RUDIGER DORNBUSCH MassachusettsInstitute of Technology PAUL KRUGMAN MassachusettsInstitute of Technology Flexible Exchange Rates in the Short Run CONSIDERABLE FLEXIBILITY in exchange rates has marked the seven- ties. A series of events, startingwith the appreciationof the deutsche mark in 1969, and includingthe realignmentin the SmithsonianAgree- ment in 1971 and a second realignment,have broughtthe world into a periodof controlledflexibility of rates.Flexible rates were the mechanism economistshad long advocatedfor attainingexternal balance,' but the experiencein the last few years has led many to reconsiderit. That reconsiderationis stimulatedby some surprisesin the performance of flexible rates. First among these are the recurrent,large cycles in ex- Note: We are indebted to Pedro Aspe, Roger Hankin, and Jay Helms for valuable assistance. Helpful comments from Karl Brunner,Jerry A. Hausman, and members of the Brookings panel are gratefully acknowledged.Financial support was provided by a grant from the Ford Foundation. 1. The literature on flexible rates goes back to Milton Friedman, "The Case for Flexible Exchange Rates," in his Essays in Positive Economics (University of Chi- cago Press, 1953). Subsequent writing includes Egon Sohmen, Flexible Exchange Rates: Theory and Controversy (University of Chicago Press, 1961); Richard E. Caves, "Flexible Exchange Rates," American Economic Review, vol. 53 (May 1963), pp. 120-29; Harry G. Johnson, "The Case for Flexible Exchange Rates, 1969," Federal Reserve Bank of St. Louis, Review, vol. 51 (June 1969), pp. 12-24; Herbert Giersch, "On the Desirable Degree of Flexibility of Exchange Rates," WeltwirtschaftlichesArchiv, vol. 109, no. 2 (1973), pp. 191-213; Edward Tower and Thomas D. Willett, The Theory of Optimum Currency Areas and Exchange- Rate Flexibility, Special Papers in International Economics 11 (Princeton Univer- sity, International Finance Section, 1976); and Richard N. Cooper, "Monetary Theory and Policy in an Open Economy," Scandinavian Journal of Economics, vol. 78, no. 2 (1976), pp. 146-63. 537 538 Brookings Papers on Economic Activity, 3:1976 changerates. The dollar-markrate, for example,fluctuated more than 10 percent in less than six months, though there was no comparabledis- crepancyin the movementof pricelevels. Both the fluctuationsin rates and their side effectssoon causedgovern- ments to adopt exchange-ratetargets and to intervene in the market. Governmentsrealized that exchange-ratemovements had real effects: they alteredrelative prices and real incomes and they causedinflation or could serve to reduce inflationarypressure. Exchange-rate targets were also an attractivealternative to the disciplineof externalbalance imposed by a flexiblerate, which was far from appropriatein face of the real dis- turbancesof 1973-74. Confrontedwith a transitorydecline in real in- come, policymakersmuch preferredusing reserves and borrowingto a free adjustmentof exchangerates. The worldwiderecession of 1974-76 demonstrated,with the benefitof hindsight,that even under flexible rates there is scope for coordinated stabilizationpolicy. A coordinatedexpansion in economicactivity would have allowed every country to experiencesome export-ledrecovery at stableexchange rates. Against this background,we proposeanother look at flexiblerates and ask how they fit into conventionalmacroeconomic thinking. The paper seeks primarilyto pull together existing knowledge,factual and theo- retical.The focus is entirelyon the shortrun. We are concernedwith two relatedquestions: How are exchangerates determined,and what role do they play in a short-runmacroeconomic context?The answersare developedin part from theoryand in partfrom empirical evidence. The next major section covers theory, from pur- chasingpower parity and the Keynesianmodel. The theoryof purchasing power parity emphasizesthe relation between price levels and the ex- change rate but is not a theory of exchange-ratedetermination and has little to say about the macroeconomicrole of exchangerates. Keynesian theory, by contrast,places the exchangerate in the center of macroeco- nomics. The exchangerate is identifiedwith the relativeprice of goods and thus is a determinantof the allocation of world spendingbetween domestic and foreign goods. Under conditionsof capital mobility there is an importantlink betweeninternational interest rates and the exchange rate. The Keynesianmodel servesas a startingpoint for a realisticmodel but requiresseveral extensions. Hence, we incorporatethe implicationsof Rudiger Dornbusch and Paul Krugman 539 exchange-rateexpectations, both as part of the adjustmentprocess and as an independentsource of macroeconomicdisturbances. The model is furtherextended by a look at the impact of relativeprices on the saving rate. It is arguedthat an exchange-ratemovement changes relative prices, or the terms of trade, and thereforechanges real income. The changein real income,in turn,may altersaving and spendingat an unchangedlevel of output.The firstsection ends with a discussionof capitalmobility and the suggestionthat exchange rates are determinedin asset markets.We conclude that asset-marketviews and balance-of-paymentsviews of exchange-ratedetermination differ little provided proper emphasis is placedon the capitalaccount in the latter. The model that emergesemphasizes the link betweeninterest rates and exchangerates, identifiesthe exchangerate with the terms of trade, and assumesthat movementsin exchangerates will bring about adjustments in the compositionof worlddemand. The second sectionpresents empiri- cal evidence on some of these issues. First, we find that exchangerates indeed have a strongimpact on relativeprices. A depreciationwill raise importprices relative to exportprices, and thus improvea country'scom- petitiveness.For the case of manufactures,a country'sshare in world ex- ports is significantlydetermined by its relative price position, but the adjustmenttakes several years. We also find that, given domesticlabor cost, competitors'prices are an importantdeterminant of export prices, thus contradictingthe simpleKeynesian model. However, this responseis only partial,thus leavingsome terms-of-tradeeffect. In the third section, we find that importprices have an importantim- pact on domesticconsumer prices-an effectthat is not consideredin the simple Keynesianmodel. Also, that section combinesthe importantfea- turesof exchange-ratemovements in a short-runmacroeconomic context. It draws on the Keynesianmodel in emphasizingthe relation between interestrates and exchangerates and the impactof exchangerates on im- port prices. It goes beyond the Keynesianmodel in emphasizingthe ad- verse short-runeffects of expansionarymonetary policies. The exchange- rate depreciationthat is inducedwill only slowly affectthe compositionof demand;it will immediatelyraise import prices and thereby consumer prices. It follows that, given flexiblerates and capitalmobility, countries that seek an expansionin aggregatedemand face an adverseinflation- unemploymenttradeoff. The exchange-rateadjustment speeds up the inflationaryimpact of monetaryexpansion. We concludethat underthese 540 Brookings Papers on Economic Activity, 3:1976 conditions,an individualcountry will wantto use a monetary-fiscalpolicy mix and that, duringa worldwiderecession, there is a compellingcase for coordinatedexpansion. These policy considerationsare taken up in the fourthand finalsection. Theoryof FlexibleExchange Rates This section provides an overviewof the main approachesto flexible exchange rates and lays out the elements of an integratedapproach. Traditionally,there have been three quite differentviews of the role and determinationof exchangerates. A monetaryapproach developed in the wake of the experienceafter WorldWar I holds that domesticmonetary upheaval will be reflectedin external depreciation.A second strandof an'alysis,again originatingin the twenties,emphasizes relations between relativeprice levels and the exchangerate-the famous doctrineof pur- chasingpower parity (PPP). Finally, Keynesianmacroeconomics of the open economy under flexiblerates was developedin the late forties and elaboratedin the sixties and stresses the interactionof output and ex- change-ratedetermination. PRICES AND EXCHANGE RATES Under the skin of any internationaleconomist lies a deep-seatedbelief in some variantof the PPP theoryof the exchangerate.2 According to the law of one price, a commodityshould sell for the same price (freightand duties apart) in variouslocations. An exchangerate that leaves an inter- nationalprice discrepancywill soon lead to arbitrageand therebyto an adjustmentin prices or the exchange rate or both. Given enough time, therefore,the domesticprices of internationallytraded goods will corre- spondto worldprices converted at the going exchangerate. Even if the arbitrageof traded-goodsprices is generallyaccepted, a numberof substantiveissues remain. Does spatialarbitrage have any fur- ther implicationsfor exchangerates, and, specifically,does it imply that 2. The theory and applications of PPP have been extensively reviewed in Law- rence H. Officer, "The Purchasing-Power-ParityTheory of Exchange Rates: A Review Article," International Monetary Fund, StaglPapers, vol. 23 (March 1976), pp. 1-60. RudigerDornbusch and Paul Krugman 541 exchangerates are a functionof price levels?Given domesticand foreign price levels, does the doctrineassert that the exchangerate will attain a particularlevel? If all goods weretraded with no transportcosts or dutiesand with