04.25.16 Exchange Rates and Industrial Wage Inequality in...
Transcript of 04.25.16 Exchange Rates and Industrial Wage Inequality in...
ExchangeRatesandIndustrialWageInequalityinOpenEconomies
By
DelfinaRossiandJamesGalbraithTheUniversityofTexasInequalityProjectLyndonB.JohnsonSchoolofPublicAffairs
TheUniversityofTexasatAustinApril26th,2016
UTIPWorkingPaper71
Abstract: We show that movements in exchange rates are a principal determinant ofmovementsinindustrialpayinequality–andthereforealsoinhouseholdincomeinequality–inopeneconomieswith floatingexchangerates.Wedemonstratethis forawideselectionofcountries for the years from 1971 to 2011, a period characterized by fluctuating exchangeratesandfinancialmarketliberalizationinmanyopeneconomies.Exchangeratesarerelatedto the domestic pay distribution by the simple fact that export and non-export sectors areaffected differently by devaluations: the home-currency revenue in export-oriented sectorsrises automatically while the home-currency revenue in other sectors does not. Given therecent large devaluations in Latin America, Asia and elsewhere, we expect increases ininequalitytobeobservedinthedatafor2015and2016.
KeyWords:Inequality,ExchangeRates,
JELCode:
2
I.Introduction
This paper compares the movement of inter-industrial pay inequality with exchange ratesmeasuredagainsttheU.S.dollarforasubstantialsetofcountries.Wefindthattherehasbeenastrong association between themovement of pay inequality and exchange rates after openeconomiesliberalizetheirfinancialsectorsandallowforfloatingexchangerates.Inpracticallyall cases, currencydevaluationsareaccompaniedbya rise in inequality levels,andcurrencyrevaluations are associatedwith falling pay inequality. Further, since cross-sector industrialpayinequalityisanimportantdeterminantofhouseholdincomeinequality,itfollowsthatthesimple mechanics of exchange rate determination have an important, and in many casesdominating,influenceonhouseholdincomeinequalities.
Thisfindingmarksadeparturefromtheprevailingliteratureontheevolutionofinequalities,which has emphasized such difficult-to-measure factors as the evolution of technology, theprogressofeducation,andtheprevalenceofinternationaltrade.ItcomplementsandadvancesthepreviousworkoftheUniversityofTexasInequalityProject,whichhasidentifiedcommontransnationalpatternsinthemovementofinequalitieswithout,uptonow,havingbeenabletospecify the transmission mechanism. The evidence presented here suggests a very cleartransmissionmechanismandanunmistakablearrowofcausation,runningfrommovementsofthe exchange rate, to changes in intersectoral industrial pay inequality, and from there tochanges inhousehold income inequality.Ourestimates suggest that for countrieswithopeneconomies and floating exchange rates, substantial changes in the dollar exchange rate canhaveadominatingeffectonthemovementofgrosshouseholdincomeinequalitiesintheshortrun.
SectionIIexplainstherelationshipbetweenexchangeratesandindustrialpayinequalities,andthat between industrial pay inequalities and gross household income inequality. Section IIIintroducesourdatasourcesandtheselectionofcountriesanalyzedsofar.SectionIVpresentstheevidence,mainlyingraphicalformat; it issufficientlycompellinginthatdescriptiveformalone.SectionVpresentsestimatesofthesensitivityofindustrialpayinequalitytochangesinexchange rates. Section VI presents conclusions and some policy implications for globaleconomyandfinancialmarkets.
II.OpenEconomies,ExchangeRatesandInequality
The 1944 Bretton Woods conference established an international monetary system thatpeggedwesterncurrenciestothedollar,whichinturnwasfixedtogold.Fixed-but-adjustableexchangerates,supportedandmanagedbytheInternationalMonetaryFund,wereintendedtofoster financial stability, to prevent competitive devaluations, and to permit autonomouseconomicpolicyaimedatpromotingfullemploymentandfosteringeconomicgrowth.
ThecollapseoftheBrettonWoodssystemin1971entailedtheendoffixedexchangeratesasageneral principle, and the major western economies subsequently allowed their exchange
3
ratestofloat.Twentyyearslater,thefalloftheSovietBlocin1991ledtotheintroductionofmany more countries into the international financial system, and to their adoption also offloatingexchangerates.
Inopeneconomies, exchange rateshavea simple,powerful effecton thedistributionofpayacrossindustrialsectors.Considerthatinallcases,industrieseitherproduceprimarilyforthedomestic market or primarily for export. Therefore when a country devalues or revalues,export and non-export sectors feel different effects. In the export sector, in the case ofdevaluation,thehome-currencyrevenuerisesinstantly,forthesimplereasonthattheforeignrevenue–whichdoesnotchange–translatesintomorelocalcurrency.Thisoccursinstantly,and does not depend on any changes in volume. In the non-export sector, there is no sucheffect.Thosewhoarepaidinthelocalcurrencycontinuetobepaidexactlywhattheyearnedbefore;anyadjustmentstooffsetanincreasedcost-of-livingwillcomelateriftheycomeatall.
Considerfurtherthatintheoverwhelmingmajorityofcases,countriesexportfromtheirmost-advanced, best-paid sectors. Therefore, the increase in home-currency revenue following adevaluationproducesgreater inequalityacrosssectors.The increasedrevenue in thebetter-paidsectorshastobepaidoutsomewhere,evenifonlytothebetter-paidelementswithinthesector.Whatever the internal (within-sector) distribution of the increased revenuemay be,someofitinvariablywillberecordedasincreasedpayandincomeinthesector.
Further,ifthenon-exportsectorreliesonimportedgoods,suchasfoodorfuel,itmaysufferadeclineinitsprofitabilityandasqueezeonitsintrasectoralpay.Again,thesqueezewillhavetoberecordedsomewhere,reducingtherelativepositionofthenon-exportsectors.
This simplemechanismworks automatically. It does not require any change in patterns ofdemandorvolumesoftrade.Anditworksessentiallyatonce;itdoesnotrequireanylag.
Think for example of a worker in the automobile industry of Canada, an exporter tradingmainlywith theU.S.. Think alsoof a schoolteacher inNorthBayorPrinceGeorge.After the2008crisisandthecollapseofoilprices,theCanadiandollardepreciated.Thus,theautomobileindustryexperiencedincreasedsalesinCanadiandollars,evenassumingconstantvolume,andautomobile sector incomes would have to reflect this increase somewhere. Meanwhile, theincomeof theschoolteacherremainsconstant.Thus,weexpectan increase in inequalities inCanadafromthiseffect.
Or consider the worker in glass or sulphur in Mexico and the taxi driver in Tijuana orMonterrey.Assume theMexicanpesodevalues.Theglass and sulphur industries experiencerising peso incomes, some of which are paid to executives, engineers, and even perhaps toordinaryworkerstocompensatethemfromthedeclineintheirrealwage.Thesearereflectedintherelativepayoftheserelatively-well-paidsectors.Atthesametime,thepesoincomeofthe taxidrivers isunchanged.Again, inequality rises.Movements inexchange ratesare thusrelatedtointer-industrialinequalitylevelsinasimplemechanicalfashion.
4
TheUniversityofTexasInequalityProjecthasalsoshownthatchangesininter-industrialpayinequalitycorrelateverycloselywithchangesingrosshouseholdincomeinequality.TheUTIPEstimated Household Income Inequality (EHII) is a global dataset built largely on thiseconometricrelationship(GalbraithandKum,2005,Galbraithetal.2014).Repeatedestimatesshow that the ratio of changes in aTheilmeasureof industrial pay inequality to changes ingrosshousehold income inequality ison theorderof10:1– thatachangeof tenpercentagepoints in the industrialpay inequalitymeasureswillyieldachangeofonepercentonaGiniindex of household income inequality before taxes. This reflects the greater volatility of theintersectoralmeasureof industrialpayinequality,andthefactthatvariationsinrelativepayacrossindustrialsectorsareamajordriverofdifferencesintheincomeofhouseholdsworkingwithinthosesectors,orinothersectorswhosepayisrelatedtotherelativepayofparticularindustries.
Therefore, if themovementofexchangerates isastrongfactorexplainingthemovementsofinter-industrialpayinequality,itcanalsohelptoexplainthemovementofhouseholdincomeinequalityinopeneconomieswithfloatingexchangerates.
III.DataSourcesandCountrySelection
TheUniversity ofTexas InequalityProject (UTIP)produces a global pay inequality data set,basedontheIndustrialStatisticsdatabasepublishedannuallybytheUnitedNationsIndustrialDevelopment Organization (UNIDO). The UTIP-UNIDO data set measures inequality usingbetweengroups componentof theTheil’sT statistic,measuredacross industrial sectors, for167countriesovertheperiod1963-2008(Galbraith,Halbach,Malinowska,Shams,andZhang,2014).Forthispaper,weupdatedthedatasetthrough2011,wheredatawereavailable,forthecountriesunderstudy.
Theequationbelowsummarizes theTheil’sTstatistic foragivencountryandyear.𝑝! is thenumberofemployeesofasectorioverthetotalemploymentinthecountry.𝑤! istheaveragepayinthesector,and𝑤 istheaveragepayforalljobsinthecountry.Thus,intersectoralwageinequalityisafunctionoftherelativesizeofeachsectorandtheiraveragewage,relativetotheaveragewageforthepopulationasawhole.
𝑇 = 𝑝!𝑤!𝑤
!
!!!
𝑙𝑛𝑤!𝑤
As noted above, the UTIP-UNIDO data set is the basis for the Estimated Household IncomeInequality (EHII) dataset. EHII is a comprehensive and consistently adjusted data set thatpresentsmeasuresofgrosshouseholdincomeinequality,nowcovering149countriesfortheyears1963-2008(GalbraithandKum,2005;Galbraith,Choi,Halbach,Malinowska,andZhang,2015).
5
The International Monetary Fund reports data on exchange rates relative to the USD onaverageperyear.Theperiodanalyzedhereis1971to2011.The1971thresholdwaschosentoreflecttheendoftheBrettonWoodssystem,asdiscussedabove(IMF,2016).
The thirteencountriespresentedhereareopeneconomies thatadopteda floatingexchangeratefollowingtheendofeithertheBrettonWoodssystemorthecollapseoftheSovietBloc,orcountriesthatundertookthisreformduringthe1990s.Hence,thetimeperiodofexaminationdiffersslightlyforeachcountry.Wepresenttimeseriesforthetwovariables,scatterplotsandcorrelation coefficients. Appendix I provides data on twenty-three additional countries forwhicharelationshipbetweenexchangeratesandinequalityisalsoapparent.
Table1:CorrelationCoefficientsbetweenPayInequalityandExchangeRates
Country Years Corr.Coef
Australia 1973-2011 0.7839
Singapore 1971-2011 0.8137
Canada 1971-2011 0.8034
Chile 1973-2011 0.686
Mexico 1994-2011 0.9443
Croatia 1992-2011 0.7293
Hungary 1971-2011 0.9262
Poland 1991-2011 0.8587
Romania 1991-2011 0.9341
India 1991-2011 0.7665
Russia 1993-2011 -0.1805
SouthAfrica 1999-2011 0.0917
U.K. 1971-2011 0.5264
AsTable1shows,withthesoleexceptionsofRussiaandSouthAfrica,thecorrelationbetweenUSDexchangeratemovementsandmovementsofindustrialpayinequalitiesisveryhigh.ItisthehighestinMexicoafter1994,whichshouldsurprisenoone.Butvaluesarealsoveryhighforthe East European countries, and they are entirely respectable in other parts of the world,includingforexampleIndia,avastcountrywhichhasonlyminordirecttradewiththeUnitedStates.SectionIVtakesupthecases.
6
IV.Findings
AustraliaandSingapore
Australia implemented a successful model of trade protectionism and promotion of itsagriculturalexportsuntil1970.Thesubsequentdecreaseofinternationalcommoditiespricesatthat time forced the country into a process of structural reforms that started in 1973. Onedrastic measure was a 25% cut in all tariffs. The 1980s “had also seen the floating of theAustralian dollar (facilitating subsequent adjustment to tariff liberalization) followed bysignificant liberalizationof the financesector, including theremovalofexchangeand interestratecontrols.”(Banks,2004)
Thesepolicydecisionsexplainwhytheexchangeratemovementscorrespondmorecloselywithchangesininequalitylevelsafter1973.In1993,inflationtargetingwasintroducedalongwithafloatingexchangerategovernedbytheReserveBankofAustralia.
Singaporeisanexampleofasmallandrichopeneconomy,withoneofthehighestlevelsofGDPpercapitaintheworld.Singaporehasthehighesttrade-to-GDPratiointheworldandreceiveslargeamountsofForeignDirectInvestment(FDI)(WorldTradeOrganization,2016).
TheMonetaryAuthorityofSingapore(MAS)wasestablishedin1971andsince1980therehasbeen a policy of management of the exchange rate. This policy prevented the country fromexhibiting large jumps in its exchange rate from 1971 to 2011. Moreover, due the growingpositionofSingaporeintheinternationalmarkets,itsnationalcurrencyappreciateduntil1996,withasimultaneousreductioninindustrialpayinequality.IntandemwiththeAsianfinancialcrisis,Singapore’scurrencydevaluedandinequalitytrendsshiftedafter1996.
7
Canada
Canadacanbeconsideredabaselinemodel for thecloserelationshipbetweenmovements inexchange rates and industrial pay inequality, due to five main characteristics of Canada’seconomy: (1) the Canadian dollar is considered a “pure” floating currency; (2) Canada has a“dualopeneconomy”:theexportindustriescanbequitedistinctfromthenon-exportones;(3)Canada’sexportsrelyinlargepartoncommodities(especiallyoilandtimber)withpricessetbyinternationalmarkets; (4) Canada has one big trading partner, theUnited States; 74%of itsexportswenttotheUSin2013;and(5)CanadianfinancialmarketsarehighlyintegratedwithWall Street. Thus, Canada is a case of free capitalmobility and floating exchange rates in anintegratedtrademarketwherethemaintradingproductsarecommoditieswhosepricesaresetattheinternationalfinancialmarkets.Intheseconditionsthecorrelationofinter-industrialpayinequalitytotheCanadianDollar/USDexchangerateisveryhigh.
ChileandMexico
LatinAmericancountriesduringthe1980sand1990sswitchedfromfloatingexchangeratestopegpolicies– inArgentinatoaone-to-oneconvertibilityscheme–oreventodollarization inthecaseofEcuador.Thisfactmakesouranalysismorecomplex.Howevertherearetwoopeneconomies in the regionwith important levels of international trade and capitalmovement:ChileandMexico.
InChile,themovementofinequalityand exchange rates is consistentbeginning in1973, afterthePinochetcoupd’état,exceptfortheabruptriseininequalitythatfollowedthebankingcrisisin1982,andwhichwaspartlyreversedfrom1985to1988.Thegovernmentdevaluedtogaincompetitiveness until 2000, and in this period industrial pay inequality again followed themovementoftheexchangerates.
MexicohadapeggedexchangeratewiththeUSDupto1994,whentheTequilaCrisishitandforced the country out of the peg. This is an example of how free capitalmovements and ahighly integrated financial sectorwithapegsystemcanbeverydestabilizing.After the1994crisis, the Mexican peso devalued by 200% and it continues to depreciate today. Wageinequality across sectors in Mexican industries rose to its highest level in 2008, when theMexicaneconomywashitbytheU.S.financialcrisis.
8
Croatia,Hungary,Poland,andRomania
In each of the “transition economies” of Croatia, Hungary, Poland and Romania, we seeapositivecorrelationofdollarexchangeratesandinequalityafterthecollapseofsocialismandtheliberalizationofexchangeratepoliciesin1991.
TheHungarianPeople'sRepublicwasgovernedbyasocialistregimefrom1949to1989.Beforetheendofcommunism,Hungaryhadverylowinter-industrialinequalityandastableexchangerate,madepossiblebystrictcontrols.In1990,Hungarystageditsfirstdemocraticelectionsandtransitionedtowardamarketeconomy,lettingtheforintfloat.HungaryjoinedtheEUin2004anditisalsoexpectedtojointheeurointhecomingyears.Withintheperiodstudied,thereisapositivecorrelationbetweenexchangeratesandinequality.
CroatiawaspartoftheformerYugoslaviaandfoughtawarforindependencein1991,gaininginternational recognition in 1992. Upon gaining independence, Croatia adopted a managedexchangerate,firsthavingtheDeutscheMarkasareferenceandthentheEuro.Despitehavingamanagedfloatingexchangeratetiedtoadifferentcurrency,thereisapositivecorrelationofthedollarexchangeratewithindustrialinequality.CroatiaenteredtheEuropeanUnionin2013anditisexpectedtojointheEurozoneinthecomingyears.
In 1989, Poland had its first parliamentary democratic election, and the new governmentundertook steps to transform the socialist-style planned economy into a market economy.Poland’sfloatingexchangeratehasbeencorrelatedwiththelevelofindustrialwageinequalityeversince.PolandenteredtheEUin2004anditisexpectedtojointheeurobefore2020.
9
In Romania the communist regime ended in 1990. Since 1991, when the country adopted aflexibleexchangeratesystem,therehasbeenapositivecorrelationbetweenexchangeratesandindustrialwageinequality.RomaniajoinedtheEuropeanUnionin2007anditisalsoexpectedtojointheEurozoneinthenearfuture.
10
India
TheIndianeconomyfacedadeteriorationofitsexternalconditionsandacrisisin1991,whichled thegovernment toundertakeaprocessof liberalizationandadoptionofstrongeconomicreforms including the greater global integration of money, securities and foreign exchangemarkets.Followingthisdevelopment,industrialwageinequalityhashadapositivecorrelationwiththeexchangerate,whichwasnotthecasebefore.
RussianFederation
In the1990s, after the collapseof the SovietUnion,Russia liberalized itsbanking sector andintroduced the convertibility of the Ruble. Russian external debt, in combination with fixedexchangeratesandexternalrestrictionpressures,forcedRussiaintocrisisanditdefaultedonitsdebtin1998.
Theabruptdevaluation inRussia in1998corresponded toa sharp increase in industrialpayinequality,whichhoweverwaslargelyreversedafter2002,whenfavorableconditionssetinforexportprices.Formuchof the2000s, theexchange rateappreciatedand inequalitydeclined,althoughthecorrespondenceisnotcloseinstatisticalterms.
SouthAfrica
SouthAfricaformallyendedthefixedpegbackinthe1980s,buttheinternationalcommunityimposedeconomic sanctionsdue to the apartheid regime, forcing the country to adopt traderestrictions.Followingthefirstdemocraticelectionsin1994,SouthAfricamaintaineditsdual-
11
exchangeratesystem:onerateforfinancialtransactionsandanotherfortransactionsoccurringintherealeconomy.Thisfactcreatesdistortionsforouranalysis.
In2000,SouthAfricaopeneditseconomytoforeigninvestment,relaxedrestrictivelaborlaws,beganprivatization,andcut interestratessharply from1998 levels.Asaresult,wecanseeapositivecorrespondencebetweenexchangeratesandinequalitybeginningin2000,sometimeswithinequalitymovementslaggingbehindoneortwoyears.
UnitedKingdom
IntheUnitedKingdom,in1979,MargaretThatcherwaselectedPrimeMinisterandthecountryundertookaperiodofliberalizationthatincludedtheadoptionoffloatingexchangerates.Asaconsequence,afterthisperiodthereisaclearcorrespondencebetweeninequalityandexchangeratemovements.
Between 1990 and 1992, the U.K. entered the European Exchange Rate Mechanism, whichaimed to create stability to increase trade among the members of the European EconomicCommunity. However, two years later, financial speculators forced the country out of ERM.(Connolly,1995)
Since this episode, the pound has had a floating exchange rate with the USD and thecorrelationbetween the exchange rate and inequality has been positive, although bothvariablesarequitestablebyinternationalstandards.
12
V.Estimatingtheexchangerateselasticityofinequality
Inordertomeasurethesensitivityofindustrialpayinequalitytochangesintheexchangerate,weproposealog-logmodelwithcountryandyearfixedeffectsasfollows:
1 log 𝐼𝑛𝑒𝑞𝑢𝑎𝑙𝑖𝑡𝑦!" = 𝛽! + 𝛽!log𝐸𝑥𝑐ℎ𝑎𝑛𝑔𝑒𝑅𝑎𝑡𝑒!" + 𝛽!log 𝑆ℎ𝑎𝑟𝑒𝐸𝑚𝑝𝑙!" + 𝛿!𝐶𝑜𝑢𝑛𝑡𝑟𝑦! + 𝛿!𝑌𝑒𝑎𝑟! + 𝜀!"
Where i are the countriesand t are theyearsbetween1970and2011when those countrieshavehad floating exchange rates. Thus, thedata set is treated as unbalancedpanel data.Weseparatetwogroupsofcountries:thoseinthetextpresentedaboveandthoseintheappendix.
InequalityismeasuredbyTheilstatisticsreportedintheUTIP-UNIDOdataset;ExchangeRateisthenominalexchangeratetoUSDreportedbytheIMF;andShareEmplistheshareofindustrialemploymentoverthetotalpopulationcalculatedusingUNIDO’sindustrialemploymentandtheWorldBank’spopulationmeasures.Model(2)controlsforrobuststandarderrorsbyclusteringbycountryandmodel(3)excludesthevariableShareEmpl.
Table2showsninevariationsofmodel(2).Thebaseregressionisacountryfixed-effects-onlyregressionofthirty-fivecountries,thetwelvepresentedaboveandthetwenty-threepresentedin theappendix.Theothermodelsarepresented ingroupsof three,withregressions for justthe twelve countries presented above, for just the twenty-three in the appendix, and for allthirty-fivetakentogether.Robuststandarderrorsarecalculatedusingclusteringbycountries.
Theresultsarebroadlyconsistentacrossthemodels. Ingeneral,weseethatadevaluationof10%ofthenationalcurrencyimpliesbetween2-3%increaseinindustrialpayinequality,withapreferred estimate, in our view, toward the high end of the range, since the time trendwillcapturecommonmovementsofexchangeratesacrosscountries,reducingacoefficientthatinfactdoescapturetheeffectoftheexchangerate.Thissimpleformulationtendstocaptureabouthalfofallvariationinintersectoralindustrialpayinequalities.
The coefficient on the variable LogExchangeRate expresses the elasticity of industrial payinequality measured across sectors to the variations in the exchange rate. The share of thepopulationemployedinthemanufacturingsectorisnotstatisticallysignificantvariableinanyof the variations proposed, probably due to the fact that countries’ fixed effects capture thatinformation. Excluding the variable from the analysis (variations 5 and 9) maintains thesignificanceandthelevelofexchangerateselasticityofinequality.
Thecausalchain inthis instancecanonlyrun inonedirection: fromtheexchangeratetothemeasuresof inequality.There isnoplausible reverse causation,underwhicha change in themeasure of inequality would tend to affect the exchange rate. So we have here powerfulevidence that exchange-rate variations, governed in large part by common movements ofcapitalacrosstheentireglobe,haveastrongeffectonamajordeterminantofinequalitywithincountrieswithopeneconomiesandfloatingexchangerates.
13
Table2:summarycoefficientsfrom9variationsofmodel(1)
HowimportantthenistheeffectofexchangeratemovementsongrosshouseholdinequalityasmeasuredbytheGinicoefficient?AccordingtotheworkofGalbraithetal.(2014),theelasticityofasetofmeasuredincomeinequalityGinicoefficientstotheUTIP-UNIDOTheilstatisticalisonthe order of 0.10; thus a ten percentage point movement in the Theil for inter-sectoralindustrial pay produces a one percentage pointmovement in the Gini for household incomeinequality.Giventhecoefficientestimateof0.3fortheelasticityoftheTheilwithrespecttotheexchangerate,wehaveacombinedelasticityof0.03oftheGinicoefficientwithrespecttotheexchangerate.Thus,a50percent increase(depreciation)of theexchangecanbeexpected toyield,onaverage,a15percentincreaseintheTheilforindustrialpaybetweensectorsanda1.5percentmoveintheGinicoefficientforincomes.ForaGinicoefficient initially intherangeof45,thisworksouttoanincreaseofroughly0.67ortwo-thirdsofaGinipoint.Consideringthatthis is a short run effect, thatmovements of the Gini coefficient are damped, and thatmanydepreciations (that is, increases in the local currency exchange rate measured against thedollar)areconsiderablygreaterthan50percent,thisisasubstantialeffect.
Exchangeratesbetween2011and2016
The figures below show the evolution of exchange rates in selected OECD countries and inBrazil,Russia,India,ChinaandSouthAfricafrom2011to2016.Weincludethemtoshowthedramatic recent developments. With the exceptions of China and United Kingdom, we canobserveahighdegreeofrecentdevaluation,includinganalmost25%devaluationoftheeuroagainsttheUSD.Recentdevelopmentsalsoincludea154%devaluationoftheArgentinePesoand a 38% devaluation of the Mexican Peso. Among the BRICS the devaluation has beenexceptionallystrong;Brazildevalued130%,Russia125%,India45%,andSouthAfrica102%.AgainthesoleexceptionsofarisChina,whichrevaluedtheRMBby2%inthisperiod.HoweverthemorerecentdevaluationoftheRMBmaysoonreversethattrend.
Theevidenceofthispapersuggeststhatdevaluationswilltransferdirectlytothestructureofindustrial pay and therefore to the incomedistributionof the affected countries. Thismeansthat in the coming years – indeed beginning now – pay and income inequality will rise in
14
EasternEurope,LatinAmerica,AsiaandAfrica,andthegainsagainstinequalityevidentinmanyoftheseregionssince2000maysoonbereversed.
Figure1:EvolutionofforeignexchangeratestotheUSDforselectedOECDCountriesFrom1stquarter2011to4thquarter2015.Exchangeratesinpercentagechangewithbaseyear2011setto100.
Source:IMF
Figure2:EvolutionofforeignexchangeratestotheUSDforBRICCountriesplusArgentinaFrom1stquarter2011to4thquarter2015.Exchangeratesinpercentagechangewithbaseyear2011setto100.
Source:IMF
15
VI.Conclusion
Thispaperhasexploredtherelationshipbetweenexchangeratesandindustrialpayinequalityin thirteen different countries, each ofwhich either has experienced floating exchange ratesthroughouttheperiodoradoptedsuchapolicyatadefinitemomentduringtheperiodunderexamination.We find that the local currency exchange rate against the dollar is a dominantdeterminant of the movement of inter-industrial pay inequality once exchange rates arepermittedtofloat.
The underlying logic is straightforward: export-oriented sectors are relativelywell-paid, andtheirearningsinlocalcurrencyriseautomaticallywhenthenationalcurrencydevalues,henceinflatingthepaygapbetweenexport-andnon-exportsectors.Sincewealreadyknowthatinter-industrialpayinequalityisastrongdeterminantofhouseholdincomeinequality,itfollowsthatthe exchange-rate is a uniquely powerful determinant of income inequality aswell, inmanydevelopedanddevelopingcountries.
This finding leads toward a number of general conclusions about the nature of economicinequalities inaglobalizedworld.Themost importantof these is tounderscore that inmostcountrieseconomicinequalityisnotunderthecontroloflocalpolicy,anddoesnotdependontechnologyoreducationintheshortrun.Itisratherdrivenbyinternationalfinancialandglobalmacroeconomicforces;anditcanbeheldatbayonlybyeitherwallingoneselfoff fromthoseforces,throughcapitalcontrols,orbybringingthevolatilityandtheasymmetricalbiastowarddevaluationindevelopingcountriesundersomenewformofinternationalcontrol.
16
References“Canada’sStateofTrade:TradeandInvestmentUpdate2014.”2016.AccessedJanuary30.http://www.international.gc.ca/economist-economiste/performance/state-point/state_2014_point/index.aspx?lang=eng.
“EstablishmentoftheBrettonWoodsSystem-ADetailedEssayonanImportantEventintheHistoryoftheFederalReserve.”2016.AccessedJanuary30.http://www.federalreservehistory.org/Events/DetailView/28.
“History,BankofEngland.”2016.AccessedJanuary30.http://www.bankofengland.co.uk/ABOUT/Pages/history/default.aspx.
“Issues-OrganisationforEconomicCo-operationandDevelopment.”2016a.AccessedJanuary31.http://www.oecd.org/forum/issues/oecd-forum-2015-income-inequality-in-figures.htm.
“MonetaryPolicy|MonetaryPolicy|БанкРоссии.”2016.AccessedJanuary31.http://www.cbr.ru/Eng/DKP/.
“MonetaryPolicy.”2016.AccessedJanuary31.http://www.mas.gov.sg/Monetary-Policy-and-Economics/Monetary-Policy.aspx.
“ReserveBankofIndia-Projects.”2016.AccessedJanuary30.https://www.rbi.org.in/Scripts/Project1.aspx.
“TradeProfiles.”2016.AccessedJanuary31.http://stat.wto.org/CountryProfile/WSDBCountryPFView.aspx?Country=CL&Language=S.
Banks,Gary.2004.“StructuralReformAustralian-style:LessonsforOthers?”OECD,EconomicSurveyofAustralia.http://www.oecd.org/australia/39218531.pdf.
Connolly,Bernard.1995.TheRottenHeartofEurope.TheDirtyWarforEurope’sMoney.London,U.K.:FaberandFaberLimited.
Galbraith,JamesK.,andHyunsubKum."Estimatingtheinequalityofhouseholdincomes:Astatisticalapproachtothecreationofadenseandconsistentglobaldataset."ReviewofIncomeandWealth51,no.1(2005):115-143.
Galbraith,JamesK.,andHyunsubKum.2005.“EstimatingtheInequalityofHouseholdIncomes:AStatisticalApproachtotheCreationofaDenseandConsistentGlobalDataSet.”ReviewofIncomeandWealth51(1):115–43.
Galbraith,JamesK.,BéatriceHalbach,AleksandraMalinowska,AminShams,andWenjieZhang."TheUTIPGlobalInequalityDataSets1963-2008."UniversityofTexasInequalityProject,(2014).
Galbraith,JamesK.,DeepshikhaRoyChowdhury,andSanjeevShrivastava."PayInequalityinManufacturing,1979-98."EconomicandPoliticalWeekly(2004):3139-3148.
Galbraith,JamesK.,JaeheeChoi,BéatriceHalbach,AleksandraMalinowska,andWenjieZhang."Acomparisonofmajorworldinequalitydatasets."UniversityofTexasInequalityProject,(2015).
Roberts,KennethM.2012.“MarketReform,Programmatic(De)alignment,andPartySystemStabilityinLatinAmerica.”ComparativePoliticalStudiesSAGE.
SouthAfricanReserveBank.2015.“ExchangeRatesandExchangeControl.”http://www.reservebank.co.zaz
17
Datasources“IMFData-HomePage-Slide1.”2016.AccessedJanuary30.http://data.imf.org/?sk=7CB6619C-CF87-48DC-9443-2973E161ABEB.
“Population,Total|Data|Table.”2016.AccessedJanuary30.http://data.worldbank.org/indicator/SP.POP.TOTL.
“UNIDOStatisticsDataPortal.”2016.AccessedJanuary31.https://stat.unido.org/.“UTIP-UniversityofTexasInequalityProject ”2016.AccessedJanuary30.http://utip.gov.utexas.edu/data.html.
18
AppendixI:Furtherevidencefromothercountries
Africa
19
Asia
20
Europe
21
22
LatinAmerica
23
MiddleEast