SWP793 The Effects of OECD Macroeconomic Policies on Non-Oil Developing Countries A Review Rudiger Dornbusch WORLD BANK STAFF WORKING PAPERS Number 793 SERIES ON INTERNATIONAL CAPITAL AND ECONOMIC DEVELOPMENT Number 5 81.5 '57 7 .793 .3 WORLD BANK STAFF WORKING PAPERS Number 793 SERIES ON INTERNATIONAL CAPITAL AND ECONOMIC DEVELOPMENT Number 5 The Effects of OECD Macroeconomic Policies on Non-Oil Developing Countries A Review Rudiger Dombusch INTERNATIONAL MONETARY FUND /UGJ I VJ IJ.Jj RLCONS'IT'. The World Bank Washington, D.C., U.S.A. Copyright (© 1986 The Intemational Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing January 1986 This is a working document published informally by the World Bank. To present the results of research with the least possible delay, the typescript has not been prepared in accordance with the procedures appropriate to formnal printed texts, and the World Bank accepts no responsibility for errors. The publication is supplied at a token charge to defray part of the cost of manufacture and distribution. The World Bank does not accept responsibility for the views expressed herein, which are those of the authors and should not be attributed to the World Bank or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank; they do not necessarily represent official policy of the Bank. The designations employed, the presentation of material, and any maps used in this document are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates conceming the legal status of any country, territory, city, area, or of its authorities, or conceming the delimitation of its boundaries, or national affiliation. The most recent World Bank publications are described in the annual spring and fall lists; the continuing research program is described in the annual Abstracts of Current Studies. The latest edition of each is available free of charge from the Publications Sales Unit, Department T, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from the European Office of the Bank, 66 avenue d'lena, 75116 Paris, France. Rudiger Dombusch is professor of economics at the Massachusetts Institute of Technology and a consultant to the World Bank. Library of Congress Cataloging-in-Publication Data Dornbusch, Rudiger. The effects of OECD macroeconomic policies on non-oil developing countries. (World Bank staff working papers ; no. 793. Series on international capital and economic development ; no. 5) Includes bibliographical references. 1. Developing countries--Economic conditions. 2. Developing countries--Commerce. 3. Debts, External-- Developing countries. 4. International economic relations. I. Title. II. Series: World Bank staff working papers ; no. 793. III. Series: World Bank staff working papers. Series on international capital and economic development ; no. 5. HC59.7.D669 1986 330.9172'4 85-31747 ISBN 0-8213-0704-5 FOREWORD This paper is one in a special series of World Bank Staff Working Papers on international capital and economic development. Prepared as background papers for World Development Report 1985, the series provides more detailed treatment and documentation of the issues dealt with in the Report. The papers cover a range of topics including a historical perspective on international capital and economic development; the effects of policies in industrial and developing countries on international capital flows, external debt, and economic development; and the role of official assistance, commercial bank lending, securities markets, and private direct investment in developing countries. Several studies of individual developing countries are also included in the series. The background papers draw on a large number of published and unpublished studies of individual researchers, on World Bank policy analysis and research, and on reports of other organ- izations working on these issues. The papers are the work of individuals and the views and interpretations expressed in them do not necessarily coincide with the views and interpretations of the Report itself. I hope these detailed studies will supplement World Development Report 1985 in furthering understanding of the relationship between international capital and economic develop- ment. A complete list of the papers appears on the overleaf. Francis X. Colaco Staff Director World Development Report 1985 Papers in the Series Aliber, Robert. Financial Intermediation and the External Debt Crisis Batista, Paulo Noguiera, Jr. International Financial Flows to Brazil Since the Late 19609: An Analysis of Debt Expansion and Current Payments Problems Blanchard, Olivier, and Lawrence H. Summers. Perspectives on High Real Interest Rates Throughout the World Codippily, Hilarian M.A. Ethiopia: International Financial Flows, 1967-83 Dornbusch, Rudiger. The Effects of OECD Macroeconomic Policies on Non-oil Developing Countries: A Review Fishlow, Albert. Capital Markets During the 19th Century and the Interwar Period: Lessons from the Past Hooper, Peter. The International Repercussions of the U.S. Budget Deficit Ibrahim, Tigani E. Kenya: Use of External Resources, 1963-1983 Ishiyama, Yoshihide, and Keiko Atsumi. Capital Outflows from Japan to Developing Countries Iqbal, Farrukh. Korea's Debt Accumulation, Use, and Management Strategies Johnson, John H. The Role of International Finance in Argentine Development, 1965-84 Lessard, Donald. International Financing for Developing Countries: The Unfulfilled Promise Llewellyn, David. International Financial Intermediation and the Role of Banks in Balance of Payments Financing Martone, Celso L. Macroeconomic Policies, Debt Accumulation, and Adjustment in Brazil, 1965-84 Muller, Patrice, and Robert Price. Public Sector Indebtedness and Long-Term Interest Rates O'Brien, Richard, and John Calverley. Private Banks and Developing Countries Pant, Chandrashekar. Yugoslavia: Economic Development and External Capital, 1970-1984 Rybczynski, T. M. Internationalization of Financial Arrangements and the Developing Countries: The Evolving Relationship Sachs, Jeffrey, and Warwick McKibbin. Macroeconomic Policies in OECD Countries and External Adjustment in Developing Countries Saini, Krishan G. Capital Market Innovations and Financial Flows to Developing Countries Sherbiny, Naiem A. Arab Financial Institutions and Developing Countries Swoboda, Alexander, and Hans Genberg. The Stages in the Balance of Payments Hypothesis Revisited Weigel, Dale, and Robert R. Miller. Foreign Direct Investment in Economic Development Wyplosz, Charles. International Aspects of the Policy Mix in Six OECD Countries Zietz, Joachim, and Alberto Valdes. The Costs of Protectionism to Developing Countries: An Analysis for Selected Agricultural Products - iv - Abstract The effects of OECD macroeconomic policies on non-oil developing countries are analyzed by examining the well-known theoretical channels of interdependence and some of the available empirical evidence. The current concern with international linkages --which is attributed to the debt crisis of the late 1970s--has a short-term orientation. The interesting question today, it is argued, is whether external growth, through the different cyclical channels by which it affects developing countries, makes it possible to service debts and restore creditworthiness. Another pertinent question is whether external developments are promising enough to make it possible for debtor developing countries to resume growth. Three variables are essential in assessing the impact of OECD policies: aggregate growth rate, dollar interest rates, and the value of the dollar. All debtor non-oil developing countries have an interest in some combination of a strong OECD growth, low real interest rates, and weak dollar. Evidence on external aggregates of developing countries (commodity prices, the terms of trade, export volume, and interest rates) indicates that growth in the developed countries favorably affects the export earnings of developing countries, although the precise quantitative link is not known. Another conclusion drawn from the evidence is that a depreciation of the dollar affects the real prices of commodities. Finally, a rise in real short-term interest rates raises debt service burdens directly and via increased spreads between LIBOR and T-bills; the cost of debt service (as a fraction of exports) increases with the size of the ratio of bank debt to exports. The paper concludes with a review of alternative policy-mix scenarios, highlighting the role of interest rate-growth-dollar combinations for developing countries. Table of Contents Page 1. The Linkages ........... ......... .......... 2 2. Commodity Prices ll...... . .... . .. . . 11 3. The Terms of Trade and Real Export Price. .. 15 4. Export Volume .....19 5. External Debt and Interest Rates . .... .... 24 Spreads ..................... ...27 The Real Interest Rate ........................ .... .30 6. The Implications of Alternative Mixes .....................32 - vii - The linkage between the growth performance of developing countries and economic prosperity of the more developed countries has long been recognized. Discussion of this linkage has traditionally centered on the view that trade is an engine of growth. That is to say, it is widely believed that trade has the potential to transmit some growth from rich to poor countries over the long term but the adequacy of the transmission and its persistence have at times been questioned. Lewis (1980) has documented the transmission of growth but has also argued that, because of its prospective slowing down, auxiliary engines of growth are required if developing countries are to continue growing at satisfactory rates. 1/ The linkages between developed and developing countries are also relevant in the short-term cyclical context of the debt problem. Many argue that the 1980-83 macroeconomic policies of the Organisation for Economic Co- operation and Development (OECD) played a significant, if not dominant, role in the debt crisis. At the same time, policymakers and international organizations believe the empirical evidence suggests unambiguously that sustained OECD growth of 3 percent a year, combined with no further increases in LIBOR and some fall in the dollar, would make it possible for debtor developing countries to improve their external position significantly even as they enjoy some recovery from their present state of depression. 2/ 1/ See W.A. Lewis, "The Slowing Down of the Engine of Growth," American Economic Review (December 1980); I. Kravis, "Trade as the Handmaiden of Growth," Economic Journal (December 1970); R. Nurkse, Patterns of Trade and Development (Almquist and Wiksell, 1959); L. Reynolds, "The Spread of Economic Growth to the Third World," Journal of Economic Literature (September 1983); and J. Riedel, "Trade as the Engine of Growth in Developing Countries, Revisited," Economic Journal (March 1984). 2/ See, for example, International Monetary Fund, World Economic Outlook TWashington, D.C., April 1984); or W. Cline, International Debt (Cambridge, Mass.: MIT Press, 1984). - 2 - This paper reviews the effects of OECD macroeconomic policies on non- oil developing countries by examining the theoretical channels of interdependence along with some of the available empirical evidence. It should be noted at the outset that these channels are well known and that the stylized facts are thought to be firmly established, even if the empirical evidence does not altogether warrant that confidence. The discussion closes with an evaluation of alternative OECD policy mixes. 1. The Linkages The broad relation between the economic performance of developing countries and that of developed countries can be summarized as follows: sustained, high economic growth at the center promotes growth at the periphery. The stability of that link over long periods of time--whether decades or a century--has occupied Lewis (1980) and others, but need not concern us here; we simply record the fact that such a link exists. More to the point here is the relation between the growth of real GDP in the non-oil developing countries and the growth rate in developed countries (Table 1). The two equations differ in that they use three alternative measures of growth in developed countries as explanatory variables--namely, the growth rates of industrial production, real GDP, and import volume. In each case, the explanatory variable is entered as a distributed lag. These regressions confirm that sustained growth in developed countries spreads to developing countries and raises growth there. The precise magnitude and the timing are quite open and so is the question of which variable--GDP growth, growth in industrial production, or growh in import volume--is the best indicator of the -3- spread of prosperity. Import growth is certainly too limited a measure because it leaves out entirely the role of international capital flows to developing countries that are a byproduct of the spreading prosperity. As growth spreads to developing countries, their external balance improves or their creditworthiness is enhanced. As a result, they gain access to more capital, which in turn makes it possible to finance growth-enhancing policies of investment even if these projects have large import content. Table 1 Trade as the Engine of Growth, 1960-83 Growth in Constant Developed Countries Rho R DW 1. 1.45 0.69 0.70 0.67 2.10 (1.44) (3.70) 2. -0.43 1.31 0.64 0.63 2.08 (-0.28) (3.50) 3. 2.47 0.34 0.48 0.58 1.85 (3.14) (3.32) Note: Let gD and gI be the growth rates in developing and industrial countries respectively. The regression is: gD = a+bgI. Ultimately, we are concerned with growth and prosperity in developing countries. The approach that considers trade an engine of growth addresses that issue in a broad manner. A more narrow perspective would focus on the specific short-term cyclical linkages between OECD countries and the external balance of developing countries rather than the growth of their GDP. -4- The principal linkages between macroeconomic developments in the industrial countries and the external balance of developing countries exist in three areas: export earnings from goods and services; import spending for goods and noninterest services; and debt service. These aggregates in turn are determined by: trade prices for commodities, manufactures, and services; trade volume; and interest rates. These three key variables are directly affected in a variety of ways by the level of OECD economic activity, inflation, and the monetary-fiscal mix within and between countries; and by interest rates, the dollar, and the composition of world demand. Added to these straightforward macroeconomic linkages are further channels. First, trade restrictions and financial regulations may interact with macroeconomic developments, or be triggered by them, and assume particular importance. Second, fiscal difficulties of OECD countries, perceived or real, will influence their willingness to make aid and/or credit available to developing countries. Third, to the extent that oil and other material prices are endogenous and are influenced by these OECD variables, there are important indirect effects via import spending of oil producers and, as the case may be, oil export revenues or oil import bills. Finally, important third-country effects may be set off by cyclical or interest-rate developments originating in the OECD. Thus a recovery in the OECD will directly increase demand for an individual country's exports to the OECD, but will also generate increased trade from within the developing country as foreign exchange tightness is reduced. The question of the macroeconomic dependence of the periphery on the policy successes and failures of the center is, of course, an old one. History here includes the terrible experience of the 1930s that turned many developing countries toward inward-looking industrialization, especially in -5- Latin America. But history also includes the long expansion of world trade in the post-World War II period and in the 1970s, which started with a massive commodity boom and deeply negative real rates and ended in a debt crisis unprecedented since the 1930s (see Table 2). Table 2 Key Macroeconomic Variables of the World Economy (average annual percentage rates) Inflation Rate Growth in Indus- Prime Rate in World Trade trial Countries 1929-33 4.8 -17.0 -8.2 1970-73 6.7 12.4 5.9 1979-82 15.5 4.4 -1.7 Source: U.S. historical statistics, International Monetary Fund, League of Nations. To study the various channels through which the macroeconomic policies of countries at the center affect developing countries, we need a conceptual framework that poses the questions that need to be answered. The standard question in this context is presumably how cyclical developments in the industrial countries affect the terms and volume of trade of developing countries, and, more generally, how the gains from trade are shared between rich and poor countries. These themes were developed particularly in the - 6 - context of the Prebisch-Singer challenge to classical trade theory. 3/ The terms-of-trade effects pertain in principle to the commodity terms of trade, the relative price of imports with respect to exports, and also the intertemporal terms of trade relevant to debtor developing countries, that is, the real rate of interest. Until recently, the intertemporal terms of trade were not central to the discussion because no major shock occurred until the late 1970s, and for a long time before that date there was no significant external debt, much of it having disappeared in the late 1930s or during World War II. The classical approach to the linkgaes is to examine the terms of trade, and to assume away mercantilist or Keynesian elements that pay attention to export volumes or abilities to meet debt service or to finance desirable levels of imports. The concern with international linkages today has an immediate short-term orientation. The external credit rationing and the need to service debts, as has been the case after other debt crises in history, forces developing countries to live from hand to mouth and to take a very short-term view which focuses on cyclical spillover effects from the developed to the developing countries, rather than on trends in growth linkages. The interesting question now is whether external growth, through the different cyclical channels by which it affects developing countries, makes it possible to service debts and restore creditworthiness in the medium term. Beyond creditworthiness, of course, lies the further question of whether 3/ See H. Singer, "The Terms of Trade Controversy and the Evolution of Soft Financing: Early Years in the U.N.," in Pioneers of Development, ed. C. Meier and D. Seers (Oxford University Press, 1984); and Report by the Secretary General of UNCTAD, Toward a New Trade Policy for Development (New York: United nations, 1984). See also J. Spraos, Inequalising Trade (Clarendon Press, 1983). -7- external developments are promising enough to make it possible for debtor developing countries to resume growth, which may now be limited by the lack of imports. With externaL debt and external bottlenecks to growth as the frame of reference, the current account, the noninterest current account, import volume, or growth rates of debt/export ratios become the center of attention. If, for example, it was felt that expansion in the developing countries was limited by import availability, it would be natural to focus on the volume of imports that are implied by export revenue, debt service, and "new money" or increased bank lending. 4/ (1) M - (P /P )X - (P/P )Kd; K E (i - v)/(l + p) x m m where Px, Pm = dollar prices of exports and imports, P, p = price level and inflation in industrial countries, X, M = import and export volume, d = real (end of previous period) debt outstanding, i = effective interest, and v = growth rate of exposure accepted by lenders. In equation (1), the key variables are the relative price of exports, PX/Pm, the price of imports relative to OECD country prices, Pm/P, OECD nominal interest rates and inflation, the export volume of developing countries, and the growth in bank exposure. Questions that would arise within this framework have to do with the impact of OECD cyclical movements on the terms of trade 4/ The equation is derived from PmM = P xX - iD + new money, where new money = vD and D is the nominal debt of the previous period in dollars. By dividing by Pm and noting that D/Pm=(D/P_ d(l+p)(P/PM)=d/(l + p), we reach (1). - 8 - ands;export volume and also with the effective real debt service cost, i-p, in developing countries. Questions about exchange rates would also arise because most of the debt is denominated in dollars and thus it is important to know how movements in the dollar affect the level of prices in industrial countries, and hence the real debt. Equation (1) determines the growth rate of import volume residually. An alternative approach is to focus on the debt-export ratio of a country. The pertinent question here is whether, with a balanced noninterest current account, the growth rate of debt determined by the effective nominal interest rate exceeds the growth rate of export revenue. If so, the debt/export ratio is declining over time, and this decline is taken to signify that a debtor country is improving its external position. The percentage change in the debt/export ratio is: 5/ (2) Percentage change in (D/PXX) = (i - Px x where px and x are the growth rate of price and volume of exports. Accordingly, a debtor position is improved if export revenue growth exceeds the interest rate. Alternatively, improvement occurs if export volume grows faster than the real effective interest rate measured in dollars. OECD macroeconomic conditions would be relevant as determinants of export price behavior, growth in export volume, and interest rates. The relevance of particular sums, differences, or ratios of variables that are affected by OECD cyclical developments depend on the question being 5/ The equation is derived by differentiating the debt/export ratio, noting that the noninterest current account is assumed to be balanced so that debt grows by an amount equal to debt service, iD. The formulation assumes that there are no direct investment flows. considered. For example, finding that an extra percentage point -OECD' grogth generates in export revenue significantly more than it costs in extra-debt service because of higher interest rates assures us that a country is able to service its debt. But it does not show that a debtor country is better off as a result. The debtor country will be better off only if improved terms of trade, not export volume, more than paid for the extra interest. A common misunderstanding about the welfare economics of interest rate changes and export earnings--we refer to it here as the Bergsten-Cline- Reagan proposition--was summarized by President Ronald Reagan in his recent address to the IMF: We sometimes hear complaints about U.S. interest rates, particularly by debtor countries, which are legitimately concerned about the additional debt service costs they must bear. But not enough mention ismade of trade and the far greater benefits developing countries receive from renewed economic growth and open market policies of the United States. For the U.S. alone, imports from the non-OPEC LDCs during the first seven months of this year increased by more than $12 billion over the amount during the same period last year. By comparison, a 1 percent increase in interest rates would increase net interest payments by the non-OPEC LDCs by only $2.5 billion. 6/ Welfare economics cannot be judged by netting export revenues and extra interest payments, however, since a great deal depends on whether the boom is in the terms of trade or volume. If the standard welfare economics of international trade and lending are used, the welfare effects of changes in world interest rates, prices, and trade volume can be summarized as follows. 6/ Reported in the New York Times, September 26, 1984, p. D5. - 10 - A country benefits: -- from an improvement in the terms of trade, the improvement being proportional to the level of imports; -- from a reduction in world real interest rates, the improvement being proportional to the stock of debt; -- from an increase in the net debtor position whenever, in the presence of credit rationing, the current value of resources in consumption exceeds the cost (in terms of future utility foregone) of an extra dollar of debt; -- from an increase in export volume only when exports have no alternative costs (that is possible only when they are generated from unused resources rather than by contraction of consumption or contraction of the production of other goods). Thus, welfare will be improved if the gains in the terms of trade more than pay for the increased interest bill, not if export volume rises. Unfortunately, the boom of 1983-84 was primarily one of volume (hard work) and had very little to do with terms-of-trade gains. That means debt service burdens were in fact aggravated, even though, in financial terms, cyclical developments made it possible to meet the increased debt service by extra export volume or import compression. The question of financial feasibility should therefore be separated from the welfare economics and politics perspective of terms of trade and real interest rates. Indeed, short-term financial feasibility may still imply political costs in the debtor countries that will ultimately deteriorate the quality of loans. We now turn to the evidence on various external aggregates of developing countries: commodity prices, the terms of trade, export volume, - 11 - and interest rates. The evidence is subsequently used to discuss the implications of alternative macroeconomic scenarios. 2. Commodity Prices The real price of commodities (see Figure 1) is measured by the Economist's index of commodity prices deflated by the unit value of industrial countries' exports. The latter is taken as a broad index of manufactures in world trade. (Table A-1 shows annual data for various series of prices.) Figure 1 shows the peaks (P) and troughs (T) of the international business cycle for seven large industrialized countries. 8/ One of the striking features of Figure 1 is that real commodity prices have not deteriorated in the longer perspective. The real price in 1984 is at the same level as it was in the early 1960s, or toward the end of the 1960s. A sharp decline or negative trend is evident only if 1971-74 is taken as the starting point. The second point is that movements in real commodity prices do correspond to the business cycle dating broadly, but the correspondence is not complete in the timing. Moreover, in some instances, as was the case in 1978, a dip in the real commodity price may have no counterpart in the business cycle. In fact, 1978 was a year of strong expansion in world industrial production. Table 3 shows the trough-to-peak cumulative change in world industrial production and real commodity prices. Real commodity prices are measured alternatively by the Economist and the IMF index of commodity prices, in each case deflated by industrial countries' export unit values. 8/ The dating is reported in G. Moore and V. Zarnowitz, "The Development and Role of the National Bureau's Business Cycle Chronologies," NBER Working Paper no. 1394 (July 1984). - 12 - co~~~~~- 0~~~~ E-4 C., X 0~~~~~~ cv,~ ~ ~ ~ ~~~~~~~c i S C~~~~~~~~~~~~~~~C co 0~~~ w 0 03 00~~~~~~~~~~~~~~~~~~~~ be 3 0%~~~~~~~~~~~~~~~~~~~~~ - 13 - Table 3 Cyclical Behavior of Real Commodity Prices (cumulative percentage change, trough to peak) 1 Cycle Industrial Real Commodity Prices Trough Peak Production Economist IMF 58:1 to 60:1 21.5 -1.9 -1.0 61:1 to 62:1 10.4 -2.7 -1.6 63:1 to 66:1 26.3 18.0 10.8 67:4 to 69:3 14.3 -3.4 4.7 71:3 to 73:4 19.5 74.0 40.9 75:2 to 80:1 30.6 10.1 15.9 Equation (3) summarizes evidence on the determinants of the rate of change of real commodity prices. 9/ The regression uses quarterly data for the period 1960:1 to 1984:3. The commodity price is represented by the IMF index of all commodities. The equation explains the movements in real commodity prices by three variables: depreciation of the dollar in terms of the currencies of the major industrial countries--Japan, Germany, France, the United Kingdom, Italy, and Canada (depreciation); growth in the industrial production of these countries plus that of the United States relative to the 9/ For empirical work on commodity prices, see B. Bosworth and R. Lawrence, Commodity Prices and the New Inflation (Washington, D.C.: Brookings Institution, 1982); R. Cooper and R. Lawrence, "The 1972-75 Commodity Boom," Brookings Papers on Economic Activity 3 (1975); C. K. Chu and T. Morrison, "The 1981-82 Recession and Non-Oil Primary Commodity Papers" IMF Staff Papers (Washington, D.C., March 1984); and A. Feltenstein, M. Goldstein, and S. Schadler, "A Multilateral Exchange Rate Model for Primary Producing Countries," IMF Staff Papers (Washington, D.C., September 1979). - 14 - growth average of the period (growth); and the lagged level of the real price relative to the period average level (level). 10/ (3) Change in Real = 96.6 + 0.97 Growth - 0.96 Level + 0.68 Depreciation Commodity Prices (4.43) (3.21) (4.50) (2.06) R = 0.37 DW = 1.88 The equation shows the significant but small role of world growth. An extra percentage point growth above trend leads to an increase of 1 percentage point in the growth rate of real commodity prices. Since growth refers to industrial production, which is much more cyclically volatile than CDP, the estimate is a plausible representation of a strong cyclical effect. The interesting point about equation (3) is the role of the dollar exchange rate. Dollar depreciation relative to a basket of major currencies leads to a rise in the real price of commodities. This fact, which has been already noted in Dornbusch (1983), comes out clearly in the present analysis. 11/ Moreover, this result does not change when the 1980-83 period is excluded from the sample. The effect is considerable: a 10 percentage point depreciation of the dollar raises the real commodity price by nearly 7 percent. The short-term role of such a large effect is perhaps even more 10/ The level and depreciation variables were entered as distributed lags; the growth variable worked best as current growth, however. The level variable is defined as Level = RPM*100/AVG, where RPM is the real price of materials and AVG is the average value of the RPM index in the period 1960 to 1983. Thus, when RPM = AVG, the lagged level effect offsets the intercept term. We note that results similar to those in (3) are obtained when the real commodity price is measured by the Economist's index of commodity prices deflated by industrial countries' export unit values. (3A) Change in Real = 44.34 + 0.88 Growth - .44 Level + 0.67 Depreciation Commodity Price (2.84) (2.83) (2.82) (2.00) R = 0.31 DW = 1.93 11/ See R. Dornbusch, "Flexible Exchange Rates and Interdependence," IMF Staff Papers (Washington, D.C, March 1983). - 15 - important than the cyclical variable that is well established. A possible explanation for this effect is that commodities are priced, in Okun's terms, in auction markets, whereas manufactures are priced in customer markets. Exchange rate movements appear to exert a relatively persistent effect on the pricing of manufactures and accordingly affect the real prices of commodities in world trade for some time--up to two years. This effect is, of course, interesting in the context of policy-mix questions that imply alternative courses for the dollar exchange rate. 3. The Terms of Trade and Real Export Price In the previous section, developing countries were assumed to be exporters of primary commodities. Of course, that is no longer the case and several of the large exporters--the newly industrialized countries (NICs)--now export predominatly manufactures. The determinants of the terms of trade of non-oil developing countries should therefore be examined more closely. (Figure 2 shows these terms of trade, in import relative to export prices.) Equation (4) presents the determinants of the terms of trade of non- oil developing countries. Using quarterly data for the period 1962-83:3, we regressed the level of the log of the terms-of-trade index (TOT) on the log of the real materials price (RMP), the log of the real price of oil in terms of industrial countries' export unit values (ROP), and the residual (RES) of a semi-log regression of industrial country output on a simple time trend. (4) TOT = 5.36 + 0.07 ROP - 0.23 RMP - 0.59 RES (21.42) (7.07) (-4.38) (-3.10) Rho = 0.59, R2 .93, DW = 1.97 - 16 - Z : V96T ^ C: E0961 Z:E961 In z C961 v.4 Z:196T 1:Z161 It: I G61 _: 1961 1:1961 1 :0861 :0861 Z:0o6e 1:0961 t:616T E:6Z61 Z:616T J :8161 > I @t~~~~~:961 t:BL6T _:ZZ61 1:116T I: 961 E:9161 E:9Z61 1:9161 4 61E:C61 L ~~~~~Z:SL6T vP4 'E1:E161 >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Z /t Zt61 cc I Z tZ661 4^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~c E:CL6T O Z: 0 61 C.) >s f0 Z 66T CZ Z:Z961 O ,S' E~~~~~~~~~~~~~~~~~~~~~~~~:9616 ri : EL Z6 T QJ 9: 1161 0 --'-,,_ : 996T bu E:99~~~~~~~~~~~~~~161 r. ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ _,, V' T:96T cs E :59~~~~~~~~~~~~~~~~~~~161 e~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~9 11 t61 o~~~~~~~~ 6:1161 l E~~~~~:E961 0 1:1161 44 _:0161 lant f :6961 6:6961 E-4 0 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Z:6961 p Go T:~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~696 1 4.4 Z : 961 o E:996T I:Z961 1:9961 C :961. E4 1~~~~~~~~~~~~~~~~~~~~~~~~~:196T 1: 1961 A 0~~~~~~~~~~~~~~~~~~~~~~~~~: 996T E-4 E~~~~~~~~~~~~~~~~~~~~~~~~:996T V:9961 T:9961 cc ly~~~~~~~~~~~~~~~~~~~~~~~~~: 6961 C14 VS~~~~~~~~~~~~~~~~~~~~~~~~:961 1:696T T:S961 9: 196T 6:9961 Z: 91961 T9:V961 6C961 1:E96T 1:6961 9:1961 6 1961 E: 1961 1: Z961T 017 - (Y ti: ES61 Z:t961 1: Z961 2 :2361 T2236T 9: Z161 T2 Z961 t t~~~~~~~~~I: T961 ~~~~~ ~~~~Z: T961 1: 196T 2 :0,61 E:0961 ,:601. £2-0616 V:GL61 1:616! 1 *~~~~~~~:9161 z ~~~~~~~E: St6T t:9161 11 C4 VLL~~~~~~~961 41 C st S6T U~~~~~~~~~~~L O- I:St~~~~~~~~~~~~~~£61 a N < T ~~~~~~~~~~~~~~~~~~~~~l96T t1 _> E ' tZ~~~~~~~~~~~~~~~C96TT O 0 Z:9i61 V r Z:CL6T a T:fL61 > r- Z:VL61 '-4 ECG O 0 TE:016T X co E: 996T 44i Z: T S6T O 1 : Z96T n~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~V cTL Z6T @ X <_ ~~~~~~~~~~~~~~~~~~~~~E:99L61 W X )~~~~~~~~~~~~~~~~~~~~~~Z TL96T (rL61 0 Z% :S9 61 0~~~~~~~~~~~~~~~ T:596T 9:6961 E:696T K ~~~~~~~Z:V96T \~~~~~~~~~~~ T:696T (a V~~~~~~~~~~~~~~~~~~~:096T 2E 961 5 T~~~~~: E96T 4)~~~~~~~~~~~~~~~ T :Z961 f I 0 E: Z96T 00~~~~~~~~~~~~~~~ Z:Z96T .0 (N~~~~~~~~~~~~~ T:Z96T C4.v4 V996 oc~~~~~~~~~~~~~~~ t: T96T ) E: 1961 ~~~~~~~Z: 196T £: 1196! £ 0961 2 0961 E9 696! X 2 6~~~~~~~~~~~~~~~~~~:961 6:961 16: 961 '-4~ ~ ~ ~ ~ ~~~~~~~~~~~~~~~~~~:16 Z£ 7961 - 18 - The equation shows that, for non-oil developing countries, a rise in the real price of oil causes the terms of trade to deteriorate (the elasticity being 0.07). By contrast, a rise in the real prices of materials significantly improves the terms of trade (the elasticity being 0.23). This reflects the fact that even when developing countries export manufactures, these exports have an important materials content--as in the case of frozen orange juice, steel sheets, and the like. Cyclical conditions in developed countries, as was pointed out earlier, affect real commodity prices and thus are already captured in that term. But there is an additional influence in the residual from trend growth in industrial countries. A 1 percentage point output above trend improves the terms of trade by nearly one-third of a percentage point. Poor growth performance in the developed countries influences the terms of trade for developing countries. We turn next to the determinants of the real export prices of non-oil developing countries. Their real export prices are measured by the export unit value deflated by the export unit values of industrial countries. Equation (5) explains the growth rate of real export prices in terms of growth in industrial countries' production (growth), the change in the real price of oil (DROP), and the lagged level of real export prices relative to the average level during the sample period (level). (5) Growth of Real = 42.0 + 0.53 (Growth) + 0.06 DROP - 45.0 Level Export Price (2.17) (3.12) (3.54) (-2.38) = 0.61 DW = 1.98 The equation shows a precisely estimated, small effect of OECD growth and a significant reversion-to-mean effect. The impact of real oil price changes is significant presumably because in some instances (for example, Mexico) these countries export oil even if they are primarily exprters of manufactured goods. - 19 - Cline (1984) has found that the lagged change in growth significantly affects real export prices. We have not found that to be the case for one- period lagged growth. 12/ However, two-period lagged growth is highly significant, for whatever reason, and does have a negative sign. This explains why the lagged increase in growth appears significant, although, of course, it does not justify introducing an acceleration term. 4. Export Volume The evidence on the link between the export volume of developing countries and economic activity in industrial countries is not definitely settled. The fact that such a link exists is not in question, but the precise estimates are. It is useful to start with a comparison of average growth rates. As Table 4 shows, the growth-import relation in industrial countries differs between subperiods. It also differs between developing country exporters. Furthermore, significant differences appear in the trend growth performance of different exporting areas. Thus factors other than industrial countries' growth may be important determinants of the export growth of developing countries. Table 4 Average Growth Rates Non-Oil Developing Countries' Industrial Country Growth Export Volume Growth Industrial Import Western GDP Production Volume All Asia Hemisphere 1960-83 3.7 4.3 6.4 6.1 8.6 a/ 5.5 1970-83 2.8 2.6 4.2 7.4 10.4 6.1 a. 1964-83. 12/ See Cline, International Debt, p. 247. - 20 - The Bergsten-Cline regression is frequently used for back-of-the- envelope predictions of the export growth of developing countries. 13/ The equation explains the import volume growth of industrial countries by the growth rate of their GDP. The regression shown in equation (4) uses annual data for the period 1960-83. (4) OECD Import Growth = -3.21 + 2.55 GDP Growth (-2.52) (8.39) R = 0.75 DW = 2.19 The estimate here is slightly lower than the original Bergsten-Cline 3:1 link, but the regression clearly confirms their finding. Of course, the data refer to total import growth from all sources, not to imports from non-oil developing countries. This distinction is important since the exports of non- oil developing countries account only for 17 percent of the total imports of industrial countries. Contrary to what (4) might lead us to believe, the evidence from regressions explaining the growth of export volume of different groups of non- oil developing countries does not point to a large cyclical effect. Table 5 shows a regression for the export volume growth of all non-oil developing countries. This regression is based on annual data for the 1960-83 period. The explanatory variables are the growth rate of industrial countries (World Growth) and the change in the relative price or competitiveness of this group's exports (Comp). The relative price is measured as the export unit value deflated by the export unit value of industrial countries. Growth in industrial countries is measured in equations 1 and 2 of Table 5 by the growth of industrial productions and of real GDP, respectively. 13/ See W. Cline, ed., Trade Policy (Cambridge, Mass.: MIT Press, 1983), p. 74; and Cline, International Debt, p. 241. - 21 - Table 5 The Determinants of the Export Growth of Non-Oil Developing Countries World Constant Comp. Growth Rho RI DW 1. 2.31 -0.28 0.75 0.70 0.36 1.99 (1.10) (-2.18) (4.46) 2. 0.94 -0.61 1.32 0.58 0.12 2.13 (0.36) (-0.61) (2.44) Note: Export volume growth = a + b Comp + c World Growth. It is immediately obvious from Table 5 that much better results are obtained when industrial production growth rather than growth in the real GDP of industrial countries is used as the explanatory variable. But even when real GDP growth is used, the estimated elasticity--1.32 in equation 2--is far below the number that the Bergsten-Cline regression proposes. The other interesting point to note concerns equation 1. Here we see that changes in competitiveness among non-oil developing countries does influence their export performance. The loss in export volume is nearly one-to-one with respect to export growth. Regressions for the growth of export volume for non-oil developing countries in the Western Hemisphere and in Asia are shown in equations (5a) and (5b), respectively. (5a) Export Growth -0.89 - 0.49 Comp + 0.88 World Growth + 6.3 Dum i V o } W-Hem (-0.42) (-3.56) (3.25) (2.80) - = 0.39 DW 2.00 - 22 - Here Dum is a dummy that assumes the value 0 for 1960-72 and 1 for 1973-83 (representing the increased trade orientation of Latin America in the 1970s). Once again, we note that the effect of external competitiveness is highly significant and the impact of world growth quantitatively small. The same effect holds for Asia, where the regression refers to the period 1964-83. 14/ (5b) Export Growth = 2.12 - 0.69 Comp + 1.25 World Growth Asia (0.54) (-3.88) (4.41) Rho = 0.74 R2 = 0.33 DW = 1.93 The absence of any large coefficient for world growth is worrying. It is, of course, possible that the country aggregation conceals such an effect. Since these results on cyclical effects differ substantially from the Bergsten-Cline estimates, we should briefly look at some other evidence. The most comprehensive work appears to have been done by the International Monetary Fund. The 1982 World Economic Outlook, for example, reports estimates of the elasticity of various aggregates with respect to industrial country growth (see Table 6). In the 1984 World Economic Outlook, the elasticity of export revenue with respect to industrial country growth is reported as 2; this estimate is in line with the view that volume elasticity is more nearly in the neighborhood of unity. Further evidence is reviewed by Goldstein and Khan (1982), who distinguish between manufactures and other commodities and between subgroups of exporting countries. 15/ Their estimates of the elasticity in 14/ The recent IMF Yearbook does not provide earlier data, and the data on the IMF tape at present do not appear to make sense. 15/ See M. Goldstein and M. Khan, "Effects of Slowdown in Industrial Countries on Growth in Non-Oil Developing Countries," IMF Occasional Paper no. 12 (Washington, D.C., August 1982). - 23 - Table 6 Non-Oil Developing Countries: Elasticities of Real Exports, Terms of Trade, and Purchasing Power of Exports with Respect to Real GNP in Industrial Countries Purchasing Power of Exports Estimates Based on Real Exports Terms of Trade Survey Estimates for Econometric Survey Econometric Real Exports and Estimates from Estimates Estimates from Econometric Estimates Data for 1963-79 for 1984-86 Data for 1963-79 for Terms of Trade Net oil exporters 0.9 1.1 1.4 2.5 Net oil importers Major exporters of manufactures 1.2 1.9 1.1 3.0 aow-income countries (excluding India and People's Rep. of China) 2.6 1.9 2.5 4.4 Other net oil importers 2.0 1.4 3.0 4.4 Source: International Monetary Fund, World Economic Outlook, 1982 (Washington, D.C.), table 45. - 24 - the 1973-80 period are as high as 2.3 in some cases. They do not give the particular equation used, however. High elasticities for manufacturing exporters are also identified in a study by Bond (1984), who disaggregates trade flows by import market. 16/ According to Bond, the cyclical elasticity of manufacturing exporters with respect to GDP in industrial countries and OPEC is 2.91 and 1.88, respectively. Here, too, the question arises whether GDP growth in part services as a proxy for structural change on the demand side or the supply side. 5. External Debt and Interest Rates A large part of the debts incurred by developing countries consists of bank debts, denominated in dollars with floating rate debt service, geared primarily to LIBOR but also to the U.S. prime rate in some cases. Developments in short-term interest rates in the United States therefore have immediate relevance for debtor countries. The distinction between gross and net debt is not considered here, although that distinction is worth noting because some developing countries have large external assets. In fact, some major borrowers incurred their external debts in the very act of financing capital flight by domestic residents. Hence, at least in accounting terms, public external debt is at least in part matched by private external assets. Those private assets are beyond the control of the authorities, however, and thus make the gross external debt the relevant measure. 16/ See M. Bond, "An Analysis of Export Demand and Supply for Groups of Non- Oil Developing Countries" (Washington D.C.: International Monetary Fund, May 1984; processed). - 25 - There are no solid data on the currency composition of debts. According to OECD estimates, the share of dollar debt in the total debt is about 50 percent in the case of long-term liabilities and about 70 percent for short-term debts. 17/ This high concentration of denomination, as we shall see, plays an important role in the context of a policy mix that involves high U.S. interest rates and dollar appreciation. Some evidence has been compiled on the effective interest rates paid by different groups of borrowers (see Table 7). The effective interest rate is calculated as the ratio of total interest payments to total debt outstanding. Total debt includes debt to official agencies at low long-term or outright concessional rates as well as debt to commercial banks with floating rates. Therefore the effective rate of interest is a weighted average of rates on bank debt and rates on concessional loans. According to World Bank estimates, in 1982, 21 percent of the total outstanding debt of developing countries carried concessional rates. Table 7 reflects the differing mix of debt among groups of countries. The poorest countries in particular (because they had no extensive access to bank borrowing), showed an effective interest rate of only 4 percent in 1982 whereas major Latin American borrowers--the upper-middle class of the developing countries--paid more than 12 percent. 17/ See Organisation for Economic Co-operation and Development, External Debt of Developing Countries (Paris, 1983). - 26 - Table 7 Effective Interest Rates (percent)- 1979 1980 1981 1982 1983 U.S. prime rate 12.7 15.3 18.9 14.9 10.8 LIBOR 12.1 14.2 16.8 13.2 9.6 Total debt of developing countries 7.7 9.0 9.7 10.0 8.7 Fixed interest 5.8 6.0 6.0 6.3 6.7 Floating rate 12.3 5.5 17.4 17.1 12.7 Lower-income countries 3.8 3.8 3.8 4.0 3.7 Lower middle-income countries 7.3 8.7 9.0 9.8 8.3 Upper middle-income countries 9.3 11.0 12.1 12.2 10.6 Source: Organisation for Economic Co-operation and Development. From the ratio of external debt to GNP and the ratio of interest payments to GNP (see Table 8), we can see that the debt burden is signifi- cantly larger and has increased much more in the case of major borrowers and Latin America than in the case of poor countries. This is once more a reflection of the fact that poor countries have primarily fixed-interest debt and were unable to splurge in the way upper-middle-class countries did. 18/ 18/ On the splurge, see R. Dornbusch, "Budget Deficits, Disequilibrium Exchange Rates and External Debt," in International Debt and the Developing Countries, ed. J. Cuddington and G. Smith (Washington, D.C.: World Bank, 1984). - 27 - Table 8 Debt Burdens Low-income Major Western Countries Borrowers Hemisphere 1977 1983 1977 1983 1977 1983 Debt/GDP ratio 14.0 15.7 27.6 42.5 27.9 54.8 Interest payments Percentage of exports 4.7 5.4 7.2 18.6 10.0 32.2 Percentage of GDP 0.4 0.5 1.3 4.1 1.4 6.1 Source: International Monetary Fund. Spreads The debt service of developing countries is linked to interest rates in the world capital market via LIBOR. The interest rate charged on debt is LIBOR plus a "country" spread. In 1983, average spreads on new credits were 180 basis points. In recent reschedulings, these spreads declined to 112 basis points. But a second spread should be noted, that between LIBOR and the "risk-free" interest rate, say, the U.S. T-bill rate. The average spread between LIBOR and T-bills for the 1971-84 period was 187 basis points (see Figure 3), but it dropped to 75 basis points in 1984. This "banking" spread reflects both the costs of bank intermediation and the risk premium that banks have to pay in funding their own Lending. Equation (6) explains the Libor/T-bill rate spread in terms of the level of the T-bill rate, the log of the real price of oil and a dummy for the Herstatt episode in 1973 (see Figure 3). The data are quarterly for the period 1971-84:2. The real price appears as a proxy for the effect of increased placements by oil producers in the Euro-dollar market. 19/ 19/ The Herstatt dummy assumes a value of unity for 1974:2 to 1974:4 and zero elsewhere. - 28 - I:V861 L V~~~~~~~:Z861 £:1861 T:861 1:1861 V:0861 E:0861 Z:Z86T T:Z861 V:T161 £:T861 Z:186T X ~~~~~~~~~~1:61861 7:0861 £:0861 Z:0961 1:0861 V7:861 Z:6L6T T:8L616 Z:9161 l >V1:9161 '-. 7 :9L61 1: IL 6T Z:9L61 ii: ) 1:9L61 Eq IV-VSL~~~~~~~~~~~~~~~~~~7761 OO £:7161 Z:9L61 T:9L61 @~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ o V£161 < @ E:SL6T 00 Z:£L61 Q Q ~~~~~~~~~~~~~~~~~~~~V ZL61 V:IL6T Q: Z IL61 A) . . . I IL61 - 29 - (6) Spread = 2.68 + 0.25 T-Bill - .71 Oil Price + 1.86 Herstatt R = .67 (4.65) (6.14) (-4.03) (4.92) DW = 2.0 Fortour purposes, the interesting point is that the level of the T-bill rate influences the spread. A I percentage point increase in the T-bill rate raises the LIBOR/T-bill rate sprad by 10 basis points in the short run and by nearly 20 basis points in the long run. Interest rate developments in industrial countries thus influence developing countries both via the level of T-bill rates and, in a minor way, via the level of the banking spread that rises with T-bill rates. In a model proposed by Dantas (1984), banking spreads are influenced by the level of interest rate. 20/ As in the Jaffee-Russell credit model, a rise in interest rates increases the debt service burden relative to the costs of default and therefore raises the probability of default. The capital market recognizes that bank lending is exposed to this risk and hence charges banks a risk premium that reflects the possibility of default among their customers. In that framework, according to Dantas, higher interest rates imply an increase in the banking spread. There is no clear and simple link in quarterly data between LIBOR and the U.S. prime rate, except that the prime always exceeds LIBOR. The discrepancy is due to the cost differential of offshore banking, which is affected by regulatory and tax policies. Between 1971 and 1984 the average excess of prime over LIBOR was only 0.6 percentage points. But during 1980-84 the discrepancy averaged 1.4 percentage points. 20/ See D. Dantas, "Bank CD Rates and LDC Default Risk" (Fundacao Getulio Vargas, August 1984; processed). The key equation of the Dantas model links the cost of bank funding v to the risk-free rate, i, by the equation v(l-L(v+x))=i where L() is the probability of default, which is an increasing function of the borrowing rate, which is v plus a country spread x. -4 - 30 - The Real Interest Rate Welfare questions, of course, are concerned with the real rate of interest, not the nominal rate. The real rate can be defined as U.S. prime rate less the rate of export price inflation in non-oil developing countries (see Figure 4). The notable periods are 1972-73, when negative real rates were 24 percent for two years, and 1981, when the real rate was plus 24 percent. There is some question as to what deflator is appropriate in calculating the real interest rate: export prices, import prices, or some other price index. The answer is clearly that we should use the consumer price index (CPI), but unfortunately a dollar-CPI series is not readily available for non-oil developing countries. If CPIs were used, the real interest rate would emerge as the series represented in Figure 4, with two adjustments. Letting px' Pmy Pn be the rates of dollar price inflation of exportables, importables, and nontraded goods, we have: (7) Real Interest Rate = (i-px) + a(px-pn) + b(P xPm)9 where a and b are the expenditure shares of nontraded goods and importables. Changes in the external and internal terms of trade therefore also enter a welfare-based measure of real interest rates. 21/ Depending on the disturbance and the adjustment response, these relative price changes may increase or decrease real rates beyond what is indicated by the term i - 21/ See R. Dornbusch, "Home Goods, Real Interest Rates and Optimal External Borrowing," Journal of Political Economy (February 1983). - 31 - Figure 4 The Real Interest Rate Percent 25 23.9 0 +4 .4 tt. .. 4.. + + + -25 -24.6 as o H cs m -< uN %, r-0 ON a oA r, cs tn %. ruN oo ax o -i cs m un %- sD %O 10 14 %0D %D 1 0 0 %Q | rv 1- t rv rN f- X0 Co 0 X ON C a% a o% as ON oN ON cr ON a% oN o % 0% C% o~ 0% a Note: The real rate is defined as the U.S. prime rate adjusted by the rate of inflation of the unit export value for nonoil developing countries. - 32 - 6. The Implications of Alternative Mixes The preceding evidence implies that all debtor non-oil developing countries have an interest in some combination of strong OECD growth, low real interest rates, and a weak dollar--which amounts to a reenactment of 1972-74. The precise trade-offs between OECD growth, interest rates, and the dollar differ between countries. Here, the debt structure and the trade structure make a considerable difference. A few conclusions can be drawn from the overall evidence: a. Growth in the developed countries favorably affects the export earnings of developing countries. The precise quantitative link is uncertain, but the estimates reviewed above suggest that growth of industrial production by 1 extra percentage point will raise real export revenue by about 1.5 percent. An extra point in GDP growth in the developed countries would contribute about 2 to 3 percent to export revenue growth. b. A depreciation of the dollar affects the real prices of commodities. Over a two-year period, depreciation of 10 percent would add somewhat more than 5 percentage points to the growth in real commodity prices. A dollar depreciation therefore raises real revenues of net material exporters and raises real import bills of net material importers. c. A rise in real short-term interest rates raises debt service burdens directly and via increased spreads between LIBOR and T-bills. The cost of debt service (as - 33 - a fraction of exports) increases with the size of the ratio of bank debt to exports. d. Developing countries are no longer uniformly exporters of commodities, nor do they all have large bank debt-to- export ratios, as can be seen from the four typical cases presented in Table 9. 22/ Table 9 Structural Characteristics of Non-Oil Developing Countries Ratio of bank debt to exports High Medium Low Net exporters Brazil Peru Sri Lanka of materials Chile Net importers Korea Hong Kong of materials Depending on the economic structure of a country, a combination of, say, 3 extra points in developed country growth, 1 percentage point higher LIBOR rates, and a 10 percent dollar appreciation could amount to a net financial benefit or loss. The higher the bank debt relative to exports and the larger the concentration in commodity exports, the less likely that a country would gain. But if the same growth was the result of reduced interest rates and was accompanied by a decline in the dollar commodity, exporters and debtors alike would gain. To give another example, when a recession in a 22/ On trade structure, see especially W. Branson, "Trade and Structural Interdependence between the U.S. and the NICs" (Princeton University, 1983; processed). - 34 - developed country leads to a cyclical decline in interest rates, a large debtor might benefit because the lower interest rates more than offset the disadvantage of reduced growth. According to recent evidence, for large debtors, interest rates are by no means far less important than growth. The monetary-fiscal policy mix in industrial countries involves alternative packages of the value of the dollar, OECD growth, and dollar interest rates. The mix has two dimensions: one has to do with the convergence or divergence of Europe-Japan and the United States with respect to growth and inflation; the other with monetary-fiscal policy. In the past year, there were strong divergences in every respect: the United States and other industrialized countries were coordinated with respect to the recovery; but the speed of recovery in the United States was above average, whereas few signs of a sustained pickup were evident in Europe. The mix is also divergent with respect to the monetary-fiscal mix. In Europe, inflation and cyclically adjusted budgets are moving toward surpluses or even toward increasing surpluses; in the United States, structural deficits are widening. There is perhaps some similarity in monetary policy in the sense that real rates are being kept positive and high everywhere. But even here there is some divergence in that Europe, faced with the need to do something, favors an easing of interest rates over transitory fiscal expansion. In the United States, however, the strength of real money demand growth, driven by fiscal expansion, has not been sufficiently accommodated in monetary policy, and hence interest rates are rising in real terms. Blanchard and Dornbusch (1984) and Layard et al. (1984) have made a case for a monetary-fiscal policy mix that involves a transitory European fiscal expansion, a long-term tightening in the United States, and an - 35 - accommodating monetary policy. 23/ That continues to be the first best recommendation in that it ensures a continuing recovery together with a declining dollar. By no means does it solve the debt problems of developing countries by the stroke of a pen. But it does help alleviate them through dollar depreciation and continuing growth. At the moment, however, Europe does not appear to be interested in coordinated fiscal expansion, even of an admittedly transitory nature. Indeed, a U.S. fiscal correction is being called for almost universally, and yet any Jeuropean action is limited to allowing nominal interest rates to decline in the aftermath of a reduction in inflation induced by the accompanying dollar depreciation. (It is worth noting that many now believe that fiscal correction will lower the dollar, whereas a year ago it was predicted that the dollar would collapse because of continuing deficits.) World attention has become focused on the U.S. budget deficits because it is widely believed that cutting the deficits will remedy many, if not all, of the world's economic problems. This is doubtful. There are a number of possible scenarios to explore, but not all are equally attractive for developing countries, or indeed attractive at all: a. Without changing monetary targets, the United States raises income taxes to eliminate, say, one-half of the deficit. The immediate effect would be a sharp slowdown in U.S. growth, perhaps to the point of a recession. Such a slowdown would be accompanied by a decline in interest rates and a depreciation of the dollar. 23/ See 0. Blanchard and R. Dornbusch, "Europe and the Dollar," Banca Nazionale del Lavoro, Quarterly Review (February 1983); and R. Layard et al., "The Case for Unsustainable Growth," Center for European Policy Studies Discussion Papers (Brussels, April 1984). - 36 - b. The United States reduces future budget deficits by phased-in income tax increaess and de-indexation, thereby reducing long-term interest rates. At the same time, monetary policy supports continued expansion by turning easier. In this case, the dollar would fall and therefore U.S. inflation would raise, at least transitorily. c. U.S. fiscal policy is corrected by a drastic and immediate fiscal package that restores a balanced budget. Monetary policy would turn steeply expansionary to rescue the continued recovery. d. Future deficits are corrected in a credible fashion, reducing long-term real rates. Monetary policy stays within the announced more limited targets for 1985, potentially forcing up short-term rates as investment responds strongly to the reduced cost of capital. The dollar would remain firm. e. Recognizing that short-term real rates are high because the stock of real money balances is unusually low relative to activity, the Fed produces a money blip like the one of 1983. The once-and-for-all increase in real balances takes short-term rates toward zero, and thus ensures a continued expansion of demand, a depreciation of the dollar, and the removal of all appearances of deflation. Removal of tax indexation and a superboom would dampen concerns about the budget. - 37 - These scenarios all allude to one important point: the lower (short- term) interest rates hoped for as a consequence of U.S. fiscal correction would come about primarily through a slowdown. Sensible fiscal correction would (perhaps?) reduce long-term rates, but these would not be of interest to -developing countries whose debts are geared to LIBOR. The short-term rates are high because of unaccommodating monetary policy and strong growth, behind which lies the fiscal expansion. Developing countries should therefore be calling for easy money in the United States, not for tight fiscal policy. Indeed, developing countries can only benefit in the short run from the most reckless spending and money printing imaginable. Option d, which has all the appearance of orthodoxy and soundness, is, at best, in the interest of developing countries when one looks far ahead and argues that this is the only avenue of securing sustained growth. Option e best satisfies the immediate concerns of liquidating debt and raising real commodity prices. The United States and hence the developing countries are caught in an uncomfortable position in which the U.S. monetary-fiscal mix can be changed only at the cost of provoking renewed inflation or recession. A shift toward tight fiscal policy and lower real interest rates can, of course, be brought about by raising taxes and expanding the real money stock. Moreover, we can make the change in a manner that sustains the level and rate of expansion of total demand in developed countries. But what we cannot easily do is avoid the depreciation of the dollar that would almost inevitably accompany such a change of the mix. It is the dollar depreciation, if it occurred on a - 38 - significant scale, that would cause inflation to accelerate. 24/ After having invested in fighting inflation, and having won to some extent because of appreciation, the United States would have to induce a new recession to fight the spillover of inflation into wages. At the same time, however, it is recognized that, even when real rates of interest are high or money is tight, there is a reluctance to advocate a decisive "liquification" of the economy. One explanation for this reluctance might be that accelerated money growth itself creates inflationary expectations, which immediately find their way into actual wage and price inflation. But there is no evidence of such a direct spillover. On the contrary, the U.S. experience in 1983 suggests that acceleration in money growth and disinflation can coexist, at least at high levels of unemployment. The evidence from Germany runs in the same direction. The reluctance to follow a more aggressive monetary policy can be attributed to two other factors. One, as noted earlier, is the strong possibility that reduced interest rates in the United States would bring about the depreciation of the dollar, which in turn would immediately speed up inflation. The other is that, near full employment, expectations arguments must receive more consideration. Under these circumstances, a shift toward tight fiscal policy and easy money, or simply easy money, is not likely to appeal to the United States. 24/ The judgment that a sizable dollar depreciation would have a considerable impact on inflation, particularly near full employment, is based on the evidence reported in R. Dornbusch and S. Fischer, "Open Economy Aspects of U.S. Monetary and Fiscal Policy" (National Bureau of Economic Research, 1984; processed). - 39 - There is one obvious difficulty in reconciling the interests of developing countries and those of industrial countries. Thus far, the United States has had the best of all worlds: strong recovery and growth without a significant increase in inflation. Because of fiscal expansion, growth has been able to take place despite high interest rates. At the same time, the high interest rates have prevented the collapse of the exchange rate and hence have warded off the inflationary pressure from importables. It might seem unreasonable to suggest that the dollar be kept overvalued, until further notice, simply to delay the day of reckoning with respect to U.S. inflation. It might seem equally unreasonable to enact wage controls merely to avoid an extra 2 or 3 percent inflation stemming from the impact of an exchange rate collapse on wages. On the other hand, it might be entirely reasonable to pursue first-best industrial country policies merely because in the long run developing countries can benefit only if stability and prosperity become a more permanent feature of the advanced countries. In that case, one would correct U.S. fiscal policies over the next few years in a credible fashion, pursue at least a transitory fiscal expansion in Europe, and make sure that the dollar comes down gently and in an undisruptive manner. This is the soft- landing option. The only trouble with this option is that, worldwide, monetary and fiscal policies would have to become sufficiently responsive, to achieve the multiple targets of recovery, stable dollar depreciation, and the nonacceleration of inflation. The political support for this sensible option is negligible so far. The World Bank Headquarters European Office Tokyo Office 1818 H Street, N.W. 66, avenue d'lena Kokusai Building Washington, D.C. 20433, U.S.A. 75116 Paris, France 1-1 Marunouchi 3-chome Telephone: (202) 477-1234 Telephone: (1) 47.23.54.21 Chiyoda-ku, Tokyo 100, Japan Telex: WUI 64145 WORLDBANK Telex: 842-620628 Telephone: (03) 214-5001 RCA 248423 WORLDBK Telex: 781-26838 Cable Address: INTBAFRAD .t WASHINGTONDC 0 S, ft ft ISSN 0253-2115/ISBN 0-8213-0704-5 O