Eastern European Export Performance during the Transition

Board of Governors of the Federal Reserve System International Finance Discussion Papers Number 562 September 1996 EASTERN E~OPEAN EXPORT PERFORMANCE D~G THE TRANSITION Nathan Sheets and Simona Boata Note:International Finance Discussion Papers are preliminary materials circulated to stimulate discussion and critical comment. References inpublications to International Finance Discussion Papers (other than an acknowledgement that the writer has access tounpublished material) should be cleared with the author or authors.

Abstract Duringthepastdecade,EasternEuropeanexportshaveundergoneadeeptransformation, ascommunistbloctradingrelationshipshavecollapsedandtradewiththeWesthasincreased. The extent of this geographicalre-orientationhas generallyexceeded the predictionsof equilibriummodelsdevelopedbyHamiltonandWinters(1992)andCollinsandRodrik(1991), suggestingtheprospectforincreasedexportactivityamongthetransitioneconomiesasaggregate demandinthesecountriesstrengthensandpaymentsystemsmature. Significantchangesinthe productcompositionof EasternEuropeanexportshave accompaniedthe geographicalreorientation. Exportsof manufacturinggoodsto formercommunistcountrieshavedeclined sharply,butexportstotheECacrossanarrayofgoods– includingheavymachinery--have grownrobustly. Evidencesuggeststhattheobservedchangesinexportcompositionreflectthe redirectionof physicaIgoods throughprice competition and the emergenceof marketdeterminedcomparativeadvantage. KeyWords:TransitionEconomies,InternationaTl rade,ComparativeAdvantage,European Integration.

EasternEuropeanExportPerformanceduringthe Transition NathanSheetsandSimonaBoata” 1. Introduction Duringthepastdecade,the sticture of exportsfromEasternEuropeancountries-- Poland, Hungary, Czechoslovakia, Romania and Bulgaria -- has undergone a deep transformation. Boththe geographicalorientationand the productcompositionof Eastern Europeanexportshavechangeddramatically.Thefist partofthispaperusesIMFDirection of TradeData to examinethe declinein EasternEuropeanexportsto Councilfor Mutual fionomic Assistance(CMEA)countriesandtheaccompanyingincreaseinexportactivitywith WesternEurope.Notably,themagnitudeofthegeographicarle-orientationhastendedtoexceed thepredictionsofequilibriummodelsdevelopedbyHamiltonandWinters(1992)andCollinsand Rodrik(1991). It is thusreasonableto anticipatethat tradingactivitybetweenthe former communistcountrieswill rebound,as aggregatedemandin thesecountriesstrengthensand paymentsystemscontinuetomature. The secondpart of this paper studiesthe changingproductcompositionof exports, UnitedNationsdataindicatethatEasternEuropeanmanufacturingsubsectorshaveexperienced large declinesin exportsto the CMEA,as well as substantialincreasesin exportsto the EuropeanCommuni~(EC). Twotier conclusionsarealsowarranted.First,thedatasuggest %e authorsare staffeconomistandresearchassistant,respectively,intheDivisionof IntematiomlFinance,Boardof GovernorsoftheFederalReserveSystem. Theauthorswish to acknowledgeJayBryson,DavidHoward,PrakashLoungan.iS,tanislavPolakandDavid Wongforhelpfil commentsanddiscussion.Thispaperrepresentstheviewsof theauthors andshouldnotbe interpretedasreflectingtheviewsoftheBoardofGovernorsofthe FederalReserveSystemor othermembersof itsstaff.

that exports from the Eastern European countries, particularly Czechoslovakia and Poland, are adjustinginamanner consistentwithunderlying comparative advantage. Second, there issome evidence that declines in manufacturingexportsto the CMEAin a givenindustryare significantly related to contemporaneous increases in EC exports by the same industry, suggesting that physical goods have been redirected horn CMEA to EC markets. 2. Geographical Orientation ofEastern European Exports Through the 1980sandearly 1990s,theexport patterns oftheEastern European countries were significantly influenced by the structure of the trading system established by the Council for Mutual Economic Assistance.1 Two characteristics of the CMEA trading structure were particularly significant.2 First, the CMEA discouraged trade with the West. Second, Schrenk (1992) notes that “monopolistic domestic and regioml sellers’ markets removed incentives for producers to keep up with international standards of product and process technology.” As a result,Eastern European countries became increasingly dependent on their CMEA partners, particularly the USSR, to purchase manufactured goods that were not competitive in world markets. IDuringthe1980s,theCMEA includedPoland,Hungary,Bulgaria,Romania, CzechoslovakiaE,astGermany,USSR,Mongolia,Vietnam,andCuba. Yugoslaviaw, hich wasanassociatemember,isalsogroupedwiththeCMEAinthenumbersreportedbelow. 2ForadetafiedassessmenotftheCMEAtradingsystemseeSchrenk(1992)andLavigne (1995). 2

Table2.1,whichreportsthegeographicadlistributionofEasternEuropeanexportshorn 1985-88,clearlyreflectstheimpactof theCMEA.3TradewithWesternindustrialcountries accountedfor onlyabout35 percentof totalexportsfromPoland,Hungary,Romania,and Czechoslovakia.Lessthana quarterof Bulgarianexportswentto industrialcountries. By contrast,the CMEAaccountedfor 55 percentof the exportsfromHungary,Bulgaria,and Czechoslovakiaandover40percentoftheexportshornPolandandRomania.4Roughlyhalf oftheEasternEuropeancountries’CMEAexportswentto theUSSR. SeveralauthorshaveattemptedtodeterminewhatthegeographicadlistributionofEastern Europe’stradewilllooklikeonceimpedimentstoWesternexportshavebeenremovedandthe artificialdependenceontheCMEAhasbeenreversed. Twomethodologicaalpproacheshave beenusedtoaddressthisissue. Thefirstapproach,pursuedbyHamiltonandWinters(1992) andothers[e.g., HavrylyshynandPritchett(1991),WangandWinters(1991),andBaldwin (1994)]involvesestimatinga “gravitymodel”usingdatafromacross-sectionofcountries.The secondapproach,proposedbyCollinsandRodrik(1991),estimatesequilibriumtradepatterns usingdatafrom1928asa baseline. ThegravitymodelpostulatedbyHamiltonandWinters(H’Wa)ssumesthatexportsfrom agivencountrytoanothercountryrisewiththeGDPofeachcountry. Exportsdeclineasthe populationofeachcountryrisesandasthedistancebetweentheWOcountriesincreases. The 3ThedatainTable2.1 areobtainedfromIMFDirectionofTradeData. TheIMF convertstradeflowsdenominatedin foreigncurrenciesintoU.S.dollarsusingperiodaverage exchangerates. 4Rodrik(1992)observesthatthesedatamayoverstatethemagnitudeofCMEAtrade flows. CMEAtradewasdenominatedintheso-called“transferableruble,”whichwas maintainedat artificiallyappreciatedlevelsvis-a-visthedollar. 3

gravity model may be interpreted as follows. Domestic GDP proxies for a country’s export supply capacity, while foreign GDP is a proxy for the foreign country’s demand for domestic exports. The population variables proxy for the size of a country’s economy. Larger economies, holding all else equal, are generally less open than smaller economies, perhaps becauselarger economiesare abletoachievemore intemaldiversification. Thedistancebetween the two countries is a proxy for transactions costs, particularly the cost of transportation. HW observe thatthe absence of prices andexchange rates in the gravi~ model does not suggestthat such variables are not important; instead, HW argue that the gravity model is a reduced form description of a long-run equilibrium, where prices and exchange rates are themselves detetied by the variables included in the model. Hamilton and Winters use data from 76 market economies (19 industrial countries and 57 LDCS)averaged over 19W-86 to estimate their gravity model. They assume that themodel that best describes the trading patterns of the 76 market economies in their sample will also describe the long-run trading relationships ofthe former CMEA countries. A weakness ofthis approach is that the results depend on estimates of USSR and Eastern European GDP; such estimates may vary by as much as five times. Hamilton andWinterscompromiseonthisissue by workingwith Heston-SummersGDP figures,whichfdl in the middleof the rangeof publishedestimates. A secondapproachto determinigngthelong-runtradingpatternsof Eastern European countrieswasdevelopedby CollinsandRodrik(1991),who generate estimates using trading patternsin1928asabaseline.CollinsandRodrik (CR) focus on 1928for two reasons. First, they note that 1928is perhapsthe last year that “typical”tradingrelationsexistedbetween 4

EasternEuropeancountriesandothercountries. Afier 1928,tradewasdistortedby increasing protectionism,theGreatDepression,andthe spreadofcommunism. Second,CRchoose1928 becausethe data happentobeavailable. Aworldtradematrixfor 1928wascompiledbythe Leagueof Nations. (Thesedata are summariezd in Table2.2.) Notably,65 percentof Bdgarianexportsin1928weredirectedtowardtoday’sECcountries,whileonly25percentof Hungarianexportswentto the EC. Conversely,EasternEuropeaccountedfor one-thirdof Hungarianexportsbut only 12percentof Bulgarianexports. The SovietUnionwas not a significanttradingpartneroftheEasternEuropeancountriesin 1928. Usingthis data, Collinsand Rodrikexaminethe followingquestion:If the Eastern Europeancountrieshadremainedpartoftheworldtradingsystem,howwouldtheir1989trading patternshavedifferedfromtheir1928tradingpatterns?Toaddressthisissue,CRexaminethe evolutionofthetradingpatternsof six “comparator”countries- Austria,Finland,Germany, Italy,Portugal,andSpain. CRregressthesecountries’1989tradeshareswith36countrieson thecorresponding1928tradesharesanddummyvariablesforeachpartner. Thepartnerdummy isincludedtocapturethechangingimportanceofsomecountriesinworldtrade. Forexample, Japan’sprominencehasincreasedsignificantlysince1928,whiletheUK’srolehasdeclined. CRusetheresultingregressionequationtodetermineequilibriumtradesharesfortheEastern Europeancountries.ThisprocedurehastheadvantageofnotdependingonGDPestimatesfor thetransitioneconomies.Itmaybe flawed,however,ifthecomparatorcountriesarenottruly comparableor if 1928tradedataarenota goodbasisforpredictinglong-runtradingpatterns. WenotethatneitherHWnorCRattemptto sayanythingaboutthedynamicsofadjustmentto thelong-runequilibrium. 5

Tables 2.3-2.6 present the Hamilton-Winters and Collins-Rodrik predicted export shares for the five Eastern European countries with various ~ading partners. Actual 1994 export shares, derived from IMF Direction of Trade Data, are also reported.50GTable 2.3 indicates several conclusions. First, boti H’wandCRpredict significantincreases inthe share ofEastern European exports to EC countries relative to the shares reported in Table 2.1. This suggests, as hypothesizedearlier, that CMEAstructuresdivertedexportsawayfromWesternEurope. Second,HWpredictsignificantlylowerexportsharestotheECforBulgariaandRomania(due totheireastwardgeography)thandoCR. Indeed,Bulgaria’sstrongexportorientationtoward WesternEuropein 1928leadsCRtopredictthatBulgariawillshipalargershareofitsexports to the EC thanany other EasternEuropeancountry. Third, Table2.3 showsthat the 1994 exportsharesforPolandandCzechoslovakiahavesi~lcantiy overshottheequfiibriumlevels forecastbyHWandCR. Moregenerally,there-orientationofEasternEuropeanexportsfrom EasttoWestappearstohavebeenrapidandsubstantial.RelativetothedatainTable2.1,the EC’s share of EasternEurope’sexportsin 1994morethandoubledfor all countriesexcept Romania,reflectingrapidgrowthinthevolumeofexportsto theEC. The data in Table2.4 generallysupportthe threeconclusionsdrawnhorn Table2.3. BothHWandCRaccuratelypredictadeclineinEasternEuropeanexportstotheformerCMEA countries;however,HWandCRbothtendtounderestimatethemagnitudeofthedecline. The 5TheCollins-Rodrikestimatesandthe 1994actualdataincludePoland,Hungary, CzechoslovalciaB, ulgaria,Romania,theformerYugoslaviaandtheformerSovietUnionas CMEAcountries. TheHamilton-WintersestimatespresentedheredifferslightlyfromHW (1992)becausetheseestimateshavebeenadjustedto includeEastGemy aspart oftheEC andYugoslaviaas partof theCMEA. 61994datafor Czechoslovakiaare thesumofdatafortheCzechRepublicandtheSIovak Republic,excludingtradewitheachother. 6

exportsharesof Poland,Romania,and Czechoslovakiato the formerCMEAcountriesare substantiallybelowtheirpredictedvalues. Table2.5 decomposesthedifferencebetweenthe equilibriumexportsharespredictedby CR and 1994actual export sharesinto a portion attributableto EasternEuropeanda portionattributableto the formerSovietUnion. (H’W estimatesarenotsufficientlydisaggregatedtoallowthisdecomposition.)Theaverageshortfall toEasternEuropeis4.2percentoftotalexports,whiletheaverageshortfalltotheFSUis3.8 percentoftotalexports. Thissuggeststhepotentialforexpandedtradingactivity,bothwithin EasternEuropeandbetweenEasternEuropeandtheformerSovietUnion,asaggregatedemand inthesecountriesstrengthensandpaymentsystemsmature. Finally,the 1994tradingpatternsofCzechoslovakiaandPolandarestrikinglysimilar. Thesecountrieshavepursuedrapideconomicreform,includingliberalizationof theirtrade regimes,andappeartohavesuccessfullyrestructuredtheirtradingrelationships.Bythesame token,the1994tradingpatternsofBulgariaandRomaniaarealsosimilar.Thesecountriesshare eastwardgeographyandhavereformedmoreslowly.No@blyT, able2.6showsthatBulgaria’s andRomania’s1994tradeshareswith“OtherCountries”(i.e., thoseoutsidetheECandthe formerCMEA)are significantlylargerthanfor the otherthreeEasternEuropeancountries. BulgariaandRomaniaexportsignificantlyto Turkeyandto countriesintheMiddleEastand Asia. 3. The ProductCompositionof EasternEuropeanExports In the previoussection,we documentedthe significantand rapid geographicalreorientationofEasternEuropean exportsthathasoccurredinrecentyears. Inthissection,we 7

examine the changing product composition of Eastern European trade, in an effort to better understand the characteristics of the geographical re-orien~tion. The discussion in this section is organized around three benchmark hypotheses. Each of the hypotheses offers a contrasting perspective on how the collapse of CMEA trading structures and economic liberalization potentially could affect the composition of Eastern Europe’s exports. We assess the empirical validity of these hypotheses using United Nations trade data. Hypothesis1. It is widelybelievedthatCMEAtradingregulationssignificantlydistortedthe compositionoftradebetweenthe communistcountries. SujanandSujanova(1995)note,for example,that“tradewithintheformerCMEAwaspredominantlybasednotonmturalmarket principles,butratheronbureaucraticcommandsforcinganartificialdivisionoflabor.” Several researchershavesuggested,however,thatthecompositionoftradewiththeWest,whichdiffered significantlyhorn the compositionof CMEAtrade, may have been less affectedby these regulations.’Inotherwords,evidencepresentedaboveindicatesthatCMEAstructuresreduced the volume of trade with the West; these structures,however,may not have significantly distortedthecomposition of that trade. Collins and Rodrik (1991)notethat “probablythebest indicatorwe haveof EasternEurope’scomparativeadvantagepatternis that reflectedin its currenttradewiththe West.” It maythusbe reasombleto hypothesizethat the equilibrium compositionofEasternEuropeanexportswillbesubstantiallymorelikeitspre-reformtradewith theWestthanitspre-reformtradewiththeCMEA. Thishypothesis,whichwewilladoptasone benchmark,hastwoempiricalimplications.First,ifEasternEurope’stradewiththeECduring ‘See~ “ ki (1993)andBohata(1995). a

the1980swasreflectiveofcomparativeadvantage,weshouldnotobservesharpchangesinthe compositionofEasternEuropeanexportstotheECfollowingtheliberalizationoftrade. Second, overtime,thecompositionofEasternEurope’sexportstotheCMEAshouldbecomemorelike its exportsto the EC, i.e., trade with the East shouldadjust in a manner consistentwith comparativeadvantage. Hypothesis FollowingthecollapseoftheCMEAtradingstructuresandthecontemporaneous 2a contractionin aggregatedemandintheformercommtist countries,manyEasternEuropean exporters,whoformerlysoldgoodstotheCMEA,mayhaveattemptedtomarkettheirproducts in the West. Sincegoodssoldin the CMEAwereotin of lowerqualitythancomparable Westernproducts,thisscenariosuggeststhateithergoodsweresoldat a discountrelativeto Westerngoods(“pricecompetition”)or thatexportersrestructuredproductionto improvethe qualityof theirproducts(“industrialrestructuring”).Thishypothesisimpliesthatgrowthin exportsto theEC shouldbe concentratedin industriesthatexperienceddecliningsalesto the CMEA. Hypothesis3. Theanalysisinthispaperhas,thusfar,emphasizedthedistortionsgeneratedby the CMEAtradingregime. There were, of course,manyotherdistortionspresentin the centrallyplannedEasternEuropeaneconomies.Forexample,relativepricesgenerallydidnot reflectunderlyingscarcity,and the policiesimplementedby centralplannersofiencreated incentivesthatwereinconsistenwt itheconomicconsiderations.Asaresult,evenifthetrading regimewith the Westdid not itself introducesizeabledistortions,the economiesof these 9

countries were sufficientlydistorted inother respects that Western trade maynot have reflected — comparativeadvantage. This hypothesis suggests that Eastern European exporters, in response to the implementationofeconomicreformsandintematiomlcompetition,willbe requiredto search outtheirmarketdeterminedcomparativeadvantage,whichmaynotresembleprevioustradewith eithertheECortheCMEA. Underthishypothesis,significantchangesinthestructureoftrade withbothregionsshouldbeexpected,asindustriespossessingcomparativeadvantageexpandand develop. Inaddition,thecompositionoftradewiththeECandtheCMEAwouldlikelybecome moresimilar,astradewithbothpartnersbecomesmorereflectiveofcomparativeadvantage. IndexofSimilarity.Asameasureofthesimilarityoftheproductcompositionofexportsacross tradingpartnersandyears,weintroducethefollowingindex: S(jl, j2)(&i,tz)=l-+~ Ujlk&x- ‘j2kt2 (1) k-l (2) Equation (1)definesthesimilaritybetweenagivencountry’svectorofexports to countryjl at timetl anditsvectorofexportsto attime~. Thecoefficient~jbis commodityk’s COuntry j2 shareintheexportstocountryj attimet. Thisindexmapsthesimilaritybetweentwoexport vectorsintothe[0,1]intervai. Specifically,ifthetwoexportvectorsare identical,thenS= 1. If the two exportvectorsare completelydisstiar, i.e., all non-zeroentries in eachvector 10

correspondto zero entriesinthe othervector, then S=0. We notethat Equation(1)is similar in spirit to an indexof similarityused by Rodrik (1992)andKaminski(1993). Compositionof TradeData. Theempiricalanalysisinthissectionusesdataontheproduct compositionoftradedrawnfromtheUnitedNations“SeriesD“tradedatabase.Inaneffortto identi~ broaddevelopmentsin the compositionof EasternEuropeanexports,we startthe analysisbyexaminingannualdataforonedigitindustries.Asthepaperprogresses,wefocus ontwodigit manufacturingindustries.8 Theavailabilityofdatavariessomewhatacrosscountries. For Hungary,thedata~ horn 1985-94,for Polandhorn 1985-93,andfor Romaniafrom1989-94. Reliabledatafor Czechoslovakiaareonlyavailablefor1989-90and1994,whiledataforBulgariaarenotincluded intheUNdataset. Thedataarereportedf.o.b., expressedinU.S.dollars,andaregenerally compatiblewiththeIMFdirectionoftradedatausedpreviously.Thedatahavebeendeflated bytheU.S.consumerpriceindexandarestatedintermsofconstant1990dollars. (Thechanges inEastern Europeanexportpatternshavebeensodramaticthatthepaper’sresultsarebroadly unchanged,independentof whethertheunderlyingdataare expressedin currentor constant dollars.) Tables3.1-3.4presentonedigitcompositionoftrade&ta--bothlevelsandshares– for eachofthefourcountriesforwhichdataareavailable.Tables3.5and3.6presentcoefficients ofsimilaritybetweenvectorsofexportstotheECandtheCMEA.Inallfourcountries,expom totheCMEAhavefallensharplysincetheimplementationofmarketreforms,decliningbyover 80nedigit industriesandtwodigitmanufacturingsubsectorsarelistedintheAppendix. 11

60 percent. This decline has been concentrated in the three onedigit industries principally engagedinmanufacturing- ManufacturedGoods, Miscell~eous Manufacturesand, particularly, Machinery and Transport Equipment. The contraction inexportshorn these industries accounts for more than 80 percent of the decline in each country’s total exports to the CMEA. On the otherhand, Tables 3.1-3.4 alsoshowstrikinggrowth inexportstotheEC horn Hungary, Poland and Czechoslovakia, increasing by more than 150percent over the period. Notably, over 80 percent of the net increase in exports to the EC is attributable to the same three manufacturing industries that registered sharp declines in exports to the CMEA. We note that Romania, while experiencing reduced exports to the CMEA, similar to the other countries, did not achieve an offsetting increase in exports to the EC. This partially reflects a sharp decline in exports of Mineral Fuels but, in addition, Miscellaneous Manufactures was the ody Romanian onedigit manufacturing industry that experienced meaningful growth in EC exports. Evaluation ofHypothesis 1andHypothesis3. Wenowusetheindexofsimilaritypresented above- Equation(1)– toevaluateHypothesis1andHypothesis3.9 Underbothhypotheses, thecompositionof exportsto the ECandCMEAshouldbecomemoresimflarovertime. In addition,bothhypothesessuggestthatthecompositionofexportstotheCMEAwillchange,as trading patterns respond to the forces of underlyingcomparativeadvantage.l” The two vests ofHypothesis2willbedevelopedandpresentedbelow. 1%testingthesehypotheses,wechoosea fairlystrongempiricalanalogof comparative advantage,i.e., werequiretheproductcompositionofEasternEuropeanexportsto theEC andtheCMEAto becomemoresimilar. Underthisinterpretation,comparativeadvantage requiresnotonlythatEasternEuropeexportthesamesetofgoodstobothpartners,butalso thatthegoodsbe exportedinroughlyequalproportions. 12

hypothesesdiffer,however,intheirpredictionsaboutthebehaviorofexportstotheEC. Under Hypothesis 1, the compositionof exportsto the EC shouldnot changesignificantly,sinceEC exportswerenotsubstantiallydistortedbytheCMEAtradingsystem. Bycontrast, Hypothesis 3envisagessignificantchangesinECexports;inefficienciesinthedomesticeconomytendedto distorttrading patterns, even if the trade regimeitselfwas not distortionary. The coefficients of similarity reported in Table 3.5 indicate that the compositionof exportstotheCMEAandtheECbecamemoresimilarinCzechoslovakiaP,olandandHungary, consistentwiththepredictionsofHypothesis1andHypothesis3. Theconvergencebetween CMEAandECtradeisobservedmostclearlyinthecaseofCzechoslovakiaw, herethesimilarity measuremovedfrom0.505in1989to0.860in1994.Czechoslovakia’EsCandCMEAexports movedhornbeingtheleastsimilaracrossthesecountriesatthebeginningoftheperiodtobeing the mostsimilarat the end of the period. Polandalsoregistereda substantialincreasein similarity,movinghorn 0.549 to0.728. Atthebeginningoftheperiod,Hungary’sECand CMEAtradeweremoresimilarthanthatoftheotherthreecountries,perhapsdueto thefact thatHungarywasthef~sttoimplementmarketreforms.11Duringsucceedingyears,however, thecompositionofHungary’stradewiththeECandtheCMEAhasbecomeonlyslightlymore similar,possiblyreflectingthegradualnatureofHungarianreform. ThecoefficientsinTable3.6suggestthatthestructureofCzechoslovakia’EsCtradewas relativelystableduringthe period, registeringa similaritycoefficientof 0.851, whilethe structureofitsCMEAtradechangedsi~lcantly, withacoefficientof0.588. Czechoslovakia’s llRodrik(1992)notesthatinHungary“considerabledecentralkationandmarket-oriented reform”beganasearlyas 1968. Bythelate1980s,centralplanningwas“largelydiscarded” andenterpriseshada “largedegreeofautonomy”. 13

CMEA trade changedmorethanthatofanyothercountry,whileitsECtradechangedlessthan anyothercountry’s. Bycontrast,thecompositionofPoland’sECtradeactuallychangedslightly morethanthe compositionof its CMEAtrade, registeringa coefficientof 0.677 for the EC versus 0.703 fortheCMEA. Basedontheseresults,thetradingperformanceofCzechoslovakiaseemswell explained byHypothesis1. ThecompositionofitsCMEAandECexportshasbecomemoresimilar,while its EC tradehas remainedrelativelystable. This suggeststhat the underlyingstructureof Czechoslovakia’seconomymay havebeen less distortedduringthe CMEAperiodthanthe economiesof the other countries. Poland’sexport performanceappears consistentwith Hypothesis3. The similaritybetweenCMEA and EC exportshas increased,along with significantchangesin the structureof both EC and CMEAtrade. This resultsuggeststhat Poland’seconomyisrestructuring,withexportersrespondingtomarketincentivesbysearching out comparativeadvantagein new industries. Hungary’sexportperformanceseemsweakly consistentwithHypothesis3,registeringa slightincreaseinsimilaritybe~een ECandCMEA tradeand substantialchangesin the compositionof exportsto bothregions. Neitherof the hypothesesseemsto describeadequatelythechangingpatternsofRomaniantrade;exportsto boththe ECandtheCMEAchangedsubstantially,butthe compositionofexportsto the two regionsbecamelesssimilarbetween1989and 1994. AnalysisofTwo-DigitManufacturingData. Ouranalysisoftheexportperformanceofonedigitindustrieshasshownthatthemanufacturingindustriesexperiencedthelargestdeclinesin exportsto theCMEAandthestrongestincreasesin exportstotheEC. Giventhesedynamic 14

changes,we believethatthe manufacturingsubsectorsare the appropriateplaceto lookfor furtherevidenceto evaluateHypothesis2 andHypothesis3. In this section,therefore,we initiate a more detailed examinationof the export performanceof Eastern European manufacturers. Thisexaminationusesdatafromthe 19twodigit manufacturingsubsectors includedintheUnitedNationsdata. ThesesubsectorsarelistedintheAppendix.Thedataare againexpressedintermsofconstant1990dollars Table3.7 presentsa simplecross-tabthatcategorizes76 country-subsector(si.e., 19 manufacturingsubsectorsineachofthefourcountries)basedonwhethertheyhaveexperienced increasesordecreasesinexportstotheCMEAandtheECduringthetransitionperiod.12Before discussingTable3.7, however,weprovidesomeinterpretationofeachofthecross-tab’sfour quadrants. Twosignificantfactorsmaygenerateobservationinthenortheastquadrantofthecrosstab. First,asnotedabove,CMEAtradingstructurestendedtoartificiallyencouragetradewith theCMEAandrepresstradewiththeWest. Astheseeffectsarereversed,exportstotheWest shouldriseandexportstotheCMEAshouldfall,holdingallelseequal. Second,overthelast severalyears,economicactivityinmanyCMEAcountries– particularlythoseinthefomler SovietUnion- has contractedsharply,causinga correspondingdeclinein CMEAexport demand.Inresponse,EasternEuropeanproducershavehadincentivestoredirectexportstothe West. WeakdemandintheCMEAanddiversionofexportstowardstheECwouldalsocause industriestofallintothenortheastquadrantofthecross-tab.Inthediscussionbelow,wewill 12Givendataconstraints,the “transitionperiod”isdefinedas 1985-94forHungary,1985- 93forPoland,and 1989-94forCzechoslovakiaandRomania. 15

dividethe industriesin the northeastquadrantinto two groups: “Group 1“ includesthose industriesforwhichtheincreaseinexportstotheECisgreaterinmagnitudethanthedecrease in exportsto the CMEA,and “Group2“ consistsof industriesfor whichthe declineto the CMEAis largerthantheincreasetotheEC. Group1industrieshaveexperiencednetgroti duringthetransitionperiod,whileGroup2 industrieshavecontracted. Industriesmaydeviatehorn the northeastquadrantfor a numberof reasons. First, dynamicindustries--includingthosewithemergingcomparativeadvantage– maysuccessfully “swimup stream”andincreaseexportstotheCMEA,despitethedeclineineconomicactivity andother factorsthathavetendedto reducetrade withthe East. Thesedynamicindustries, whichare consistentwithHypothesis3, willfall intothe northwestquadrant. Alternatively, observationsmayfallintothesoutheastquadrant– declinesinbothCMEAandECexports– ifindustrialrestructuringrequiressomeindustriestodownsize,inordertofreeresourcesforuse inmoreefficientsectors. Anotherexplanationforobservationsinthesoutheastquadrant,which maybe particularlyapplicableto Romania,is thatadministrativecontrolsimplementedunder centralplarmingmayhaveencouragedexcessiveexternaltrade,repressingconsumption.13The removalof suchcontrolsmayleadto acrossthe boardreductionsin exports,as consumption rises. Finally,thereisnocompellingreasonforobservationstofti intothesouthwestquadrant. Undercurrentcircumstances,anindustrythatissufficientlydynamicto increaseexportstothe CMEAshouldalsobe sufficientlydynamicto increaseexportsto theEC. The datain Table3.7 indicatetit 50 of the 76 manufacturingindustriesfallintothe northeastquadrant.Oftheseindustries,28experiencedincreasesinexportstotheECthatwere 13SeeWilliamson(1991),pp. 76-78. 16

greater than the decrease in exportsto the CMEA;whilein 22 indusrnes, the increasein EC exportswas lessthan the declinein CMEAexports. As mentionedabove, we call the former setof industriesGroup 1andthe lattersetof industriesGroup2. Theseindustriesare listedin thef~st and secondcolumnsofTable3.8. Group1industriesconstitutea fairlydiverseset. Clothingexportsofallfourcountries registeredstrongincreasesto the EC, far offsettingdeclinesto the CMEA. Specifically, Poland’sclothingexportstotheECincreasedby$1.2billion.HungarianandRomanianclothing exportsto theEC eachincreasedby over$500million,andCzechoslovakiaexperiencedan increaseof$250million.OthercommonGroup1industriesareFurniture,NonmetallicMineral ManufacturesM, etalManufacturesa,ndMiscellaneouMs anufacturing.Group1includeseleven Hungarianindustries,eightPolishindustries,eightCzechoslovakindustriesandoneRomanian industry. Group2 industriestendto beinheavymanufacturingandsophisticatedproducts. For example, exports of Power Machinery,Electric Equipment,Transport Equipment,and Instruments,Watchesand Clocksaccountfor twelveof the twenty-twoGroup2 countryindustries.TheseindustriesgenerallyconstitutedlargesharesofEasternEurope’sCMEAtrade andtendedtobeoverbuiltduringcentralplanning. Adeclineinexportsfromsuchindustries may be indicativeof economicrestructuring. This observation,however,shouldnot be overstated. Poland’sexportsto the EC of TransportEquipmentgrewby over$800million, exceedingthecorrespondingdeclineinexportstotheCMEA. Moreover,someof theheavy manufacturingindustriesin Group2 registeredsizeableincreasesto the EC. Czechoslovak exportstotheECofElectricalMachineryandTransportEquipmentincreasedby$420million 17

and$600million,respectively,duringthetransitionperiod,andHungarianexportsofElectrical Machineryrosebyover$500million.Theseincreases,althoughsomewhatsmallerinmagnitude thanthecorrespondingdeclinestotheCMEA,werestillsubstantial. Fifieensubsectorsfallintothenorthwestquadrant,achievingincreasesinexportstoboth theECandtheCMEA. AsshowninthethirdcolumnofTable3.8, theseindustriestendtobe in natural resourcesand light manufacturing.14For example,three countriesexperienced increasesineachofWoodandCorkManufactures,PaperandPaperboardManufactures,and Plumbing,HeatingandLightingEquipment.lsTwocountriesexperiencedincreasesineachof LeatherandDressedFurandNon-FerrousMetals. Thestrongperformanceofthesenorthwest Q quadrantindustriesduringthetransitionsuggeststhattheymaypossesscomparativeadvantage andholdtheprospectforfiture exportgrowth. PolandandtheCzechoslovakiaeachhavefive subsectorsinthenorthwestquadrant,whileRomaniahasthreeandHungaryhastwo. Finally, wenotethatallelevenobservationinthesouthernhalfofthecross-tabareRomanianindustries. CoefficientsofSimilarity.Coefficientsofsimilaritycalculatedusingtwodigitmanufacturing data are reported in Tables 3.9-3.10. As before, Czechoslovakia’s export performance is consistentwith the predictionsof Hypothesis1. Czechoslovakexports show a substantial 14DatainTables3.1-3.4suggestthatsevenonedigit country-industriesachieved increasedexportstoboththeCMEAandtheEC, i.e., FoodandAnimais(Poland, Czechoslovakia),BeverageandTobacco(Poland,Czechoslovakia),AnimalandVegetable Oils(Czechoslovakia,Romania),andOtherCommodities(Poland). Theseindustriestendto be involvedwithagriculture,food,andagriculturalprocessing. IsThePlumbing,HeatingandLightingEquipmentindustryincludesa varietyofproducts. Theplumbingproductsare mainlyceramic,ironandsteelplumbingfutures. Heating productsare varioustypesofcentralheatingequipment. Lightingproductsincludelamps, metalbases,andglassware. 18

increaseinthesimilarityof EC andCMEAtrade,withthecompositionofECtraderemaining quite stableand CMEAtrade changingbroadly. Poland’spetiomance again seemsbest explainedbyHypothesis3. Polandexperiencessomeincreaseinthesimilarityof itsECand CMEAexports(althoughnotasstrikingasinTable3.5)andsignificancthangesinthestructure ofbothitsECandCMEAtrade. Hungaryregistersonlya slightincreaseinthesimilarityof itsECandCMEAexports. ThecompositionofitsexportstotheEC,however,remainsmore stable than any other country’s. Hence, Hypothesis1 may best describe Hungarian manufacturingexportPerformance.lbRomaniahasthesamepatternofcoefficientsasinTable 3.5, againsuggestingthatitstradingpatternsarenotwellexplainedbyeitherHypothesis1or Hypothesis3. Evaluation ofHypothesis2. WenowevaluateHypothesis2usingthetwodigitmanufacturing data examinedabove. In the analysisthat follows,we attemptto distinguishbetweentwo scenarios: * Price Competition. In response to declining CMEA demand, Eastern European exportersmayhaveattemptedtomarketessentiallythesamegoodsintheWestthatthey previouslysoldintheCMEA.Thisscenariosuggeststhatphysicalgoodswereredirected fromCMEAmarketsto EC marketsandthat EasternEuropeanexporterscompeted primarilyonthebasisofprice,i.e., lowerpricesweresufficienttocompensateforlower quality.Thesharpdevaluationoftheexchangerate,wh.ichaccompanietdheintroduction l%e resultsforHungarypresentedinTable3.5,bycontrast,wereweaklyconsistent withHypothesis3. 19

of reform programs in many of these countries, may have facilitated this sort of price competition. ~ ~ndus~ia~Restructuring. Eastern European exporters may have concluded that the goods they sold in the CMEA were not of sufficient qualityto be competitive inthe EC, motivating exporters to improve thequalityof their products. In this scenario, declining CMEA exportdemandwouldmotivatefm restructuring that, inturn, wouldprepare the way for increased exports to the EC. While there is no deftitive method for differentiating between these scenarios, we implementthefollowingtest,whichisbasedontheassumptionthattherestructuringnecessary to raisethe qualityof exportsto Westernstandardscouldnot occurinstantaneously.To the extentthat tidamental restructuringhas occurred,therefore,the declinein CMEAexports shouldbe relatedtotheexpansionofECexportswithadiscerniblelag– thetimenecessaryto completerestructuring.Conversely,ifthere-orientationofEasternEuropeantradehasprimarily consistedofredirectingphysicalgoodsfromonemarketto anotherthroughpricecompetition, thereshouldbe a roughlycontemporaneouscorrelationbetweenthedeclineinCMEAexports and the increase in EC exports, since the same goods have been shipped to different destinations.*7 In Table3.11, wereportresultsobtainedfromthe followingregression: I’Giventhenecessityofestablishingdistributionchannelsandmarketing,redirectionof exportsthroughpricecompetitionshouldgeneraterapid(i.e.,withina year)butperhapsnot instantaneoucshangesinthestructureofexports. 20

AEC&=$0+al ACME4ti +a2ACMEAti-1+a3ACMEA&-2 (3) The dependentvariableis the change in the exportsofcountryi totheECingoodk andyear t. The independentvariables are the current and lagged changes in good k exports to the CMEA, as well as NO lags of the dependentvariable. The regressionalso includesdummy variablesfortheTransportEquipmentandClothingindustries. Otherindustrydummies,aswell ascountrydummies,wereinsignificantinpreliminaryspecificationsandweredroppedfromthe regression. TheresultsinTable3.11suggestthatcontemporaneoucshangesinexportstotheCMEA arenegativelyandsignificantlyrelatedto increasesinexportstotheEC,evenafiercontrolling forlaggedchangesinECexportsandindustryeffects. ThelaggedchangesinCMEAexports, however,are notsignificant. Theseresultsare clearlyconsistentwiththepricecompetition hypothesisdiscussedabove. Thisconclusion,however,requiressomequalification.First,the coefficientoncontemporaneouesxportstotheCMEAsuggeststhat,holdingallelseequal,a$1 decreaseinexportstotheCMEAleadsto anincreaseinexpo~ totheECofonly15cents.18 Moreover,theadjustedR-squaredcoefficientsfortheregressionsarelessthan0.2. Theseobservationssuggestthattheredirectionofexportsthroughpricecompetitionis apartialexplanation- butonlyapartialexplanation- forEasternEuropeantradeperformance duringthetransition.Thisconclusion,however,shouldnotbesurprising.First,inlightofthe significantqualitydifferentialsbetweenWesternandCMEAproducts,aportionofEasterngoods l%e sizeofthiscoefficientis fairlyrobust,varyingbetween-0.12and-0.19ina variety ofspecificationsandacrossa numberofdatasubsamples. 21

were probably unmarke~ble in the West, regardless of theprice. For example, Rodrik (1992) points to computer productsand manufacturinggoodsdesignedto satisfySovietproduction processes. Second,the regressionspecificationis unableto accountfor the deep structural factorsthatdeterminelong-runcomparativeadvantage,whicharediscussedinHypothesis1and Hypothesis3. To the extentthat suchfactorshaveinfluencedthe re-orientationof Eastern Europeanexpo~, theexplanatorypoweroftheregressionwouldbe expectedto be low. summary 4. andConclusions ThediscussioninthispapersuggestsseveralconclusionsaboutEasternEuropeanexport performanceduringthetransition.First,substantiaslharesofexportshornthesecountrieshave beenre-orientedfromthe CMEAto WesternEurope. The extent of the geographicarleorientationhasgenerallyexceededthepredictionsofmodelsdevelopedbyHamiltonandWinters and Collinsand Rodrik. It is thusreasombleto anticipatethattradingactivitybetweenthe transitioneconomieswillrebound,asaggregatedemandinthesecountriesstrengthensandtheir paymentssystemsmature. Second,evidenceindicatesthatCzechoslovakexportstotheEC in the late 1980swere broadlyreflectiveof underlyingcomparativeadvantage,suggestingthat Czechoslovakia’seconomyintheyearsprecedingtheintroductionofeconomicreformsmayhave beencomparativelyfreeofdistortions.Poland’sexportperformanceindicatesthatitseconomy is in the process of dynamicrestructuring,with the emergenceand developmentof new industries,apparentlyreflectingcomparativeadvantage. Third, the behaviorof Romanian exportshasdifferedsignificantlyhornthatoftheothercountries.Forexample,overhalfof Romania’stwodigitmanufacturinigndustriesexperienceddecreasedexportstotheECduring 22

the transition period, but none of the twodigit manufacturing industries in other Eastern Europeancountriesregisteredsimilardeclines. Fourth, EasternEurope’sexportsofmachinery and equipmentto the EC have grown rapidly during the transition. Althoughthese increases generallyhavebeensmallerinmagnitudethanthecorrespondingdeclinesinsuchexportstothe CMEA,theyhavestillbeensignificant.Finally,pricecompetitionappearstobeanimportant -- butfarfromcomplete- explanationfortherapidgeographicalre-orientationandchanging product composition of Eastern European exports. Deep structural factors, identified in Hypothesis 1 and Hypothesis3, also appear to have playeda major role in affecting~stem Europeanexport performancein recent years. 23

References Baldwin,Richard. ~ow~ . CentreforEconomicPolicy Research, 1994. Bohata, Marie. “ChangingPatternsof CzechForeignTrade.” Mimeo,GrantAgency of tie CzechRepublicandKIASAN, ovember1995. . . . Collins,SusanandDaniRodrik. ~ope _ SOVWUrllMmtheWo~ EC- Institute for Intematioml Finance, Policy Analyses in Intematioml Economics, No. 32, 1991; Hamilton,Carl and AlanWinters. “OpeningUp InternationalTradewith EasternEurope.” . . ~ 9April 1992,77-104. Havrylyshyn,OlehandLantPritchett. “EuropeanTradePatternsaftertheTransition.” World Bank,WorkingPaperNo. 748,August1991. Kaminski, Bartlomiej. “The Foreign Trade Dimension of the Market Transition in Poland.” World Bank, Working Paper No. 1144, June 1993. . . Lavigne,Marie. F= so~v to ~ . St.Martin’s Press: New York, 1995. Rodrik,Dani. “ForeignTradeinEasternEurope’sTransition:EarlyResults.” NationalBureau ofEconomicResearch,WorkingPaperNo.4064,May 1992. Schrenk,Mb. “CMEATradeandPayments:Conditionsfor InstitutionalChange,” in B. . . . . . . . MilanovicandA. Hillman,eds., me T~ne. D~ . WorldBank,1992. Sujan,IvanandMilota Sujanova. “CzechRepublic: StructuralChangesinExternal Trade 1989- 94.” Mimeo, Czech Republic Statistical OMce, September 1995. Wang, Zhen and Alan Winters. “The Trading Potential of Eastern Europe. ” Centre for Economic Policy Research, WorkingPaperNo. 610, November1991. . Williamson, John. On- of F- E- Institutefor International Finance,PolicyAnalysesinInternationalEconomics,No. 31, 1991. 24

o o“ o 0 o“ o r e n 25

Table 2.2: Export Shares In 1928 Poland Hungary Bulgaria Romania Czechos. (percent) Total 100.0 100.0 100.0 100.0 100.0 EC 55.9 25.0 64.5 53*9 43.9 EasternEurope 16.6 33.6 11.8 22.3 20.6 U.S.S.R. 1,7 0.4 0.0 0.O 1.3 OtherCountries 25.8 41.0 23.7 23.8 34.2 ,

Table 2.3: Eastern European Exports to EC Countries ~ Actual 1994 (percent) Poland 44.1 51.2 62.7 Hungary 47.9 37.2 48.9 Bulgaria 35.8 57.1 45.6 Romania 38.5 50.0 46.0 Czechoslovakia 53.7 46.3 63.8 Table 2.4: Eastern European Exports to Fomer CMEACountries HWPredicted CRPredicted Actual1994 (percent) Poland 24.5 23.2 13.9 Hungary 14.4 33.0 26.2 Bulgaria 27.8 16.9 15.9 Romania 25.0 25.4 14.0 Czechoslovakia 22.2 25.1 13.2 27

Table 2.5: PredictedExport Shares to CIUIEAless 1994 Actual Shares HW 1/ CR (percent) Poland 10.6 9.3 Eastern Europe 4.7 Fsu 4.6 Hungary -11.8 6.8 Eastern Europe 4.4 FSU 2.4 Bulgaria 11.9 1.0 Eastern Europe 3.2 FSU -2.2 Romania 11.0 11.4 Eastern Europe 4.6 FSU 6.8 Czechoslovakia 9.0 11.9 Eastern Europe — 4.3 FSU 7.6 1/HWdataarenotsufficientlydisaggregate toallowdecompositionbetweenEastern EuropeandtheformerSovietUnion. Table 2.6: Eastern European Exportsto Other Countries HW Predicted CR Predicted Actual 1994 (percent) Poland 31.4 25.6 23.4 Hungary 37.7 29.8 24.9 Bulgaria 36.4 26.0 .38.5 Romania 36.5 24.6 40.0 Czechoslovakia 24.1 28.6 23.0 28

Table 3.1: Product Composition of Polish Exports EC-85 EC-93 CMEA-85 CMEA-93 (Millions,1990 dollars) Food&Animals 469 784 193 249 Beverages&Tobacco 6 10 44 102 CrudeMaterials 394 457 242 89 MineralFuels 855 588 715 190 Animal&VegetableOils 11 10 14 3 Chemicals 170 408 458 241 ManufacturedGoods 676 2154 770 286 Machinery&Transport 318 1622 3230 395 Misc.Manufactures 252 1987 520 133 Other 0 6 0 2 Total 3152 8026 6186 1690 Product Shares: EC-85 EC-93 CMEA-85 CMEA-93 (percent) Food&Animals 14.9 9.8 3.1 14.7 Beverages&Tobacco 0.2 0.1 0.7 6.o CrudeMaterials 12.5 5.7 3.9 5.3 MineralFuels 27.1 7.3 11.6 11.3 Animal&VegetableOils 0.3 O.’l 0.2 ().2 Chemicals 5.4 5.1 7.4 14.3 ManufacturedGoods 21.5 26.8 12.4 16.9 Machinery&Transport 10.1 20.2 52.2 23.4 Misc.Manufactures 8.0 24.8 8.4 7.9 Other 0.0 0.1 0.0 ().1 29

Table 3.2: ProductComposition of Hungarian Exports EC-85 EC-94 CMEA-85 CMEA-94 (Millions, 1990 dollars) Food &Animals 464 701 927 576 Beverages&Tobacco 19 30 219 140 Crude Materials 139 308 116 71 Mineral Fuels 65 59 67 134 Animal &Vegetable Oils 15 7 54 65 Chemicals 279 418 554 340 ManufacturedGoods 259 914 511 222 Machinery &Transport 189 1198 2679 488 Misc.Manufactures 220 1169 728 132 Other 16 0 75 0 Total 1666 4804 5930 2168 Product Shares: EC-85 EC-94 CMEA-85 CMEA-94 (percent) Food &Animals 27.9 14.6 15.6 26.6 Beverages &Tobacco 1.1 0.6 3.7 6.5 Crude Materials 8.4 6.4 2.0 3.3 Mineral Fuels 3.9 1.2 1.1 6.2 Animal &Vegetable Oils 0.9 0.2 0.9 3.0 Chemicals 16.7 8.7 9.3 15.7 Manufactured Goods 15.5 19.0 8.6 10.2 Machinery &Transport 11.4 24.9 45.2 22.5 Misc.Manufactures 13.2 24.3 12.3 6.1 Other 1.0 0.0 1.3 0.0 30

Table 3.3: Product Composition of Czechoslovak Exports EC-89 EC-94 CMEA-89 CMEA-94 (Millions,1990dollars) Food&Animals 244 305 89 214 Beverages&Tobacco 12 38 35 52 CrudeMaterials 186 487 174 161 MineralFuels 221 206 225 275 Animal&VegetableOils 5 9 1 9 Chemicals 399 792 411 404 ManufacturedGoods 1027 2950 1130 851 Machinery&Transport 387 1787 5061 681 Misc.Manufactures 317 1182 951 389 Other 22 8 253 4 Total 2820 7765 8329 3041 Product Shares: EC-89 EC-94 CMEA-89 CMEA-94 (percent) Food&Animals 8.6 3.9 1.1 7.0 Beverages&Tobacco 0.4 0.5 0.4 1.7 CrudeMaterials 6.6 6.3 2.1 5.3 MineralFuels 7.8 2.7 2.7 9.0 Animal&VegetableOils 0.2 0.1 0.0 0.3 Chemicals 14.1 10.2 4.9 13.3 ManufacturedGoods 36.4 38.0 13.6 28.0 Machinery&Transpoti 13.7 23.0 60.8 22.4 Misc.Manufactures 11.2 15.2 11.4 12.8 Other 0.8 0.1 3.0 0.1 31

Table 3.4: Product Composition of Romanian Exports EC-89 EC-94 CMEA-89 CMEA-94 (Millions, 1990dollars) Food &Animals 182 74 146 114 Beverages &Tobacco 43 10 28 6 Crude Materials 65 55 78 28 Mineral Fuels 1137 147 48 104 Animal &Vegetable Oils o 1 0 2 Chemicals 169 147 210 83 Manufactured Goods 547 578 664 103 Machinery &Transport 544 238 2075 198 Misc. Manufactures 685 1243 664 96 Other o 0 0 1 Total 3372 2492 3914 736 Product Shares: EC-89 EC-94 CMEA-89 CMEA-94 (percent) Food &Animals 5.4 3.0 3.7 15.4 Beverages&Tobacco 1.3 0.4 0.7 0.9 Crude Materials 1.9 2.2 2.0 3.8 Mineral Fue!s 33.7 5.9 1.2 14.2 Animal &Vegetable Oils 0.0 0.0 0.0 0.2 Chemicals 5.0 5.9 5.4 11.3 ManufacturedGoods 16.2 23.2 17.0 14.1 Machinery&Transport 16.1 9.5 53.0 26.9 Misc. Manufactures 20.3 49.9 17.0 13.0 Other 0.0 0.0 0.0 0.2 32

Table 3.5: Similarity between EC and CMEA Trade(one-Digit Industries) * Countries CMEAOECO1/ CMEA1ECI 2/ Change Poland 0.549 0.728 0.179 Hungary 0.633 0.674 0.041 Czechoslovakia 0.505 0.860 0.355 Romania 0.620 0.540 -0.080 > Table 3.6: Similarity of EC and CMEA Trade (One-Digit Industries) Countries EC~EC, CMEAOCMEA1 Difference Poland 0.677 0.703 -0.026 Hungary 0.719 0.699 0.020 Czechoslovakia 0.851 0.588 0.263 Romania 0.623 0.670 -0.047 1/ Period “zero”, i.e., the first observation,is 1985for Poland and Hungary and 1989 for Czechoslovakiaand Romania. 2/ Period“one”,i.e.,thelastobsewation,is 1993for Polandand1994fortheothermuntries. 33

i Table 3.7: Manufacturing Exports to EC and CMEA - Crosstabs AGGREGATE upI 15 I 50 EC ‘own- POLAND HUNGARY CMEA CMEA up down up 2 17 EC ‘pm ‘ownm EC down o 0 CZECHOSLOVAKIA ROMANIA CMEA CMEA up down up down I up 5 14 up 3 5 EC EC down o 0 down 1 10

i u o

Table 3.9: Similarity between EC and CMEA Trade (TwO-Digit Industries) Countries CMEAOECO1/ CMEA1 EC, 2/ Change Poland 0.516 0.608 0.092 Hungary 0.557 0.596 0.039 Czechoslovakia 0.505 0.793 0.288 Romania 0.648 0.516 -0.132 Table 3.10: Similarity of EC and CMEA Trade (Two-Digit Industries) Countries EC~EC, CMEAOCMEA1 Difference Poland 0.656 0.672 -0.016 Hungary 0.809 0.721 0.088 Czechoslovakia 0.801 0.663 0.138 Romania 0.650 0.735 -0.085 1/Period“zero”, i.e., the first obsewation, is 1985forPolandand Hungaryand 1989for Czechoslovakiaand Romania. 2/ Period“one”, i.e., the lastobservation,is 1993forPolandand1994fortheothercountries. 36

Table 3.1 1: Trade Diversion Regressions Dependent Variable: AECifl Variable (1) (2) (3) I 1 I I I Constant .— 1.4 — 5 . 0 1.260 1.200 I (2.60)”” (2.78)** (2.68)** ACM~ifl -.164 -.162 -.130 (-1.91)* (-2.12)** (-I.90)* ACMEAti~.l .024 .063 (.25) (.94) AcMWif.*l .042 (.55) fiECifl_l .243 .289 .284 (2.96)** (4.02)”’ (3.96)** iECijt_z -.003 (-.03) rranspodEquip. 4.960 4.060 3.810 (2.22)** (2.17)** (2.05)”” >Iothing 12.800 10.500 10.500 (5.81)** (5.73)** (5.74)** \djustedR2 .169 .161 .161 )bsewations 294 351 351 Note:Asing[e(doubje) asteriskdenotessignifj=nm atthe10per~nt(5 percent)ieve~ 37

Appendix Onedigit commodities O-Food and live animals l-Beverages and tobacco 2-Crude materials, inedible, exceptfuels 3-Mineral fuels, lubricants and relatedmaterials 4-Animal and vegetable oils,fats and waxes 5-Chemicals and related products 6-Manufactured goods classifiedchiefly by materials 7-Machinery and transport equipment 8-Miscellaneous manufactured goods 9-Commodities not classifiedby kind Two-digit commodities 61-Leather, dressed fur,etc 62-Rubber manufactures 63-Wood,cork manufactures 64-Paper and paperboard 65–Textile yam,fabric etc 66-Nonmetal mineral manufactures 67-iron and steel 68-Non-ferrous metals 69-Metal manufactures 71-Machinery, non-electric 72-Eiectricai machinery 73-Transport equipment 81-Plumbing, heating,lighting equipment 82-Furniture 83-Travel goods,handbags 84-CIothing 85-Footwear 86-lnstruments,watches, clocks 89-Miscellaneous manufactured goods 38

International Finance Discussion Papers IFDP Number Titles Authotis] 1996 562 EasternEuropeanExportPerformanceduring Nathan Sheets theTransition SimonaBoata 561 Inflation-AdjustedPotentialOutput JaneT. Haltmaier 560 TheManagementof FinancialRisksat German Allen B. Frankel NonfinancialFirms: TheCaseof Metallgesellschfi DavidE. Palmer 559 BroadMoneyDemandandFinancialLiberalization Neil R. Ericsson inGreece SunilSharma 558 StockholdingBehaviorof U.S. Households: Evidence CarolC. Bertaut tim the 1983-89Surveyof ConsumerFinances 557 Firm Sizeandthe Impactof Profit-MarginUncertainty VivekGhosal on Investment: Do FinancingConstraintsPlaya Role? PrakashLoungani 556 Regulationandthe Costof CapitalinJapan: A Case JohnAmmer Study MichaelS. Gibson 555 TheSovereigntyOption: TheQuebecReferendumand MichaelP. Leahy MarketViewson the CanadianDollar CharlesP. Thomas 554 RealExchangeRatesand InflationinExchange-Rate StevenB. Karnin BasedStabilizations: An EmpiricalExamination 553 MacroeconomicStateVariablesasDeterminants JohnAmmer of AssetPriceCovariances 552 TheTequilaEffect: TheoryandEvidencefrom MartinUribe Argentina 551 TheAccumulationof HumanCapital: Alternative MuratF. Iyigun Methodsand WhyTheyMatter Ann L. Owen 550 Alternativesin HumanCapitalAccumulation: MuratF. Iyigun Implicationsfor Ewnomic Growth Ann L. Owen Pleaseaddressrequestsfor copiesto InternationalFinanceDiscussionPapers,Divisionof InternationalFinance,Stop24, Boardof Governorsof theFederalReserveSystem, Washington,DC 20551. 39

International Finance Discussion Papers — IFDP Number Titles 1996 549 MoreEvidence on the Link between Bank Michael S. Gibson Health and Investment in Japan 548 The Syndrome of Exchange-Rate-Based EnriqueG.Mendoza Stabilizationand the Uncertain Duration of MartinUribe Currency Pegs 547 Geman Unification:WhatHaveWeLearned JosephE.Gagnon horn Multi-CountryModels? PaulR.Masson WarwickJ.McKibbin 546 Returnsto ScaleinU.S.Production:Estimates SusantoBasu andImplications JohnG. Femald 545 Mexico’sBalance-of-PaymentCs risis: AChronicle Guillemo A.Calvo ofDeathForetold EnriqueG.Mendou 544 TheTwinCrises: TheCausesofBankingand GracielaL.Kaminsky Balance-of-PaymentsProblems CarmenM.Reinhart 543 HighRealInterestRatesintheAfiermathof GracielaL.Kaminsky Disinflation:Is ita Lackof Credibility? Leonardo Leideman 542 Precaution~ Portfolio Behaviorfrom a Life-Cycle Carol C. Bertaut Perspective MichaelHaliassos 541 UsingoptionsPricesto InferPDF’sforAssetPrices: William R. Melick AnApplicationto OilPricesDuringtheGulfCrisis CharlesP. Thomas 540 MonetaryPolicyintheEnd-Gameto Exchange-Rate Steven B. Kamin BasedStabilizations:TheCaseofMexiw John H. Rogers 539 ComparingtheWelfweCostsandtheInitialDynamics Martin Uribe ofAlternativeTemporaryStabilizationPolicies 538 LongMemoryinInflationExpectations:Evidence JosephE.Gagnon fromInternationalFina.ncialMarkets 537 UsingMeasuresof Expectationsto Identi&the AlIanD.Brunner Effectsofa MonetaryPolicyShock 40

I International Finance Discussion Papers IFDP Number Titles Authods\ 1996 536 RegimeSwitchinginthe DynamicRelationship ChanHuh betweentheFederalFundsRateandInnovationsin NonborrowedReserves 535 TheRisksandImplicationsof ExternalFinancial EdwinM.Truman Shocks: Lessonsfrom Mexico 534 Cumncy Crashesin EmergingMarkets: An Jefiey A. Frankel EmpiricalTreatment AndrewK. Rose 533 RegionalPatternsinthe Lawof OnePrice: The CharlesEngel Rolesof GeographyVs. Currencies JohnH. Rogers \ 41