美元的国际地位 Goldberg-The International Role of the Dollar and Trade Balance Adjustmentw12495

NBER WORKING PAPER SERIES THE INTERNATIONAL ROLE OF THE DOLLAR AND TRADE BALANCE ADJUSTMENT Linda Goldberg Cédric Tille Working Paper 12495 http://www.nber.org/papers/w12495 NATIONAL BUREAU OF ECONOMIC RESEARCH 1050 Massachusetts Avenue Cambridge, MA 02138 August 2006 We thank Joseph Gagnon, Peter Kenen, and participants of the Group of Thirty for valuable comments. The views expressed in this article are those of the authors and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System. The views expressed herein are those of the author(s) and do not necessarily reflect the views of the National Bureau of Economic Research. ©2006 by Linda Goldberg and Cédric Tille. All rights reserved. Short sections of text, not to exceed two paragraphs, may be quoted without explicit permission provided that full credit, including © notice, is given to the source. The International Role of the Dollar and Trade Balance Adjustment Linda Goldberg and Cédric Tille NBER Working Paper No. 12495 August 2006 JEL No. F1, F3, F4 ABSTRACT The pattern of international trade adjustment is affected by the continuing international role of the dollar and related evidence on exchange rate pass-through into prices. This paper argues that a depreciation of the dollar would have asymmetric effects on flows between the United States and its trading partners. With low exchange rate pass-through to U.S. import prices and high exchange rate pass-through to the local prices of countries consuming U.S. exports, the effect of dollar depreciation on real trade flows is dominated by an adjustment in U.S. export quantities, which increase as U.S. goods become cheaper in the rest of the world. Real U.S. imports are affected less because U.S. prices are more insulated from exchange rate movements pass-through is low and dollar invoicing is high. In relation to prices, the effects on the U.S. terms of trade are limited: U.S. exporters earn the same amount of dollars for each unit shipped abroad, and U.S. consumers do not encounter more expensive imports. Movements in dollar exchange rates also affect the international trade transactions of countries invoicing some of their trade in dollars, even when these countries are not transacting directly with the United States. Linda Goldberg Research Department, 3rd Floor Federal Reserve Bank-New York 33 Liberty Street New York, NY 10045 and NBER linda.goldberg@ny.frb.org Cédric Tille International Research Function Federal Reserve Bank of New York 33 Liberty Street New York, NY 10045-1003 cedric.tille@ny.frb.org 1 The International Role of the Dollar and Trade Balance Adjustment Linda Goldberg∗ Federal Reserve Bank of New York and NBER Cédric Tille Federal Reserve Bank of New York May 8, 2006 I. Introduction. A number of things must happen before exchange rate changes lead to trade balance adjustments. First, the exchange rate change must lead to changes in the border prices of goods imported by the destination market. Next, the change in border prices must lead to a change in the price of goods charged to consumers. Finally, the consumer must react to the relative price of import goods, substituting away from imports if they have become more expensive, or increasing demand for imports if they have become relatively cheap. If the exporters adjust their own profit margins to insulate foreign prices from exchange rate fluctuations, nominal trade balance implications arise purely due to revenue effects of exchange rates on existing quantities of goods traded. In the current debate over global imbalances, contributors differ in their views of how much trade balance adjustment is achievable through adjustments in global aggregate demand versus specifically through relative prices, as in Feldstein (2006), who advocates substantial depreciation of the U.S. dollar as a key adjustment instrument. Most of these studies share the key premise that exchange rates play a role in influencing the relative prices of domestic and foreign goods, so that a U.S. dollar depreciation reduces U.S. imports and spurs its exports to foreign markets.1 Whether or not the demand for traded goods is sensitive to exchange rates has long been recognized as a central issue in international transmission and adjustment, and was the theme of the first Occasional Paper published by The Group of Thirty, by Kenen and Pack (1980). ∗ We thank Joseph Gagnon and Peter Kenen for valuable comments. The views expressed in this article are those of the authors and do not necessarily reflect the position of the Federal Reserve Bank of New York or the Federal Reserve System. 1 A recent NBER volume nicely overviews some alternative positions in this debate. See G-7 Current Account Imbalances: Sustainability and Adjustment, edited by Richard Clarida (forthcoming, NBER and University of Chicago Press). 2 Immediately after the Bretton-Woods system ended, one concern was whether real exchange rates were truly flexible in a way that supported real adjustments. At that time, the debate focussed on the point that if both nominal exchange rates and domestic prices were flexible, real exchange rates would not move enough to correct external imbalances. As a consequence, trade adjustment would need to occur solely through changes in the aggregate demand of trading partners and nominal exchange rate movements would be ineffective in addressing real imbalances. Kenen and Pack (1980), however, were early proponents of the view that changes in nominal exchange rates would induce relative price adjustments, and would help economies adjust their international imbalances. Now, more than twenty-five years later, the strength of the exchange rate instrument in trade-balance adjustment is again hotly debated. We argue that the debate needs to pay heed to the special international role of the dollar as a vehicle currency and as a unit of account. The theory and evidence on the use of dollars and euros as invoicing currencies in international trade transactions are consistent with evidence on low pass through of exchange rate movements into the import prices of the United States and the higher pass through into import prices of other countries. A consequence of this distinction in exchange-rate pass through into prices is that the trade balance effects of exchange rate depreciation can be dramatically different for the United States compared with other countries. A U.S. dollar depreciation is expected to lead to little change in U.S. demand for imports while leading to potentially large increases in world demand for U.S. exports, which will have become cheaper in foreign currency terms. The competitive effects of dollar depreciation are strongest for U.S. exporters and foreign producers competing with these goods. While the U.S. trade balance will respond to changes in the value of the dollar, the special international role of the dollar and the related low degree of exchange-rate pass through into U.S. import prices will lead the trade balance response to be primarily through increased quantities of U.S. exports. Section II begins by reviewing the theory of currency selection in invoicing international trade transactions. Our recent contributions to this area build on the rich history of thought on this topic, from the 1960s through the present.2 These contributions advanced a range of explanations for why a currency becomes a vehicle for pricing and international transactions, including transaction costs in foreign exchange markets, inertia and market depth, industry 2 Kenen (1983) overviews the early evidence and arguments in Group of Thirty Occasional Paper no. 13. 3 structure, and macroeconomic volatility. We then provide evidence on the extent to which exporters use dollars and euros as invoicing currencies. The U.S. dollar and the euro are the main currencies used for transactions in global markets. The dollar is dominant because of the large weight of the United States and dollar bloc countries in international transactions, and because of the importance of the dollar in invoicing trades of commodities and other homogeneous goods. These findings are consistent with survey results reported in the recent Conference Board and Group of Thirty project on “Do Exchange Rates Matter?” (2005). That paper observed that “one of the most interesting results of the survey – one with significant policy implications – is the extent to which global businesses manage exchange rate risk by reducing the number of currencies that account for the majority of foreign exchange exposure. Global companies may produce in dozens of countries and may sell in a hundred or more, but two-thirds of the companies indicate that five currencies or less account for all of their foreign exchange exposure.” (page 8). We argue that this dominant role for the dollar in international transactions is likely to persist. In Section III we discuss exchange rate pass through into the prices of goods as they enter a country, and reasons why this pass-through is typically incomplete. It is now well- established that exchange-rate pass through into import prices is incomplete, variable across types of goods, and lower for the United States than for other industrialized countries. The special role of the U.S. dollar as a key vehicle currency in international transactions helps to insulate U.S. prices from exchange rate fluctuations. While this insulation is immediately apparent in the dollar prices of U.S. imports, additional damping of exchange rate effects on prices occurs by the time the imported good reaches consumers. The damped effect on consumption prices of imported goods occurs because local country content arising from the distribution sector dilutes the share of the total product cost that is attributable to import content. Moreover, distributors may actively absorb some exchange rate movements within their own profit margins. In Section IV we examine how exchange rate movements affect the trade balance. When consumption prices of imported goods are responsive to exchange rates expenditure-switching occurs when a home currency depreciation raises the price of foreign goods sold in home markets relative to the prices of local substitutes. A real trade surplus develops as foreign consumers increase their demand for home exports and home consumers reduce quantities 4 imported of the foreign good. This real surplus does not necessarily translate to a nominal surplus in home currency terms, as the higher price of imports (a worsening of the terms of trade) tends to increase the dollar cost of imports. Our twist to this well-known adjustment story hinges on the special role of the dollar as an invoicing currency in international trade, and the related observation that exchange-rate pass through into import prices is lower for the United States than for other countries. This feature generates an asymmetry across countries in trade balance effects of exchange rates. When import prices are not affected much by exchange rate movements, which is the case when exchange-rate pass through into prices is very low, expenditure-switching does not occur in the home market. A domestic currency depreciation does not lead to much of a decline in import quantities. At the same time a depreciation of the dollar raises quantities of U.S. exports through the expenditure effect, as foreign consumers substitute away from other suppliers and toward U.S. goods. The U.S. dollar depreciation reduces the trade deficit of the United States, but mainly by stimulating U.S. exports. As U.S. import prices do not increase substantially and its exports are priced in dollar terms, the United States is shielded from a depreciation of its terms of trade. In addition to affecting trade adjustment between the United States and the rest of the world, the dollar’s prominence in invoicing international trade transactions even affects the trade balances of countries engaging in transactions that do not involve the United States. The pattern of consequences depends on the extent to which different countries use dollar for invoicing and pricing their bilateral transactions. U.S. dollar depreciations can raise the competitiveness of products invoiced in dollars elsewhere, both relative to produces invoicing in other currencies, and relative to their domestic goods. We conclude this paper with a discussion of the policy implications of our analysis. Some of the lessons drawn from recent research were already anticipated twenty-five years ago, in early seminal work on this topic, including the first Occasional Paper written for the Group of Thirty by Kenen and Pack (1980). They had concluded, “a change in the nominal exchange rate does produce a change in the real exchange rate. The latter will be smaller than the former and may be very small for small open economies, but it need not be trivial or transitory. Trade flows are moderately sensitive to a change in the real exchange rate; price elasticities are not negligible. Taken together, these conclusions say that a depreciation of one country’s currency will increase its exports and reduce its imports, though long lags may be involved.” Now, with 5 the advantage of considerably more data and decades of development of theoretical tools, this theme from the early literature is validated. We add the key observations that the special role of the U.S. dollar in invoicing international trade transactions, and evidence on partial and asymmetric exchange-rate pass through into import prices across countries, imply that real trade adjustment may be asymmetric. The U.S. might not adjust demand for imported goods when the dollar depreciates. By contrast, the dollar depreciation increases the competitiveness of U.S. goods in foreign markets, so that real quantities adjusted would be limited to exports of U.S. products and the foreign sales that these exports displace. II. Invoice Currencies in International Trade Transactions The theoretical determinants of invoicing-currency choice include arguments that emphasize the roles of transaction costs, industry structure, and macroeconomic volatility. Here we briefly review the key theoretical contributions before turning to the empirical evidence on invoicing, and the key roles currently played by the dollar and the euro. Determinants of trade invoicing. Early contributions to this literature focused on the role of currencies as mediums of exchange. The importance of using currencies with low transaction costs in foreign exchange markets, still discussed in current analysis, was stressed by Swoboda (1968, 1969). Such low costs could, for instance, reflect a high degree of liquidity in the foreign exchange markets for the currencies in question (Portes and Rey, 1998). Rey (2001) argued that there are “thick market externalities” associated with a currency’s large presence in global international trade, which yields low transaction costs of exchange. In theoretical work by Devereux and Shi (2005), a vehicle currency through which all trading takes place can emerge as a way to minimize trading-post setup costs. Another argument for invoice currency selection centers on industry characteristics, and in particular stresses that single currencies may be selected for use in the pricing and invoicing of homogeneous goods. McKinnon (1979) argued that industries where goods are homogeneous and traded in specialized markets are likely to have transactions invoiced in a single low transaction cost currency. Krugman (1980) pointed to the presence of inertia in the choice of currency used for this pricing and discussed the disincentives against deviating from the industry 6 norms. Once a currency acquires prominence, perhaps because of low transaction costs, it may keep this role even if another currency with similarly low costs emerges. A third line of argument emphasizes that macroeconomic policy and volatility could influence the emergence of vehicle currencies. Firms have an incentive to invoice their trade in the currency of a country where the volatility of shocks, such as those arising from monetary aggregates, is moderate, as this will ceteris paribus minimize the fluctuations in the exchange rates associated with the currency in questions (Giovannini 1988, Wilander 2006). Recent research contributions by Bacchetta and van Wincoop (2005), Devereux, Engel and Storegaard (2004), and Goldberg and Tille (2005) take these alternative arguments further. Bacchetta and van Wincoop (2005) stress the role of structural features of economies, such as price elasticities of demand and the convexity of production costs. They also point to the presence of a herding effect: if several currencies are available for invoicing, the exporter has an incentive to invoice in the currency used by the majority of her competitors, as doing so limits the volatility of output and marginal costs. Devereux, Engel and Storegaard (2004) focus on the role for monetary fluctuations in the invoicing decision, showing that exporters choose to set their prices in the currency of the country where monetary shocks are the least volatile. The research by Goldberg and Tille (2005) expands on the multi-currency analysis of Bacchetta and van Wincoop (2005) and contrasts the role of the herding motive with other motives, such as the desire to hedge macroeconomic volatility. Herding considerations are found to play a sizable role in choosing the single currency to be used for invoicing international transactions, and are expected to be the dominant consideration for industries where goods are more homogenous. Invoicing of transactions for highly differentiated goods, like manufactured products, is likely to be more evenly spread across the various currencies and may be more responsive to macro-economic volatility. If macroeconomic volatility is an input into invoice currency decisions, it would enter because of patterns of correlations between exchange rates and marginal costs, as opposed to the volatility of the exchange rate per se. Our model also confirms the findings in the early literature that a currency is more likely to be used in invoicing if it has low transaction costs. While herding means that exporters will use a common currency, it does not indicate which currency should play that role. Introducing transaction costs complements the analysis, as we can expect exporters to coordinate their choice 7 of invoicing currency on the currency with low transaction costs. As in the Krugman (1980) inertia arguments, this mechanism can feed on itself if the low costs are driven by the liquidity of the foreign exchange market. Such a virtuous cycle mechanism can explain why a particular currency can persist as a vehicle currency for a substantial period. The evidence on currency invoicing in international trade Goldberg and Tille (2005), the follow-up study by Kamps (2005), and the ECB (2005) report provide extensive cross-country evidence on the currency invoicing of international trade transactions.3 These analyses show that both the U.S. dollar and the euro are extensively used in international trade transactions, with the dollar still the dominant currency. Across countries, “Grassman’s Law” still seems to hold, i.e. countries tend more extensively use their own currency on export invoicing than on import invoicing (Grassman 1973). Table 1, using data reported in these studies, shows the use of dollars in invoicing country exports