W4S i6 o

POLICY RESEARCH WORKING PAPER

1410

MacroeconomicEffects
of Terms-of-Trade Shocks
The Case of Oil-ExportingCountries
Nikola Spatafora Andrew Warner
N.kola S- t'fora

Theexperience of 18oi
exporting countries(from

--1973to 1989)suggeststhat
permanent terms-of4rade shodcshavea significant positiveeffe'z on :consumptionrinvestment

.-;Ade Wae nontradables. Thereisalmost -{

no effecton savings, and -at. advEnerese ect on the trade. balance.Thereis no evidence -6of'Dutch disease effects.

and output, particularlyof

The World Bank

IntrntioalEconomics Depatrtment Intermaional EcoDnomic Analysiand Prospects Divison January 1995

Poi.icy RESIEARC(I-I WORKIN;PAI'hiR 1410

Summary findings
Spatafora and Warner investigate the impact on economicgrowthiand development of long-run movementsin the external tcrms of tradc, with special referenceto the experience of 18 oil-exporting countries between 1973 and 1989. They argue that this sample approximates a controlled cxperiment for examining the impact of unanticipatedbut permanent - shocks to the terms of trade. They analyzethe sample econometricallyusing panel data techniques. They find that permanent ternis-of-tradeshocks have a stronglysignificantpositive effect on investment,which they justifytheoretically on the grounds that countries in the sample import much of their capital equipment. The shocksal;o havea significantpositive effect on consumption. Government consumption responds almost twicc as strongly as private consumption. The shockshave no effect on savingsand adversely affect the trade and current account balances. There is a significantpositiveeffect on the output of all main categoriesof nontradables. But Dutch disease effects arc strikinglyabsent. Agricultureand manufacturingdo not contract in reaction to an oil price increase. Dutch diseaseeffects may be absent in part becauseof policy-inducedoutput restraints in the oil sector, or becauseof the "enclave" nature of the oil sector, which does not participate in domestic factor markets.

Thispaper-a product of the InternationalEconomic Analysis and ProspectsDivision, International Economics Departmentis one in a seriesof backgroundpapers preparedin support of the analysesand scenariosin GlobalEconomic Prospects 1994. Copiesof thispaper are available free fromthe World Bank,1818 H StreetNW, Washington,DC 20433. PleasecontactJackie Queen,room S8-216,extension33740 (37 pages).January 1995.

The PolicyResearch Working PaperSeries disseminates the fidings of work miprogressto encourage the exchange of ideasahout development isue An objecaeof theseries isto getthefindings out quickly, evenif thepresentations arekessthanfly polished The papers carrythe names of t authors andshould beused andcitedaccor&ngly. Thefindings. inteapretations, andcondusions arethe authors'oumandshould not beattributedto theWorldBank.its Exwtive Board of Directors.or an of its moneber countries.

Producedby the Policy ResearchDissemination Center

Macroeconomic Effects of
Terms-of-Trade Shocks

The Case of Oil-Exporting Countries

Nikola Spatafora and Andrew Warner

MACROECONOMIC EFFECTS OF TERMS-OF-TRADE SHOCKS: THE CASE OF OIL-EXPORTING COUNTRIES.

Nikola SPATAFORA& Andrew WARNER I. Introduction
Most developmenteconomists agree that variations in world prices are an imnportant source of risk and instabilityfor developingeconomies. Developingcounaies derive about half their export earnings from primarycommodities,whose world prices are extremelyvolatile. Nor is such volatilitypurely short-term: mauchdata suggest that commodity prices undergo long periods of rise and decline. To provide one exampleof the magnitudesinvolved,the World Bank's index of non-oil commodity prices has exhibited a trend decline of about 1.5%per annum since 1948,cumulatingto a 50% decline over 45 years. Over the next decade, the World Bank is forecastingthat this trend will reverse and there will be a rise of about 0.7% per annum. What is the impact of these long-run movements in the external terms of trade on economic.growth and development? We think that the world has provided a natural experiment over the past 25 years that can be used to analyzethis issue. Oil-exportingdevelopingcountriesexperienceda major rise in the world price of their main export between 1973and 1981, followedby a smaller,but still substantial,decline between 1981and 1989. Nominaloil prices rose from $2.70 in 1973 to $34.31in 1981,and then fell back to $16.31in 1989; deflatingthis by US producerprices, real oil prices rose from an index of 1.0 in 1973 to 5.83 in 1981 and 2.42 in 1989. From a researchstandpoint,we think that this episode represents a fortunate opportunityto better understandthe impact of terms-of-tradeshocks, for the following reasons. First, it seemsimportantto distinguishbetween permanentand tansitory, or anticipatedand unanticipated changes in the terms of trade, and we believe that oil prices come closer than any other data to measuring an unanticipated permanentchange. The 1973 rise was not anticipated several years before, and it was quickly and widely perceived as a permanent feature of the economic landscape. The long decline

page 2 beginning in 1981was also notwidelyanticipated; it probablytooklongerfor it to be acceptedas a longbut at leastby 1986,whenoil prices fellby 25%,the 1981levelswere longconsidered runphenomenon, a thing of the past. Hence,it is crediblethat the majorlong run changesin oil prices werepreviously once theyoccurred. eventsthat werewidelyviewedto be permanent unanticipated Second, the timespanof the riseanddeclinein oil pricesis fairlylong-8 or 9 years-so thatit is credible effects. Third,the movements in oil priceswere large. Hence, to believethat we can observelong-run there is much statisticalvariation,and if effects are present,we shouldbe able to estimatethem more efficiently. Fourth,terms-of-trade changeswere so dominantfor thesecountriesover this periodthat decidingwhat variablesto controlfor, so as to avoid omittedvariablebias, should constituteless of a problem in the Fifth,we have both a rise and a declinein oil prices, so that we can try to detect asymmetries problemsassociatedwith index numbersare less of an issue for oil responses. Sixth, measurement betweenthe termsof trade homogeneous product Seventh, simultaneity exporters, sinceoil is a relatively and domestic economic variables, whilenot alwaysabsent,is at leasta tractableproblembecausethereis and for manyof the countries Further, aboutthe causesof themajoroil pricemovements. a fairconsensus to viewthe termsof tradeas predetermined. mostof the variableswe examine, it seemsreasonable Our broadconclusionis that an examination of the data for oil exporters leads to a differentpictureof terms-of-trade effects than can be obtainedfrom an unbiasedsampleof terms-of-trade articlesin the literature.We find that the impacton investmentis crucialto understanding the responseof the current is filledwithcurrent shocks; yettheliterature growth effectsof,terms-of-trade account to, andthe long-run trademodelsthat treat the capitalstockas an constant, and long-run account modelsthathold investment endowment whichis unaffected by terms-of-trade shocks.Conversely, Dutchdiseaseeffectsareexamined extensively in the literature, but wefailto findanyevidencethattheDutchdiseaseis a majorphenomenon. Wealsofind thatthe simpleinsights fmmthe tradables-nontradables modelare wellsupported by thedata. shocksis not verysensitiveto whether to terms-of-trade of expenditure Finally, we findthat the response the expenditure comesfromthe publicor privatesector.

page 3

The rest of the paperis dividedas follows. Section II summarizes the relevanttheoretical literatureand empiricalstudies. Sectionm containsa description of the data. SectionIV sets out the econometric framework andanalyses someimportnt econometric issues. Section V presents themainresults. Section VI discussessomecriticisms.SectionVII concludes.

II. TheoreticalJiterature and empiricalstudies

For expository purposes, wefind it usefulto dividethetheoretical literature on theimpactof terms-of-trade
shocks into two broad groups. One class of modelsis microeconomnic in nature and stresses the differing

effects of terms-of-trade shockson differentsectorsof the economy. Examplesincludethe tradablesnontradables modelsoriginated by Meade,Salter(1959),and Swan;the DutchDiseasemodelsfoundin, say, Wijubergen (1984, 1984b)and Neary& Wijnbergen (1984),and summarized in Corden& Neary (1982)and Corden(1984); and the computable generalequilibrium modelof Bruno& Sachs(1982). A second classis concermed withthe behavior of broadmacroeconomic aggregates, particularly savingand the currentaccount, andrecently has tendedto stressintertemporal issues. A partialreferencelist for this literatureincludesLaursen& Metzler (1950),Harberger(1950),Obstfeld(1982),Persson& Svensson (1985),and Bean(1986). A. Sectoral Effects Manyof the insightsfromthe earlierliterature stressing sectoraldisaggregation canbe obtained, following Corden,by thinldng in termsof a thee-sector,perfectlycompetitive neoclassical model,with fixedtotal factorendowments. Letthe firstsector(0) produce oil for export;let the second,'DutchDisease'sector (D)produce all othertradables, both exportables and importables; and let the thirdsector(N) produce nontradables. Thefirst twosectorsproducetradablesat givenworldprices,and the difference betweenthe quantity supplied and demanded is made up by exports or imports; the price of nontradablesis endogenously detwrmined by the condition thattheirsupplyequaltheir demand. Assumefirst that output in eachsectoris produced bya factorspecific to tluht sector,andby labor,whichis mobile betweenall three sectorsand movesbetweensectorsto equalizeits wage in all threeemployments.

page 4

Nowlet therebe an oil priceshock. Thedirecteffectis to increasethe aggregateincomesof the factors initiallyemployedin 0, which in turn has two effects. First, if some partof the extra incomeis spent, whetherdirectlyby factorowneisor indirectlythroughbeingcollectedin taxes and then spent by the govenunent, andprovidedthe incomeelasticityof demandfor N is positive,demandfor N rises. Given less than perfectlyelasticsupply. the priceof N relativeto tradablesmustrise. That is, thereis a real appreciation whichdrawsresourcesoutof 0 and D into N. Thisis the spendingeffect. In addition,the price increaseraisesthe marginalvalueproductof factorsin O so that, at a constantwagein tenns of tradables,the demandfor laborin 0 rises, inducinga movement of labor out of D and N. This is the resource movenen effect. Theoutputof D must finallybe lowerthanbeforethe shock,whilethe output of N couldbe higheror lower,the spendingeffecttendsto increaseit, the resourcemovement effectto decreaseit. Assume now,as is common, thatthe oil sectordoesnotin factemployany factorsthat canbe usedby the rest of the economy.Here,the oil sector is an 'enclave'whichdoes not participate in domesticfactor markets.Thereis thenonly a spending effect,and the key mechanism of resource reallocation is the real appreciation. Providedspending on nontradables initially goesup, outputof N mustfinally be higherthan in the pre-shock situation. Assumeinsteadthat morethan one factoris mobilebetweenat leasttwo of ourthreemainsectors. For instance, saythatcapitalis also mobilebetweenD andN, so that thesetwo sectors,employing laborand capitalin different proportions, makeup a mnini-Heckscher-Ohlin economy.Nowthe resourcemovement effect can haveparadoxical result. Because of thepriceshock,labormovesout of thismini-economy into 0. If D is thecapital-intensive industry, its outputcouldon balanceexpand. If N is capital-intensive, the shockcouldcausea realdepreciation. At this stageit is usefulto comparethe predictions of this neoclassical modelwith a Keynesian model. In the neoclassical model,with factorsfixed in the shortrun, and full employment determining output throughthe production functions, thereis no scopefor 'aggregate output' to changemuch:the shockcan at most changethe composition of output. However, in a Keynesian economywith stickyprices, the demandstimuluscan raiseaggregateoutputevenin the shortrun.

page5

the neoclassical framework predictsthatthesehingecrucially longer-term supply-side effects, Regarding factors,suchas physicaland humancapital,will accumulate on the relativepricesignals. Reproducible the relevant insectors whoserelative pricesrisefollowing the terms-of-trade improvement (moreprecisely, priceis a productpriceadjusted for movements in inputprices,thatis, a valueaddedprice). As we have seen, the relativepricesin 0 and N typicallyrise in the short run, and thus factorsaccumulatein these sectors,expandingoutput in the longerrun. By contrast,capitalwill shift out of, or be allowedto depreciate withoutreplacement, in D. B. Macroeconomic effects Tuning nowto the macroeconomic literature, earlyanalyses, conducted withintheframework of thestatic Keynesian savingand investment functions, include Laursen & Metzler(1950)andHarberger (1950).The drivingmechanism is thatterms-of-trade shocksaffectreal income, andhencesavings.Giveninvestment, this determines the evolution of thecurrentaccount.Laterresearch,typifiedby Obstfeld(1982),Persson models,typicallyinvolving & Svensson(1985),and Sen & Turnovsky (1989),is basedon neoclassical income,intratemporal and intertemporal dynamic optimization.Terms-of-trade shocksalterpermanent relative prices, and henceaffect consumption, saving,and investment. Oneconclusion to emergefromthis and literatureis that the effectsdependcriticallyupon whetherthe shockis permanentor temporary, anticipated or urnanticipated. in the termsof tradeon unanticipated increase Letus first considerthe possibleimpactof a permanent, bothcurrentand permanent income. All modelsthereforepredictan consumption. Theshockincreases increase in aggregate consumption. In sectoral terms,consumption of oilis probablyinsignificant for our sample, and consumptionof other tradables must rise through income effects. Consumptionof whichwasdiscussedabove. nontradables mustequal production, modelsoftenpostulatea by muchof the literature.OlderKeynesian Investment is treatedas exogenous simple acceleratormechanismwherebythe shock raises expecteddemand and hence investment. Optimizingmodels usually concludethat aggregateprivate investment is determined by the ratio of investment products of capital)to the aggregate valueof futuremarginal Tobin'sq (thepresentdiscounted price deflator. Assumingthat capitalgoods have a strongimport content,the shockboosts domestic

page 6 investment-Schmidt-Hebbel & Serven(1993). Yet mostmodelstreatinvestment goodsas dcmestically produced,and henceobtainfar more ambiguous result-Persson & Svensson(1985).Sen & Turnovsky (1989). For developing economies, the formerassumptionappearsto be correct. Threefactorsmayhowever reducethe incentive to invest. First,profitsmaybe capturedby unions(rentsharing). Yetmanyof thesecountrieslack an organizedlabormovement.Second,in OPECcountries production quotasmaybe seenas a permanent ceilingon oil production, leavinglittle reasonto investin theoil sector'sproductive capacity.On theotherhand,an increase in suchcapacity mightstrengthen one's hand in the periodicquotanegotiations; also, there is stil an incentivefor cost-reducing investment. Finally,increased wealthmayencourageconsumption of leisure,and a contraction in labor supplywill reduce the marginalproductof capital;see Bean (1986),Sen & Turnovsky (1989). Yetmanyof these countriesface an extremely elasticsupplyof potential immigrant workers. Regardingthe currentaccount,Laursen& Metzlerpostulatedthat the shock by raising real income, increases savings. For a given investment,this increases the current account. In an optimizing intertemporal framework, with domestically producedinvestment goods,Obstfeld(1982)and Svensson & Razin(1983)showedthatwitha constant rateof timepreference theeconomy should jump immediately to its new steady state, with no effecton the currentaccount. If insteadthe rate of time preference risein savingsand currentaccountsurplus. On the otherhand, thereis a transitory decreases withutility, if investmentrequiresimportsof capital machinery, it may be encouraged by the shock, leadingto a temporary currentaccountdeficit. C. Empirical Studies Warner(1992)examinedwhetherthe international debt crisis whichbegan in 1982could explainthe investmentdeclinein 14 heavilyindebtedcountries. He foundthat equationswhichomittedall debtrelatedinformation, butincorporated theeffectsof fallingexportpricesand highworldrealinterestrates, of the termsof tradein determining could forecastthe fallin investment.Thissuggeststhe importance investment. Warner(1994)proposedand estimateda microeconomic investment model to determinethe relative importance for Mexico'sinvestment decline in the early1980'sof threeexplanations: theoil pricedecline,

page 7 the terriination of capitalinflows, anddebt-overhang/uncertainty effects. Usingquarterly investment data for68 private-sector industries between1981and 1985, he foundthatthemainmirroeconomic mechanism drivingthe investment declinewasthe risein therelative priceof imported investment goods,and further that the termsof tradedecline(drivenby fallingworldoil prices)explainsmuchof the increasein this relativeprice. Morley(1992)examined stabilization programs in a broadsampleof LDC's. Usingpaneldata,he found thatthe termsof tradehad a significant positiveimpacton investnent and output. The samewas true of real appreciations, whichas we have arguedare a likelyconsequence of terms-of-trade shocks. Etherington& Yainshet(1988),usingtime-seriesmethodology, foundthat the price of coffee,the key Ethiopian export,had a significant positiveimpacton Ethiopian capitalgoodimportsanddomesticcapital formation. De Gregorio(1992)analyzedgrowthdeterminants in twelveLatin Americancountriesover the period 1950-1985.Controlling for investment and macroeconomic stability,he did not find a significant effect of the termsof tradeon growth.However, he did notconsiderthat investment itselfmightbe affected by the termsof trade.

III. The Data

We considerthe period 1965-1989. Our sampleconsistsof the following18oil-exporting countries: Algeria Bahrain Congo Ecuador Egypt Gabon Indonesia Iran Iraq Kuwait Mexico Nigeria Oman SaudiArabia Syria TrinidadandTobago UnitedArabEmirates Venezuela

For each country,on the expenditure side we examineddata on consumption and investment(by the privatesector, by the government, andin the aggregate), savings, thetradebalance,andGDP. At a sectoral level, we analyzedvalue addedin the oil sector and the non-oilsector. The non-oilsectorwas further brokendownintothe following categories: agriculture, manufacturing, construction, publicutilities,and

page 8 services.Serviceswerein tum disaggregated into transportation & communications, wholesale & retail trade, and other services. Tradablesare probablymost appropriately identifiedwith agricultureand manufacturing, nontradables withthe othernon-oilcategories. Allthe abovevariableswerein per capita, constantlocalcurrency terms.As explanatory variables, we useddata on the termsof trade,a debtcrisis dummy, and the worldreal interestrate.All variableswereat an annualfrequency.A fulldescription of the data usedand its sourcesis in Appendix1.

IV. Econometricframeworkand issues.

A. Estimation Throughout, we usea panel ierthana time-series methodology since,if the underlying assumptions are

satisfied, this allowsfor morepreciseestimatesand rendersomittedvariablebias a lessseriousproblem (seenextsub-section). Wehaveno strongpriors,anddteoryprovides littleguidance on thespeedat which our variables adjust to terms-of-trade shocks. Therefore we take whatwe thinkis a flexibleapproach to the issueof dynamics.As a first pass,we try to pick out long-ranrelationships by selectingyearswhen sufficient time has passed under a given regime for the countriesto be in long-runequilibrium. Specifically, we notethat 1973wasthe endof a longperiodof stableoil prices;in 1981,aftereightyears of risingprices,agentsgeneallyaccepted the higherpricesas prmanent, and hadhad sometimeto adjust to them;in 1989,afteran eight-year periodof fallingprices,agentslikewiseaccepted and had had some timeto adjustto the lower prices;and,after 1989,oilpricesrosr.sharplyaroundthe timeof the GulfWar. Thus in 1973, 1981,and 1989the countriesin our samplecould arguablyhave achievedlong-run equilibrium.We thereforeuse thesethreebenchmark yearsto cartyout fixed-effects estimation of the following panel regression equation: y
-t at,+

TOT,+ en

(1)

wherey denotesthe naturallogof eachof the variableswe study(exceptfor the tradebalance, whichis leftin levels), TOTdenotesthelogof the termsof trade,the i subscriptrefersto countries, and t refersto

page 9
time. The elasticityof each variable withruspect to t toerms of trde is givenby P. The regressionresults

are givenin Appendix 11,table 1. In addition,we also use the full dataset to carryout fixed-effects estimation of the following regression equation,whichallowsfor laggedresponses: y g + ITOT1 + P2TOTIr, + yCO) OL 1 , + e(,

!2)

whereCONTROL denotesthe controlvariables, whichpotentially includethe debt crisisdummyand the worldreal interestrate. Becausethereis littletheoreticalguidanceavailableon whenand whetherwe we excluded it whereit wasinsignificant. Theresultsfor the world shouldinclude the debt crisisdummy, real interest rate were hard to interpret,and we alwaysexcludedit. The only exceptionwere the investment equationswhere,giventhe strongtheoretical priors,we alwayscontrolledfor the debt crisis with respectto the tens of and for the worldreal interestrate. Theshort-runelasticityof eachvariable by (P, + N3).The regressionresultsare givenin Appendix the steady-state elasticity tradeis givenby J,, II, table2, and are discussedin the nextsection. Notethatourspecification of equations (1) and (2)makes e important assumptions.First,the random efrorses, havezeromeanand arei.i.d.overtimeandacrosscountries.Second, theintercept a variesacross to the cffects. I subjectthis hypothesis countries(but not over time),so that there arecountry-specific & Pagan(1980). Let a be a randomtermwithcrossfollowing Laange Multiplier test,dueto Breuisch countryvariancea.. Let the null hypothesis be that there are no country-specific effects,that is, Ho: a. =0, versusHI: a. > 0. Let e,tdenotethe residualsfrom the standardOLS regressionwith homogeneous intercepts.UnderHo,the test statistic

LAM-

NT
2(T-1) iS

I3

d X21).

page 10

The asymptotic P-values arereportedin Appendix . In all regressions, LM is significant at the .0001% level; we thereforereject the null. Note that, since the test statistichas a block-diagonal information matrix,thispre-testdoes not affectthe standard errorsof the otherestimatorswecompute. Third, we assumethat the elasticities with respectto the termsof trade(the p coefficients) are constant over time and acrosscountries. Wetest bothof thesehypotheses.To investigate temporalstability, we split up the sampleinto the subperiods1965-1980 and 1981-1989; sincetheyroughlycorrespond to the periodsof rising andof failingtermsof trade,thisalso allowsus to checkwhether theresponses to positive and to negative terms-of-trade shocksdiffer. As detailedin AppendixIL table 3, the nullhypothesis of identicalbcoefficients acrossthe two subperiods is generally supported by the F-testdescribedin Hsiao (1986),chapter 2.2,equation2.2.20;wherethe differences acrosssubperiods are statistically significant, they are typically quantitatively unimportant.Thissuggeststhereare no significant asymmetries in the responseto risingand to fallingtermsof uade. In contrast, as detailed in Appendix m, thenullhypothesis of stabilityacrosscountriesis alwaysrejected at the 1%significance levelby the F-testdescribedin Hsiao(1986),chapter2.2,equation2.2.15. The rejection is hardlysurprising, giventhat we aredealing withsucha broadsampleof countries, but it does imply that we must look at country-specific regressions to determinehow representative the panel estimates are. Thesecountry-by-country regressions are accordingly presentedin Appendix m. Finally,notethat ourfixed-effects estimator is onlyBLUEif we interpretthe country-specific effectsa1 as fixedregressors.To forecastan out-of-sample response to a terms-of-trade shock,we musthowever treat the a%'s as randomerrors;our estimator is then in generalnot BLUE,but is still preferableto the random-effects estimatorsinceit is consistent even when the az's are correlatedwith the explanatory variables. B. Omitted variable bias Theissueof bias fromomittedvariablesneedsto be treateddifferently in a panelcontext.To seethis,it is convenient to definethe mean acrosstimefor any givencountryand variable,x, as:

page 11

Xi

(4)

and the deviations from country-means as:

Xs
X. -X. (5)

Re-written in terms of deviation data, equation (1) then becomes

YU,-

pTOT,+

(6)

Say that the true model is instead given by .= .

PTOT,, yOAM

+ X, -

(7)

If equation (5) were estimated by OLS for each country, the bias b; in our estimator of regressors as fixed, would be
T

tating the

E T6Tt , y OMEDI( ,bf

T

(8)

E ToT2 Hence, a necessaryand sufficientconditionfor the countrty-by-county OLS estimatesof i to be consistent is that asymptoticallythe terms of trade be orthogonalto any omitted variables:
Al. lim(T--) b,r = 0.

page 12

Now saythat equation(5) is estimatedin a panelcontext. Thebias in ouLr estmator of P is then

Ir r

E 7T&T * y 0MJZTAD,, b;
I1 .1 Ir

(9)

6T

II t*ib-

(10)

Hence, a necessary and sufficientconditionfor the panel esfimatesof p to be consistentis that asymptotically a weightedaverage of the country-specific biasesbe zero: A:2. lim(T--) F,,aTUb r = 0. Clearly, Al impliesA2 butnotthe reverse: therearemanywaysfor assumption A2 to be satisfied without any of the countriesindividually satisfying assumptionAl. The importantpoint is that for the panel estimates to be consistent, it is unnecessary for the termsof tradeto be asymptotically orthogonal to the omittedvariables in anycountry: whatmattersis thatthe country-specfic asymptotic biasesbe negatively correlatedacrosscountriesso that theyaverageto zero. This latterassumption is morecredibleto the extentthat the relationship betweenthe omittedvariablesandthe terms of tradeis idiosyncratic to each countryand notpositivelycorrelated acrosscountries. Of course, to theextentthatthe omittedvariablesare worldvariables whichaffectall countries, thenthe b;termswill tend to have the samesignsand assumption A2 mayfail. We thereforecontrolfor the debt crisisandfor worldinterst rates. Etmightalsobe desirable to controlfor worldoutput,but sinceperiods of risingoil prices coincided with changesin OECDactivitythat wouldhave depressed growthin our sample, andvice-versa, thiswouldprobably strengthen the linkswe find betweenterms-of-trade changes and activityvariables.

page 13
The additionalissue of the weights, a,, is unlikelyto be important for our sample. Countries have larger than average weights to the extent that the sample variability of their terms-of-tradedata is larger than average. Since our sample consists of oil exporters,the variability of their terms of trade is similar. C. Non-stationarity There are two ways to inltpret our econometricmodel. Frst, we may view the terms of trade (which are driven mainly by the oil price) as a fixed regressor. Equations (1) and (2) are then the mechanism generatingthe rest of the data. The presence or absence of stationarity is not an issue in this framework. Altemvatively, we may treat the terms of trade as being generated by some time-series statistical model. Here, the order of integrationof our data is an issue. Unfortunately,little is known about how to test for and deal with nonsaionarity in a panel data setting. When estimating equation (2) for each country in isolation,we carryout an EngleGranger co-inwgrationtest, using the augmentedDickey-Fuller statistics. For most countriesand many variables, we cannot reject the null hypothesisthat the regression residuals are integratedof order 1. However, given that such tests are well known for their very low power, and given the paucity of observationsat the single-country level, we do not feel there is much strong evidence that our data is non-stationaryin the sense that it trends over time.

V. Results
A. The importance of Investment Overall,we find fairly strong evidencethat the relationshipbetweenterms-of-tradeshocks and investment is positive. The panel regression suggests the 95% confidenceinterval (0.500,0.675) for the long-run elasticity. In countryby countryregressions,14 of the 18 countrieshave positiveelasticities and of the 14, 13 are stastically significant This sectionexamines what the availableevidence says about the sectoral breakdown and the causes of this investment effect, and how this investment effect is crucial to understandingother effects such as those on the trade balance and on growth.

page 14 It is likelythat muchof this investmnent occurredin the nontradablesector of the economy. Although data on investmentby sectoris simply not availablefor mostof these countries,' and thereforethis claim cannot be supported directly,much of the otberdata we do have is supportive. First, thereis strong evidence that the relativepriceof nontradableswere positivelyassociatedwith the terms-of-tradeshocks. So there seems to have been a clear price incentivefor investmentin these sectors. Second, we do have evidence that in the long run output expanded in nontradable sectors such as construction and transportation. It seems reasonable to think that this rise was made possibleby a higher capital stock in these sectors. This investmenteffi- in nontradablescan be explainedformally in a variety of intertemporaloptimizing models of investment (recent examples in the literue include Schmidt-Hebbel& Serven (1993) and Warner (forthcoming, 1994)). These models usually conclude that aggregate private investment is determined by the ratio of Tobin's q (the present discounted value of future marginal value products of capital in a given sector) to the relevant price deflator for investment goods. Since the countries in our sample import manycapitalgoods, and the world pricesof capital goods did not changedramaticallyover the period, the price of nontradables relative to physicalcapital probably rose in the 1970's and declined in the 1980's, causing similar changes in investmentin the nontradable sector. Tuming to the oil sector,one might think that relativeprice changeswould also stimulateinvestmentthere. Althoughthis would probably hold for countriesexperiencingexogenous changesin their export prces, it is important to rememberthat OPEC raised oil pricesin the 1970's partly through a deliberate policy of output restriction,so that for much of our period the oil sector was engineering price changes rather than passively respondingto them. In such a setting, where countries exercise their power in the world oil market, it is a prioriunclear whichway the incentiveswork for investmentin the oil sector. Oil production quotas may be seen as a permanentceiling on oil production, leaving little reason to invest. On the other band, an increase in productive capacity may strengthen one's hand in the periodic quota negotiations.

XAn exception isWamer(1994),whereit is shownthatinvestment in Mexican private non-tradables diddecline

over 1981-85.Thatstudyalso suggests that an important mechanism for the investment declinewas the fall in the ratioof nontradable pncesto imported equipment prices, andthat theterms-of-trade declinecanexplainmostof the reduction in thisrelative price.

page 15 Therefore, we cannot deduce from theoly whether investment should have risen or declined in the oil sector. We do knowthat real oil output per capita did not rise over the long run between 1973and 1981. It may be that countries invested in modernizingthe oil sector rather than in expanding its capacity. It may also be that they did invest in expandingcapacity but just did not use this capacity fully over the period. We simply do not have the data to pin this down very precisely. However, it is clear that overall investment responded positively to the terms of trade, and this helps explain why the trade balancewas strongly negatively related to terms-of-trade shocks, even though the export price effect in isolation should cause a higher trade surplus. B. Consumption, Savings, and the Current Account As discussed above, tory suggests that after a permanent positive terms-of-trade shock, consumption

should rise, and savings should as a first approximation be unaffected. The data broadly confirms these implications. The only exceptionswere Ecuador,Egypt, Syria,Kuwait, and the Emirates, whichdisplayed an insignificant or indeed negative consumption response to the terms-of-trade shock. However, the formerthree countries were not really oil exporters before the firstoil price shock, whichconsequentlyhad a smaDler positive impact on their wealth. For Kuwait and the Emirates,the shock seems to have induced a largeincrease in the consumptionof leisure,made up for by extensive immigration which pushed down averagelevels of consumptionper capita. The data also suggests government consumptionmay respond almosttwice as stronglyas private sectorconsumption,perhaps because govemmentssystematicallyview shocks as more permanent than is the case. Numerically,panel estimates suggest the 95% confidence interval (0.310, 0.426) for the elasticity of consumption with respect to the terms of trade. Given the lack of an impact on savings, and the strong response of investment, the trade balance and currentaccount should be expectedto decrease,and withoutexception thesewere the estimatedresponses. C. Nontradable Prices and Quantities shocks are verystronglyassociated with real exchange rate appreciation, i.e., increasesin Terms-of-trade the relativeprice of nontradables. Panel estimates of the elasticitysuggest the 95% confidence interval (-

page 16

0.606,-0.695), andonlyCongodisplayed an insignificant response. The shocksare also associatedwith significant increase in valueaddedwithineachcategoryof nontradables.Thisconfirmsthe importance of spendingeffectsas a key mechanismin the transmission of terms-of-trade shocksto the rest of the economy: higherwealthleadsto greater demand for nontradables, andhenceto an increase in theirrelative pricesandquantities.At a countrylevel,the mainexceptionwasEcuador, whichdisplayedinsignificant responses in the value addedof eachnontradable sector. D. DutchDiseaseEffects Theliterature stronglyfocuseson DutchDiseaseeffects;here,the two non-oiltradables(agriculare and manufacturing) are the closestapproximation we have to a DutchDiseasesector. Yetwe failedto detect clearcontractions in eitherof themin response to a rise in the priceof oil;the onlyexception wasNigerian agriculture. Oilvalueaddeddidrespond negatively overthesubperiod1965-198 1, but this simplyreflects the impositionof OPEC quotas:the oil sectoris best not viewedas facingexogenouslygiventermsof trade,sincepartof OPEC'sstrategyat the timeclearlywasto engineer pricerisesby restraining output. Theabsence of DutchDiseaseeffectsmaybe partlyexplainedby eitherthe compression in the outputof the oil sector,or by its typically beingan 'enclave'sector whichdoes not participate in domesticfactor markets.Bothfactors wouldact to reduce the pullof resourcesawayfromothersectors,as mentioned in the theoretical section.Nevertheless, asdiscussed above,the spending effectwasoperational, increasing relativeprices and outputin the nontradable sector. To the extentthat the nontradablesector maybe competing withagricultureand manufacturing for scarcefactorsof production, or if nontradables are themselves intermediate inputs into othersectors,one mighthave expecteda small bout of the Dutch Disease. E. Growth Giventhattherelativepricesbetween oil andotherproducts changedso dramatically overthe period,and that theoil sectoris so importantin oursample,extremely seriousindexnumberproblemsarisein even
definingand measung aggregateoutput. We therefore feel that it is not analyticallyuseful to talk about

the aggregate economy, and we insteadseparatethe oil fromthe non-oilsector. The formerhas already been mentioned.In the non-oilaggregate, we do see an effecton GDP,drivenmainlyby the expansion

page 17

responses. Givenour in nontradables; Ecuadorand Nigeriawere the only countrieswith insignificant this response is best seenas beingdrivenby capitalformation in the nontradable evidence on investment, sector. Numerically, panel esimates of the elasticityof non-oilvalue addedwithrespectto the termsof by outliers;the trade suggestthe 95% confidenceinterval(0.206,0.297). This is howeverinfluenced medianelasticityfrom the countryregressions is 0.381.

VI. SomeWarnings regarding Extrapolation

When interpreting and aboveaUextrapolating our results,severalissuesmustbe considered. First,the causes of the oil priceshock were differentfrom the causes of the forecastfuture increase in primary commodityprices. IJ our sampleexport prices rose because of oligopolistic coordinationaimed at reducing output. Someresponsesmightwellbe differentfor a truly exogenous terms-of-trade shock.In particular,both investment and outputin the favoredsector wouldprobablyrise; since outputwas no longerconstrained, we shouldalso expecta biggerincreasein wealthandhenceconsumption. Second,ourresultsneedonlybe validfor a permanent changein the termsof trade. A temporary terms-oftradeshockmay lead to someintertemporal substitution in consumption andinvestment, but is unlikely to causea permanent changein output. Third,an economy's response to an increase in thepriceof primary commodities willclearly hingeon how responsewill alsodependon what shareof its capital dependent it is on primaryexports. Its investment goodsis imported. Fourth, we mightexpectthe availability of external financeto be a crucialfactorinteracting with changes in the terms of trade,and oil exportersin the 1970'sprobablyenjoyedbetteraccessto the worldcapital markets an is trueof the developing worldtoday.On the otherhand,this may wellchangeif the trends in commodity priceschange. wasteda lot of theirwindfallrevenueson prestigeprojects Fifth, in the 1970'sand 1980'soil-exporters with little impacton output. Presumably, developing countrieshaveby nowlearntthe lesson,and next timearoundwillmakebetteruse of theirluck.

page 18

VII. Conclusions
Weconsidered a panelof oil-exporters thatimporta significant fractionof their capitalgoods. We found thatpermanent terms-of-trade shockshaveno impacton savings, a strongpositiveimpacton investment, and a negativeimpacton the currentaccount. There is also evidenceof a long-runeffecton output, particularly of non-tradables.Realexchange-rate appreciations are a key mechanism in triggering the resourcereallocation.We failedto find any evidencethat the Dutchdisease is a majorphenomenon. Finally, the responseof expenditureto terms-of-trade shocks is not very sensitiveto whetherthe expenditure comesfromthepublicor privatesector. As discussed in sectionVLestimating an in-sample response to a terms-of-trade shockis mucheasierthan forecasting an out-of-sample response.Whileour qualitative resultsmayhold in quitegeneralcontexts, our quantitative estimates almostcertainly will not. Let us nevertheless try to obtainsomeidea of the potential impactof theforecast increase in commodity priceson the growthrateof commodity exporters. As a lower-bound scenario, assume thatthecommodity sectordoesnotgrowat all,thatthenon-commodity sectoraccountsfor 75%of GDP,andthat it respondslikethe non-oilsectorin our sample. An increase in the trendgrowthrateofcommodity pricesfrom-1.5%p.a to 0.7%p.a. then impliesan increase in the orderof 0.6%in the trendgrowthrateofper capitaGDPat constant prices. Tothe extentthatcommodity outputincreases, andthatthisdoesnot pullresources awayfromthe restof the economy, the growth effect
is magnified.

page 19

Appendix I: Full Data Description and Sources.

We consider the period 1965-1989. Our sample consists of the 18 oil-exporterslisted below. Algeria Bahrain Congo Ecuador Egypt Gabon Indonesia hran Lrq Kuwait Mexico Nigeria Oman Saudi Arabia Syria Trinidad and Tobago United Arab Emirates Venezuela

For all countries, fuels accountedfor over 50% of total exports over at least half the sample period; for most countries, fuels accounted for over 70% of total exports over at least three quarters of the sample period. We used data on the following variables: (1) Termsof Trade (TOT), computed as Merchandise exports deflator US Dollar (USD) I merchandise imnports deflator USD, base 1987= 1.0. (2) A Debt Crisis dummy (DEBT),set to unity for Ecuador,Mexico,Nigeria and Venezuelaover 19821989, and to zero elsewhere.

(3) World Real Interest Rates (INTRATE),as computed by Barro and Sala-i-Martin. (4) The Real Exchange Rate (REALEXRA), or relative price of tradables to nontradables. This is computed as USA GDP derlator USD I local GDP deflator USD. index 1987. (5) Gross DomesticProduct (GDP),per capita, constant 1987marketprices,Local CurrencyUnit (LCU). (6) Gross National Product (GNP),per capita, constant 1987 marketprices, LCU. (7) Consumption (CONS), per capita, constant 1987 prices, LCU. (8) Private Consumption (CONSPRIV),per capita, constant 1987prices, LCU. (9) General GovernmentConsumption(CONSGOVT),per capita, constant 1987 prices, LCU. (10) Gross Domestic Savings (SAVINGS),per capita, constant 1987prices, LCU. (11) Fixed Investment (INVT), per capita, constant 1987prices, LCU. (12) Private fixed Investment (INVTPRIV),per capita, constant 1987prices, LCU. (13) General Government fixed Investment (INVTGOVT),per capita, constant 1987 prices, LCU. (14) Balance of merchandise Trade (TRADBALA),per capita, constant 1987 prices, LCU. (15) Value added in the Oil sector, constant prices (OILVA),LCU. (16) Value added in the Non-oil sector (NONOILVA), constant prices, LCU. (17) Value added in Agriculture(AGRICULT),constant prices, LCU.

page 20 (18) Valueadded in Manufacturing (MANUFACT),constant prices, LCU. (19) Valueadded in Construction (CONSTRUC),constant prices, LCU. (20) Valueadded in Public Utilities (UTILITY),constant prices, LCU. (21) Value added in Services (SERVICES),constant prices, LCU. (22) Value added in Transportation and Communications(TRANSPORT),constant prices, LCU. (23) Valueadded in Wholesale and Retail Trade (TRADE), constant prices, LCU. (24) Valueadded in Other Services (OTHRSERV),constant prices, LCU. (25) Total Labor force (LABOR). Variables 1, 4.. 14 and 25 came from the World Ban's DAD database (except for Mexico's terms of trade, obtained from Mexico's central bank), and variables 2 and 3 are as described above, They are available for all countries for the whole period. Variables17, 18 and 21 were from the WorldBank's STARS database, and all other variables came from World Bank CountryEconomic Memoranda. They were not available for Iraq, nor before 1969. In all cases, the ultimatesources are nationalcentral banks, national statistical services, and estimates by World Bank missions.

page 21

Appendix I: Full Panel Regression Results.

Table 1. Panel regression: In yi, = a, + Dependent variable Ln Real exchange rate Ln Consumption Ln Private Consumption Ln Govt Consumption Ln Savings Ln Investment Ln Private Investment Ln Govt Investment Trade Balance Ln GDP Ln Oil value added Ln Non-oil value added Ln Agriculture Ln Manufacturing Ln Construction Ln Public Utilities Ln Services Ln Transportation Ln Trade Ln Other services

1B InTOTi, + e, . Sample period: 1973, 1981, 1989.

Terms of trade Tratio -6.8338 4.4355 4.2327 4.1049 0.2177 4.3860 3.5540 4.5936 -2.8571 -0.1613 -1.7479 4.4454 0.7940 1.2890 2.7997 1.5724 1.0117 1.7262 0.8500 3.7993
2

Terms of trade Coefficient -0.5617 0.3594 0.3413 0.4085 0.0312 0.4780 0.4895 0.5731 -15.1420 -0.0123 -0.3862 0.2211 0.0706 0.1710 0.3314 0.2553 0.0951 0.1958 0.1419 0.2704

N 0.3973 0.3917 0.3898 0.4810 0.6939 0.5267 0.6656 0.6030 24.9415 0.3689 0.9133 0.2055 0.3787 0.5432 0.4852 0.6039 0.4003 0.4323 0.5207 0.2713 54 54 54 54 54 54 54 54 53 54 23 23 46 36 25 21 46 22 21 22

0.5716 0.3598 0.3386 0.3250 0.0014 0.3547 0.2652 0.3761 0.1936 0.0007 0.2763 0.7118 0.0220 0.0767 0.4655 0.2610 0.0353 0.2986 0.0936 0.6734

page 22
Table 2. Panel regression: In Yi In TOTi, + 12 In TOTi,.I + .. + E,. 0 = cc + 3,B

Terms of trade sum Tenns of trade TR2 I N of coefficients ratio -0.6508 -28.7336 0.6747 0.2488 431 Ln Real exchange rate Ln Consumption 0.3681 12.5055 0.2766 0.3248 431 Ln PrivateConsumption 0.2902 10.2829 0.2105 0.3114 431 Ln Govt Consumption 0.5206 12.9296 0.2892 0.4442 431 Ln Savings -0.0679 -1.1228 0.0636 0.6660 429 Ln Investment 0.5875 13.1991 0.3217 0.4706 413 Ln Private Investment 0.6831 11.7508 0.2902 0.6146 413 Ln Govt Investment 0.5970 12.3281 0.2874 0.5120 413 Trade Balance -15.4053 -9.2591 0.2145 15.9832 395 Ln GDP -0.0201 -0.7378 0.0158 0.3000 431 Ln Oil value added -0.3640 -6.3850 0.2470 0.3548 178 Ln Non-oil value added 0.2519 10.8723 0.4316 0.1441 178 Ln Agriculture 0.0701 2.2620 0.0691 0.2155 293 Ln Manufacturing 0.2306 3.7061 0.0803 0.4022 240 Ln Construction 0.3967 9.1806 0.3581 0.2709 193 Ln Public Utilities 0.3163 6.3042 0.3226 0.2787 158 Ln Services 0.0965 2.7620 0.0276 0.2438 293 Ln Transportation 0.2332 5.4244 0.1758 0.2467 175 Ln Trade 0.2331 4.6300 0.1578 0.2382 158 Ln Other services 0.2606 9.7659 0.3833 0.1531 175 For the real exchangerate, the trade balance, non-oil value added, agriculture,construction, utilities, and trade, the regressor matrix also includes DEBTj,;in all the investment equations, it also includes DEBT 21 and In LNTRATE,. Sample period: 1965-1989. Dependentvariable

page 23 Table 3. Panel Regression: In y, 1 = oc + 01 In TOT,, + 132In TOT,,., + .. + ell. unrestrictedacross the subperiods 1965-1980and 1981-1989. Dependent variable coeficients

N R2 Terms of trade Terms of trade P-value for sum of coef. coef. sum of coef. 1965-1980 1981-1989 stability 0.0235 0.6807 0.2471 431 Ln Real exchange rate -0.6746 -0.5484 0.3791 0.3110 0.25 0.2815 0.3245 431 Ln Consumption La PrivateConsumption 0.3042 0.2276 0.403 0.2140 0.3115 431 Ln Govt Consumption 0.5395 0.4211 0.0931 0.2974 0.4428 431 Ln Savings -0.1439 0.1884 0.0894 0.0747 0.6637 429 Ln Investment 0.3124 0.3510 0 0.4505 0.4253 413 Ln Private Investment 0.4647 0.1246 0 0.3689 0.5819 413 Ln Govt Investment 0.2590 0.6564 0 0.4226 0.4627 413 Trade Balance -16.9939 -10.9972 0.419 0.2181 15.9906 395 -0.0488 0.0699 0.0893 0.0274 0.2990 431 Ln GDP Ln Oil value added -0.4854 -0.0789 0.0258 0.2807 0.3491 178 Ln Non-oil value added 0.2449 0.2710 0.703 0.4341 0.1448 178 Ln Agriculture 0.1125 0.0087 0.259 0.0784 0.2153 293 Ln Manufacturng 0.2021 0.2877 0.281 0.0908 0.4018 240 Ln Construction 0.3898 0.4083 0.76 0.3601 0.2721 193 Ln Public Utilities 0.3213 0.3142 0.713 0.3258 0.2802 158 Ln Services 0.0842 0.1241 0.399 0.0341 0.2439 293 Ln Transportation 0.1601 0.4157 0.0798 0.2020 0.2445 175 Ln Trade 0.2396 0.2250 0.936 0.1585 0.2400 158 0.53 0.3883 0.1536 175 0.2597 0.2674 Ln Other services For the real exchangerate, the trade balance,non-oil value added, agriculture,construction, utilities, and trade, the regressor matrix also includesDEBTk;in all the investmentequations, it also includes DEBT,, and In INTRATE. Sample period: 1965-1980, and 1981-1989. The P-value denotes the minimum significancelevel at which we can reject the null hypothesisof identicalterms-of-tradecoefficientsacross subperiods,using the F-test describedin Hsiao (1986), ch. 2.2, equation 2.2.20.

page 24

Table 4. Share of capital goodsin totalinportsin 1973, in 1981, andinthe latestavaiableyearupto 1989 inclusive. Country Algeria Congo Ecuador Egypt Gabon Indonesia Iran
Iraq

Shareof capital goodsin theyear 1973 1981 Latest 0.3720 0.3834 0.2570 0.3868 0.3874 0.3620 0.4117 0.4582 0.3397 0.2477 0.2816 0.2303 0.4066 0.4094 0.4009 0.4104 0.3536 0.3774 0.3779 0.2832 0.3884
0.3297 0.5328 0.4478

Latestavailable year1989if notstated 1986

1983 1983
1983

Kuwait Mexico Nigeria Oman SaudiArabia Syria Trinidad andTobago UnitedArabEmirates Venezuela

0.3442 0.4438 0.4012 0.3101 0.3524 0.2369 0.1313 0.3801 OA688

0.4103 0.4707 0.4402 0.3908 0.4047 0.2171 0.2242 0.3568 0.4344

0.2961 0.3260 0.3812 0.3685 0.3870 0.2592 0.2729 0.3025 0.4682

1986

1986 1986 1988

page 25

Results. Regression AppendixIII: Country-by-country

Sampleperiod: 1965-1989. For eachdependent variable, two regressions arecarriedout.
+ P(L)In xk, + e,1, j identicalacrosscountries.Resultsare in the form: Regression 1: Inyj, = uxi TOTCoef TOTT-stat DebtCoef DebtT-stat IntRate Coef IntRateT-stat Y variable 2 P-value R a N

unrestrctedacrosscountries.Resultsare in the form: Regression 2: Inyt = a; + 3P(L)Inxt + ej,,,B Country TOTCoef TOTT-stat DebtCoef DebtT-stat IntRateCoef IntRateT-stat N P-value R2 a of no countryThe P-valuein regressionI refersto the Lagange Multipliertest of the nullhypothesis across ofcoefficient stability 2 to theF-testofthe nullhypothesis in regression specific effects; the P-value variablewas not are not reported,the corresponding and its T-statistic countries. Wherea coefficient included in the regression. REALEXRA Algeria Bahrain Congo Ecuador Egypt Gabon Indonesia hran Iraq Kuwait Mexico Nigeria Oman SaudiArabia Syria Tobago Emirates Venezuela CONS Algeria -0.651 0.675 -0.447 -2.53 -0.0493 -0.263 -0.818 -0.629 -0.467 -0.691 -0.866 -0.728 -0.637 -0.731 -0.853 -0.783 -0.214 -1.14 -0.577 -0.359 0.804 0.368 0.277 0.488 -28.7 0.249 -6.82 -13.6 -0.257 -1.89 -3.69 -6.29 -5.25 -11.5 -12.3 -13.4 -2.29 -8 -15.5 -14.8 -1.55 -5 -9.97 -5.48 0.203 12.5 0.325 6.03 0.325 431 6.07 4.21e-17

0.166

1.87

0.178 0.489

1.53 4.99

0.354 431 431

3.28 2.7e-14 5.81e-07

page 26 Bahrain Congo Ecuador Egypt Gabon Indonesia Iran Iraq Kuwait Mexico Nigeria Oman SaudiArabia Syria Tobago Emirates Venezuela CONSPRIV Algeria Bahrain Congo Ecuador Egypt Gabon Indonesia Iran Iraq Kuwait Mexico Nigeria Oman SaudiArabia Syria Tobago Emirates Venezuela 0.917 0.145 -0.261 -0.64 0.682 0.596 0.292 0.186 -0.00163 0.276 0.327 0.671 1.09 -0.513 0.794 -0.0429 0.225 0.609 0.29 0.21 0.48 0.538 0.202 -0.23 -0.666 0.539 0.544 0.256 0.0561 0.00356 0.297 0.312 0.425 0.99 -0.502 0.856 -0.232 0.23 0.581 0.521 0.289 4 0.613 -1.57 -2.34 5.54 5.45 3.95 2.15 -0.0244 0.993 3.25 9.9 16.7 -3 2.84 -0.603 3.32 0.25 10.3 0.311 6.25 2.47 0.898 -1.45 -2.56 4.6 5.23 3.64 0.682 0.056 1.12 3.27 6.59 16 -3.09 3.21 -3.42 3.57 0.237 12.9 0.444

431 431

0 1.94e-10

431

CONSGOVT

431

3.82e-08

page 27 Algeria Bahrain Conglo Ecuador Egypt Gabon Indonesia Iran Iraq Kuwait Mexico Nigeria Oman Saudi Arabia Syria Tobago Emnirates Venezuela 0.51 2.21 -0.0562 -0.454 -0.501 1.03 0.995 0.418 0.371 -0.0119 0.106 0.414 1.03 1.23 -0.557 0.617 0.329 0.195 0.567 4.34 6.62 -0.163 -1.88 -1.26 5.74 6.25 3.89 2.95 -0.123 0.263 2.83 10.5 13 -2.25 1.52 3.18 1.97 0.363 -1.12 0.666 2.37 4.5 o.0264 -2.83 -0.322 2.65 5.31 0.419 -1.24 -4.72 0.383 -0.768 1.42 -1.82 -2.27 1.62 -2.01 -1.66 0.601 13.2

431

0 9.34e-16

SAVINGS

-0.0679 0.0636 0.462 Algeria 2.76 Babrain 0.015 Congo -1.14 Ecuador -0.212 Egypt 0.787 Gabon 1.4 Indonesia 0.0746 Iran -0.258 Iraq -0.76 Kuwait 0.256 Mexico -0.186 Nigeria 0.232 Oman -0.288 SaudiArabia -0.931 Syria 1.09 Tobago -0.345 Enmirates -0.271 Venezuela 0.303 INVT 0.587

429

429 -0.368

9.18e-11 --3.34 -o.00073

1 -11

page 28 Algeria Bahrain Congo Ecuador Egypt Gabon Indonesia Iran Iaq Kuwait Mexico Nigeria Oman SaudiArabia Syria Tobago Emirates Venezuela 0.322 0.743 1.76 0.282 -o.246 -1.3 1.41 1.41 0.401 1.03 0.316 1.01 0.624 0.68 1.33 -0.554 1.77 -0.0034 0.25 0.661 0.683 0.29 0.756 1.74 1.08 -0.191 -0.749 1.29 1.36 0.524 1.62 0.0359 0.448 0.495 1.15 1.74 -0.54 1.54 -0.0215 0.00987 0.701 0.471 6.04 5.09 0.618 -0.798 -2.81 7.3 8.5 3.57 7.63 3.19 1.89 3.67 6.84 13.7 -2.09 3.87 -0.0322 2.11 0.359 11.8 0.615 5.12 4.2 1.97 -0.516 -1.35 5.57 6.82 3.89 9.97 0.301 0.702 2.43 9.62 14.9 -1.69 2.8 -0.169 0.0694 0.432 413 4.16e-10 -0.00298 -0.00417 0.00402 -0.00117 0.00772 -0.00653 0.00233 -0.00636 -0.00198 -0.00202 o.oD0483 -0.00697 -0.00163 -0.000509 0.00413 -0.000137 -0.00318 -0.00265 -1.42 -1.99 1.61 -0.39 3.83 -3 1.13 -3.04 -0.955 -0.975 0.19 -2.69 -0.785 -0.245 1.98 -0.0675 -1.53 -1.05

0.0597

0.248

-0.0707 -0.389

-0.264 -1.78

-0.487 413 -0.675 413

-1.93 3.61e-27 -4.7 3.Ole-10

i'VTPlRV Algeria Bahrain Congo Ecuador Egypt G3abon Indonesia Inn Irq Kuwait Mexico Nigeria Oman Saudi Arabia Syria Tobago Emirates Venezuela

8.03e-05 -0.00581 -0.00449 -0.000525 -0.00218 0.00223 -0.00215 0.00169 -0.00478 0.00611 -0.000757 0.00053 -0.0115 0.0094 -0.00121 0.00603 0.00132 -0.00389 -0.0044

0.0938 -2.31 -1.78 -0.175 -0.603 0.92 -0.825 0.683 -1.9 2.45 -0.305 0.173 -3.71 3.77 -0.486 2.41 0.545 -1.56 -1.46

0.0579

0.2

-0.141 -0.635

-0.439 -2.42

-0.488 413

-1.61 7.07e-38

page 29 lNVTGOVT Algeria Babrain Congo Ecuador Egypt Gabon Indonesia Iran Iraq Kuwait Mexico Nigeria Oman Saudi Arabia Syria Tobago Emirates Venezuela TRADBALA Algeria Bahrain Congo Ecuador Egypt Gabon Indonesia Iran Iraq Kuwait Mexico Nigeria Oman Saudi Arabia Syria Tobago 0.597 0.287 0.737 1.8 -0.747 -0.332 -1.88 1.51 1.5 0.289 0.973 0.444 2.05 0.699 0.583 1.1 -0.537 2.06 0.00568 0.606 0.616 -15.4 0.214 -2.71 -2.55 -1.92 -3.2 -0.615 -5.58 -85.3 -2.31 -1.26 -1.37 -1.96 -5.06 -4.06 -0.522 -0.784 -9.82 12.3 0.512 5.3 4.61 -1.45 -0.953 -3.59 6.9 7.98 2.27 6.36 3.96 3.41 3.64 5.18 9.97 -1.79 3.98 0.0475 4.52 0.406 -9.26 16 -1.21 -0.362 -0.267 -0.666 -0.0809 -1.64 -28.1 -1.03 -0.506 -0.582 -0.207 -1.62 -0.781 -0.277 -0.138 -1.27 -0.173 413 -1.44 6.48e-15 -0.00084 0.00205 -0.0035 0.0096 0.000259 0.0115 -0.0139 0.00315 -0.00791 -0.00273 -0.00245 0.000309 -0.00326 -0.004 -0.000543 0.00311 -0.00255 -0.00271 -0.00213 -1.18 0.865 -1.48 3A1 0.0761 5.03 -5.68 1.35 -3.34 -1.16 -1.05 0.107 -1.11 -1.7 -0.231 132 -1.12 -1.15 -0.747

0.0624

0.229

0.072 -0.266

0.238 -1.08

-0.471 413 -11 395

-1.65 4.86e-14 -3.17 4.07e-13

-0.246

-0.081

0.179 1.09

0.0449 0.323

Emirates
Venezuela

-1.73
-77.1 0.868

-0.676
-32.2 6.93 -1.03 395 -0.281 3e-108

page 30 GDP Algeria Bahrain Congo Ecuador Egypt Gabon Indonesia Iran Iraq Kuwait Mexico Nigeria Oman SaudiArabia Syria Tobago Emirates VJenezuela GNP Algeria Bahrain Congo Ecuador Egypt Gabon Indonesia hran Iraq Kuwait Mexico Nigeria Oman SaudiArabia Syria Tobago Emirates Venezuela -0.0201 0.0158 0.249 0.051 -0.065 -0.478 -0.776 0.607 0.516 0.0491 -0.198 -0.589 0.261 -0.0206 0.352 0.17 -0.35 0.69 -o.324 -0.142 0.419 0.0163 0.0112 0.255 -0.189 0.00967 -. 419 -0.753 0.589 0.495 0.0917 -0.146 -0.472 0.274 0.000775 0325 0.288 -0.336 0.797 -0.313 -0.126 0.399 -0.738 0.3 3.18 0.231 -0.284 -2.97 -2.94 5.1 4.88 0.687 -2.38 -9.13 0.972 -0.212 5.37 2.69 -2.12 2.55 -4.71 -2.17 0.241 0.621 0.289 3.34 -0.871 0.0432 -2.66 -2.92 5.06 4.79 1.31 -1.78 -7.48 1.04 0.00816 5.07 4.68 -2.09 3.01 4.66 -1.96 0.236

431

431

431

2.04e-14

431

page 31 OILVA Algeria Bahrain Congo Ecuador Egypt Indonesia Iran Kuwait Mexico Nigena Oman Saudi Arabia Tobago Venezuela NONOILVA Algeria Bahrain Congo Ecuador Egypt Indonesia Iran Kuwait Mexico Nigeria Oman Saudi Arabia Tobago Venezuela -0.364 0.247 -0.122 1.1 1.09 -0.138 1.94 0.00744 -0.776 -0.537 -0.173 0.187 -0.202 -0.198 -1.16 -0.559 0.462 0.252 0.432 0.336 0.235 0.92 -0.0364 0.754 0.591 0.28 0.091 0.417 -0.00423 0.847 0.381 0.836 0.0778 0.721 0.0701 0.0691 0.183 0.607 0.00972 -0.108 0.001 -6.39 0.355 -0.916 1.51 1.01 -0.398 2.38 0.0191 -6.38 -2.16 -0.46 0.524 -0.326 -2.01 -1.04 -4.93 0.33 10.9 0.144 7.39 0.952 2.51 -0.355 2.71 3.74 6.75 1.07 2.48 -0.0332 4.01 11.3 2.21 1.72 0.113 2.26 0.216 2.08 1.36 0.0466 -0.631 0.00387 178 4.32e-43

178 -0.0682 178

0.00347 -1.72 8.75e-44

0.137

2.32

0.102 -0.117

1.48 -0.938

-0.246 178 -0.167 293

-3.29 3.54e-10 -3.39 2.33e-18

AGRICULT Algeria Babrain Congo Ecuador Egypt

-0.00651

-0.0722

page 32 Gabon Indonesia Ian Kuwait Mexico Nigeria Oman SaudiArabia Syria Tobago Emirates Venezuela MANUFACT Algeria Bahrain Congo Ecuador Gabon Indonesia Iran Kuwait Mexico Nigeria Oman SaudiArabia Tobago Emirates Venezuela -0.0547 0.151 0.0964 0.17 0.228 -0.193 -0.203 0.0405 0.63 0.955 0.14 -0.0647 0.304 0.231 0.0803 0.298 -0.227 -0.0879 0.232 0.278 0.918 0.199 0.233 0.193 0.188 -0.571 0.102 0.922 1.53 0.0121 0.332 0.397 0.358 0.714 0.382 2.4 0.00434 0.945 0.707 0.053 0.133 -0.208 1.29 1.18 2.21 0.744 -2.11 -1.25 0.533 2.41 4.13 0.709 -0.756 0.2 3.71 0.402 1.84 -0.239 -0.188 0.737 0.578 4.26 1.33 0.73 0.468 1.2 -1.92 0.733 2.18 4.24 0.0833 0.368 9.18 0.271 9.07 0.893 3.78 0.026 1.96 3.08 0.739 0.906

-0.079 -0.216

-0.601 -2.06

-0.081 293 240

-0.768 0.000138 3.25e-28

240 -0.306 193

5.92e-05 -4.19 1.44e-2B

CONSTRUC Algeria Babrain Congo Ecuador Egypt Indonesia Iran Kuwait

-0.0965

-1.03

page 33 Mexico Nigeria Oman Saudi Arabia Syria Tobago Venezuela UTILIIY Bahrain Congo Ecuador Egypt Indonesia Iran Kuwait Mexico Nigeria Oman Saudi Arabia Tobago Venezuela SERVICES Algeria Bahrain Congo Ecuador Egypt Gabon Indonesia Iran Kuwait Mexico Nigeria Oman Saudi Arabia Syria Tobago 0.475 0.64 1 0.704 0.79 1.9 0.232 0.732 0.316 0.323 0.074 0.235 -0.271 1.02 0.492 0.354 0.138 1.56 0.491 1.25 0.406 1.17 0.202 0.536 0.0965 0.0276 0.114 0.357 0.588 0.0865 -0.449 0.277 0.497 0.0653 -0.296 0.17 0.316 0.0162 0.188 0.985 0.489 1.63 2.91 2.75 12.1 1.96 3.93 2.95 0.195 6.3 0.279 0.131 0.28 -1.22 1.6 1.62 3.73 0.709 2.2 1.68 2.59 4.94 1.83 1.94 0.259 2.76 0.244 1.16 0.713 2.52 0.45 -1.55 0.942 3.76 0.713 -3.43 0.676 3.31 0.0887 2.2 3.36 1.88 0.0674 -0.488 0.567 -2.26

-0.825 193 0.259 158 0.603

-6.37 2.5e-15 3.21 1.85e-50 4.86

0.429 0.0379

2.46 0.133

0.11 1SS 293

0.639 0.00748 1.22e-20

page 34

Emirates Venezuela TRANSPORT Bahrain Congo Ecuador Egypt Indonesia hban Kuwait Mexico Nigeria Oman Saudi Arabia Syria Tobago Venezuela TRADE Bahrain Congo Ecuador Egypt Indonesia Iran Kuwait Mexico Nigeria Oman Syria Tobago Venezuela OTHRSERV Bahrain Congo

0.19 -0.0328 0.277 0.233 0.176 0.962 1.18 -0.00442 1.63 0.636 0.428 0.106 1.58 0.37 1.46 0.144 0.507 1.25 0.0341 0.46 0.233 0.158 1.1 1.59 -0.0376 1.25 0.304 0.569 0.16 0.344 0.00838 1.04 0.443 0.914 -0.0189 0.61 0.261 0.383 -0.554 0.634

0.859 -0.369 0.225 5.42 0.247 1.99 1.64 -0.0236 2.99 2A6 5.28 0.641 2.73 1.55 3.53 2.18 1.12 2.3 0.451 0.22 4.63 0.238 2.75 2.69 -0.242 2.8 1.42 8.51 1.17 1.27 0.0409 3.06 1.18 2.03 -0.26 0.182 9.77 0.153 -2.02 1.56

293 175

2.01e-05 5.47e-52

175 -0.132 158

0.000119 -2.05 4.43e-47

0.111

1.27

0.0616 -0.0632

0.556 -0.315

-0.373 158 175

-3.09 6.55e-10 1.14e-50

page 35 Ecuador Egypt Indonesia hran Kuwait Mexico Nigeria Oman Saudi Arabia Syria Tobago Venezuela 0.0369 0.816 0.437 0.305 -0.0263 0.389 0.00246 0.576 0.42 0.7 0.861 0.0801 0.663 0.347 2.65 2.98 6.64 -0.28 1.19 0.0182 2.46 11.3 2.72 2.78 1.87 0.125

175

1.53e-08

page 36

References.
Backus,D. (1993), 'Interpreting comovementsin the tradebalance and the terms of trade,' Journal of InternationalEconomics34, p. 375-387. Bean,C.R. (1986), 'The Termsof Trade, Labour Supplyandthe CurrentAccount,' EconomicJournal Supplement96, p. 3846. Breusch,T.S. & Pagan, A.R. (1980), 'The Lagrange MultiplierTest and its Applicationsto Model Specificationin Econometrics,' Review of EconomicStudies 47, p. 239-253. Bruno, M. & Sachs J. (1982), 'Energy and Resource Allocation: A Dynamic Model of the Dutch Disease,' Review of EconomicStudies 49, p. 845-859. Corden,W. M. (1984),'Booming Sector and Dutch DiseaseEconomics: Survey and Consolidation,' Oxford EconomicPapers 36, p. 359-380. Corden, W. M. & Neary J.P. (1982), 'Booming Sector and Deindustrialisationin a Smal Open Economy,' EconomicJournal 92, p. 825-848. De Gregorio,J. (1992),'EconomicGrowthin Latin America,'Journalof DevelopmentEconomics39, p. 59-84. Etherington,D. & YainshetA. (1988), 'The Impact of IncomeTerms of Trade for Coffee on Capital Goods hnports and Investmentin Ethiopia,' EasternAfrica Economic Review 4, p. 48-52. Gavin,M. (1990), 'StructuralAdjustmentto a Terms of Trade Disturbance,' Journal of International Economics28, p. 217-243. Harberger, A.C. (1950), 'Currency Depreciation, Income, and the Balance of Trade,' Journal of Political Economy58, p. 47-60. Hsiao, C. (1986), Analysis of Panel Data. Laursen,S. & MetzlerLA. (1950),'FlexibleExchangeRatesandthe Theoryof Employment,'Review of Economicsand Statistics 32, p. 281-299. Matsuyama,K (1988), 'Terms-of-Trade, Factor Intensitiesand the Current Account in a Life-Cycle Model,' Review of Economic Studies 55, p. 247-262. Morley,S. (1992), 'On the Effectof Devaluation during StabilisationPrograms in LDCs,' Reviewof Economics& Statistics74, p. 21-27.

page 37 Neary,J. P. & S. Wijnbergen (1984), 'Can an Oil DiscoveryLead to a Rcccession?A Comment on Eastwoodand Venables,' EconomicJournal94, p. 390-395. Obstfeld,M. (1982),'AggregateSpendingand the Terms of Trade:Is There a Laursen-Metzler Effect,' QuarterlyJournal of Economics97, p. 251-270. Persson, T. & Svensson L.E.O. (1985), 'Current Account Dynamics and the Terms of Trade: Harberger-Laursen-Metzler TwoGencrations Later,' Journalof PoliticalEconomy93, p. 4 3 -65 . Salter, W.E.G. (1959), 'Internal and External Balance: the Role of Price and Expenditure Effects,' EconomicRecord 35, p. 226-238. Schmidt-Hebbel,K & Serven L. (1993), 'Dynamic Responsesto External Shocks in Classical and KeynesianEconomies,' CEPR ConferencePaper. Sen, P. & Turnovsky S.J. (1989), 'Deteriorationof the Termsof Trade and Capital Accumulation:a Re-Examination of theLaursen-Metzler Effect,' Journalof Intemational Economics26, p. 227250. Svensson, L.E.O. & Razin A. (1983), 'The Terms of Trade and the Current Account: the LaursenHarberger-MetzlerEffect,' Journal of Political Economy91, p. 97-125. Warner, A. (1992), 'Did the Debt Crisis Cause the Investment Crisis?', Quarterly Journal of Economics 107,p. 1161-1186. (forthcoming,1994), 'Did the Debt Crisis or the Oil Price Decline Cause Mexico's Investment Collapse?'. Wijnbergen, S. (1984),'The "DutchDisease": A Disease AfterAll?', EconomicJournal 94, p. 41-55. (1984b), 'Inflation,Employment,and the Dutch Disease in Oil-ExportingCountries: A Shortrun DisequilibriumAnalysis,' QuarterlyJournal of Economics,p. 233-250.

Policy Research Working Paper Series Contact for paper G. Langton 38392

Title WPS1384 Fiscal Decentralizationand IntergovemmentalFinances in the Republicof Albania WPS1385 FiscalFederalismDimensionsof Tax Reformin Developing Countries

Author David Sewell ChdistineI. Wallich

Date November1994

Robin Boadway Sandra Robens Anwar Shah BrentBorrell

November1994

C. Jones 37754

WPS1386 EU BananaramaIll

December1994

G. llogan 33732 C. Jones 37699

WPS1387 Fiscal Decentralizationand the Size of Govemment: An Extensionwith Evidencefrom Cross-CountryData WPS1388 Does VoiceMatter? For Public Accountability, Yes

Jaber Ehdaie

December1994

SamuelPaul

December1994

B. Moore 35261 L. Suki 33974 J. Ngaine 37947 S. Falon 38009 C. Jones 37699

WPS1389 Ruble Overhangand Ruble Shortage: Patrick Conway WereThey the SameThing? WPS1390 Regional Integrationand the Baltics: Pirita Sorsa WhichWay? WPS1391 Where in the World Is Population Growth Bad? Jeff Kling Lant Pritchett

December1994

December1994

December1994

WPS1392 Some EconomicConsequences of Jean-PaulAzam the Transitionfrom Civil War to Peace David Bevan PaulColrier Stefan Dercon Jan Gunning Sanjay Pradhan WPS1393 The Interfaceof Trade, Investment, J. Luis Guasch and CompetitionPolicies: Issues and Sarath Rajapatirana Challengesfor Latin America WPS1394 Macroeconomic Reformand Growth in Africa: Adjustmentin Africa Revisited WPS1395 A Typology of Foreign Exchange Auction Marketsin Sub-Saharan Africa LawrenceBouton Christine Jones Mi;uel Kiguel Janine Aron Ibrahim Eibadawi

December1994

December1994

J. Troncoso 37826

December1994

R. Martin 39026

December1994

R. Luz 39059

WPS1396 ForeignExchange Auction Markets Janine Aron In Sub-SaharanAfrica: Dynamic IbrahimElbadawi Models for Aucton Exchange Rates

December1994

R. Luz 39059

Policy Research Working Paper Series Contact for paper J. Queen 33740

Title WPS1397 Are PrivateCapital Flowsto DevelopingCountriesSustainable?

Author Uri Dadush Ashok Dhareshwar Ron Johannes RaymondS. Haytman David Wheeler ManjulaSingh

Date December1994

WPS1398 The Cost of Air PollutionAbatement

December 1994

E. Schaper 33457

WPS1399 How Importantto India's Poor is the MartinRavallion Urban-RuralCompositionof Growth? Gaurav Datt WPS1400 Technicaland MarketingSupport Systems for Successful Small and Medium-SizeEnterprisesin Four Countries

December1994

P. Cook 33902 D. Evans 38526

Brian Levy with December1994 Albert Berry, Motoshige Itoh, Unsu Kim,Jeffrey Nugent, and ShujiroUrata December1994

WPS1401 Colombia'sSmalland Medium-Size Abert Berry Exportersand Their SupportSystems Jose Escandon WPS1402 Indonesia'sSmall and Medium-Size Albert Berry Exportersand Their Support Systems Brian Levy WPS1403 Small and Medium-SizeEnterprise Support Policiesin Japan WPS1 404 The Republicof Korea's Smalland Medium-SizeEnterprisesand Their Support Systems WPS1405 Growth and Povertyin Rural India Motoshige Itoh Shujiro Urata LinsuKim Jeffrey B. Nugent

D. Evans 38526 D. Evans 38526 D. Evans 38526 D. Evans 38526

December1994

December1994

December1994

MartinRavallion Gaurav Datt Javier Le6n RaimundoSoto RashidFaruqee

January 1995

WDR 31393 R. Luz 31320 F. Willie 82262

WPS1406 StructuralBreaks and Long-Run Trends in CommodityPrices WPS1407 Pakistan's Agriculture Sector Is 3 to 4 PercentAnnual Growth Suslainable? WPS1408 Macroeconomic Managementand Intergovemmental Relations in China WPS1409 RestructuringUganda'sDebt The CommercialDebt Buy-Back Operation WPS1410 MacroeconomicEffects of Termsof-TradeShocks:The Case of OilExporting Countries

January 1995

January 1995

Jun Ma

January 1995

C. Jones 37754

Kapil Kapoor

January 1995

E.Spano 35538

NikolaSpatafora AndrewWaner

January 1995

J. Queen 33740