The Misuse Of Regional Economic Models - Cato Institute

1 Economicanalysisissometimesusedselective lyandprejudi-ciallytosupportpositionsmot ivatedbyself-interestorideologyishardlyn ewstomost importantexample ofthemisuseofeconomic areusedinwaysthatsystematically exaggeratethepublicbenefitsofproposedgov ernmentprojects, thusbiasing governmentdecision-making inthedirectionofexcessivegovernmentspend ingandexpansionintoareasthatshould stateandlocalgovernmentsrequirethatecono micimpactstudies ofeconomicimpactstudiesissimilarto thatof environmentalimpactstudies:tomeasure theposi-tiveandnegativeeconomicimpactsof aproposed ;mostscholarswouldagreethatnoimportantgo vernmentproj-ectshould beundertaken withoutaprior alsoimportantthatsuch studies be carriedoutobjectivelyandwithmodelsthatar eappropriatefor the anacronym forRegionalEconomicModelsGatoJournal, , (SpringlSummer1993).

2 Copyright JOURNALI ncorporated,a firminAmherst, ,itiswidelyused,perhapsmorewidely ,it consultingfirmshaveregionalmodelsofoneki ndoranother,butmostconsultantskeeptheirm odelsproprie-taryandthereis ownedbyprofit-makingconsultingfirms. ToREMI scredit,theREMI modeliscompletely (1986)andtheJournalofRegionalScience,vol . 25 (1985).RogerBolton s(1985) ,readingofdozensofpub-lishedmodeldocumen ts,andreadingofmanyimpactanalysesunderta kenwithproprietarymodels,REMI appearstobeamongthevery researchandhardworkhave goneintoREMI sformulationandestimation; muchofitscontentisbasedonpublicationsins cholarly journals. Thus, myattimesseverecriticismisofthewaysREMI ismisused,notofREMI bedividedbetweenREMI staffandthegovernmentofficialswhouseit, Ihavenowaytoascertain, , ,counties, , moredetailisavailable forsomeplaces , relatingsectoraloutputintheregiontosecto rallabor,capital,andfuelinputs, of inputs purchasedlocally,andaredetermined inthesameway similarly setbysharecoeffi-cients,whichdependonend ogenousregionalprices consumergoodandserviceisproportionaltolo caldispos-30 MISUSEOFREGIONALMODELS ableincome,allowing setbylocalproductioncosts,butdonotaffect thecompositionoflocaldemand.

3 Disposableincomeiscorrectlydefinedasearn ings plus propertyincomeplus governmenttransfers minustaxespaid. Sectoralinvestmentequationsbringsectoral capitalstockstotheirdesiredlevels,determ inedbymarginalproductivityconditions. Localwageratesaredeterminedbydemandsforw orkersinvariousoccupations,byoveralldema ndandsupply forlocallabor, ,adjustedforlaborforceparticipationrates bycohorts. Cohortsareadjustedbybirths, afunctionofincomeandamenitiesintheregion relative srepresentation of personalincome(lesstrans-fers)equalnatio nalaveragetaxesperdollarofpersonalincome (simi-larlyadjusted)multipliedby adjusted sixequations,oneeachfor:federalcivilian; federalmilitary;stateandlocaleducation;s tateandlocalhealthandwelfare;stateandloc alsafety; sshareofnationalpopulation, expenditureonproposedprojectsandpolicysi mulations arecarriedoutbyaddingtermstorelevantequa tionstorepresent ,spendingon aproposedstateorlocalgovernment projectwouldberepresentedbytermsaddedtot hedemandequationsforinputsneeded srepresentation userfeesplustransfersfromother governmentsplus increasein indebtedness ,thereisnoexhaustivelistof ,thereis regionatatime, ,andthereisofcoursenowaytoanalyzeanexhau stivesetofstateandlocalproposalsinall , ,acountygovernmentdecisiontobuildadomeds tadiumtobefinancedbycountygovernmentfund slogicallyimpliessomecombinationofreduct ioninotherspendingbythecountygovernment.

4 Increasedtaxesinthecountyorincreasedinde btednessbythecounty partlybythemunicipalgovernment s fundsandpartlybystategovernmentfundsrequ iressometaxincreases,orotherchangesamong thoselistedabove,inthemunicipality. Usingstategovernmentfundsrequiresstateta xincreases, (ifthatisthemethodoffinancing) remainder of therequiredstatetaxincreaseispaidbystate residentsoutsidethemunicipality,buttheya renotrepresented aboveisthatREMI makesitappearthatallincrementstogovernme ntspending,federal,stateandlocal,forproj ectsREMI analyzes ,themodelintroducesnoextrataxes,nocutsin othergovernmentservicesortransfers,andno increaseingovernment ,capitalcosts aretobepaidbygovern-mentsandthefacilityi sexpectedtogeneraterevenuesfromuserfeest hat ,REMI evaluates theprojectsasthough noopportunitycostofgovernment nofederal, ,Indeed,itispossiblethatstateand/orlocal spendingisbasicallydrivenbypopulation ,ifaproposed governmentprojectwouldincrease populationandemployment, ,notoverallstateand/orlocalgovernment modelrequiresthatgovernment project.

5 Thenaproposedgovernmentprojectmayreduce intotalrealprivateincomefromaproposed , ,totalemploymentandprivateincomearepredi ctedbyREM!toriseas aresultofanygovernment project anylaborresponsebecause ofinmigrationorreducedunemployment, opportunitycostofstateandlocalgov-ernmen tfunds , ,atransferofanactivityfromtheprivatetoth egovernmentsectoravoidstheprivatesector scapitalcosts,whileatthesametimefailingt oregistergovernments (sincetheprivatesector had tochargeprices thatcoveredcostswhereasthegovernmentcang ivethegoodorservicetoconsumers).Totalear ningsareunaffected,sincethesame employmentandwageratesarerequiredinthego vernmentasintheprivatesector,Noneofthe areinherentinREMI anexhaustivelistofapplicationsofREMI. REMIhas kindlyfurnishedalistofREMI projectsandREMI clients,butIhavecomeacrossseveral economicimpactstudiesdoneforproposedstat eand/orlocalgovernmentprojectsthatindica tethatREM!

6 Wasemployed, ,betweenthelistsfurnishedbyREMI andapplicationsthathavecometomyattention inotherways(mostlyreportssenttomeinconne ctionwithotherwork),itismyestimatethatIh aveseennearly100applicationsofthemodel.( Someanalysesmayappearinmorethanonesource .) ,projectsthatIknowaboutplusdescriptionso fanalysesfur-nishedbyREMI makeclearthat,inthelastfivetotenyears,RE MI hasbeenusedtoanalyze projectswhosecostsare ,itis impossibletoknowhowmuch influenceREMI has ,governmentagencystatementsinsupportofpr oposedprojectsfrequentlyquoteREM! !usesare:analysisoftheeconomicimpactofpr oposedexpansionoftheMcCormickConven-tion CenterinChicago;analysisofproposedexpans ionofMASS-PORT/LoganAirportinBoston;econ omicimpactofexpansionofFortDruminNewYork ;several Economic impactanalysesofpro-posed ofREM!

7 ,itshould bestatedthatREMI willdoanalysisforclients,rentREMI programstoclients, hasnocontrolovermodificationstothemodelm adebyclientsinsomeofthe above !applicationsappeartobeentirelyappropria teandmayprovidebetteranalyses thananyalternativemodel available,However,thedescriptionofREM!in theprevioussectionimpliesthatuseofREM!to analyzeeconomicimpactsofproposedgovern-m entinvestmentsleadstoexaggeratedestimate sofprojects netpublic benefits. A typical economicimpactanalysisconcludes that(x)jobs,(y)dollars of privateincome,and(z) aregeneratedbysimulationswith REMI thatweredescribed inthe ,onesimulationisdoneintheabsence oftheprojectand another ,which impactanalyses generatedincludes bothincomepaidtoworkersandcontractorsont heprojectandalso !takes accountoflimitationsonlabor supply and ofleakagesoutsidethejurisdictionsstudied ,whichsomeKeynesianmacromultiplieranalys esdid !

8 Simulations stems preciselyfromtheincomplete budgetconstraintswereincludedinthemodel, REM! , increasedspendingbyrecipientsoftheincrea sedgovernmentspendingwouldhavetobeweight edagainstreducedspendingbythosewhosetaxe swereincreasedor ,all costsofagovernment projectnotfinancedbychargesforuseofthefa cilityshould beregardedasbeing financedbytaxeswith forexample,a tax onhotelandrestaurantbillsinthevicinityof aproposed convention theextentthatsuchtaxesarepaidby conventioncenterusers,theyare simplyindirectchargesforuseofthefacility andshouldbe,buttypicallyarenot,added ,suchtaxesarepaidby allthosewhoconsumethetaxedservices,wheth ertheyuse ,theyarejustone kindoftaxleviedonthepopulationandhavethe samenegative multipliereffectasanyothertaxtopayfor ,Ihaveindicated inMills(1991)

9 Thattaxesostensiblyleviedonfacilityusers seldom cover whetherothertaxesaretobeincreasedorother governmentexpendituresaretobereducedtopa yforthe tax-financedpartofafacility ,itcouldhavebeenreducedintheabsence ofthefacility, ,theanalysisdoesnotdependonwhetherthefac ilityisbondfinancedor ,discountedatthe government sborrowingrate,equalsthesalepriceofthe discountrates exceedthegovernmentborrowingratethen ofcoursethepresentvalueofthedebtservicec osts islessthanthesalepriceofthe bond, ,forexample,isassumedtodrawpatronsfromou tsidethejurisdictioninwhicheffectsare thatmanypatronsofalarge conventioncenterorsportsstadium , meals, thelocal economy,andmultipliersarecalculatedthata repreciselyanalogoustothose abilliondollar expansion ofMcCormickplace,Chicago sconvention seconomicimpactstudy,KPMGPeatMarwick(199 0),concludedthattheoutsidemoneymultiplie rwouldresult inapermanentnetcreationof6,000jobs.

10 (Outsidemoneymultiplier jobsarepermanent multiplier jobsresultfromaone-timeinjectionofconstr uctionmoneyanddisappear afterthecon-structionmultiplierhasworked itselfout.)Virginia Carison(1991)redidthecalculationsusingal ltheconsultant sassumptionsexceptthatshetookaccountofjo bsthatwouldbedisplacedbytheMcCor-mickexp ansionandoftheimpactoflocal ,335jobswillbecreatedbydirectand ,799jobslostbecause ofbusinessdisplacementand884jobslost36 MisusEOFREGIONALMODELS because of negative ,buttheyareunlikelytoraisemorethanenough tooffset thecenter ,thePeatMarwickanalysiswasundertakenwith a sanalysis thatsellsgoodsorservicesoutsidethelocala rea apensionmanagementcompanyorautoassemblyp lant, for example bringspreciselythesamekindsofoutsidemone ybenefitstothelocal ,stateandlocalgovernmentsrecognizethisbe nefitandprovidetemporaryanddecliningtaxf or-giveness,lowinterestloans,orothersubs idiesthatareintendedtostimulate investmentsintheprivate businesses.