how do we improve the quality of tax data for research purposes?

Post on 13-Jan-2017

279 Views

Category:

Government & Nonprofit

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

How do we Improve the Quality of Tax Data for Research Purposes?

WilsonPrichardInterna(onalCenterforTaxandDevelopment

TiangbohoSanogo

UniversityofAuvergne,Clermont-Ferrand,FranceUniversityFelixHouphouet-Boigny,Côted’Ivoire

The ChallengeHighqualitytaxdataforresearchpurposescanbeacri(caltool:•  Fortaxadministra(ontoassess,andimprove,theirperformance

•  Forresearcherstoanalyzetaxperformance,andtheconnec(onsbetweentaxandbroaddevelopmentoutcomes,inordertoinformpolicyandprac(ce

However,availabletaxdatahashistoricallybeenincomplete,andoIenhighlyinaccurate,andtheICTDhasinvestedheavilyinseekingtoaddressthisgap.

Overview

1.  TheICTDGovernmentRevenueDataset

2.  Improvingtheuseofdata–andnextsteps

3.  StrengtheningSub-Na(onalTaxData

THE ICTD GRD

The ICTD GRD - Motivation•  Ini(atedin2010,andlaunchedin2014,respondstomajorgapsandanaly(calinaccuraciesinexis(nginterna(onaldatasets

•  Constructedbymergingdatafromallmajorinterna(onaldatasetsandIMFAr(cleIVreports

The ICTD GRD – DataTheICTDGRDdataisnotablerela(vetoothersourcesinfivemajorways:

1.  Muchimproveddatacoverage,par(cularlyfordevelopingcountries

2.  Eliminatesimportanterrorsandinaccuracies,includingrelatedtoGDPdata

3.  Moresystema(ctreatmentofrevenuesfromnaturalresources

4.  Systema(ca\en(ontodifferencesbetweencentralizedandfederalstates

5.  Greatertransparencythanotherresearcherdatasets

The ICTD GRD - ImpactTheICTDGRDhasalreadyhadademonstratedimpactonresearchfindingsandprac(ce:1.  Thedatasetisnowaleadingdatasetfortaxresearchers,

andexpanding§  Over5000downloadsoverpastthreemonths§  MajorresearchconferenceatUNU-WIDER

2.  Earlyfindingsprovideevidencethatemployingimproveddatayieldsnewandbe\erresearchresults

3.  Providesaclearerandmorereliablepictureofcross-countryandwithincountrytrendsforpolicymakers

Recent Data Initiatives

1.  OECD/ATAFdatacollec(onforAfrica:Importantini(a(vefordeepeningdataqualityinpar(cipa(ngcountries

2.  IMFWoRLDdataset:Majorstepforwardindatatransparency

3.  IMFworkonnaturalresourcerevenues:Acri(calini(a(veforthefuture

Sustainability and AdoptionResearchershavelongsoughttoimprovetaxdata–butsustainability,andwideadop(on,havebeenkeybarriers.1.  Shortterm:TheICTDhascompletedafirst

updateofthedata,andac(velypublicizedthedata,whichisnowinwideuse

2.  Longterm:WehaveenteredintoapartnershipwithUNU-WIDERwhowillhost,andconsistentlyupdate,thedata,inpartnershipwiththeICTD.

Next Steps – Data UseKnowingthedatabeingusedforresearch,andusingitappropriately

1.  Understandingwhatthenumberscapture,andtheirimplica(ons,cri(caltointerpre(ngresultsandappropriateanalysis

2.  Cau(oninusingtaxra(os,andusingcross-countrycomparisonsappropriately•  ProblemswithGDPdatacanmakecross-country

comparisonsmisleading•  Evalua(ngtaxperformancedependsonrevenue

collected,butalsounderstandingthetaxbase

Next Steps – Resource Revenues

TheICTDGRDa\emptstodis(nguishresourcerevenuesfromothertaxrevenue,butbe\erdis(nc(onsareneeded

§  Clearerdefini(ons§  Be\er,andpubliclyavailable,data§  IMFini(a(veonthisareaholdssignificantpoten(al

Next Steps – Deeper DataWhileICTDGRDtellsushowmuchrevenueiscollected,understandingoftaxsystemsalsorequiresunderstandofwhatexactlyispaid,bywho,andrela(vetowhatispossible

§  Howistaxperformancerela(vetopoten(al?§  Howfairisthetaxsystemintermsofoverallstructure?

§  Howcompleteiscompliance,ordosomegroupsavoidpayment?Whatisthestructureofthetaxbase?

Local Government Dataset in Cote D’Ivoire

Overview•  TheCôted’IvoireLGRdataaimsatfillingthegaprela(vetoabsenceofdatasetonlocalgovernments’revenue;

•  This data represents the most complete and most accurate dataset on localgovernmentrevenue,covering12yearsfor90%ofthe196exis(ngmunicipali(es;

•  Itisanimportanttoolforsta(s(calanalysisandtaxpolicyreforms;•  However,thedatasetsuffersfromtwomajorlimita(ons:Missingdataremainandsomevariablesneedtobedisaggregated;

•  Thereisaneedfortraininglocaladministra(onsandtosetupelectronicsystemfordatacollec(on;

•  Itisvaluabletoundertakesuchprojectindevelopingcountrieswherenotexis(ng.

Outline

1.  Mo(va(ons

2.  Methodology

3.  Outputs

4.  Limita(ons

5.  Usingthedataandlessons

Motivations1.  Crucialneedtoundertakeresearchtodeepenprocessesof

state-building:§  Localrevenuedatasetallowtoundertakeresearchthatfoster

administra(vereforms

2.  Absenceofaggregated,reliabledatacoveringallmunicipali(esoveralongperiod.§  Theexis(ngna(onalsourcessufferfromlimita(ons:Inconsistency

withthedefini(onofvariablesbetweenmunicipali(es,presenceofoutliers,significanterrors

3.  Acomprehensivedatasetcontributestowidenthedebateonlocalgovernments'responsibili(es.

Methodology - AccessContac4ngthedifferentsourcesofdataanddealingwithconstraintsofgainingaccess

§  Sendingformalle\ersforclearexplana(ons;§  Organizingseveralroundsofmee(ngs;§  Interviewingeightselectedmunicipali(es.

•  DepartmentinchargeofDecentraliza(onandLocalDevelopment,MinistryofInterior(Datarequested:Theadministra(veaccountsformunicipali(es(localtaxandnon-taxrevenuesandexpenditures)

•  MinistryofEconomyandFinance(Datarequested:Thena(onalaccountforlocalgovernmentsgrants)

•  Financialservicesof8municipali(es(Datarequested:Localaccounts)(Bouake(Centre);Korhogo(North),Daloa(West);Kanakono(North);Abobo,Yopougon,Plateau(South),Tengrela(North)

Methodology –Data Cleaning

Makingsuitableforresearch(Datacleaning)

§ Manualdataentering§  Dealingwithoutliers,possibleerrorsoftranscrip(onorrepor(ng

§  Runningsomestandardsta(s(caltests§  Addingnoteforclearexplana(on§  Findingallabruptchangesandtrea(ng§ Mergingdataformthedifferentsources

Outputs (Data)

Outputs (Data)

Limitations

1.  Missingdataremain

2.  Outliersduetodiscre(onarydecisionsformcentralgovernment

3.  Needtodisaggregatesomevariables

Using the Data and LessonsExample•  SanogoandMoummi(2015).Localgovernmenttaxa(on

andaccesstobasicservicesinCôted’Ivoire,CERDI,Workingpaper.

Lessons•  Needfortraininglocaladministra(onsandtosetupelectronicsystemfordatacollec(on;•  Needtoupdatetheexis(ngdata(from2012-2015);•  Needtoextendthedata(1990to2000);•  Valuabletoundertakesuchworkindevelopingcountrieswherenotexis(ng.

Thank you for your a-ention!

Commentsandsugges(onsarehighlyappreciated

top related