9/19 SA-TIED.WIDER.UNU.EDU RESEARCH BRIEF mispricing inSouth Africa -motivated transfer Revenue lossfrom ZAR1 billion. or US$73 million to approximately year 2012isequivalent US dollarsperyear. Forexample,a0.1%loss infiscal total taxrevenue,thelostis stillseveralmillionsof from transfermispricingofgoods,issmall,at0.1% While theproportionatelosstoSouth Africa, resulting transacted goodsbyatleast8.59%. engage intax-motivatedtransfermispricingofinternally concludes thatmultinationalfirmsinSouth Africa effects), itisreducedto8.59%. Thestudytherefore for allotherpossibleexplanationsofmispricing(fixed average of31%.Whenthisfigureisadjustedtoaccount ofgoodsfromtheirownsubsidiariesbyan In astandardanalysis,thestudyfindsthatfirmsoverprice South Africa) reducingitsrecordedincome. the firmsoveralltaxbillbyartificially(andillegallyin tax-motivated transfermispricing. The practicereduces overcharging itselfforinternaltransactions,itisknownas rate toanaffiliateinacountrywithlowertaxrate,by shifts profitsfromanaffiliateinacountrywithhighertax for asimilargoodfromanexternalparty. Whenafirm transacting internally, definedasthepricetheywouldpay Africa requiresfirmstousethearm’s-length pricewhen The currentlegislationontransfermispricinginSouth mispricing? What istax-motivatedtransfer strong evidenceoftax-motivatedtransfermispricing. systematic overpricingofimportsisinterpretedas firms actuallychargethemselves.Intheanalysis, prices (knownasthearm’s-length price)totheprices the differencebetweenlegitimateestimatesofmarket data fromthetaxauthority, theanalysiscalculates country. Usinghighlydetailedfirm-levelcustoms tax-motivated transfermispricinginadeveloping New researchprovidesthefirstdirectevidenceof transfer mispricingishighlydependent onthedifference a multinationalfirm’s willingnesstoengageinillegal Past researchontransfermispricing hasfoundthat firms’ overall responsetoatargetedpolicymeasure. (BEPS) initiative. The research, therefore,isabletotrack under theinternationalBaseErosionandProfitShifting mispricing inaccordancewithOECDrecommendations to South Africa’s legislationon tax-motivated transfer The periodofanalysisincludesa2012amendment The impactofpolicyreforms arm’s-length price,by2.5%. their internalcross-bordertransactions,relativetothe lower taxratejurisdictions,firmsincreasedthepriceof increase inthetaxdifferentialbetweenSouth Africaand The studyalsoestimatesthatforeverypercentagepoint FINDINGS goods byupto31%whenimportsare Firms over-chargethemselvesforimported tax-motivated transfermispricing jurisdictions, themorefirmstendtoengagein The greaterthedifferenceintaxratesbetween digital taxenforcementscheme worth millionsofrandannually, througha It maybepossibletoreducethisrevenueleak, authority everyyear multinationals isnotcollectedbythetax of nearly2%thetotaltaxreceiptsfrom estimated at8.59%perimportsothatavalue The averagerateofgoodsmispricingis transfer mispricingforserviceimports jurisdiction. Theanalysisdoesnotestimate purchased fromanaffiliateinalowertaxrate between its perceived risk of being caught Figure 1: Estimated tax loss of transfer mispricing of goods (% of versus its potential gain. More specifically, receipts) by study Figure 1: Estimated tax loss of transfer mispricing of goods (% of corporate tax receipts) a willingness to use transfer mispricing should depend on both the likelihood of audits and corresponding penalties from the tax authority — or the risk of being penalized — and the tax differential between the countries the multinational is operating in — or the size of the potential financial gain.

The study finds that the South African legislative amendment of 1 April 2012 on transfer mispricing dramatically, but only temporarily, reduced the rate of transfer mispricing. The estimated price deviation per fell from over 70% before the reform to under 40% in 2014 after the reform.

This is interpreted as firms responding Note: The graph shows the estimated tax loss caused by transfer mispricing of goods in prior studies as a share to the reform by expecting greater of total corporate tax receipts. The tax loss is based on one direction of (imports or ) in all studies enforcement from tax authorities. When but Vicard (2015) where the averagedata across and a imports model and like exports the one is used. used See for section the study 7 for toa full flag description. greater enforcement did not materialize, theySource: resumed Author’s their own literature reviewfirms ( seeengaging online a inppendix transfer). mispricing for further audits. prior behavior, with the estimated price deviation rising to Tax authorities can use the econometric method applied in this paper as an automated digital pre-reform levels by 2015. It took two weeks to set up the data in South Africa flagging system. Such a system would alert tax authorities when firms are systematically divergent in their external and internalsuch price that-setting it could behavio automaticallyur. For many flag governments,companies that the data is already If this analysis is correct, a credible threat of thereincreased and used when firms aresystematically audited. The deviate next natural from developmentthe estimated is arm’s-length to use the full data source enforcement would have the effect of decreasingin an automatedtax- flagging modelpricing. to guide The coststhe selectio of doingn of this firms are for in audits.the thousands This would be a feasible, motivated transfer mispricing and increasing lowSouth-cost Africa’s, and easily implementedof dollars digital while intervention. the potential4 The tax cost gain of is doing in the this tens is of in the thousands fiscal revenue. of dollars while the potentialmillions tax gain of isdollars. in the tens of millions of dollars. Such an intervention is an example of the potential for digital tax enforcement, which the OECD (2016a) and the IMF (2017) are promoting. To myFor knowledge, 2012, the tax no losstax authorityto South Africahas yet from implemented transfer such a system. However, the fact that I (and others) find systematic mispricing using this methodology implies mispricing was estimated at US$70 million or ZAR1 Policy options that there should be some scope to pursue this further. billion. The OECD and the IMF both promote the useThe of paperdigital willtax proceed as follows. In Section 2, I give an overview of the previous literature. In enforcement to plug revenue leaks. One exampleSection would 3, I describebe the SouthAs the African study context did not andinclude transfer multinational pricing legislation. firms’ use Section 4 gives a the establishment of an automatic flagging systembrief theoretical at the motivation.of Section transfer 5 mispricingpresents the for data service used, and the intellectual identification strategy, and the tax authority as a cost-effective means of increasingmain empirical the results. Sectionproperty 6 evaluates rights transactions, the transfer pricingthe total reform real revenue that took loss place in 2012. expectation of greater enforcement by the taxSection authority. 7 presents For a systematicdue reviewto transfer of prior mispricing transfer may mispricing be higher. estimates Still, this and compares these estimates to the South African case. Finally, I conclude and discuss the findings in Section 7. such a system, tax authorities can use their own recommendation, which would only impact mispricing of goods, can still plug a substantial leak in South Africa’s revenue collection.

This brief is based on WIDER Working Paper 2018/123 ‘Tax-motivated transfer mispricing in South Africa: Direct evidence using transaction data’, by Ludvig4 ItWier. took two weeks for me to set Whilethis up intax-motivated South Africa. transfer mispricing of goods is a small leak in the state’s revenue collection abilities in relative terms, it is 3not insubstantial in absolute terms

Increasing the flagging and auditing capacity of the tax authority might boost South African revenue by hundreds of millions of rand every year

IMPLICATIONS Continued collaboration with international experts may reveal other practical opportunities to improve the tax enforcement and revenue collection capability of tax authorities

Other emerging economies can also benefit from South Africa’s innovative approach to data and partnership management