SSRN：金融科技與主權財富基金-How Disruptive is Financial Technology

HowDisruptiveisFinancialTechnology?

DouglasCummingt,HishamFarag?,SantoshKoirala§andDannyMcGowan*

Abstract

WestudywhetherFintechdisruptsthebankingsectorbyintensifyingcompetitionforscarcedepositsfundsandraisingdepositrates.Usingdifference-in-differenceestimationaroundtheexogenousremovalofmarketplaceplatforminvestingrestrictionsbyUSstates,weshowthecostofdepositsincreasebyapproximately13%withinsmallfinancialinstitutions.However,thesepricechangesareeffectiveinpreventingadrainofliquidity.SizeandgeographicaldiversificationthroughbranchnetworkscanmitigatetheeffectsofFintechcompetitionbysourcingdepositsfromlesscompetitivemarkets.ThefindingshighlighttheunintendedconsequencesofthegrowingFintechsectoronbanks.

19August2024

JELCodes:D26,G21,G23

Keywords:fintech,banking,deposits

WearegratefulforhelpfulcommentsandsuggestionsreceivedfromAlinAndries,PiotrDanisewicz,ElenaLoutskina,HuyenNguyen,EnricoOnali,TomaszPiskorski,ArisyFarizaRaz,KlausSchaeck,ChendiZhang,MandyZhangandseminarandconferenceparticipantsattheFinancialManagementAssociationAnnualConference,CardiffUniversity,DublinCityUniversity,KathmanduUniversitySchoolofManagement,LundUniversity,SwanseaUniversity,UniversityofExeter,UniversityofNottingham,andtheUniversityofSydney.Noneoftheauthorshaveaconflictofinterestorfinancialandpersonalrelationshipswithotherpeopleororganizationsthatcouldinappropriatelyinfluence(bias)theirwork.

tEmail:

cummingd@.

FloridaAtlanticUniversityandUniversityofBirmingham.?Email:

h.farag@bham.ac.uk.

UniversityofBirmingham.

§Email:

s.koirala@bham.ac.uk

.UniversityofBirmingham.

*Correspondingauthor:DurhamUniversityBusinessSchool,MillHillLane,Durham,DH13LB.Email:

danny.mcgowan@durham.ac.uk.

DurhamUniversity.

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

HowDisruptiveisFinancialTechnology?

Abstract

WestudywhetherFintechdisruptsthebankingsectorbyintensifyingcompetitionforscarcedepositsfundsandraisingdepositrates.Usingdifference-in-differenceestimationaroundtheexogenousremovalofmarketplaceplatforminvestingrestrictionsbyUSstates,weshowthecostofdepositsincreasebyapproximately13%withinsmallfinancialinstitutions.However,thesepricechangesareeffectiveinpreventingadrainofliquidity.SizeandgeographicaldiversificationthroughbranchnetworkscanmitigatetheeffectsofFintechcompetitionbysourcingdepositsfromlesscompetitivemarkets.ThefindingshighlighttheunintendedconsequencesofthegrowingFintechsectoronbanks.

19August2024

JELCodes:D26,G21,G23

Keywords:fintech,banking,deposits

1

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

1.Introduction

Overthepastdecade,newfinancialtechnologies(Fintech)haverevolutionizedcredit,insurance,andpaymentsmarkets.Theseentrantshavecapturedmarketsharefromtraditionalincumbentsandactasanewsourceofcompetitionforincumbents.WhileFintechpromisestoincreaseefficiencyindeliveringfinancialservices,andwidentheirrange,thesetechnologiesalsoposeapotentialthreattofinancialinstitutionsthatwarrantscrutiny.Forexample,extantresearchsuggeststhatwheretheFintechsectorincreasesbanks’fundingcoststhismayleadtocontractionsincreditsupply(LoutskinaandStrahan(2009),Duquerroyetal.(2021)),andpromptgreaterrisktakingasfinancialinstitutionsreachforyield(MatutesandVives(1996),BoydandDeNicolo(2005),Dell’AricciaandMarquez(2006),Razetal.(2022)),anddiversifyintobusinesslinesoutsidetraditionalretailbankingactivities.HowlargeisthedisruptiveeffectofFintechonbanks?Throughwhichchannelsdoesitarise?Thesequestionsareoffirst-orderimportancebutremainlargelyunanswered.

Motivatedbytheseconcerns,weevaluatehowthecostofbankdepositsrespondtotheentryofmarketplacelendingplatforms,whicharenowubiquitouscreditprovidersthroughoutdevelopedanddevelopingcountries.WeconjecturethatthisformofFintechratchetscompetitionandprovokesahigherequilibriumdepositinterestrateasbanksmustcompetewithplatformsforfundingandattempttopreventadrainofliquidity(Lietal.(2019),McGowanetal.(2024)).Furthermore,smallfinancialinstitutionsarelikelytobemorestronglyaffectedduetotheirgreaterrelianceondepositstofinancetheiractivities.

LocalUSbankingmarketsareanidealsettinginwhichtostudyFintech’sunintendedconsequences.UndertheSecuritiesActof1933andtheSecuritiesExchangeActof1934,USstatesecuritiesregulatorshaveauthoritytodeterminewhethermarketplacelendersmaysolicitfundsfromtheircitizensandbusinessesheadquarteredinthestateonacase-by-casebasis.Regulatorsimposemarketplaceinvestingrestrictionsduetoconcernsthatborrowers’loanapplicationsmaycontainfraudulentinformationthatposesarisktoinvestors.Obtainingregulatoryapprovaltosourcefundsfromin-stateinvestorsrequiresthatamarketplacemeetsthedemandsofastateregulator’s‘meritreview’processby

2

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

demonstratingthatitsdataprotectionandverificationmeasuresprotectinvestorsfromfraudulentclaimsinmarketplaceborrowers’creditapplications(ChaffeeandRapp(2012)).Theremovalofmarketplaceinvestingrestrictionsisduetoregulators’concernsaboutprotectinginvestorsfromfraudandlossesandareplausiblyexogenouswithrespecttobanks’depositcosts,andconditionswithinthebankingindustrymoregenerally(ChaffeeandRapp(2012)).Additionally,thecomprehensiveclassificationofbanksintolargeandsmallfinancialinstitutionsbytheFederalDepositInsuranceCorporationhelpsusneatlyidentifysmallbanksforwhomfintechcompetitionislikelytobeacompetitivethreattodepositsourcing

.1

Ourempiricalanalysisexploitstheentryofmarketplacelendersacrossstatesandtimefollowingtheremovalofentrybarriers.Usingdifference-in-differenceestimationappliedtobankbranch-leveldepositratedata,wefindrobustevidencethatallowingplatformstosolicitfundswithinthestateleadstoa12.85%increaseinthecostofdeposits.Withintheuniverseofsmallbanks,relativelylargerinstitutionsthataremorereliantonwholesalefunding,areaffectedtoalowerdegree.Similarly,thecostofdepositsincreasesrelativelymoreamongbanksthatoperatealimitednumberofbranches,consistentwithbranchnetworksmitigatingcompetitionforfundsbysourcingdepositsfromregionswheremarketplacesdonotoperate.Moregranulartestsrevealthatdepositratesincreaseacrossdepositproducts,butthattheeconomicmagnitudeislargestforcertificatesofdepositsandmoneymarketaccounts.FurtheranalysesrevealsthattheremovalofFintechbarriersdonotleadtocontractionsinthesupplyofbankdeposits.Hence,whileFintechintensifiescompetitionforfunding,settinghigherinterestratesstemsdepositoutflows

.2

Ourresearchdesignexploitsthepanelstructureofthebranch-leveldatatoensurethatchangesindepositcostsandquantitiesarenotdrivenbyconfoundingforces.Specifically,weincludebank-quarter-yearfixedeffectsintheestimatingequations.WethusidentifyFintech’seffectsthroughcomparisonsbetweenbranchesownedbythesamebankatthesamepointintime.Inessence,wecomparehowdepositcostsevolvebetweenabranchinastatethatremovesmarketplaceinvestingrestrictionsversusabranchinastatethatdoes

1FDICpublishesquarterlylistofbigbanksthathaveassetsizeover$300millionUSD(

/releases/lbr/

)

2Ourresultsarepotentiallyexternallyvalidasthebusinessmodulesofthemarketplaceplatformsoperatingduringthesampleperiodresemblethoseinothercountries.

3

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

not,wherethebranchesbelongtothesamebank.Thisapproachrulesoutalltimevaryingforcesatthebanklevelaswellasmacroeconomicfundamentalsthathavebeenfoundtoinfluencedepositdemandelsewhereintheliterature(SaundersandSchumacher(2000)).

Aseriesofrobustnesstestsruleoutconfounds.Diagnosticchecksshownopre-emptiveanticipatorytrendsinthecostofdepositspriortotheremovalofmarketplaceinvestingrestrictions,theparalleltrendsidentifyingassumptionholds,andthetreatedandcontrolunitsarecomparablealongobservabledimensions.Placebotestsindicatethecostofdepositsdoesnotsimultaneouslyincreaseamongbanksinstatescontiguoustothosethatremovemarketplaceinvestingrestrictions.Thismakesitunlikelythefindingsreflectconfoundingobservableorunobservableomittedvariablessincebanksinneighboringstatesoperateinsimilarenvironments.Inaddition,shockstobanksoundness,monitoringbycreditors(Danisewiczetal.(2018,2021)),regulatorymonitoring(Agarwaletal.(2014)),equitycrowdfunding,creditrisk(McGowanandNguyen(2023)),andchangestocompetitionandmarketpowerwithinthebankingindustry(FocarelliandPanetta(2004),Bergeretal.(2020),Duqietal.(2021),McGowanetal.(2024))donotdrivetheinferences.Furthertestsshowthederegulationofcrowdfundingrestrictionsdonotconfoundtheresults,whilemethodologicalchecksdemonstratethatstaggeredtreatmentsdonotexplainthefindings(SunandAbraham(2021)).

Ourpapercontributestoarapidlyevolvingbodyofresearchonnewfinancialtechnologies.WeprovidenovelevidenceonhowthederegulationofFintechmarketsspillsovertoanddisruptsthebankingsector,particularlycommunityandsmallbanks.Sofar,studiesinthisareahavesoughttounderstandwhethermarketplacelendingaffectsthecreditsupplybehaviorofcommercialbanks,andwhethertheseplatformscomplementorsubstitutebanklending.Cornaggiaetal.(2018)finda1.2%decreaseinthevolumeofunsecuredpersonalbankloansfollowingaonestandarddeviationincreaseinmarketplacelending.Tang(2018)presentsevidencethatpeer-to-peerlendersencroachontobanks’marketsharebutcomplementbanklendingamongsmallloans.OtherworkinthisliteratureshowshowFintechlendersdisrupttheshadowbankingmarket(Buchaketal.(2018)).

4

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

AnotherbodyofresearchexamineswhetherFintechlendershavelowerfinancialintermediationcostsrelativetobanks(Philippon(2015),Buchaketal.(2018)),whileBartlettetal.(2021)reportthatFintechlenderschargeminoritiessystematicallyhigherinterestratesonmortgageseligibleforsaletotheGovernmentSponsoredEnterprises.Fusteretal.(2019)documentthegrowthinFintechlenders’marketshareinthemortgagemarketattheexpenseofbanksandnon-banksandthatFintechsprocessmortgageapplications20%fasterthanotherlenders.Cummingetal.(2022)andDanisewiczandElard(2023)provideevidenceontherealeffectsofmarketplaces.Incontrasttothesestudies,wedocumentanunintendedconsequenceoftheFintechsectoronbanks’depositcosts.Wealsoextendtheliteratureonthebanking-marketconsequencesoftechnologicalinnovationsthatfocusonSWITFandpre-Fintechtechnologies(Scottetal.(2017)).

Ourevidencemattersforpolicymakers.Asthenascentmarketplacelendingindustrycontinuestoexpandrapidly,andhasbecomeamajorsourceofcreditforbothhouseholdsandfirmsbothintheUSandinternationally(Claessensetal.(2018),Cummingetal.(2022)),theeffectswedetectmaybecomelargerinfuture.Recentprojectionsforecastmarketplacelendingwilloriginateover$1trillionofcreditby2023.Marketplaces’disruptiveeffectsmayalsowarrantregulatoryscrutinywheretheydestabilizebanks’operationsandtriggerchangesinfundingcoststhatarerelevantfrombothmacroprudentialandmonetarypolicyperspectives(Claessensetal.(2018),Thakor(2020)).

Thepaperisstructuredasfollows.Section2providesanoverviewofthedataset.Weprovidedetailsoftheregulatoryenvironmentsurroundingmarketplaceinvestments,andthelegalbackgroundtomarketplaceinvestingrestrictionsinSection3.WeoutlinetheidentificationstrategyinSection4,andpresenteconometricresultsinSection5.Section6dealswithalternativeexplanationsandrobustnesstests.Finally,wedrawconclusionsinSection7.

2.DataDescription

Theeconometricanalysisreliesonbranch-leveldatafromtwosources.TheFDICSummaryofDepositsdatabasereportsannualinformationonthegeographicallocation

5

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

ofeachbankbranchthroughouttheUS.Thisallowsustoobservethequantityofdepositsheldbybranchbbelongingtobankilocatedinstatesduringyeart.

DepositcostinformationistakenfromR.Thissourceprovidesweeklydepositandloanratesforeachdepositandloanproductthatabranchoffers.Rreportsdepositratesbasedonthefundingrate(FR)andannualpercentageyield(APY).ThemeasuresprovidestronglysimilarvaluesasshowninTable1.Usingthisinformation,wecalculatethedepositrate(FR)andannualizedpercentageyield(APY)whichmeasuretheaveragequarterlydepositrateacrossalldepositproducts,foreachbranchover2004Q1to2019Q4

.3

Usingthegranularproduct-levelinformation,wealsocalculatethequarterlydepositratepaidon,interestchecking(IC),regularsavings(SAV),money-market(MM)and12-monthcertificatesofdeposits(CD)accountstoprovidedetailedinsightsintosomeofthemostimportantdepositproductsbanksoffer.

Fintechcompetitionisunlikelytohavesubstantialimplicationsonlargebanks’depositbaseduetotheirscale,geographicalreach,andaccesstowholesalefunding.Rather,itissmallfinancialinstitutionsthataremostlikelytoexperiencemoreintensedepositcompetitionasdepositorsreallocatefundstofintechplatforms.Ourteststhusfocusonsmallbanks.WeusethelistoflargebankspublishedbytheFDIConquarterlybasis,andremoveanylargebankstoobtainthesample

.4

Wemergeadditionaldatatakenfromseveralsources.Weretrievequarterlybank-leveldatafromtheFederalFinancialInstitutionsExaminationCouncil031ConditionandIncomeReports(callreport)database.Thisprovidesinformationfrom2004Q1to2019Q4onseveralbankvariablesincludingbanksize(totalassets),returnonassets(ROA),totalliabilities,anddepositliabilities.Tocapturelocalbusinesscyclesanddemand-sidedeterminantsofdepositcosts,weusethestate-levelpercapitaincomegrowthrate(BureauofEconomicAnalysis),populationgrowthrate(BureauofEconomicAnalysis),unemploymentrate(BureauofLaborStatistics),andthenumberofbusiness

3Weaggregatethemonthlydatatothequarterlylevelbecausewemergeinbankleveldatathatisavailableat

quarterlyintervals.Ourchoicetobeginthesamplein2004Q1ismotivatedbythefactthatProsperandLendingClubwereincorporatedin2005and2006,respectively.Settingthestartingpointat2004Q1thereforeprovidessufficienttimetotesttheparalleltrendsassumption.

4Asarobustnesstest,wealsorunthebaselinemodelsremovingthetop100banksbyassetsizeratherthanusingtheFDIC’slist.Theresultsaresimilar,albeittheeffectsizesaresmaller.

6

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

establishmentspercapita(CountyBusinessPatterns).Table1providesadefinitionofeachvariableinthedataset.Table2reportssummarystatistics.

[InsertTable1:VariableDescription][InsertTable2:DescriptiveStatistics]

Aswedetailbelow,investingrestrictionsonLendingClubandProsperwereremovedatdifferenttimesbyeachstate.Wecontactedbothplatformsandeachstatesecuritiesregulatortoverifythedatewheninvestingrestrictionswereremoved.Usingthisinformation,weconstructthevariableDeregulationindexstwhichisacountofhowmanyplatforminvestingrestrictionshavebeenremovedinstatesinquartert.

3.InstitutionalBackground

LendingClubandProsperarethemostprominentmarketplacelendersintheUSandoperatesimilarbusinessmodels.Prospectiveborrowersregisterwithaplatformandcompleteanonlineloanapplication.Usingdigitalscreeningalgorithms,theplatformsassigneachapplicationacreditriskratingthatdetermineswhethertheloanislistedonthemarketplaceforfunding.Duringtheapplicationprocessplatformsscreentheborrower’scredithistory,outstandingdebt,income,employmentstatus,andotherfactors.Applicants’riskratingdeterminestheinterestrateaborrowerpays

.5

Investorsdonotmakedirectloanstoborrowers,ratheranissuingbankissuestheloantotheborrowerandthensellstheloantotheplatform

.6

Theplatformthenissuesaseparatenotetotheinvestorwithareturnontheinvestmentcontingentontheborrowerrepayingtheoriginalloan(ChaffeeandRapp,2012).Platformsdonottakeastakeineachloan,rathertheychargeservicefeesfororiginatingeachloanandontradingnotesbetweeninvestorsinthesecondarymarket.

Mostborrowerapplicationsareunsecuredconsumerloans.Theseareprimarilyusedtoconsolidateexistingdebts,althoughasubstantialshareofloansareusedforhomerepairsandtofinancepersonalorfamilypurchases.Whilebusinessloansareincreasinglycommon,theyremainaminority.Theinterestrateonmarketplaceloansrangesbetween

5Before2010Prosperoperatedanauctionforeachloanwherebyinvestorswouldsubmitbids(aninterestrate)foreachloan.Thelowestbidderswouldwintheauctionandfundsfromthosebidderswerepooledtoextendloans.From2010Prospershiftedtoamodellikethatdescribedabove.

6LendingClubandProsperhavebothusedWebBankastheissuingbank.

7

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

6.46%and29%onLendingCluband6.95%and35.99%onProsper.Loanamountsrangebetween$1,000and$40,000andthetermstructurevariesbetween12and60months.

3.1StateMarketplaceInvestingLaw

Thenotesthatareoffered,sold,andpurchasedinthemarketplacelendingmodelconstitutesecuritiesandareregulatedbytheSecuritiesActof1933andtheSecuritiesExchangeActof1934

.7

TheActsmandatethatsecuritiesareregisteredeitherwithafederalorstateregulator.Section18(b)oftheSecuritiesActof1933stipulatesthatsecuritiesthatmaybelistedandtradeonanationalmarketsystem(aregisteredexchange)areexemptfromstate-levelregistrationandmaybefederallyregistered.Asmarketplaces’notesarenotlistedortradedonanationalmarketsystem,theplatformsmustsecureapprovalfromstatesecuritiesregulatorstosolicitfundsfrominvestorsineachstate(Cornaggiaetal.,2018).

Manystatesecuritiesregulatorsmandatesecurityregistrantsmeettherequirementsofa‘meritreview’

.8

Thisrequiresthestatesecuritiesregulatorfindthat,“thebusinessoftheissuerisnotfraudulentlyconducted…thattheplanofissuanceandsaleofthesecurities…wouldnotdefraudordeceive”(ChaffeeandRapp,2012)

.9

Informationprovidedbyborrowersinloanapplicationsmaybeinaccurate,missing,ordeliberatelymisleading.Forexample,theymaymisstatetheirincome,currentemploymentstatus,oremploymenthistory.Whereamarketplaceisunabletoverifytheinformationinborrowers’loanapplications,theregulatorrulesitisunabletoconcludethebusinessisnotfraudulentlyconductedasrequiredbystatelaw.Inthesecases,theplatformisdeniedtheopportunitytoregistersecuritiesbythestateregulatorandisprohibitedfromsolicitingfundsfrominvestorswithinthestate.Marketplacesareonlygrantedapprovaltosolicitfundsinameritreviewstateoncethestatesecuritiesregulatorisconvincedthe

7Section2(a)1oftheSecuritiesActof1933andsection3(a)10oftheSecuritiesExchangeActof1934providethedefinitionofa‘security’.Bothsectionsincludewithinthedefinitionofasecuritytheterms‘investmentcontracts’and‘notes’.P2Ploansfallunderthisumbrella.

8ThestatesareAlabama,Arizona,Arkansas,Indiana,Iowa,Kansas,Kentucky,Maine,Maryland,Massachusetts,Michigan,Nebraska,NorthCarolina,NorthDakota,Ohio,Oklahoma,Tennessee,Texas,Pennsylvania,Vermont,Virginia,andWestVirginia.

9Ohioisarepresentativeexampleofthelawinmeritreviewstates(ChaffeeandRapp,2012).See,Section1701.09oftheOhioRevisedCodeandAmendmentsforfurtherdetails.

8

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

platformhasimplementedproceduresthatensureinvestorscannotbedefrauded(ChaffeeandRapp,2012).

[InsertTable3:TimingofDeregulationacrossStates]

Theremainingstatespermitmarketplacelendingwithoutrestrictions.Thisisbecausethesestates’securitieslawmirrorstheSecuritiesandExchangeCommission’sapproachtosecuritiesofferingswhichdoesnotinvolvemeritreviewbutsimplyrequiresdisclosure(GAO,2011)

.10

Asthesestateshistoricallyfollowedthisapproach,aftertheirestablishmentmarketplaceswereimmediatelygrantedapprovaltosolicitinvestmentfunds.Amongthesestates,sevenauthorizeinvestinginnotesbutonlyfor‘sophisticated’investorsthatmeetsuitabilityrequirements.Thisisthecaseforallsecurities,includingmarketplaceloans

.11

Inmostofthesestates,investingislimitedtoindividualswithanincomeofatleast$70,000andaminimumnetworthof$70,000.Californiaimposeslessstringentrequirements,andonlyforindividualswhoinvestmorethan10%oftheirwealthinnotes.Thereasonsstatesimposetheserestrictionsarethefinancialhealthofmarketplaceinvestors

.12

Table3providesanoverviewofthedateswheneachstatesecurityregulatorremovedinvestingrestrictionsforLendingClubandProsper.

4.ResearchDesign

4.1EmpiricalModel

Toisolatecausalinferences,weusedifference-in-differenceestimationthatexploitstime-varyingchangesinmarketplaceinvestingrestrictionsacrossUSstates.Wecomparethecross-timeevolutionofthedependentvariableinbranchesinstatesthatremovemarketplaceinvestingrestrictionsrelativetobranchesinstateswhereinvestingbarriersremaininplace.Weestimate

ybist=βFintechindexst+δxbist+φbst+εbist,(1)

10SeetheGovernmentAccountabilityOfficereportPerson-to-PersonLending:NewRegulatoryChallengesCouldEmergeastheIndustryGrows,supranote5.

/new.items/d11613.pdf.

11ThestateswithsuitabilityrequirementsareCalifornia,Idaho,Kentucky,NewHampshire,Oregon,Washington,andVirginia.

12Forexample,theKentuckyDepartmentofFinancialInstitutionsnotedLendingClub’sauditor’s“goingconcern”lettermentioneditsnegativeearnings.Thedepartmentopinedthatinvestmentinthesite“constitutesalevelofrisksuitableonlytoAccreditedInvestors”(ChaffeeandRapp,2012).

9

Thispreprintresearchpaperhasnotbeenpeerreviewed.Electroniccopyavailableat:

/abstract=4931232

whereybistisadependentvariable(e.g.depositcosts).Fintechindexstisacountvariableofmarketplacessolicitingfundsinastatethere.HighervaluesindicatemoreintenseFintechcompetitionasmultiplemarketplacesoperatewithinajurisdiction,therebyofferinggreateropportunitiesforinvestors.Fintechindexsttakesthevalue0ifneitherLendingClubnorProsperhavebeengrantedpermissiontosolicitfundsand1(2)ifone(two)oftheplatformshavebeengrantedpermissiontosolicitfundswithinthestate.xbistisavectorofcontrolvariables;φbstdenotesbank×quarter×yearfixedeffects;εbististheerrorterm.Owingtothemultilevelnatureofthepaneldata,wefollowVig(2013)andtwo-wayclusterthestandarderrorsbybankandquarter.

Ourreviewofthelegalliteratureshowsthestate-levelmarketplaceinvestingrestrictionsareduetoregulators’concernsaboutprotectinginvestorsfromfraud.Therestrictionsareunrelatedtothecostofbankdepositsandconditionswithinthebankingindustrymoregenerally.Changesininvestingrestrictionsaredrivenbyaplatformconvincingstatesecuritiesregulatorsthattheirproceduresaccuratelyverifyborrowers’applicationclaimsandensurethatinvestorsarenotexposedtofraud.Instatesthatdonothaveameritreviewprocess,theremovalofmarketplaceinvestingrestrictionsisduetofederalSECregulationsthatareunrelatedtothemarketplacelendingandbankingindustries.LendingClubandProsperarethereforeabletosolicitfundsfrominvestorsinthesestatesassoonastheplatformgoeslive.

Whilethereviewofthelegalliteraturesuggeststheremovalofinvestingrestrictionsareexogenouswithrespecttoouroutcomesofinterest,weconductempiricaldiagnosticchecksaswell.OnlineAppendixTable1.Areportsestimatesof

dst=βxst+φs+φt+εst,(2)

wheredstisadummyvariableequalto1inquartertifstatesremovesinvestingrestrictionsoneitherLendingCluborProsper;xstisavectorcontainingstate-levelvariables(population,theunemploymentrate,thecorporatetaxrate,themeandepositrateacrossallbankbranchesinthestate,themeanZ-scoreofallbanksoperatinginthestate,andmeanbanksize(measuredasthenaturallogarithmofassets));φsandφtarestatea