommercialbanks
Therestofthepaperisstructuredasfollows.Section2looksatpreviousresearch esontheimpactso f
non-performingloansonprofitabilityandcreditgrowth.Section3providesthemethod thatusedi n thisresearch,anddescribesthedatathatareused.Empiricalresult sarepresentedinsection4.Finally,section5containsconcludingremarks
2 Literaturereview
Intheliterature,impactofnon-performingloansonbanksprofitabilityandlendingbehaviorisindicatedthatthei ncreaseofNPLswouldleadtohigherprovisions,lowerprofitabilityandconsidera bleerosioninbankcapital.Thismaycausenegativeeffectsforfurtherlending.Th etopicattractaconsiderableattentionaccordingtothestageofbusinesscyclea ndbanks’specificcharacteristics(Le,2016; Athanasoglouetal.,2008; Demirgu¨c¸-Kunt,&Huizinga,1999;Cucinelli,2015;Hou&Dickinson,2007) 2.1 Theeffectsofnon-performingloansonbankprofitability
Doesa higherlevelo f n o n
-p e r f o r m i n g loansrefertoa lower-profitabilityf o r banks?
Ther e l a t i o n s h i p betweenNPLsandprofitabilityisoneofcentraltopicsinbankin gstudies,becauseofthe
1 See
WorldBank& IMF(2014).Financialsectorassessmentprogram–Vietnam.June2014
Trang 32 See
Moody’sInvestorsService (2014).Vietnam bankingsystem outlook.February2014.
Trang 4potentialimplicationsforregulatorypolicies.Anumberofstudiesfoundthatfailin gbankstendstohavelowerefficiencyandhighratios ofproblemloans(Berger& Humphrey,1992;Wheelock& W i l s o n , 1994).Anumberofotherstudieshave foundnegativerelationshipsbetweenprofitabilityandproblemloansevenamong banksthatdonotfail(Kwan&Eisenbeis,1995;Hughes&Moon,1995;K a r i m , 2 010)
Inaddition,studiesonbankprofitabilityrecentlyhavetakenintoaccountassetq uality,specificallynon-performingloans.Athanasogloue t al
(2008)s ho ws thatthep o o r qualityo f loansr e d u c e s interestrevenue,thusNP Lshasnegativeeffectonbankprofitability.Anumberofresearchershavef o u n
d
thatnon-performingloansleadtolowerprofitabilityinthebankingsector(Altunbasetal.,2 000,Fan&Shaffer,2004;Girardoneetal.,2004).Thefindingssupportthehypot hesisthatthee ffi c i e n c y banksarebetteratmanagingtheircreditriskasprop osedbyBergerandDeYoung(1997).B a n k e r etal
(2010)findsthatoncetheimportanceofnon-performingloansisambiguous,banksf e a r thattheirlendingbehaviorwillhave disadvantage,ifNPLsincreaseexceedingexpectedlevels,thiswillnegativelyim pactonthebankprofitability
Usingapaneldatasetfor14Koreancommercialbanksovertheperiod1995-
2005,Bankeretal.(2010)findsthatthenon-performingloansratiohasanegativeimpactonbankproductivity.Marius(2011)st
udiestheEuropeanbankingsectorovertheperiod2004-2009andfindsthatthenegativer e l a t i o n s h i p betweenNPLsandtheproducti vity.Thismeanstheincrease
ofNPLsleadstodecreaseo f
ROAandROEstrongly.Trujillo-Ponce(2013)hasthesameresultsforevaluatingdeterminantsonproductivityofS paincommercialbanksfrom1999to2009.Byusingunbalancedpaneldataan dGMMmodeltoanalysisimpactofNPLsfor89bankswith697observations,thefin dingsshowthatNPLshavenegativeeffectonROAwithsignificancelevelof5percen tandROEwithsignificancelevelo f 1percent
ByevaluatingperformancethroughcontrolofriskfactorsandassetqualityofJ apanesec o m m e r c i a l banksintheperiod1993-1996,Altunbasetal
(2000)havefoundthatNPLsratioandperformancehavenegativerelationship, andaftercontrollingofriskfactors,bankstendtosufferareductioninoperatinge
findingisconsistentwiththestudiesbyHughesandMester(1993)thatconducted onbanksintheUS,andresearchofGirardoneetal
(2004).InVietnam,Pham(2013)evaluatestheimpactofNPLsontheprofitabilityof theVietnamesec o m m e r c i a l banksintheperiod 2005-2012.The resultsindicatethat NPLshasnegativelyimpactonprofitabilityratioofthebanks Theempiricalpapershavealsoprovidedconsiderableevidencetosupportthef
ollowinghypothesesrelatingtobank-specificcharacteristicsonprofitability,suchascapital,banksize,loangrowth,and
competition.Thestructure-conduct-performancehypothesisreferstotherelationshipbetweencapital,competitio n,andprofitability.Theresultsofsuchresearchshowthatoperatingperformance issignificantlyrelatedtomarketstructure.Marketstructure,whichreferstothedeg reeo f marketconcentrationwithinanindustry,representsthedegreeofcompeti tionwithinthespecific
Trang 5(1997)findthat,withinafi n a n c i a l systemcharacterizedbylesscompetition ,firmstendtohavelargerscalesofoperation,andthisinturnleadstoahigherdegre eofmarketconcentrationandprofits(Lee&Hsieh,2013;Hannan
&Berger,1991;Neumark&Sharpe,1992;Demirgüç-Kunt&Huizinga,
sizei s
showntoyieldapositiveeffectonprofitability(Demirgu¨c¸-Kunt&Huizinga,1999;Goddardetal.,2011)
2.2 Theeffectsofnon-performingloansonbanklendingbehavior
Non-performingloanshavebeenc o n c e r n e d aso n e o f them o s t i m p o r t a n t f actorscausingr e l u c t a n c e f o r thebankstoprovidecredit.I n a highN P L con dition,banksincreasinglytendtoimplementedinternalconsolidationtoimprove theassetqualityratherthandistributingcredit.Inaddition,thehighlevelofNPLsr equiresbankstoraiseprovisionforloanlossthatleadtodecreasethebanks’reven ueandreducesthefundsfornewlending(Hou&Dickinson,2007).Thefinanciala
c c e l e r a t o r
e ff e c t alsoreferstotheeffectsofNPLsonbanks’lendingbehavior.Thistheoryr elates
tob o r r o w e r s ’ equityposition(ornetworth)whichinfluencestheiraccesst ocredit.Thisalsoexplainsbanklendingbehavioranditsrelationshipwiththecyclic alfluctuationsintheeconomy.Anetwortho f afirmisimprovedandthegreateritis,t helowertheexternalfinancepremiumaslendersassumelessriskwhenlendingtoh ighnetworthagentsduringbusinessupturn.Anadverseshockthatlowersb o r r o
w e r s ’ currentc a s h flowsl e a d s toa declinei n theirn e t wortha n d raisesext ernalfi n a n c e premium.Theincreaseinborrowers’costoffinancingwilldisco uragetheirdesirestoundertakem o r e investmentprojectsandconsequentlya ffectthedemandforcredit,andamplifyingtheeffecto f theinitialshocks(Bernan keetal.,1994;Kiyotaki&Moore,1995;Le2016)
Theempiricalstudiesontherelationshipbetweenloangrowthandbankrisk, especiallycreditlossesroundupatmacroeconomiclevelinseveralstrandsoft heliterature(Keeton,1999;Borioetal.,2002),butstillneedmorestudieswhichfoc usontherelationshipbetweenNPLsandbanklendingbehavior.Basedonasampleo fpubliclistedbanksinChina,Luetal.(2005)
discusstherelationshipbetweenbanks’lendingbehaviorandNPLs.Thefindingsi ndicatesthatthebankingsectorpresentsa biasinChina,asbanksaremorelikelyt
olendtostate-ownedfirms,eventhoughthesecanpresenta highcreditrisk.Borioetal
(2002)showsthatproblemloansincreaseasaresultoffirms’ andh o u s e h o l
d s ’ financialdistressforSpanishbanksduringrecession.Thisresearchalsoim pliesbanklendingisstronglyprocyclical,andthatinperiodsofexpansion,banksa remorelikelytolendcredittofirmswithlowcreditquality.Thisleadstofutureprobl emsanddefault,typicallyduringdownturns,withanestimatedtimelagofapproxi matelythreeyears.Tomak(2013)investigatesthedeterminantso f banklending behavioro n a sampleo f T u r k i s h banks,andfi n d s a significantr e l a t i o n s h
i p betweenNPLandbanklendingbehaviorinStateownedbanksandNPLshowane gativei m p a c t onthe growthoftotalloans
Foosetal
(2010)analyzetheeffectofloangrowthontheNPLsofindividualbanks.Theyfindth
Trang 6atloangrowthhasanegativeimpactontherisk-adjustedinterestincome,whichsuggeststhatloan
Trang 7(2013)examinetherelationshipbetweenabnormalloangrowthandbankrisk-takingbehavior.Theirfindingsshowthatabnormalc r e d i t growtho v e r a prol ongedp e r i o d o f timewouldl e a d toa n i n c r e a s e i n b a n k s ’ riskiness,ac companiedbyareductioninsolvencyandanincreaseintheratioofNPLs.Severalstu diesfindthatexcessivecreditgrowthcanleadtothedevelopmentofassetpricebu bbles.Borioetal
(2002)andB o r i o andDrehmann(2009)indicatethatexcessivecreditgrowthi sthemainfactorofafinancialc r i s i s incaseswhereitappearsthattheflowof loansremainshighfortheremainderoftheyear
Insummary,mostoftheevidencesuggeststhatbanks’riskappetiteiscompr
omisedbyexperiencesrelatedtonon-performingloans.AnincreaseinNPLisexpectedtoleadtoareductioni n banks’c reditlines,hencethenegativerelationshipbetweenNPLandloangrowthrate
3 Methodology
Thispaperappliesthetwo-stepdynamicpaneldataapproachsuggestedbyArellanoandBover(1995)andBl undellandB o n d (2000)andalsousesdynamicpanelGMMtechniquetoaddressp otentialendogeneity,heteroskedasticity,andautocorrelationproblemsi n thed ata(Doytch&Uctum,2011).Thedynamicpaneldatamodelprovidesf o r a m o r e
flexiblevariance-covariancestructureu n d e r themomentconditions.TheGMMapproachi s bette rthant r a d i t i o n a l OLSi n examiningfinancial variablemovements.Forinst ance,Driffill etal
(1998)indicatethataconventionalOLSanalysisoftheactualchangeintheshort rateontherelevantlaggedtermspreadyieldscoefficientswithsomewrongsigns
andwrongsize.TheresearchalsofollowsWindmeijer’s(2005)finite-
samplecorrectiontoreportstandarderrorsofthetwo-stepestimation,withoutwhichthosestandarderrorstendtobeseverelydownwar dbiased
Thestudyadoptsthed y n a m i c paneldataapproachandGMMtoe s t i m a t e t heparameters.Althoughthereisc orre la tion orheteroskedasticityamongthe equations,theestimatedstandarddeviationstillappearstoberobust.Therefore, theindependentvariablew i t h laggedperiodsi s i n c l u d e d i n E q s
(1)and(2),asshownbelow.B e y o n d thedynamicpaneldata,themodelthatesta blishestheimpactofNPLsonprofitabilityandlendingbehaviorisbasedontheearlie rliterature.Accordingtotheearlierliteraturediscussionandthisstudy’purposeofre search,theauthormodifiestheequationsofLe(2016),Altunbasetal
(2007),CasuandGirardone(2006),andGoddardetal
(2004)toestablishtherelationshipbetweenNPLsandprofitabilityandlendingbe havior.Theser e l a t i o n s h i p s canbespecifiedasfollows:
𝜋��=�2𝜋��−1+�2���+ 𝜆2�����+ 𝜋2���+ 𝜀2,�� (1) 𝑅
�� ��=�4�� �𝐴 ��−1+�4���+𝜆4�����+𝜋4���+𝜀4,��
(2)
Here,tandidenotetimeperiodandbanks,respectively,𝜀1,2,3,4,��=𝜂�+���and𝜂�
�isanunobservedbank-specificeffect,���istheidiosyncraticerrorterm.
Trang 8(1)and(2)aredesignedtoexaminetheimpactofNPLsonbankprofitabilityandbankl ending
behavior,respectively.Term�����is
theratioofnon-performingloansovergrossloan;𝜋��r e f e r s
totheithbank’sprofitabilityinyeart,proxiedbyreturnonassets(ROA).Here,��𝑅��
r e f e r stothe
ithbank’slendingbehaviorinyeart,proxiedbythepercentagedifferenceintotal grossloan.The
vectorofexplanatoryvariablesincludesbank-specificvariables(F),includedthecapitalproxiedbytheratioofequityontotalas sets,thesolvencypresentedbytheratioofloantodeposit,degreeofb a n k i n g c
ompetition(Fu&Heffernan2009),thedegree’sproxyCR4(thefour-
bankconcentrationratio),theHHI(Herfindahl-Hirschmanindex),bankownershipproxiedbydummyvariable,andm a c r o
e c o n o m i c factor(M).I t i s c r u c i a l toconsiderthepersistenceo f profitabili tythrought h e dynamicpanelmodelbecausebanksarealwaysaccompaniedbyt hefeatureofprofitabilitypersistence(Leeetal.,2013).Previousresearchesshowth
atbank-specificcharacteristicvariablesarelikelytobepotentiallyendogenous(Athanaso glouetal.,2008)andsomeotherindependentvariablesarenotstrictlyexogenous ByusingGMMestimation,itallowsforinstrumentingoftheendogenousvariables andprovidesconsistentestimates.Thepaperusesthelagsofrighthandsidevaria
blesintheequationsasinstruments.Thetwo-stepestimationisusedbecauseitisasymptoticallymoree ffi c i e n t than
theone-stepestimationforthepresenceofheteroskedasticityandserialcorrelation(Bl
undell&Bond,1998).Inthisestimation,theHansenJ-
testisusedtotestthevalidityofinstrumentsetsandtheArellano-
Bondtestisappliedtochecktheabsenceofsecond-orderserialcorrelationinthefirstdifferencedresiduals
Asf o r therelatedinternalc o n t r o l variables,accordingtoCasuandGirardone (2006),S h o r t (1979),LeeandHsieh(2013),andLe(2016),theyincludeequitytot otalassets(ETA),loantodeposit(LTD),loangrowth(LGR),totalassets(TA),theco mpetitionratiossuchasHHI,CR4.Thecoefficientso f ETA,TA,LDR,CR4,andHHIar eexpectedtobepositivewithprofitabilityandlendingbehavior.A highervalueof concentrationr e f e r s tolesscompetition.Thus,banksenjoya higherm a r k e t a dvantage,suchaseconomiesofscaleorscope,withtheresultofgreaterprofits.T herefore,theα1c o e ffi c i e n t shouldbepositive.Onthecontrary,NPLsisexpe ctedtobenegativewithprofitabilityandlendingbehavior
Twomacrocontrolvariablesaresetastherelatedexternalcontrolvariables:infl ation(INF),GDPgrowthrate(GDP).ThecoefficientsofINFandprofitabilityandlen dingbehaviorisexpectedtobenegativebecausebanksmaychargecustomersm
oreinhigh-inflationcountries,yetatthesametimetheyfacedueloansthatareshrinking.Ahig hergrowtheconomymayimplythatbankscangeneratem o r e profitability
lendingbehaviorareexpectedtobepositive
Trang 9Classification Variable Descriptions
Expect eds i
g n
Expectedsig n
Independent ROA Netincomeaftertaxtoaverageassets +
+ customers
Non-performingloantogrossloanETA Theratioofequityontotalassets +- +
-LDR Ratiobetweenloantocustomerdeposit + +
HHI theconcentrationofaspecificindustry
HHI=∑nMS2where
Sj
denotesthemarketshareofthejthbankusing totalassetsasaproxyformarketshare CR4 theshareoftheloanmarketcontrolledbythe
fourlargestbanks,C R 4 =∑ 4 MS + +
j=1 j OWN
Thecontrolleveloftheownershipdeno te3dummiesOWN1showsthepercent ageofbank
ownershipofanindividualororganizationof relevant
-change)
480|
PoliciesandSustainableEconomicDevelo
Table1
Summaryofexplanatoryvariables
variables LGR Percentagechangeingrossloanprovidedto
varia
ble INF Inflation,averageconsumerprice(percentage
4 Datadescription
Thisstudyanalyzesapaneldatasetcomprising34Vietnamesecommercialban
ksovertheperiod2005-2015.Thepaneldatasetisextractedfromnon-consolidatedincomestatementsandbalancesheetsofthesebanks,anditconsis tsof 357observations.Themacroeconomicdatacomefrom IMF
-IFSwebsite.SampleofVietnamesebanksincludesAnBinhCommercialbank, AsiaCommercialB a n k , VietnamBankforAgricultureandRuralDevelopment,B ankforInvestmentandDevelopmento f Vietnam,VietCapitalCommercialJointSt ockBank,VietnamBankforIndustryandTrade,EasternAsiaCommercialJointStoc kBank,VietnamExportImportCommercialJointStockBank,HousingDevelop mentCommercialJointStockBank,KienLongCommercialJointStockBank,Lien VietPostCommercialJointStockBank,MilitaryCommercialJointStockBank,M ekongDevelopmentJointStockCommercialB a n k , MekongHousingCommerci alB a n k ,MaritimeCommercialJ o i n t StockB a n k , SouthernCommercialJoin tStockBank,BACACommercialJointStockBank,Orie nt CommercialJointSto ckBank,OCEANCommercialJointStockBank,PetrolimexGroupCommercialJ o i n
Trang 10PoliciesandSustainableEconomicDevelop ment |10
t StockBank,VietNamPublicBank,SouthernCommercialJointStockBank,SaiGonJ ointStock