FinanceandEconomicsDiscussionSeries

FederalReserveBoard,Washington,D.C.

ISSN1936-2854(Print)ISSN2767-3898(Online)

NoNewsisBadNews:Monitoring,Risk,andStaleFinancial

PerformanceinCommercialRealEstate

SamuelK.HughesandJosephB.Nichols

2025-032

Pleasecitethispaperas:

Hughes,SamuelK.,andJosephB.Nichols(2025).“NoNewsisBadNews:Monitoring,Risk,andStaleFinancialPerformanceinCommercialRealEstate,”FinanceandEconomicsDiscussionSeries2025-032.Washington:BoardofGovernorsoftheFederalReserveSystem,

/10.17016/FEDS.2025.032

.

NOTE:StafworkingpapersintheFinanceandEconomicsDiscussionSeries(FEDS)arepreliminarymaterialscirculatedtostimulatediscussionandcriticalcomment.TheanalysisandconclusionssetfortharethoseoftheauthorsanddonotindicateconcurrencebyothermembersoftheresearchstafortheBoardofGovernors.ReferencesinpublicationstotheFinanceandEconomicsDiscussionSeries(otherthanacknowledgement)shouldbeclearedwiththeauthor(s)toprotectthetentativecharacterofthesepapers.

1

NoNewsisBadNews:Monitoring,Risk,andStaleFinancial

PerformanceinCommercialRealEstate*

SamuelK.HughesJosephB.Nichols

FederalReserveBoard

April2025

Abstract

Asfinancialintermediaries,bankshaveakeyroleinproducinginformationandmanagingtherisksondiverseloanportfolios.Animportantinputintothisprocessisongoingcollectionoffinancialperformancefromborrowers.Usingsupervisorydataoncommercialrealestateloans(CRE),thispaperstudiesrelationshipsbetweenthecontentandtimelinessofborrower-reportedperformance,internalbankriskratings,andsubsequentloanperformance.Banksheavilyrelyonborrowerreportingwhensettingriskratings,despitethefactthatborrowerswithstalefinancialsaremorelikelytodefault.Althoughbankscangenerallybeslowtoupdatetheirratingsasinformationbecomesmorestaleonaverage,wefindcausalevidencethattheydomonitormoreintensivelyinresponsetoloan,locationandportfoliorisks.

Keywords:bankmonitoring,riskmanagement,commercialrealestatemortgages,financialperformancereporting

JELClassification:G14,G21,G32,R33

*Email:Hughes,

samuel.k.hughes@

;Nichols,

joseph.b.nichols@

.WethankDarrenAiello,CooperHowes,MehdiBeyhaghi,RobertKurtzman,DavidGlancy,RavenMolloy,MarkPaddrik,TengWang,EmilyRoss-Johnston,JasonAlpert,andMarkCareyfortheirthoughtfulfeedback,aswellasattendeesofOfficeofFinancialResearchandFederalReserveBoardseminars,BocaFinanceandRealEstateConference,andAREUEA-ASSA2025.TheviewsexpressedinthispaperaresolelythoseoftheauthorsanddonotnecessarilyreflecttheopinionsoftheFederalReserveBoard,ortheFederalReserveSystem.Allerrorsandomissionsareourown.

2

1Introduction

Bankshaveakeyroleinthefinancialsystemasintermediaries.Theirroleincludesscreeningandsubsequentlymonitoringthecreditworthinessofborrowers.Monitoringisessentialforpromptlyidentifyingandmanagingrisksandvulnerabilitiesastheyemergewhentheunderlyingcollateralforloansisverysensitivetoshiftsineconomicconditions(Weil

2024

),asisthecasewithcommercialrealestate(CRE)loans.Monitoringispartlybasedonacquiringregularreportsoncollateralperformancefromborrowersandproducinginternalcreditriskassessments.

1

Butthemonitoringisnotfree,banksfacebothdirectcostsintermsoflaborhoursforloanofficersandriskmanagersandindirectcosts,includingimposinghigherreportingrequirementsfromborrowerswhichcouldaffectfuturebank-clientrelationships.Studyinghowbanksoptimizetheirmonitoringgiventhesecostsisessentialtounderstandinghowsystemicallyimportantbankslearnaboutandmanagerisksduringtimesofrisingfinancialstress.

Thispaperaimstostudyborrowerreportingoffinancialperformanceandhowthatreportingisusedinbanks’ongoingmonitoringofCREloans.Thepaperhasthreemainfindings:first,banksheavilyrelyonborrowers’performancereportingintheirriskassessments;second,banksrespondrelativelyslowlytostaleordelayedreportingdespiteapositivecorrelationbetweenstaleperformanceanddefaultrisk;andthird,banksadjusttheirinformationacquisitioninresponsetoloan-orportfolio-levelshocks,demandingmoretimelyreportswhenthebanksaremoreconcernedaboutlossesonCREloans.Thispaperusesaloan-leveldatasetofcommercialrealestatemortgagesfromlargeU.S.banksthattrackstheperformanceoftheloans,property-levelfinancials,andthebanks’estimateoftheloans’probabilityofdefault.

Themainfindingsareconsistentwiththeoriesoffinancialintermediationinwhichbanks’compar-ativeadvantageoveratomizedinvestorsistheirabilitytomonitorloanperformance(e.g.,Diamond

1984

;BoydandPrescott

1986

).Largebanksmaintainsignificantresearch,servicing,andriskman-agementdepartmentsconsistentwiththeory,allofwhichrepresentsadirectcostofmonitoring.Givenbanks’growinganalyticalcapacityandaccesstoinformation,itissomewhatsurprisinghowmuchbanksstillseemtorelyonthetraditionalmonitoringprocessofborrowersreportingtheirperformancedirectlytothebank.Thesereportingrequirementsrepresentanindirectcosttothebank,asmoreonerousreportingrequirementsmightencourageborrowerstoconsiderlenderswithmorerelaxedmonitoring.Theinformationbanksreceiveaboutperformanceappearstobevaluabletothemwhenassessingdefaultrisk,buthowdotheybalancethisbenefitagainstboththedirectandindirectcostsofproducingit?Empiricalresultsonadjustmentsinmonitoringandinformationproductionprovidesupportfortheviewthatacquiringmoreinformationiscostly.Slowadjustmentsinbankriskratingsmaybeefficientgiventheneedtotradeoffcostsandbenefitsofmonitoringonthemargin.Banksadjusttheiracquisitionofinformationasriskexposureschange,allowingthemtohedgeagainstlossesorbuildupreservesinresponsetoeconomicorfinancialshocks.

Thefirstfindingdocumentedinthispaperisthatwhenborrowersreportanupdatetotheirnet

1Riskassessmentsorriskratingscanhaveregulatoryimplicationsforrisk-basedcapital(affectingBaselIIriskweightsoncommercialrealestatebasedonloan-to-valueratiosorobligorriskweights)andloanlossreservesorprovisioning(asofDecember2019,somebanksarerequiredtomaintainallowancesforlossesbasedonCurrentExpectedCreditLosses,aformofratingtheprobabilityofdefaultattheloanlevel).

3

operatingincome(NOI)oroccupancyrate,banksare50%morelikelytochangetheirestimateoftheloans’probabilityofdefault(PD).ThiseffectoninternalPDs,whichmapontobanks’internalriskrating,isshowningeneralizedeventstudiesaroundthequarterthataperformanceupdateshowsupinthebank’sinternalsystem.Oncethebankreceivesaperformanceupdate,thesizeofthechangeinNOIoroccupancyflowsthroughtoupgradesanddowngradesinriskassessments.A10percentdeclineinNOIincreasesthePDby6.9basispoints(anda10percentoccupancydeclineincreasesPDby16.7basispoints)onasampleaveragePDof1.1percent.

Thenextmainfindingisthatoutdatedfinancialreportingispositivelyrelatedtodefaultprob-ability,butnegativelycorrelatedwithbanks’internalcreditratings.Thepaperdocumentsthisbyconstructingaloan-levelmeasurereferredtoasthe“staleness”ofbanks’informationonperformance(alsoreferredtoas“outdated”informationorthe“age”ofreporteddata).

2

Stalenessisdefinedasthenumberofquartersbetweenthecurrentquarterandthedateattachedtothemostrecentper-formanceupdate.Figure

1

previewstheseresultsintwobinnedscatterplotsfortherelationshipbetweenthestalenessmeasureofNOIwithdefaultprobabilityintheleftpanelandtheprobabilityofadowngradeinbanks’internalestimateofPDintherightpanel.

3

Increasinglystaleinformationisstronglypositivelycorrelatedwithdefaultbuthaslittleimpactonthepropensityforbankstodowngradetheloan.Ifanything,banksarelesslikelytodowngradeloanswithstaleinformationevenasdefaultriskrises,suggestingtheyoftenemploya“wait-and-see”approachtoreporting.

HavingshownthatperformancereportingisanessentialcomponentofbankmonitoringforCREmortgages,thefinalportionofthepaperanalyzestwoempiricalexercisesthatelucidatesourcesofendogenousadjustmentinbankmonitoringactivity.Inparticular,wearguethatbanksarewillingtoincurmorecostswhenthebenefitsofmoretimelycreditriskinformationisgreater.Increasingmonitoringlikelyinvolvesbothdirectcostsintermsofprioritizingtheworkofcreditriskmonitoringstaff,andindirectcostswhereloanofficersmaydemandmoretimelyreportingfromborrowers,potentiallyimposinghasslecostsonborrowersandaffectingrelationshipcapital.Theseexercisessupportthefinalfindinginthispaperthatbanksmonitormoreintensivelywhentheyobserveshockstoloan-specificrisks,location-specificrisks,orrisksonthebank’soverallportfolio.

Thefirstempiricalexerciseusesvariationfromasharpoilpricedeclineofmorethan50%inlate2014,whichwetreatasanexogenousshockinatripledifference-styleregressionframework.Relativetocitieswithlittleornoemploymentinoilandgas-relatedindustries,banksincreasemonitoringofCREloansinhighlyexposedcities.MonitoringofCREappearstodecline,onaverage,incitieswithonlysmallormoderatelevelsofoilandgasindustryactivity.Theseeffectsusingacross-cityvariationaredecomposedusingwithin-cityvariationbymeasuringhowexposedindividualbanksaretooilandgaslendingontheircommercialandindustrial(C&I)loanportfolios.BanksthatareheavilyexposedappeartoincreasemonitoringofCREmortgagesacrossalltypesofcities,regardlessofwhetherthecityisexposedtooilandgasemployment.Incontrast,bankswithminimalC&IloanexposuretooilandgasfirmsincreasetheirCREmonitoringactivityincitiesheavilyexposedtooilandgasbutdecreasemonitoringforloanssecuredbypropertyinless-exposedcities.These

2Seee.g.,BernhardtandMiao(

2004

)foradescriptionofstaleinformationinanumericalandtheoreticalframework.

3Thebinnedscatterplotsofeachvariableresidualizethequarterlydatawithloanfixedeffects.Oneadditionalquarterofstaleinformationiscorrelatedwith5basispointshigherdefaultrisk.

4

Figure1:ImpactofStalenessofNOIOnDefaultandRatingsDowngrades

Notes:Figuresshowabinnedscatterplotofloan-by-quarterobservationsoftheprobabilityofdefault,probabilityofadowngradeinbankinternalratingsofdefaultrisk,andfinancialreportingstaleness(currentreportingquarterminusquarterthatthemostrecentfinancialswerereported‘a(chǎn)sof’).Thefigureplotsresidualsfromaregressionremovingloanfixedeffectstoisolatewithin-loanvariationinstalenessanddefaultriskratings.Amongthesefigures,theminimumnumberofloansinthesmallestbinnedpointonthescatterplotis19,780,andtheminimumnumberofbankssummarizedbyanybinnedpointis31.

Source:FederalReserveFormY-14QScheduleH.2,Authors’analysis.

resultscomplementexistingevidenceonde-riskingandre-balancingbyexposedbanks(e.g.,Wang

2021

;Bidderetal.

2021

).

Thesecondempiricalexerciseisolatesvariationinloan-specificriskbystudyingtherelativeeffectofshiftsininterestratesoninformationproductionforfloating-ratemortgagesversusfixed-ratemortgages.Difference-in-differenceregressionswithheterogeneoustreatmentvariationcomingfrominterestratechangesshowthecausaleffectsofwithin-loanchangesincreditriskonbankmonitoringbehavior.Banksincreasemonitoringactivityonfloatingrateloansrelativetofixed-rateloansinresponsetorisinginterestratesanddecreaseinformationacquisitionasratesdecline.Oneconsequenceofthisbehavioristhatbanksarerelativelymorelikelytobothdowngradeandupgradetheirinternalriskratingsonfloating-rateCREmortgagesasinterestratesrise.Thisevidenceonmonitoringimpliesthatapreviouslyunidentifiedchannelexistsbywhichinterestratescanaffectbankbehavior(e.g.,Ippolitoetal.

2018

)andrelatestodiscontinuouschangesinbehavioraftermortgagerateresets,whichhasbeenanalyzedmostcloselyintheresidentialhomeownercontext(DiMaggioetal.

2017

).

Thefollowingsectioncontainsadiscussionofthetheoryandpracticeofloanmonitoringwithasummaryofrelatedliterature.Section3thendiscussesthedatausedinouranalysisandhowwemeasurethestalenessofperformance,whichistheprimarymeasureofmonitoringactivity.Section4establishestheextenttowhichbanksuseperformanceupdatestoinformtheirinternalriskassessments.Section5providesevidencethat“stale”propertyfinancialsindicatesubsequent

5

loandistress,andbanksappearslowtocatchthatdistressearly.Section6studiesloanmonitoringintwoempiricalsettings:responsetothe2014oilandgasshock,andexposuretointerestratesonfloatingrateloans.ThepapercloseswithashortdiscussionofpotentialimplicationsforimprovingcreditriskmonitoringandthestatusofCRE.

2BankMonitoringDiscussionandLiteratureReview

ForCREmortgages,loancontractsspecifyareportingscheduleforborrowerfinancials.Thisisacorepieceofloanmonitoring,whereborrowersreportfinancialsapproximatelyonceperyear.

4

FinancialreportingisoftentreatedasaCREloancovenant,similartothecommercialloancontext—ifborrowersdonotreportontime,thelendershouldhavetheabilitytoenactsomepenaltiesorclassifytheloanasintechnicaldefault.Forloansincommercialmortgage-backedsecurities(CMBS),thestandardprocessbasedontheCREFinanceCouncil’sInvestorReportingPackagespecifiesthatifborrowersfailtoreportaftermultiple(usuallyfour)consecutivequarters,theyareplacedonaservicerwatchlistandeventuallysenttoaspecialservicer.

5

Multiplechannelsofservicercommunicationareessentialtoloanmonitoring.Bankscollectbothsoftandhardinformationaboutborrowerperformance.However,weonlyobservetheoutcomeofthiscommunicationandmonitoring—hardinformationonwhenborrowerfinancialsshowupinbankdatabases.Communicationsmayincludecalling,emailing,sendingformal‘delinquentfinancialreporting’letters,andeventuallyshowingupatthebuildingorofficeoftheborrower.Gustafsonetal.(

2021

)usesthetext-basedservicercommentarysummarizingthesecommunicationsinthesyndicatedloanmarket.ThesecuritizedCREmortgagemarkethassomeofthisstyleofcommentary,buttheCCARY-14Qsupevisorydatacollectiondoesnotincludeservicercommentary.

Byourunderstanding,bankshavesignificantdiscretioninthesecommunicationsandinhowtheyrespondtodelayedreporting.Someindustrydiscussionaboutthemonitoringprocesspointstopotentialcarrotsandsticksbywhichbankscaninfluenceborrowerstoreport:banksmaybeabletoimposeorwaivemonetarypenaltiesorassumesomeoperationalcontrol,forexamplethroughsomethinglikeaspringinglockbox(ora“cashflowsweep”);alternatively,forfaithfullyreportingandwell-performingborrowers,banksmaybeabletoreleaseadditionalreservesortenantimprovementfunds.Likelytheultimatediscipliningfactorinachievinggoodbehaviorismaintainingborrower-lenderrelationshipsandobtainingrepeatbusiness.Chodorow-Reich(

2014

)suggeststhattheserelationshipsarequitevaluable,sothisisaparticularlyimportantmechanismforborrowerswhovaluetheabilitytorolloverdebtwiththesamelender.

Inbankingandfinancialintermediationtheory,banksplayadualroleofassemblingcapitaltolendandresolvingaprincipal-agentproblembetweeninvestorsandfirmoperatorsbyspecializingininformationacquisitionandperformancemonitoring.Diamond(

1984

)andBoydandPrescott(

1986

)theorizethatbanks’roleasadiversifiedfinancialintermediarywithacomparativeadvantageincollectinginformationandmonitoringborrowerperformancecanresolvetheproblemthatwith-outaspecializedintermediary,allinvestors(i.e.,depositors)wouldhavetoengageinduplicative

4AppendixFiguresA.2&A.4indicatethat25%ofloansreportfinancialsintheaveragequarter.

5Seee.g.,

/irp

.

6

monitoringofborrowersornoinvestorswouldengageincostlymonitoringbecauseofafreeriderproblem.Inthecontextofcommercialrealestate,wemightexpectthatbankswouldhavedepart-mentsspecializinginCRE,usingtheirscaletoacquireandusemarketactivityandperformancedatawhilemonitoringtheirlendingportfolio.

Theoreticalworkonbanks’informationandscreeningbehavioroftenfocusesonloanoriginationwithlessemphasisonmonitoringorinformationgatheringduringthelifeoftheloan.Theoptimalinformationproductionbehaviorforbanksisnotentirelyclear.Someresearchsuggeststhatdecayinginformationqualitycontributestocreditboomsandbusts(e.g.,GortonandOrdonez

2014

),whileotherworkarguesthatmoreopacityininformationaboutcollateralqualitycanhelpstaveofffinancialcrisesdrivenbyshort-termdebtmarkets(e.g.,Dangetal.

2020

).

6

CREmaybeimportanttostudyinthecontextofthismacroandcreditresearchbecauseitisconsideredcyclicallysensitiveandhasarelativelyhighriskforlenders.

Empiricalworkoninformationproductionhasalsofocusedonloansatoriginationinthesame

vein—–theroleofinformationandmisreportingwasacoreareaofstudyinresidentialmortgagemarketsandhousingconstructionleadinguptotheGreatFinancialCrisis(seee.g.,Keys,Mukherjee,etal.

2010

;Keys,Piskorski,etal.

2012

;Lisowskyetal.

2017

).

7

Aiello(

2016

)andMunnekeandSmith(

2023

)analyzeresidentialmortgageservicers’useofpaymentbehavioraspredictiveinformationaboutfutureperformance.InCRE-relatedsettingssimilartothispaper,Garmaise(

2015

),KrugerandMaturana(

2021

),andGriffinandPriest(

2023

)showthatintermediaryorborrowermisreportingispredictiveofdefaultandrelativelycommoninresidentialmortgageandCMBS2.0markets,respectively.

8

Exposuretorealestateappearstobequiteimportanttothebankingsystem’sfunctioning,andbankmonitoringadjustsstronglyoverthebusinesscycle.Mostresearchonrealestatemonitoringoverrecentbusinesscycleshighlightstheroleofconstructionlending(Heitzetal.

2023

).BanksdecreasedcollectionofauditedfinancialstatementsonconstructioncompaniesduringtheboomandsharplyincreasedcollectionduringthebustaroundtheGreatFinancialCrisis,contributingtolargerloanlossesandbankstress(Lisowskyetal.

2017

).

Becauseoftherecentavailabilityofqualitysupervisorydataonbankbalancesheetlending,agrowingresearchliteraturedocumentslenders’lending,monitoring,andriskmanagementbehaviorinthecommerciallendingmarket.Thereisevidencethatbanksendogenouslyadjustmonitoringactivity,

9

andthattheanticipatedcostsofactivemonitoringinfluenceloancontracttermsatorigi-

6GortonandOrdonez(

2014

)studyendogenouscreditboomsandfinancialcrisisriskinthecontextofinformationaboutcollateralthatdecaysovertime.Dangetal.(

2020

)arguethattheincreasedopacityofcollateralqualitycanbebeneficialforshort-termdebtmarkets,atleastintermsofpreventingfinancialcrises.AhnertandKakhbod(

2017

)makesimilartheoreticalpointsaboutwheninvestorsmakeinformationacquisitionchoices—theyarguethatoptimalpolicyaroundinformationacquisitiondependsonthecollateralqualityorsolvencyofborrowers.Asriyanetal.(

2022

)studythemixofcollateralandcostlyscreeningthatlendersusetomitigatemoralhazardbyborrowers,andarguethatinacreditboom,costlyinformationisunderprovided(duetoincreasingcollateralvalues).

7Suggestiveoflimitationsintheuseorproductionofqualityinformationonnewlyoriginatedloans,Tzioumis

(

2017

)pointsoutthatco-borrowingaffectsmortgageperformance,butbanksappeartofailtopriceinterestratesbasedonnumberofborrowers;Aiello(

2016

)showsthatMBSinvestorsandratingagenciesappearnottoincorporateinformationaboutborrowers’mortgageperformancepriortosecuritizationintotheirpricingatissuance.

8Ambroseetal.(

2024

)showthatCMBSPoolingandServicingAgreements,whichprovideinformationaboutmortgagecollateralandconveyinformationaboutservicingandmonitoringbehavior,aremeaningfullycorrelatedwithexpostdefaultandloanperformance.

9Cerqueiroetal.

2016

;MinnisandSutherland

2017

;Beckeretal.

2020

;Gustafsonetal.

2021

;Branzoliand

7

nation.

10

Amongthecommerciallendingresearch,Cerqueiroetal.(

2016

)constructsameasureofinformationquitesimilartothemeasureofstaleinformationusedinthispaper.Thatmeasureiscalculatedasthetimebetweenbankemployees’reviewsofborrowersorcollateral.ThetimebetweenreviewsisusedasaproxyformonitoringinanempiricalsettingthatanalyzesaSwedishlegalreformthatcausedadeclineinloancollateralvaluesandresultedinadeclineinbankmonitoringintensity.

Thispapercontributestothebroadliteratureoninformationandthenarrowliteratureonmonitoringandriskmanagementintherealestatesectoratlargebanks.Ourfocusonmeasuringspecificpiecesofinformation,howthatinformationistransmittedtothebank,andhowbanksusethatinformationadvancesthisliterature.Weusevariationattheloan,location,andbankleveltostudyinformationproductiontotestandvalidateanewandpotentiallyusefulmeasureofinformationavailabletoregulatorsthroughexistingsupervisorydatacollection.

11

3DataonMonitoringofPropertyFinancialsandBankIn-ternalCreditRatings

ThispaperusesdataintheFederalReserve’sFormY-14QScheduleH.2CommercialRealEstateloan-levelreporting.

12

ThisreportingispartoftheFederalReserve’sannualComprehensiveCapitalAnalysisandReview(CCAR)processforevaluatingwhetherthelargestbanksoperatingintheU.S.arewell-capitalizedandabletomeettheirobligationsintheeventofeconomicstress.

13

Thispaperusesloanfacilitiesclassifiedasretail,industrial(andwarehouse),multifamily,andofficepropertytypes.

14

Banksreportcurrentnetoperatingincomeandoccupancyforpropertieswithloansontheirbalancesheets.EachcurrentNOIobservationisreportedwithan“asof”date.TheNOIreportedbythebanksinagivenCCARreportingquarterarebasedonthemostrecentlyreportedvaluesthebankhasreceivedfromtheborrower.

15

Generally,thereisaseveral-quarterlagbetweenthe“asof”

Fringuellotti

2022

;HowesandWeitzner

2023

;Claessensetal.

2023

;Haqueetal.

2024

.

10SeeChodorow-Reichetal.

2022

;Jiangetal.

2024

.Aconnectedliteratureoninformationproductioninthecontextofbanklendingstudiestheconsequencesofinformationonassetmarketsandtrading.SeeBeyhaghietal.

2023

;Haselmannetal.

2022

.Informationdisclosurebehaviorcanalsobestrategic.Borrowerschoosetodiscloseorfailtodiscloseinformationintheprocessofsearchingforornegotiatingfinancing.SeeBird,Karolyi,andRuchti

2019

;Bird,Ertan,etal.

2020

;Gustafsonetal.

2021

;Jiangetal.

2024

;Backetal.

2023

.Strategicincentivesininformationdisclosureareimportantinothercontextsaswell:T.Gormleyetal.(

2019

)andT.A.Gormleyetal.(

2022

)showthatfirmsinternalizehowstockmarketparticipantswillrespondtofinancialinformation.Informationprovisionandprocessingalsoappearstobeimportantforparticipantsinpublicmarketsandforcreditratingagencies(Grediletal.

2022

).TheremaybereasonforcautioninrelyingtooheavilyononesourceofriskassessmentgiventheroleofcreditratingagenciesintheGreatFinancialCrisis(FreixasandLaux

2011

).

11Connectedtotheempiricalsettinginthispaperonbankexposuretooilandgasindustrylending,priorworkhighlightedbankspecializationorlendingconcentrationasdeterminantsofbankinformationproduction(Bergeretal.

2017

).

12ThedataisScheduleH.2oftheFR-Y14Qdatacollectionandisactuallycollectedatthecreditfacilitylevel.Giventhatmostcreditfacilitieshaveasingleloan,wetreatitasloan-level.

13Appendix

A

containsadditionalinformationabouttheY-14QCREdataset,performancereporting,andinternalriskratings.

14Owner-occupiedpropertiesarenotcollectedintheFRY-14Q,ScheduleH.2.Wedropfromouranalysiscon-structionloansandloanssecuredbyotherpropertytypesasthefocusofthispaperistheongoingmonitoringontheinvestmentclassincome-generatingpropertieswithcomparableleasingbehaviorandsimilartenantrelationships(e.g.hotelseffectivelyhaveonenightleasesandsignificantlydifferentmarketdynamics).

15Thestructureofthisperformancereportingissimilartothereportingforsecuritizedloanswhereborrowersreporttoservicersforcommercialmortgage-backedsecurities(CMBS)usingtheCREFinanceCouncil’sInvestorReportingPackage,thoughonlyasmallamountofinformationisavailableintheCCARsupervisorydata.

8

dateandthedatethattheborrowersendsthebanktheirperformanceupdate—forexample,NOIforfiscalyear2017willbelisted“asof”Q42017(Dec.31,2017),andwilloftenbereportedinQ22018(e.g.,byJune30,2018).

Weconstructameasureofthe“staleness”ofborrowers’performancereportingasthecurrentreportingquarterminusthecurrentlyreported“asof”quarterofNOI.IntheexamplediscussedabovewithQ42017financialsreportedinQ22018wouldhaveastalenessmeasureoftwoquarters,thestalenessmeasurewouldincreasetothreequartersinQ320