V Transmission of Economic Influences from Industrial to Developing Countries
Author:
Mr. David John Goldsbrough
Search for other papers by Mr. David John Goldsbrough in
Current site
Google Scholar Close
and
Mr. Zubair Iqbal
Search for other papers by Mr. Zubair Iqbal in
Current site
Google Scholar Close
Language:
English
Print Publication Date:
01 Jan 1986
Keywords:
WEFS; money stock; U.S. dollar; wage employment; export volume; central bank discretion; unemployment rate; credit rationing; purchasing power; capital goods; export trade; developing country group; Exports; Imports; Terms of trade; Commodity prices; Africa; Europe; Asia and Pacific
Download PDF (3.3 MB)

Abstract

The impact of the recession of 1980–83 on the economies of developing countries, the effects of high interest rates in financial markets since 1979, and the reduction in new bank lending to developing countries since 1982 have been major recent topics of discussion. Nevertheless, the precise mechanisms through which these effects have been transmitted are complex and only partially understood. This paper reviews the available evidence on the principal links through which changes in macroeconomic performance in the industrial countries influence major economic variables, in particular the rate of economic growth, in developing countries.

The impact of the recession of 1980–83 on the economies of developing countries, the effects of high interest rates in financial markets since 1979, and the reduction in new bank lending to developing countries since 1982 have been major recent topics of discussion. Nevertheless, the precise mechanisms through which these effects have been transmitted are complex and only partially understood. This paper reviews the available evidence on the principal links through which changes in macroeconomic performance in the industrial countries influence major economic variables, in particular the rate of economic growth, in developing countries.

These links are numerous: changes in the pace of economic growth and inflation as well as shifts in interest rates and exchange rates in the industrial world can all influence developing countries’ economic performance through the markets for goods, services, and migrant labor, as well as through financial markets. The impact of various influences is determined simultaneously. For instance, a shift in the stance of fiscal or monetary policies in industrial countries can alter interest rates and exchange rates and shift the pattern of savings and investment; they can also cause changes in aggregate demand that result in price movements and at least short-term variations in output and employment. These developments in industrial countries in turn affect the price and volume of developing countries’ exports, the price of their imports, and the volume and terms of their capital inflows. Moreover, the resulting changes in developing countries’ economies have a feedback effect on the industrial countries by altering the demand for the latters’ exports of goods and capital and by shifting the supply schedule for their imports.

These various effects can be analyzed either by using single equation (reduced-form) estimates and other partial equilibrium studies of individual linkages, or by constructing a more comprehensive structural model that takes account of the general equilibrium nature of the links between industrial and developing countries. The former approach is used in this paper, since it complements more closely the procedures used to prepare the medium-term scenarios for developing countries in the World Economic Outlook. These scenarios rely heavily on survey estimates prepared by staff members in area departments on the basis of overall environmental assumptions about the world economy (for example, average output growth in industrial countries, prices of manufactured goods imported by developing countries, interest rates on commercial credits, scale of private lending to developing countries, and exchange rates among the major currencies). Comparisons between econometric estimates and estimates obtained from the survey can help to indicate whether, in the aggregate, individual survey data might be inconsistent with the assumed global environment. The general equilibrium framework of the scenarios is provided in two ways: first, the survey gives a set of projections (for example, of trade flows, gross domestic product (GDP), and the balance of payments) for each country that is internally consistent for that country; second, any inconsistencies between the aggregated survey estimates and the assumed global economic environment is eliminated through iterative adjustment.1

Nevertheless, the limitations of the single-equation approach taken in isolation should be borne in mind during the subsequent discussion. In particular, estimates will be biased to the extent that there are significant interactions in the world economy that are omitted from the reduced form and to the extent that variables which would be determined endogenously in a structural model are assumed to be given exogenously in the reduced form.2

As a background to the discussion in the rest of the paper, Table 47 provides information on broad trends in the growth rates of output in industrial and developing countries. Several aspects of these trends deserve special mention. First, there does appear to be a rough association between the rates of economic growth in the industrial countries and those of the developing countries. The slowdown in economic growth in the industrial countries between the 1968–72 period and the 1973–80 period was accompanied by a moderate slowdown in developing countries’ economic growth. Similarly, the 1981–82 recession in industrial countries and the subsequent recovery in 1983–85 were associated with broadly comparable outcomes in the developing countries.

Table 47.

Growth Rates of Real Output, 1968–85

(Average annual percentage changes)

article image
Source: International Monetary Fund, World Economic Outlook. Note: See Appendix 1 to this chapter for the definitions of the country groups referred to in this note. GNP is gross national product, GDP is gross domestic product.

The relationship between output growth in the two groups is not a precise one, however, and there are clearly many other influences on developing countries’ economic growth. This can be seen most clearly when the growth performance of the fuel exporters and non-fuel exporters among developing countries are considered separately. (See Appendix I to this chapter for details on these country groups.) The average rate of economic growth of the non-fuel exporters declined only moderately between the periods 1968–72 and 1973–80, even though industrial country growth slowed more substantially, and the decline in the growth of the fuel exporters was even more pronounced. Also, the recovery in industrial countries’ economic growth in 1983–85 was not reflected in renewed output growth in the fuel exporting countries, while the non-fuel exporters did recover.

A second observation emerging from Table 47 concerns the considerable diversity in the growth performances of different groups of developing countries. Among the non-fuel exporters, the exporters of manufactures managed to sustain their rate of economic growth during the 1981–82 recession while output growth was actually negative in the primary product exporters. The former group also appear to have benefited much more from the recovery in industrial countries during 1983–85. On a geographical basis, the output growth of developing countries in Asia was only moderately affected by the recent recession and has since recovered sharply, whereas the average rates of economic growth of developing countries in Africa, the Middle East, and the Western Hemisphere have been substantially lower during both the recent recession and the recovery phase. There are also substantial variations in growth performance within these groups.

A third observation is the striking divergence between the average growth performances since 1981 of those countries which have and those which have not encountered recent debt-servicing problems. These two groups had broadly similar growth records during the period 1968–80, which suggests that much of the recent sharp decline in output growth in countries with debt-servicing problems is related to those problems and the associated sharp reduction in private capital flows. In this regard, the average rate of economic growth of the market borrowers fell much more sharply between 1973–80 and 1981–85 than did the growth of the official borrowers.

Transmission of Economic Influences Through Goods and Services Markets

Macroeconomic developments in industrial countries can significantly influence the external current account and output growth of developing countries through their simultaneous effects on relative prices and volumes in foreign trade. The resulting changes in export earnings will, in turn, affect developing countries’ growth prospects both in the shorter term, through their impact on the level of aggregate demand and the availability of foreign exchange, and in the longer term, through their effects on the level of investment and the relative size of the export sector. Before discussing these various influences, and those characteristics of the trade structure of developing countries that help determine the size and nature of industrial country influences on the developing world, it is important to make a distinction between those countries that rely on oil exports for a substantial proportion of their foreign exchange earnings and those that do not. The classification of developing countries into fuel exporters and non-fuel exporters reflects this distinction and will be used where possible in this section. However, some econometric results and some statistical information are only available on the basis of the alternative analytical categories of oil exporting countries and non-oil developing countries, and these categories will be used where necessary (see Appendix I to this chapter for details on these classifications).

Developing Countries: Foreign Trade

Several features of developing countries’ foreign trade are important to any discussion of the transmission of economic influences through the goods market. First, the industrial countries as a group represent the major market for developing countries’ exports and supply the major portion of developing countries’ imports, despite some recent decline in their share of developing countries’ foreign trade (Chart 15).

Chart 15.
Chart 15.

Developing Countries: Destination of Exports and Origin of Imports, 1967–84

(In percent)

Second, export developments for the fuel-exporting and non-fuel exporting developing countries, which were broadly similar during 1968–72, have differed sharply since 1973 because of the wide fluctuations in oil prices. The total volume of exports of the fuel exporters has declined substantially since 1973, although, until 1982, this was more than offset by the rapid increase in export prices (Table 48). The volume of their exports to industrial countries has fallen even more rapidly than those to other regions, owing to a sustained decline in energy consumption per unit of output in these countries, and to rising domestic production of energy in some of them. Nevertheless, the large increases in oil prices caused the share of the fuel exporters in total imports by industrial countries to more than double between the late 1960s and the early 1980s, although their market share has since fallen as a result of the drop in oil prices and a continued decline in export volumes (Table 49).

Table 48.

Developing Countries: Changes in Export Volumes, Export Values, and Terms of Trade, 1968–85

(Average annual percentage changes) 1

article image

Compound annual rates of change.

In U.S. dollar terms.

Table 49.

Industrial Countries’ Imports From Developing Countries: Changes in Values, Volumes, and Market Shares, 1968–84

(Average annual percentage changes, unless otherwise indicated)1

article image

All values are in terms of U.S. dollars.

In percent of total value.

In contrast, the marked slowdown in the rate of growth of the volume of industrial country imports after 1973 had only a moderate impact on the growth of total export volumes by the non-fuel exporters; the latter continued to expand the volume of their exports to the industrial countries more rapidly than the total volume of industrial country imports, and also continued to expand their exports to other developing countries. Between the periods 1968–72 and 1973–80, the average annual growth rate of industrial country imports fell from 10 percent to 4 percent, but the growth rate of total imports from the non-fuel exporters slowed only moderately from around 8 percent to about 5.5 percent (Table 49). This trend has continued through the recent recession and recovery.3 However, for most of the period since 1973, the purchasing power of these exports from non-fuel exporters has been adversely affected by declining terms of trade.

A third important feature of the goods markets is the commodity composition of developing countries’ foreign trade, especially of their exports, which has shifted considerably over time. The share of manufactures in developing countries’ exports has increased markedly since the mid-1960s, while import substitution and the growing cost of fuel imports have contributed to a moderate decline in the share of manufactures in their total imports (Table 50). By 1980, manufactured exports accounted for almost 60 percent of the total exports of capital-importing developing countries, which was not much below the 69 percent share of manufactures in total world trade (Appendix Table I). Developing countries in Europe and in Asia now export mainly manufactures; in 1980, around four fifths and two thirds, respectively, of their total exports consisted of manufactures. Nevertheless, most low-income economies other than China and India still rely on non-fuel primary commodities for the major part of their export earnings; this group includes many countries in sub-Saharan Africa. Also, these commodities still represent over half of all exports by developing countries in the Western Hemisphere.

Table 50.

Developing Countries: Commodity Composition of Foreign Trade, 1965 and 1982

(Percentage shares)

article image
Source: World Bank, World Development Report, 1984, Tables 10 and 11, pp. 192–95. Note: The classification of countries is that followed by the World Bank in the World Development Report. Low-income economies are those which had a per capita GNP of less than $400 in 1983. The groups low-income economies and middle-income oil importers combined are broadly equivalent to the Fund’s category of non-fuel exporters.

The first and fourth columns cover only fuel imports.

The variations in commodity composition over time and across countries have important implications for the effects on developing countries of economic growth in the industrial world. Apart from fuels, the fastest growing component in imports by industrial countries imports since the 1960s has been manufactures (Table 51). Those developing countries that were well placed to take advantage of this growing demand for manufactures have achieved substantially higher rates of export growth. Thus, during 1973–84, the volume of exports to industrial countries from the group of exporters of manufactures grew at an average annual rate of 9.7 percent, while the export volume of the group of primary product exporters grew at a rate of only 3.2 percent (Table 52).

Table 51.

Industrial Countries: Commodity Composition of Imports, 1965 and 1982

(Percentage shares)

article image
Source: World Bank, World Development Report, 1984, Table 11, p. 195.
Table 52.

Industrial Countries: Imports and Implied Income Elasticities, 1973–84

(Average annual percentage changes, unless otherwise indicated)

article image

Export values deflated by the index of export unit values for each group’s total exports.

Average growth rate of industrial country imports from each developing country group divided by the average growth rate of real GNP of the industrial country.

In addition to their role as demanders of imports, industrial countries are also large exporters of primary commodities. Indeed, their share of global world exports of non-fuel primary commodities has grown over time and, in 1980, they accounted for 69 percent of all such exports by industrial and developing countries combined (Appendix Table II). At the level of aggregation of the single-digit SITC groups, the export market shares of industrial countries were larger than those of developing countries for all categories except beverages and tobacco and fuels. (SITC stands for Standard International Trade Classification.) Consequently, those developing countries that export primary products similar to those produced by industrial countries will be strongly affected by developments, such as changes in domestic agricultural pricing policies, that affect the volume of industrial country primary production or exports.

Finally, a fourth key feature of developing countries’ foreign trade is the wide variations in the importance of different industrial country markets for the various regional groups of developing countries. If the exports of developing countries consisted mainly of relatively homogeneous primary commodities, and if there were no significant trade barriers against these, regional variations in the direction of trade would not be of great importance. Divergent movements in the growth rates or real exchange rates of different industrial countries would cause changes in the demand and hence in the world price of various commodities, and each developing country supplier would face the same price changes, irrespective of the destination of their exports. The resulting change in the volume of their exports would depend on the domestic supply responses to these price changes. In practice, however, a large share of the exports of developing countries consists of relatively non-homogeneous manufactures for which prices can vary between markets and between suppliers. In addition, some of the industrial country markets for certain primary product exports of developing countries—such as sugar, grains, and meat—are in effect fragmented by various quantitative import restrictions. In such circumstances, the growth in demand for a developing country’s exports can be significantly affected by the geographic distribution of those exports.

In this regard, developing countries in Africa and in Europe have strong links with industrial countries in Europe; almost one half and two fifths, respectively, of their total exports are sent to that region (Table 53). In contrast, the principal industrial country markets for Asian developing countries are the United States and Japan, while developing countries in the Western Hemisphere rely heavily on the U.S. export market. These differences in the geographic orientation of exports were especially important during 1983–84, when the structure of expansion in the industrial world was unbalanced. The rapid expansion of economic activity in the United States, together with the appreciation of the U.S. dollar, caused a large increase in U.S. demand for imports; over the two years the volume of total U.S. imports rose by 47.5 percent (Table 52). At the same time, the slow pace of recovery in Europe was reflected in only a moderate increase in import demand; total import volumes grew by only 9 percent over the two years combined. The consequences of this uneven expansion for the different regional groups of developing countries can be seen by calculating the weighted average growth rates of real gross national product (GNP) and import volumes in the industrial countries, using the direction of exports of the regional groups of developing countries as weights (Table 54). During the three-year period 1983–85, developing countries in the Western Hemisphere faced an average industrial country import market that grew by some 3–5 percentage points per annum faster than did the average industrial country import market faced by African, European, or Middle Eastern developing countries, whereas the differences during the 1973–80 period had been relatively minor.

Table 53.

Developing Countries: Destination of Exports, 1970 and 1984

(As a percentage of total exports)

article image
Source: International Monetary Fund, Direction of Trade, various issues.

Derived as a residual. The data include exports to the U.S.S.R., Eastern Europe, and unspecified countries or areas.

Table 54.

Industrial Countries: Output Growth and Growth of Markets Weighted by the Direction of Exports of Developing Country Regional Groups, 1973–85

(Average annual percentage changes)

article image
Note: The data show the growth in real GNP, total import volumes, and non-oil import volumes in the industrial countries, weighted by the direction of exports for each developing country group. The weights are based on 1984 exports.

Transmission of Disturbances Through Prices

Developments in industrial countries can cause shifts in the demand and supply on world markets of goods that are exported or imported by developing countries, thereby causing simultaneous changes in both prices and volumes. Most empirical studies of these influences on developing countries’ trade flows have concentrated on the impact on demand.4 These studies typically assume infinite price elasticities of supply. While this assumption may be reasonable for the case of the world supply of imports to individual developing countries (or even to groups of developing countries, if they represent a small part of the total world market), it appears less reasonable for the supply of exports from developing countries. When the price elasticity of export supply is not infinite, an upward shift in demand (due perhaps to a rise in economic activity in the industrial countries) will result in a rise in both the price and volume of developing country exports. In such circumstances, estimates of the impact on developing countries’ export earnings of, say, higher industrial country income that were based on single-equation estimates of demand side relationships would tend to underestimate the price response and overestimate the volume response because substitution effects are omitted. This should be borne in mind during the following discussion.

The most appropriate approach would be to estimate the supply and demand relationships simultaneously. However, there have been relatively few studies of developing country exports that use simultaneous equation models comparable to those used for industrial country exports, in large part because of the difficulties involved in specifying separate supply relationships. Exceptions include the work by Bond (1985 and 1986) and various studies on the determinants of commodity price movements, including those by Chu and Morrison (1984 and 1986) and Holtham et al. (1985). Results from these three sets of studies are used later in this section.

Nominal and Real Commodity Prices

Although prices and volumes are obviously determined concurrently, it is useful to discuss them separately when considering the transmission of economic influences from industrial to developing countries. Movements in the prices of goods in world trade since the late 1960s are summarized in Table 55 and in Chart 16. The prices of non-oil primary commodities have shown large short-term fluctuations in response to shifts in both demand and supply factors, but, over the period 1968–85, have tended to increase at a slower average rate than the prices of manufactures exported by industrial countries.5 Among the factors affecting non-oil primary commodity prices have been, on the demand side, the level of economic activity, interest rates, and rates of inflation in industrial countries and, on the supply side, various changes in industrial country policies affecting levels of domestic production, particularly for agricultural commodities.6

Table 55.

Movements in Commodity Prices, 1968–85

(Average annual percentage changes) 1

article image

Compound annual rates of change. Nominal prices are in terms of U.S. dollars.

Oil export unit values of the oil exporting countries.

United Nations export unit value index for the manufactures of the developed countries.

Chart 16.
Chart 16.

Non-Fuel Exporting Developing Countries: Trends in Real Non-Oil Primary Commodity Prices and in the Terms of Trade, 1967–85

(Indices, 1980 = 100)

1 Index of market prices for non-oil primary commodities deflated by the export unit value index of the manufactures of industrial countries.

A number of recent empirical studies have confirmed that, when supply conditions are relatively stable, a decline in the level of economic activity in industrial countries exerts a downward influence on the prices of non-oil primary commodities (referred to henceforth as primary commodities or just commodities) (Table 56). Although estimates of the exact size of the cyclical effect have varied, a broad consensus estimate might indicate that the elasticity of commodity prices with respect to industrial production in the industrial countries is around 2. Substantially higher estimated elasticities are obtained by Fishlow (1985) and by van Wijnbergen (1985), but the former is based on a very short time period and the latter is derived indirectly, by estimating the relationship between real commodity prices and unemployment in the industrial countries, and then making use of the association between GDP growth and unemployment in the industrial countries—a relationship which may have changed substantially during the last decade. The studies by Chu and Morrison also indicate that the elasticity was substantially higher during the 1972–82 period than for the 1958–71 period; this apparent increase in the elasticity over time may partially account for the relatively low elasticity estimated by Dornbusch (1968) for the period 1960–84. There is also evidence that the prices of metals and agricultural raw materials are more sensitive to cyclical fluctuations in industrial production than are the prices of food and beverages. This is to be expected, since the former groups of commodities are more heavily used as industrial inputs.

Table 56.

Estimates of the Elasticity of Real Non-Oil Primary Commodity Prices with Respect to Industrial Countries’ Real Economic Activity

article image
Note: Although the form of the estimating equation varies, all of the above elasticities estimate the effect on real commodity prices of cyclical fluctuations in the rate of output growth. All of the equations either use primary commodity prices deflated by the price of manufactures as the dependent variable, or include an inflation variable among the right-hand terms. ** and * mean the data are significant at the 1 percent and 5 percent levels, respectively.

Derived as a weighted average of the coefficients for individual groups.

Derived from an estimating equation linking real commodity prices and the unemployment rate in the Organization for Economic Cooperation and Development, together with the assumption that a 1 percentage point increase in GNP is equivalent to a 0.3 percent decline in the unemployment rate.

All of these estimated elasticities, however, refer to the sensitivity of real commodity prices to relatively short-term, cyclical variations in the rate of output growth in industrial countries. The elasticity of response of real commodity prices to a shift in the longer-term, trend rate of economic growth in industrial countries would probably be lower. Not only would the share of stockbuilding—which has a high raw materials component—in the change in total demand be lower in the longer term, but the supply response would also be greater. In this regard, Holtham et al. (1985) have estimated reduced-form equations for the determination of commodity prices in which, by construction, all commodity prices (other than those for beverages) eventually return to their trend levels following a deviation of industrial country GDP from its trend growth path. Their results (estimated semiannually over the period 1967–84) suggest that such a deviation would cause commodity prices to rise, with the peak response occurring around the fourth semester (that is, up to two years after the initial rise in GDP); the price response then fades and virtually disappears after five to six years. Their results also confirm Chu and Morrison’s (1984) conclusion that cyclical fluctuations in the industrial countries have a bigger impact on the prices of metals and agricultural raw materials than on the prices of food. A 1 percentage point deviation in industrial country GDP from its trend is estimated to cause, at the peak, about a 2 percent increase in the former prices, but a smaller increase in the latter prices.7 Beverage prices are estimated to be permanently affected by the deviation of GDP from its trend—with prices about 3 percent higher as a result of a 1 percentage point deviation in industrial country growth—but this may simply reflect the greater difficulties in empirically estimating the lags in the price responses for these commodities, owing to the long lags in supply response for the crops involved.

As can be seen from Chart 16, the three most recent periods of reduced economic activity in industrial countries—in 1971, 1975, and 1981–82—have broadly corresponded with periods of low real commodity prices. However, there are clearly many other influences on commodity prices; their real level fell substantially in both 1978 and in 1985, when the level of economic activity in the industrial world was rising.8 One factor underlying the most recent weakness of commodity prices has been the geographic imbalance in industrial country growth rates. The European industrial countries account for more than one half of total world imports of non-oil primary commodities (Table 57). If world markets for all primary commodities were relatively homogeneous and free from trade barriers, then it would be the shares of the different industrial countries in total world consumption (and production) of these commodities rather than their shares in world trade that would determine the impact on prices of their GDP changes. However, as noted earlier, the markets for a number of primary commodities are fragmented. The relatively slow economic growth of the European countries in recent years has dampened the demand for commodities, and this has been less than fully offset by the more rapid growth of output in the United States and Japan.

Table 57.

World Trade in Non-Oil Primary Commodities and Manufactures

(In percent; 1979–81 average)

article image
Source: Based on the United Nations, International Trade Statistics, Series D. Some processed primary commodities such as copper, tin, aluminum and coffee concentrates are regarded as primary commodities, while all the other processed or semiprocessed goods are regarded as manufactures.

Inflation in industrial countries influences nominal commodity prices both by increasing the price of substitutes in importing countries and by pushing up the costs of production of those primary commodities exported by developing countries. Chu and Morrison (1984) found that a 1 percent increase in the wholesale price index in industrial countries was associated with an increase in commodity prices of slightly greater than 1 percent.9 Of course, such a close relationship is not surprising since components of the two indices are overlapping; moreover, the direction of causation is probably two way, since changes in commodity prices may also influence industrial country wholesale prices.

The level of interest rates on world financial markets could be expected to have several counteracting influences on commodity prices. High real interest rates discourage stock-holding demand for commodities, which would exert a downward influence on prices, but also tend to raise production costs, which would cause prices to rise. The unusually high level of real interest rates during the 1981–82 recession probably contributed to the sharp decline in commodity prices in that period by encouraging stock liquidation, and the persistence since then of relatively high real rates may have dampened the recovery in commodity prices for the same reason. Most econometric studies have not detected a strong, statistically significant impact of real interest rates on the prices of non-oil primary commodities as a group. There is, however, some evidence of a moderate inverse relationship between interest rates and the prices of agricultural raw materials and metals, demand for which is more affected by stockbuilding than that of other commodity groups (Chu and Morrison (1984), Holtham et al. (1985)).

The relationship between commodity prices and exchange rates has been of particular importance in recent years because of the substantial movements in the U.S. dollar against other major currencies. Two aspects of the relationship can have a direct impact on developing countries’ welfare. First, a depreciation of the dollar will lead to a general increase in the dollar prices in world trade, and will consequently tend to lower developing countries’ debt and debt service ratios, since around four fifths of developing country debt is denominated in U.S. dollars. The larger the rise in dollar prices, the greater will be the decline in the debt ratios. Second, some observers have argued that a change in the dollar exchange rate affects the price of primary commodities relative to the price of manufactures.

These effects can be investigated using two alternative approaches. First, the estimated price elasticities of export supply and import demand for various groups of primary commodities and manufactures, together with information on the share of the United States in world trade of the goods concerned, can be used to calculate the impact of a change in the U.S. dollar exchange rate on the prices of primary commodities and manufactures. Second, a direct econometric analysis of the relationship between nominal and real commodity prices and the U.S. dollar exchange rate can be undertaken.

Beginning with the first approach, the effect of a dollar depreciation on the dollar price of goods in world trade depends on the supply and demand elasticities for those goods in different countries and on market shares. If export supply and import demand elasticities are constant across countries then, other things being equal, the impact of a dollar depreciation on the dollar price of a good will increase: (1) the smaller is the U.S. share in world trade of that good; (2) the smaller are export supply elasticities; and (3) the smaller are import demand elasticities.10 The U.S. share in the world trade of primary commodities as a group is fairly small and also not that different from its share of world trade in manufactures (Table 57). Hence, in the absence of other factors—such as differences in price elasticities of export supply and import demand across countries and commodities—one would expect, on a priori grounds, that a given percentage dollar depreciation would raise the dollar prices of both primary commodities and manufactures by a relatively large fraction of the depreciation and that the relative price of primary commodities to manufactures would not be greatly altered. An analysis using U.S. market shares and estimated price elasticities of supply and demand for various broad commodity groups suggests that a 10 percent depreciation of the dollar would raise the dollar prices of both non-oil primary commodities and manufactures by between 6 and 7 percent.11 A similar exercise suggests that a 10 percent dollar depreciation would raise the dollar price of fuels by around 3.5 percent.

Nevertheless, the effect of a dollar depreciation on relative prices could be considerably larger at a more disaggregated level, since the relative importance of the United States in world markets varies substantially across commodities. Indeed, a dollar depreciation could tend to reduce the relative price in world trade of some primary commodities, compared with the price of other traded commodities and manufactures. For instance, U.S. production of agricultural products such as coarse grains, wheat, and fats and oils is a substantial proportion of the world total. Consequently, a dollar depreciation might be expected to have a significant downward effect on the world prices of these commodities relative to the price of other goods in world trade. This can be explained as follows: a depreciation of the U.S. dollar would raise the price of these commodities relative to the price of nontraded goods in the United States. The resulting output increase and consumption decrease in the United States would then tend to force down the world price of these commodities relative to other traded goods because of the large proportion of supply on world markets for these commodities that is accounted for by U.S. exports. The size of the relative price change would depend on the relevant supply and demand elasticities with respect to price.

Additional evidence on these issues is provided by econometric analyses of the determinants of the dollar prices of primary commodities and manufactures. Analyses of quarterly data covering the period from the early 1970s to the early 1980s find that the elasticity of dollar prices of primary commodities with respect to the exchange rate of the U.S. dollar vis-à-vis other major currencies is around –¾; a similar elasticity is obtained for the dollar price of manufactures.12 These results confirm the above finding regarding the lack of a strong impact of changes in the U.S. dollar exchange rate on the relative price of primary commodities vis-à-vis manufactures.

To a considerable extent, of course, the question of whether changes in the U.S. dollar exchange rate affects “real” commodity prices depends on the price index used to deflate commodity prices. An appreciation of the dollar would tend to lower the price of primary commodities relative to the composite price of U.S. domestic output and raise their price relative to the composite price of domestic output in other industrial countries. In this regard, several econometric studies detect a large and significant negative relationship between the real effective exchange rate of the U.S. dollar and the “real” price of primary commodities when the latter is derived by deflating commodity prices in current U.S. dollars by the U.S. GDP deflator. However, this relationship reflects, in large part, the effect of a dollar appreciation in raising the relative price of nontraded goods in the U.S. economy. This does not directly affect developing countries’ terms of trade since, by definition, developing countries do not purchase U.S. nontraded goods. Nevertheless, the effect of a dollar appreciation or depreciation on the terms of trade of individual developing countries will be strongly influenced by the degree to which their foreign trade is oriented toward the U.S. markets.

Macroeconomic developments in industrial countries will also influence the price of oil. Although the share of industrial countries in total world oil consumption has declined (from 71 percent in 1973 to 57 percent in 1984), changes in real output in industrial countries still have a substantial impact on the total world demand for oil. Most empirical estimates suggest that the elasticity of industrial country demand for oil with respect to real output is somewhat below unity, around, say, 0.8. This would indicate that a 1 percentage point decrease in industrial country output would lead to a decrease in total world demand for oil of just under ½ of 1 percentage point. The response of oil prices to such a change in demand would be strongly affected by the supply policies of the oil producers belonging to the Organization of Petroleum Exporting Countries (OPEC) particularly in the short term. If the members of OPEC reduced their oil output levels to match the decline in world demand, then oil prices would remain unchanged, although total oil export earnings would decline because of the reduced export volumes. However, recent experience has shown that the larger the fall in world demand, and consequently the larger the cuts in production needed to maintain an unchanged oil price, the greater the tendency for production quotas to be exceeded and for the actual oil price to decline as various oil exporting countries offer direct and indirect discounts.

Terms of Trade

It is evident that the consequences for developing countries’ terms of trade of the various shifts in relative prices discussed above will depend on the commodity composition of each country’s exports and imports. To illustrate this, rough estimates of the cyclical elasticity of the terms of trade of groups of developing countries with respect to gross national product (GNP) in industrial countries are given in Table 58. Although the simple estimating equations used to obtain these elasticities are obviously subject to numerous econometric problems, the results do indicate that the terms of trade of those countries that export mainly primary products benefit more from faster growth in industrial countries (and suffer more from slower growth) than do the exporters of manufactures. The terms of trade of the service and remittance countries appear to improve the least, perhaps because their exports are less oriented toward the industrial countries and because most of their exports that do go to the industrial world are destined for the more slowly growing European countries. For the group of non-oil developing countries, the terms of trade elasticity is estimated at 1.7; a similar estimate is derived by van Wijnbergen (1985) using a somewhat different approach.

Table 58.

Developing Countries: Estimated Marginal Elasticities of Export Volume and the Terms of Trade with Respect to Industrial Countries’ Real GNP, 1967–84

article image
Note: Estimates show the impact of changes in industrial country GNP within a year. Elasticities were estimated from ordinary least square regressions on equations of the forms: Δ log (volume of exports) = a0 + a1 Δ log (industrial country GNP); Δ log (terms of trade) = α0 + α1 Δ log (industrial country GNP). ** and * denote significance at the 1 and 5 percent levels, respectively.

Sum of the elasticities for export volumes and the terms of trade.

Elasticities are weighted averages of the elasticities for the analytical subgroups, using 1980 total exports as weights.

Excluding China.

However, there are grounds for believing that such estimates may be biased upward, since part of the effect attributed to changes in industrial country output may actually be due to other factors omitted in the analysis. In this context, two features of these estimates should be noted. First, they measure the rela- Table 58 relate changes in the terms of trade to the rate of output growth in industrial countries during the same year. The longer-term impact of faster industrial country growth could be smaller, because demand from stockbuilding would be less important and because supply would have longer to adjust to the increased demand. In this regard, estimates by Cline (1984) on the effect of economic growth in countries of the Organization for Economic Cooperation and Development (OECD) on the terms of trade of selected developing countries suggest that there may be an initial terms of trade improvement as OECD growth accelerates, but that no further improvement occurs once a stable growth rate is reached. According to his estimates, a 1 percentage point increase in the rate of industrial country growth would cause, on average, a rise in developing countries’ terms of trade of 1.5 percent in the first year and another 1.5 percent in the second year, but the terms of trade would then remain unchanged unless there was a further change in the rate of growth in industrial countries. However, the nature of the supply responses that would generate such a relationship are not clear.

A second reason for believing the estimates may be high is that the earlier estimates of the impact of industrial country growth on the relative price of primary commodities vis-à-vis manufactures, together with the commodity composition of developing countries’ exports and imports, suggest somewhat smaller effects on the terms of trade than those estimated in Table 58. For example, for low-income developing countries (excluding China and India) manufactures represented about 30 percent of exports and 56 percent of imports in 1982 while the shares of primary commodities were around 55 percent and 20 percent, respectively (Table 50). If a 1 percentage point increase in industrial country economic growth raises the price of primary commodities relative to manufactures by 2 percent for both exports and imports then (without taking account of the impact of industrial country growth on the relative price of fuels) this would imply an increase in this group’s terms of trade of the order of 7/10 of 1 percent. The effect on the terms of trade of other developing country groups would be even lower, because of the smaller difference between the share of manufactures in their exports and imports; for middle-income oil importers, this approach would suggest a terms of trade effect of only around ¼ of 1 percent.

It might be argued that these latter figures tend to underestimate the terms of trade elasticities because a substantial proportion of manufactured exports from many developing countries consists of processed primary products. Since the value added in processing for these goods is often relatively small compared with the value of the primary inputs, the world prices of such manufactured exports could be more strongly influenced by the prices of the primary inputs than are the prices of manufactures from industrial countries. Some evidence on the importance of this factor is provided by an econometric analysis of the elasticity of developing countries’ import and export unit values with respect to the world prices of various groups of goods (Table 59). The results suggest that the elasticity of export unit values with respect to the world price of manufactures is as important as the corresponding elasticity for import unit values. This in turn would seem to indicate that the effect on non-oil developing countries’ terms of trade of faster industrial country growth is substantially smaller than that suggested by the directly estimated elasticity of 1.7.13 Consequently, that estimate should perhaps be regarded as an upper limit of the likely true elasticity.14

Table 59.

Developing Countries: Long-Run Elasticities of Unit Values of Imports and Exports with Respect to Prices of Fuels, Non-Oil Primary Commodities, and Manufactures, First Half 1962 to Second Half 1979

article image

Source: The Fund’s World Trade Model. See Spencer (1984b).

Note: Estimated elasticities were derived from regressions for semiannual data on equations of the form: Δ log (import unit value) = a1 Δ log (price of fuels) + a2 Δ log (price of agricultural goods) + a2 Δ log (price of raw materials) + a4 Δ log (price of manufactures). Only the long-run elasticities are reported. ** and * represent significance at the 1 and 5 percent levels, respectively.

Unit values of industrial countries’ exports of agricultural goods (SITC 0 + 1).

Unit values of world trade in raw materials (SITC 2 + 4).

Unit values of industrial countries’ exports of manufactured goods (SITC 5 – 8).

World spot prices for non-oil primary commodities, weighted by the commodity composition of the exporting region.

Finally, a further econometric analysis covering the period 1967–84 finds that an appreciation of the real effective exchange rate of the U.S. dollar does not result in a significant deterioration in the terms of trade of either the group of non-oil developing countries or the non-fuel exporting countries.15 A similar investigation suggests that, on the basis of historical experience, the terms of trade of the fuel exporters among developing countries did improve when the U.S. dollar appreciated. However, this relationship probably reflects the attempts by certain oil exporting countries to maintain a dollar export price for oil unchanged for quite long periods. Consequently, this past experience may not be a good guide to the future relationship between the exchange rate of the U.S. dollar and the terms of trade of the fuel exporters.

Transmission of Economic Influences Through Volumes Changes

There is ample evidence that a faster growth of real incomes in industrial countries leads to a more rapid growth in their total imports, including those from developing countries. The magnitude of the response of developing country export volumes will depend, among other factors, on their commodity composition, the relative importance of the industrial country markets, the relative competitiveness of developing countries’ exports, and the price elasticity of industrial countries’ import demand.

The following discussion presents two sets of evidence on the relationship between industrial countries’ real incomes and developing countries’ export volumes, using the distinction between “marginal” and “average” elasticities.16 In the first set, the implied elasticities of changes in export volumes with respect to changes in industrial country GNP are estimated for broad groups of developing countries. These are marginal elasticities in the sense that they reflect the percentage point increase in the rate of growth in developing country export volumes that would result from a 1 percentage point increase in the rate of economic growth in the industrial world. These estimates do not take account of developing countries’ supply responses, nor of variations in the direction of trade. In the second set, for a sample of the non-fuel exporters among developing countries, the elasticity of the level of export volumes with respect to the level of real GNP in industrial countries is estimated.17 These are average (or conventional) elasticities in the sense that they reflect the percentage increase in the level of developing country export volumes that would result from a 1 percentage point increase in the level of real GNP in industrial countries. For each developing country, the latter index is weighted according to the geographical distribution of its exports. The estimated average elasticities also take account of supply responses to the extent that the effects of changes in developing countries’ relative competitiveness on export volumes are included.

The estimated marginal elasticities are shown in Table 58 and comparable estimates provided in other studies are shown in the first part of Table 60. Three broad conclusions can be drawn. First, the elasticity of changes in export volumes of non-oil developing countries (and of the broadly similar group of non-fuel exporters) with respect to changes in industrial country GNP appear to be within the range 1.3 to 2.3. A higher elasticity is used by Cline but this is based on an estimate for all industrial country imports (with the constant term adjusted to reflect the higher trend growth of developing country exports to industrial countries) and is not confirmed by other studies.18 Second, Goldstein and Khan’s results indicate that this elasticity increased between 1963–72 and 1973–80, perhaps because of the increasing share of manufactures in developing country exports.19 However, experiments with sub-periods during the 1970s and early 1980s (not reported in Table 58) provide no evidence that there has been any further increase in these elasticities over time. Third, the commodity composition of developing countries’ exports does seem to affect their responsiveness to changes in industrial country GNP. The exporters of manufactures recorded an elasticity of 2.4, substantially higher than the elasticity of 1.6 recorded for primary product exporters. The elasticity estimated for the service and remittance countries was even lower, perhaps because of the lesser importance of the industrial country markets for these countries’ exports.

Table 60.

Alternative Estimates of the Elasticities of Developing Countries’ Export Volumes with Respect to Real GNP in Industrial Countries and with Respect to Developing Countries’ Real Effective Exchange Rates

article image
Note: The estimates of the marginal elasticities were derived from economic investigations of the relationship between the rate of growth of developing countries’ export volumes and the rate of growth of industrial country GNP in the same year. The estimates of the average elasticities were derived from econometric investigations of the relationship between the level of export volumes and the level of industrial country GNP (both in terms of logarithms).

This is derived from an equation of the form gx = –3 + 3gind where gx is the growth rate of export volumes and gind the growth rate of industrial country GNP. This implies that the average elasticity of export volumes with respect to industrial country GNP is 2.0, when evaluated at an industrial country growth rate of 3 percent.

This is derived from an equation of the form gx = 0.9 + 1.3 gind. Therefore, at an industrial country growth rate of 3 percent, it is equivalent to an average elasticity of 1.6.

Based on a sample of 36 countries.

These are weighted averages of the elasticities estimated for analytical sub-groups, with 1980 exports used as weights.

These estimates are derived from regressions on reduced-form equations similar to those used by Bond (1985): Ln (XVOL) = ao + a1 Ln (WTDGNP) + a2 Ln (REER) + a3 Ln (REER-1) + a4t where XVOL is export volume from the respective groups of developing countries to the industrial countries, WTDGNP is industrial countries’ GNP weighted by their imports from each group, and REER is the average export-weighted real effective exchange rate of each group. Only the results for the long-term elasticity with respect to real effective exchange rates (i.e., a2 + a3) are shown in the table. These estimates are based on a sample of 32 countries.

The estimated average elasticities are shown in the second part of Table 60. In general, they are slightly higher than the estimates obtained for the marginal elasticities, but this probably reflects the use of weighted GNP indices for the industrial countries which more closely reflect the pattern of developing countries’ export trade. The results generally confirm the broad conclusions, stated above, on the elasticities of developing countries’ export volumes.20 In addition, the relatively high coefficients obtained for the elasticity of export volumes with respect to developing countries’ real effective exchange rates suggest that broad groups of developing countries, and not just individual developing countries, can expand their shares of industrial country markets by improving their competitiveness.

The effects of changes in industrial country GNP on the exports of the fuel exporters have been largely determined by the singular conditions in world oil markets. The efforts by the members of OPEC to maintain the U.S. dollar prices of their oil exports unchanged for fairly long periods caused them to act as residual suppliers of oil to importing countries for much of the period since 1973. Consequently, the volume of oil exports by these countries has been highly sensitive to fluctuations in economic activity in the industrial world—this is reflected in the high estimated elasticity of 3.9 (Table 58). Also, since the OPEC strategy has caused oil prices to be less affected by short-term, cyclical fluctuations in industrial country output, there has tended to be a negative relationship between changes in fuel exporters’ terms of trade and cyclical changes in industrial country GNP. However, the more recent emphasis of the major oil exporting countries on protecting their market shares rather than on maintaining a given price would suggest that changes in industrial country output in the future may have a smaller impact on these countries’ export volumes than in the past, but a larger positive impact on their export prices.

Finally, there are no strong reasons for expecting that a shift in the exchange rate of the U.S. dollar relative to the currencies of other industrial countries would, by itself, have a significant effect on the total volume of exports by all developing countries as a group. To be sure, if a significant number of developing countries chose to appreciate or depreciate their own currencies along with the U.S. dollar, their export volumes would clearly be affected, but such effects would be more appropriately regarded as resulting from these countries’ own exchange rate action. At a more disaggregated level, however, the shift in merchandise trade deficits among industrial countries that would result from a change in the exchange rate of the U.S. dollar against other major currencies could affect the export volumes of individual developing countries. For those countries that export a large proportion of their exports to the United States, a depreciation of the dollar could be expected to reduce U.S. demand for their exports by more than the increase in demand in other industrial countries.

Purchasing Power of Exports by Developing Countries

The combined effects of changes in the pace of growth in industrial countries on developing countries’ terms of trade and export volumes determine the overall effect on the purchasing power (in terms of imports) of their total export earnings. The results for the marginal elasticities suggest that a 1 percentage point increase in the rate of growth of real GNP in industrial countries would be associated with an increase of around 3½ percent in the purchasing power of exports by non-oil developing countries (Table 58). Because of the uncertainty concerning the estimates for the terms of trade effect, this estimate is probably toward the upper end of the range of likely overall effects. Using the lower estimates of their terms of trade elasticity (of around 0.3) derived from the effects of industrial country growth on commodity prices gives an estimated elasticity of the purchasing power of exports of around 2, which may be regarded as toward the lower end of the range of likely overall effects.

Grouping countries by the category of their dominant exports, the results indicate that a cyclical increase in the rate of economic growth in industrial countries would have the largest impact on export volumes from exporters of manufactures and the largest impact on the terms of trade of primary product exporters, but that the overall impact on the purchasing power of exports from the two groups would be similar (Table 58). However, the group of primary product exporters includes some countries that are substantial exporters of manufactures. Results for the alternative analytical category of major exporters of manufactures, which includes these countries, suggests that the purchasing power of their exports would improve by more than that of primary product exporters without significant manufactured exports. The purchasing power of exports from the service and remittance countries would increase at only about half the rate recorded for the two other groups. On a regional basis, developing countries in Asia appear to record the greatest increase in the purchasing power of exports as a result of faster industrial country growth, while developing countries in Europe record the smallest increase.

However, these estimates reflect the effects of a cyclical increase in industrial country growth. The earlier discussion has suggested that the longer-term impact on developing countries’ purchasing power may be lower, largely because the effect on these countries’ terms of trade of an increase in the trend rate of growth in industrial countries could be substantially smaller than the effect of a cyclical increase.

Transmission of Economic Influences Through Receipts from Services and Private Transfers

Discussions of the economic interdependence between developing and industrial countries have largely concentrated on merchandise trade flows. Nevertheless, earnings from services, migrants’ remittances, and other private transfers are also quite important. During the period 1982–84, earnings from services and net receipts of private transfers amounted to almost 30 percent of total foreign exchange earnings by the non-fuel exporters (Table 61). Within this group, the exporters of manufactures are the least dependent on receipts from services and private transfers—in fact, their reliance on merchandise exports has increased since the late 1960s. In contrast, merchandise exports account for a small and declining share of the total foreign exchange earnings of the service and remittance countries. Investment income earnings have increased markedly in importance for both the service and remittance countries and for the group of fuel exporters. This resulted from, respectively, the substantial accumulation of overseas assets by some oil exporting countries and the rapid expansion of the international financial services sector in a few of the service and remittance countries.

Table 61.

Developing Countries: Average Shares of Merchandise Exports, Services, Investment Income, and Private Transfers in Foreign Exchange Earnings, 1968–70 and 1982–84

(As a percentage of total foreign exchange earnings)

article image

Although little information is available on the geographic destination of developing countries’ service exports, a considerable share must be directed toward the industrial countries. In 1984, just under one third of total service earnings (other than from investment income) by the non-fuel exporters consisted of receipts from shipping and other transportation activities and one third were travel receipts, including those from tourism. The geographic distribution of the former category in particular was probably quite close to that of merchandise trade. Consequently, total demand for these services can be expected to be influenced by real incomes in industrial countries.

The change in the volume and price of services (other than investment income) resulting from a change in industrial country real GNP will depend on the income and price elasticities of demand for such services in industrial countries, the price elasticity of supply from the developing countries, and the price elasticity of supply from other sources, including from within the industrial countries. While it is not possible to estimate these effects directly, some information is available on the reduced-form relationship between the growth of the purchasing power of these service exports and the growth of industrial country real GNP. Econometric estimates suggest that a 1 percentage point increase in the real GNP of industrial countries is associated with an increase of 1.4 percent in the purchasing power of exports of services by the non-fuel exporters (Table 62). This is substantially lower than the corresponding elasticity estimated for the purchasing power of this group’s merchandise exports. As might be expected, the purchasing power of service exports from the service and remittance countries appear to be more sensitive to changes in industrial country GNP than are those of developing countries that rely mainly on merchandise trade. On a regional basis, changes in industrial country growth rates appear to have the largest impact on the purchasing power of service exports from developing countries in Europe and in the Western Hemisphere, whereas the effect on countries in Africa, Asia, and the Middle East are relatively low and insignificant.

Table 62.

Developing Countries: Estimated Elasticities of the Purchasing Power of Receipts from Services and Private Transfers with Respect to Industrial Countries’ Real GNP, 1968–84

article image
Note: Elasticities were estimated from ordinary least square regressions on equations of the forms: Δ⁢log (exports⁢of⁢servicesunit⁢value⁢of⁢imports)=a1+a1⁢Δ⁢log (industrial⁢country⁢GNP)andΔ⁢log (net⁢private⁢transfer⁢receiptsunit⁢value⁢of⁢imports)=α0+α1⁢Δ⁢log (industrial⁢country⁢GNP) ** and * indicate significance at the 5 and 1 percent levels, respectively.

Services exclude investment income earnings.

Net private transfers; consequently, equations could not be estimated for those groups of countries that recorded a net outflow of private transfers.

Weighted average of the estimated elasticities of the sub-groups, with 1980 service earnings or private transfers used as weights.

Estimated over the period 1973–84.

Excluding People’s Republic of China.

The major component of private transfer receipts by non-fuel exporting developing countries consists of migrants’ remittances. These countries have exported labor to the industrial countries as well as to the major oil exporting countries. The remittances of these workers back to the home country have been strongly influenced by the level of economic activity in the host country. Estimates by Swamy (1981) of the elasticity of remittance inflows with respect to various measures of cyclical economic activity (either GDP, government expenditures, or exports) in host countries ranged from under 1 to over 3.21 Rough estimates of the elasticity of the purchasing power of developing countries’ net private transfers with respect to industrial country GNP are given in Table 62.22 Although none of these estimated elasticities are statistically significant, they do suggest that, as for service receipts, it is transfer receipts by developing countries in Europe and the Western Hemisphere that are the most sensitive to fluctuations in economic growth in the industrial countries. The lower sensitivity of transfer receipts by other regional groups probably reflects their greater dependence on migrants’ remittances from oil exporting countries.

Effects of Protectionism in Industrial Countries

The stance of trade policies in the industrial countries, which can have significant consequences for developing countries’ export expansion, is influenced by overall macroeconomic conditions in the industrial world. Pressures to impose protectionist trade measures generally arise when individual sectors experience a loss of comparative advantage and when rigidities in labor and other markets hamper rapid structural adjustment. In such circumstances, producer interests in import-sensitive sectors of the economy often seek government action to insulate those sectors from import competition in order to protect employment and wage levels and profit margins. Although rising pressures for protection need not translate themselves into more restrictions if governments firmly refuse to accede to these demands, in practice governments have found it difficult to do so during periods of stagnating or declining economic activity and rising unemployment. Moreover, protectionist measures imposed during periods of rising unemployment often prove difficult to dismantle quickly when economic activity recovers. Consequently, there may be a tendency for a ratchet effect in the degree of protectionism, as the next cyclical downturn begins with a higher prevailing level of protectionism than did the previous downturn.

Such factors underlie the increase in protectionism in most industrial countries during recent years.23 This protectionism has largely taken the form of nontariff restrictions. A recent study by the OECD (1985) estimates that the proportion of manufactures subject to nontariff barriers in the major industrial countries rose from 20 percent in 1980 to 30 percent in 1983.24 One aspect of this drift toward protectionism that is of particular concern to developing countries is the increasing recourse to bilateral, sector-specific trade restraints. Such measures are designed to restrict import competition from countries with a comparative cost advantage and are often imposed on products with relatively labor-intensive production techniques—for which developing countries are more likely to have such a comparative advantage. Nontariff barriers against the agricultural exports of developing countries are even more prevalent than those against their manufactured exports. The World Bank estimates that, in 1983, 29 percent of developing countries’ agricultural exports to industrial countries were affected by non-tariff barriers compared with 18 percent for manufactured exports.25 Overall, around one fifth of industrial country imports from developing countries were subject to nontariff barriers compared with around one tenth of imports from within the industrial country group.26

Increased protectionism in industrial countries has a direct impact on developing countries’ export earnings by lowering the effective demand for their exports and thereby exerting downward pressure on prices and volumes. The final impact will depend on factors such as the elasticity of foreign demand for the goods concerned (in both the industrial countries and in other markets) and on domestic supply and demand conditions in the developing countries themselves. Several studies have investigated these factors. Klein and Su (1979) use the integrated econometric country models of Project LINK to estimate the worldwide effects of tariff increases (or corresponding quantitative restrictions when import equations do not depend significantly on relative prices) on the manufactured imports of 13 OECD countries. Their results suggest that a 20 percent increase in trade barriers, introduced in 1978, would have caused a cumulative deterioration in the trade balance of developing countries of about $28 billion in the first two years and a $6 billion decline in real GNP over the same period.27

Using partial equilibrium analysis, Kirmani et al. (1984) estimate the potential impact on the export volumes of ten developing countries of eliminating the tariff and nontariff barriers to trade in seven relatively highly protected agricultural or manufacturing sectors of industrial countries.28 Their results indicate that such a trade liberalization could generate an increase in the total export volumes of the sample countries in the order of 5 percent to 10 percent. (They do not estimate the terms of trade effects.) Some of the largest estimated gains would result from trade liberalization in the textiles and clothing industries, where trade restraints have been directed primarily at exports from developing countries and have tended to become more extensive and more restrictive in recent years. Indeed, a more recent study by the OECD (1985) concludes that the implementation of the third Multi-Fiber Arrangement, which came into effect in 1982, reduced the volume of textile and clothing imports from non-OECD sources by around a further 10 percent in 1982 and 1983.

However, these various estimates only capture the direct impact of protectionism on developing countries’ exports and economic growth. The longer-term dynamic effects, in terms of the lost opportunities for reaping economies of scale and the potential disincentives for investment in the export sectors, are likely to be even greater, albeit difficult to measure. Also, the spread of nontariff barriers, especially in the form of bilateral restrictions, reduces the price sensitivity of trade flows and can thereby increase the costs of balance of payments adjustment for developing countries. For instance, as industrial countries increase their trade barriers, the reduction in the real exchange rate required to achieve a given improvement in a developing country’s trade balance is likely to rise.

In a more general equilibrium framework, changes in the stance of protectionism in the industrial countries may also affect the volume and price of capital inflows into developing countries. An increase in protectionism that hampers developing countries’ export earnings may increase banks’ perceived risks of lending to these countries and lead them to tighten their credit rationing. van Wijnbergen (1985) also argues that an increase in protectionism is likely to raise world interest rates. He argues that higher protectionism is likely to lead to no ex ante improvement in the current account of industrial countries but to an ex ante deterioration in the current account of developing countries; the latter deterioration is largely due to the assumption that the demand for imported capital goods is price inelastic. These movements in the current account would imply that ex ante world investment increases relative to ex ante savings, necessitating higher real interest rates to restore the global balance. However, empirical evidence on the importance of this effect is not yet conclusive.

Developing Countries: Export Growth, Terms of Trade, and Real Income Growth

The changes in developing countries’ terms of trade and export volumes discussed in the previous sections can in turn affect output growth in developing countries through four broad channels. First, the level of aggregate demand may be affected and cause cyclical fluctuations in the degree of capacity utilization; second, changes in the ability to import scarce inputs may also affect capacity utilization; third, the level of investment may be affected by changes in real incomes and savings rates or by changes in the ability to import capital equipment; and fourth, changes in the rate of growth of the export sector may induce changes in the overall rate of growth through various production linkages. The first two of these channels can be thought of as representing the short-term or “cyclical” effects of export growth on output and the last two channels as representing the longer-term, trend effects.

A quantitative assessment of the impact through these various channels is not easy. The effects will vary substantially across developing countries according to their economic structure, including the relative importance of the foreign trade sector, and will also vary over time for individual countries, according to the degree of internal and external macroeconomic balance. In particular, it should be remembered that developing countries’ own policies will have a substantial influence on the degree to which their output growth is affected by changes in export earnings.

The initial impact on a country’s welfare of an increase in the purchasing power of its exports is likely to differ substantially according to whether the increase is due to a terms of trade improvement or an expansion in export volumes. The former represents a clear gain in real income, since an increased volume of imports could be obtained without any change in the domestic resources allocated to producing exports, whereas an increase in export volumes will require the diversion of resources to the export sector from sectors serving domestic expenditures, unless none of the resources were previously utilized. The welfare effect of such a resource shift will depend on the relative benefits accruing from increased foreign exchange earnings and the alternative production foregone.

In the following, a perspective is offered on the nature of these various interconnections and on the factors influencing their quantitative importance.

Effect on Aggregate Demand

Increased export earnings tend to raise aggregate demand. This can occur directly, as the increase in incomes from exports leads to a general rise in demand, and indirectly if higher exports lead to an improvement in the external position, and this encourages the adoption of more expansionary monetary and fiscal policies. The latter effect could occur if domestic financial policies reacted in response to various indicators of the external position—such as the current account or the debt and debt service ratios—which might be altered by higher exports. Conversely, developing countries may try to maintain their growth rates when export earnings fall by stimulating domestic demand and increasing their recourse to foreign borrowing. This was particularly the case in the years following the first large oil price increase, and led eventually to the need for subsequent harsher adjustment measures.

It is difficult to reach any precise a priori conclusions on the size of the overall change in aggregate demand that would result from a change in export growth. The effect would clearly tend to be larger the greater is the share of exports in a country’s total aggregate demand and the smaller are the propensities to import and to save out of additional income. One important determinant of the degree to which domestic financial policies react to changes in exports will be the extent to which access to international capital flows is constrained. If a country’s debt and debt service ratios are high, so that the possibilities for new external borrowing are restricted, the initial change in aggregate demand resulting from a change in exports is more likely to be magnified by a shift in financial policies than if there is no binding constraint on new borrowing. An additional important factor is the sector to which the export gains initially accrue. This factor is of particular significance for those oil exporting countries where the bulk of receipts from oil exports accrue directly to the government, since in this case the impact on domestic aggregate demand of changes in export receipts will depend almost entirely upon how government expenditures respond to the corresponding changes in fiscal revenues.

The degree to which an increase in nominal aggregate demand is, in turn, reflected in an expansion of real output rather than in increased prices will depend in part on the prevailing level of capacity utilization, the tightness of labor markets, and the manner in which price expectations are affected by changes in demand. Domestic prices are likely to adjust more slowly to changes in nominal aggregate demand when there is substantial unutilized capacity, when labor is in excess supply, and when the change in aggregate demand is unanticipated. In such circumstances, much of the increase in aggregate demand may be reflected in higher output. In contrast, when the degree of capacity utilization is high, labor markets are tight, and prices and wages respond rapidly to changes in aggregate demand, then the impact on real output of externally induced changes in nominal aggregate demand will tend to be small.

Quantitative estimates of these aggregate demand effects would require the formulation of a macroeconomic model for developing countries, including a specification of how domestic financial policies would react to shifts in the external current account. Elaboration of such a model is outside the scope of this paper, but some indicative results for a group of six Asian developing countries—India, Indonesia, Korea, Malaysia, Philippines, and Thailand—are provided by Schadler (1986). Her results suggest that a decline of 2 percentage points in industrial country output growth (say, from a 3 percent rate of growth to a 1 percent rate) that is sustained for two years might be associated with a decline of around 2 percentage points in the annual export volume growth of the six countries and with a deterioration in the terms of trade of around 2 percent in each year. These effects on export earnings imply elasticities with respect to industrial country GNP that are toward the lower end of the range discussed earlier. If financial policies in these countries are moderately restrictive, this decline in export receipts is estimated to result in a reduction in the rate of growth of these countries’ real GNP by almost 1 percentage point in the first year and by about 1.3 percentage points in the second year.29 A larger slowdown in real economic growth would occur if tighter financial policies were adopted to achieve an even more rapid adjustment of the current account.

Changes in the Capacity to Import

Many developing countries are heavily dependent on imports of capital and intermediate goods as inputs into production. Consequently, when the stance of exchange rate and other macroeconomic policies is such that it is necessary to restrict imports either through the exchange system or the trade system, an increase in the purchasing power of exports (or in the level of capital inflows) may increase domestic output by providing foreign exchange to purchase scarce imported inputs. Moreover, the lack of foreign exchange to purchase imported capital goods may directly constrain investment, and therefore long-term growth prospects; this aspect will be considered later.

The strength of the link between import availability and output will depend on the severity of the foreign exchange constraint, the relative openness of the economy, the degree of substitutability between imported inputs and domestic alternatives, and the extent to which a country’s policies permit shifts in relative prices to allocate available imports to the uses with the highest opportunity costs. Changes in the capacity to import will tend to have the largest impact on domestic production in the short term, when substitution possibilities are lowest.

Empirical estimates of the elasticity of output with respect to imports are complicated by the fact that a change in output will itself cause a change in demand for imports; consequently, the direction of causation involved in any direct estimates of the relationship between changes in output and changes in imports may be ambiguous. Nevertheless, some indication of the likely range of the elasticity may be obtained from three alternative estimates.

Leven and Roberts (1983) estimate the relationship between the growth of GDP in Latin American countries during the period 1956–80 and the change in import volumes in current and previous years. Their results suggest that a 1 percentage point rise in real imports leads to an increase in real GDP of about 0.5 percent, of which about one half takes place in the same year and one half over the next five years.30 This estimate is likely to represent an upper bound to the actual elasticity, because of the ambiguous direction of causation already mentioned and because the contribution to output growth of other factor inputs has been ignored. Alternative estimates incorporating the contributions to output growth of fixed capital and labor are derived by Bergsten et al. (1985) for selected Latin American countries. Their estimates for the period 1960–83 suggest that the elasticity of output with respect to the average of the current and previous years’ real imports varies from virtually zero for Argentina, which was not subject to significant foreign exchange constraints for much of the period, to around 0.14 for Brazil, 0.23 for Mexico, and 0.27 for Chile.31

Finally, broadly similar estimates were obtained for the Philippines by Chopra and Montiel (1986) using an entirely different approach. They incorporate a system of foreign exchange rationing of imported intermediate goods into a model in which aggregate demand policies can affect the domestic level of output only to the extent that they produce unanticipated changes in the price level, which cause the supplies of labor and domestic production to deviate from their long-term equilibrium level. In such a system, deviations in domestic output depend upon changes in the level of excess demand for imported intermediate inputs or upon unanticipated changes in either domestic money supply, foreign income, or foreign prices. Their results over the period 1959–84 suggest that a 1 percentage point increase in imports will lead to a 0.28 percent cyclical increase in real output if the increase in imports is anticipated and to a 0.18 percent increase in output if the increase in import availability is unanticipated.

These various results suggest that, for countries which are subject to significant foreign exchange rationing of some form, a 1 percentage point increase in imports is associated with an increase in output in the range of 0.14 to 0.28 percentage points.

Changes in Export Earnings and the Rate of Investment

Changes in export earnings can alter developing countries’ long-term growth prospects by affecting the rate of fixed investment. This can occur through several channels: (1) increased foreign exchange receipts from exports (or from capital inflows) may relax foreign exchange constraints on the importation of scarce capital goods; (2) a large share of the increased export earnings may accrue to central governments or to government-controlled agencies, which generally have high propensities to invest; and (3) if the increased export earnings are due to an improvement in the terms of trade, then the resulting gain in real incomes may change the desired pattern of future consumption and hence lead to changes in the rates of saving and investment.

A scarcity of foreign exchange to purchase imported capital goods would usually have the largest impact on investment in the less industrialized developing countries. These countries generally do not have a significant domestic capital goods industry and so have fewer possibilities for substitution between domestic and foreign goods in capital formation. When such countries are faced with severe foreign exchange constraints, higher export earnings may raise the volume and the speed of implementation of investment. Marquez (1985) estimates that the long-run elasticity of investment with respect to the purchasing power of exports (in terms of imports of capital goods) is around 1.25 for the group of non-oil developing countries.32

In many developing countries (and especially the oil-exporting countries), government revenues are highly dependent on export earnings—both directly through various export taxes and other receipts and indirectly through tariffs on the additional imports that are made possible by higher export earnings. Consequently, an increase in export earnings raises government revenues, which in turn is likely to result in increased public investment. If the increase in export earnings is due to higher export volumes rather than higher prices, thereby requiring a shift of resources to the export sector from other sectors, then investment is likely to fall in those other sectors. However, this fall may not fully offset the increase in public investment especially when the effective rate of taxation of export earnings is high and a large proportion of any increase in government revenues is used to expand public investment.

Finally, it is difficult to determine on a priori grounds the size, or even the direction, of the effect that an improvement in the terms of trade has on the rate of investment, since it depends upon whether the terms of trade improvement is perceived as temporary or permanent and whether it is anticipated or not. In general, the more permanent a terms of trade improvement is considered to be, the more likely it is to generate a higher rate of investment, so as to make possible the desired higher consumption path. There is substantial empirical evidence that an improvement in the terms of trade strengthens a developing country’s external current account, which reflects an increase in aggregate saving relative to aggregate investment. Khan and Knight (1983) estimate that, for a group of non-oil developing countries, a 1 percentage point improvement in the terms of trade leads, on average, to an improvement of over ⅖ of 1 percentage point in the current account balance expressed as a ratio to exports. However, there is only limited evidence concerning how this improvement in the current account is reflected in changes in the rates of domestic savings and investment taken separately. Estimates by Fry (1986b) for a group of Asian developing countries over the period 1961–83 suggest that a 10 percent permanent improvement in the terms of trade leads to an increase in investment as a ratio to GDP of about ½ of 1 percentage point, but that a temporary improvement in the terms of trade has a relatively minor overall impact.33

The Production Linkage

The transfer of factor resources from the rest of the economy to the export sector may raise a developing country’s overall rate of growth because the latter sector is often the most productive and the most open to economies of scale. Faster export growth may also have beneficial effects on the non-export sectors by encouraging the introduction of better infrastructure and a more highly trained labor force and by promoting the spread of improved production and management techniques.

Most of the recent empirical evidence on the quantitative importance of these links between export performance and output growth for developing countries has been reviewed by Goldstein and Khan (1982).34 Their broad conclusions can be summarized as follows. First, cross-country studies suggest that a 1 percentage point faster growth in exports may be associated with a faster growth in real GDP of the order of 1/10 of 1 percent. Second, countries that export relatively more manufactured goods seem to be the ones in which the link between export performance and output growth is the strongest. Third, there is some indication that the relationship between export performance and output growth has grown stronger over time, probably because of the growing share of manufactures in developing country exports.

Transmission of Economic Influences Through Financial Markets

The transmission of economic influences from industrial to developing countries through the financial markets has become increasingly important, both because the external indebtedness of developing countries has increased and because a greater share of capital flows to developing countries has taken the form of bank lending at variable interest rates. The principal financial linkages operate through the level of international interest rates and the exchange rates of industrial countries, which affect debt service burdens, and through the availability of private international lending. These in turn affect the levels of domestic saving and investment and the macroeconomic policies of developing countries. Although financing constraints and externally induced changes in debt service burdens have affected almost all developing countries, the relative importance of these forces has varied widely across individual countries. Developing countries that are most heavily dependent on financing from commercial banks were the most affected by the rise in interest rates and the sharp decline in new bank lending. Accordingly, the analysis presented in this section emphasizes the capital importing developing countries, particularly the subgroups of market borrowers and countries with debt-servicing problems.

The analysis begins with a review of the increased international financial integration of developing countries. This provides the background for a discussion of how changes in international interest rates and in the level of financial flows to developing countries affect key macroeconomic variables in these countries, including domestic saving, investment, and economic growth. The section concludes with a discussion of the links between capital inflows, financial policies, and capital flight.

Increased International Financial Integration

The developing countries became much more dependent on private external financing for their economic development during the 1970s than they were before, as both the public sector and private residents began to borrow heavily in world capital markets. From the end of 1973 through the end of 1981 the total external debt of capital importing developing countries increased at an annual rate of 21 percent. Despite the decline in capital inflows since 1982, this debt rose to $878 billion in 1985 compared with $752 billion in 1982. In real terms—deflated by the export unit value index of these countries—the external debt increased at an annual rate of 8 percent between 1973 and 1985, while export volumes rose at a rate of only 4 percent.

The most notable change in the pattern of capital inflows during the last decade was a marked shift from official flows and private foreign direct investment to international bank lending. During the 1960s, the main form of international bank lending was short-term trade credit. During the 1970s, however, institutional developments in the domestic banking systems of the industrial countries lowered the risk on deposit liabilities of the money-center banks, which enabled the major banks to become the largest recipients of international loanable funds. Furthermore, financial innovations—notably the growth of syndicated loans and the increased use of cross-default clauses, together with the greater use of variable interest rate loans—reduced perceived levels of risk in lending to developing country borrowers, resulting in a significant rise in the volume of private bank lending.35

From the perspective of the borrowing country, the demand for bank loans increased rapidly because of sizable payments imbalances in the 1970s, due in part to the two waves of major oil price increases. The attractiveness of such loans stemmed from their flexibility, convenience, and the low real interest rates prevailing for much of the period. The real three-month London interbank offered rate (LIBOR) on U.S. dollar deposits, to which interest rates charged on many bank loans are linked, averaged only around ½ of 1 percent over the period 1974–78.36

Consequently, there was a marked shift from non-debt-creating flows—official transfers and private direct investment—to debt-creating and interest-sensitive borrowing in world capital markets (Chart 17). The contribution of non-debt-creating flows to the financing required for the current account deficits of capital importing developing countries declined from 61 percent in 1970 to 28 percent in 1981, while the share of bank financing rose from 54 percent to 74 percent.

Chart 17.
Chart 17.

Capital Importing Developing Countries: Sources of External Finance, 1973–85

(In percent of exports of goods and services)

1 Mainly from banks.2 Official transfers, SDR allocations, valuation adjustments, and gold monetization.

These developments greatly increased the sensitivity of developing countries to events in world financial markets. Most bank lending was at variable interest rates, and the share of capital importing developing countries’ total external debt that was subject to floating interest rates rose from an estimated 25 percent in 1973 to an estimated 52 percent in 1985. These interest payments were due irrespective of the uses to which the original borrowing had been put, whereas profit payments on foreign equity investments were more closely linked to the returns on the underlying investments.37 Also, new flows of bank lending to a developing country were more likely to be affected by sudden and uniform shifts in the perception of the country’s creditworthiness than were other forms of capital inflows.

Interest Rate and Exchange Rate Effects

The cost of developing countries’ borrowing in world capital markets is largely determined by the financial policies of the industrial countries, since there is a close relationship between domestic interest rates in industrial countries and the rate at which banks lend to developing countries. Furthermore, since a high proportion of developing country debt is denominated in U.S. dollars, an appreciation of the dollar increases the value of debt and debt service payments in terms of the currencies of other industrial countries.

Fiscal and monetary policies in the industrial countries influence interest rates on the world capital markets. For example, a combination of an expansionary fiscal policy and tight monetary policy causes a shift from bank deposits to government securities, which has a negative effect on bank liquidity. Since banks have balance sheet objectives, which they act to secure through asset and liability management, they respond by raising their deposit rates so as to attract the necessary deposits to fund their loan portfolio. The resulting higher cost of funds implies a general hardening of lending terms, including higher interest rates and shorter average maturities on new lending. These in turn affect the volume and cost of international finance to developing countries.

Following a period of relatively high inflation and low real interest rates in the industrial world during the 1970s, which facilitated the servicing of developing countries’ external debt, these countries have faced high interest rates in the 1980s. The disinflationary policies adopted by the industrial countries have entailed historically high nominal and real international interest rates and a consequent rise in borrowing costs (Chart 18). The shift in interest rates has been especially large when compared with changes in developing countries’ export prices. For example, the real interest rate for the capital importing developing countries, measured as the LIBOR on U.S. dollar deposits minus the annual percentage change in these countries’ export unit values was 18 percent in 1981–82, in sharp contrast to the negative real rate of 14 percent in 1973–77 (Table 63). Although this measure of the real interest rate declined during 1983–85, it is still very high by historical standards. Fluctuations in these interest rates were even more substantial for the subgroups of countries with debt-servicing problems and for the fifteen heavily indebted developing countries.

Chart 18.
Chart 18.

Capital Importing Developing Countries: Interest Rates, 1963–85

(In percent per annum)

1 The real interest is measured as the London interbank offered rate on three-month U.S. dollar deposits minus annual percentage change in export unit values in dollars.
Table 63.

Capital Importing Developing Countries: Real Interest Rates and External Borrowing, 1973–85

(In percent unless otherwise indicated)

article image
Note: For classification of countries in groups shown here, see the Appendix.

Annual average of LIBOR on U.S. dollar deposits minus annual percentage change in export unit values (in terms of U.S. dollars).

Annual average. Residual. Except for minor discrepancies in coverage, amounts shown reflect almost exclusively net external borrowing from private creditors.

Annual averages of ratio of year-end debt to exports of goods and services. Does not include debt owed to the Fund.

The cost of international bank lending to developing countries is based on a formula, with the costs of funds to the borrower comprising the market interest rate—generally represented by the LIBOR or the U.S. prime rate—plus a margin or spread. The interest rate fluctuates over the maturity of the loan, but the spread is usually fixed. These spreads over LIBOR vary according to the perceived risks associated with lending to a particular country—although generally within a relatively narrow range—and also vary substantially with the degree of liquidity in international capital markets.

The LIBOR itself can be regarded as consisting of a “risk-free” interest rate, which could be represented by the U.S. treasury bill rate and an additional margin or banking spread which reflects both the costs of bank intermediation and the risk premiums that banks have to pay in funding their own lending. This banking spread has varied considerably over time and is significantly influenced by financial developments in industrial countries. Since the probability of a developing country running into debt service difficulties is increased with higher interest rates, the banking spread tends to rise to reflect this increased risk. This relationship between the LIBOR-Treasury bill spread and the level of the Treasury bill rate (T-bill rate) is shown in Chart 19. The difference between the LIBOR and the Treasury bill rate is plotted on the vertical axis and the level of the Treasury bill rate on the horizontal axis.38 The positive correlation between the two is apparent from the clustering of the observations in the upper right-hand and lower left-hand quadrants. The correlation coefficient between the two series is 0.67. Dornbusch (1985) has estimated that a 4 percentage point increase in the level of the T-bill rate raises the spread between the LIBOR and T-bill rates by a full percentage point. Therefore, financial policies in industrial countries may influence the borrowing costs of developing countries not only through the level of domestic T-bill rates but also through the level of the banking spread, which tends to rise along with T-bill rates.39

Chart 19.
Chart 19.

Capital Importing Developing Countries: Relationship Between Treasury Bill Rate and Banking Spread, First Quarter 1975–Third Quarter 1985

(In percent)

1 Three-month bill rate.

Given the large amount of the debt owed to private creditors at variable interest rates, a rise in real international interest rates has a significant effect on the interest payments ratio—interest payments divided by exports of goods and services. Furthermore, the size of this effect varies considerably across the subgroups of developing countries classified by financial criteria because of considerable differences in both the ratio of external debt to exports of goods and services and the share of external debt owed to private creditors. For the capital importing developing countries as a group, it is estimated that, on the basis of 1985 exports and end-1985 external debt, a 1 percentage point rise in interest rates would directly increase the interest payments ratio by about 1 percentage point. However, for the market borrowers and for the fifteen heavily indebted countries, the increase in this ratio would be larger, at approximately 1¼ and 2¼ percentage points, respectively. In contrast, the rise in the interest payments ratio for the official borrowers (excluding China and India) would be only about ½ of 1 percentage point, reflecting the relatively small proportion of debt owed to private creditors. The impact of the interest rate increase on the interest payments ratio also differs sharply between the subgroups of countries with debt-servicing problems and those without these problems, with increases of approximately 2 percentage points and ½ of 1 percentage point, respectively. In addition to the short-term, direct effect of higher interest rates on debt service ratios, there would also be a further, longer-term impact as debt incurred at fixed interest rates matured, and was replaced by new borrowing at the higher rates.

Moreover, when nominal interest rates move with inflation there are important short-run effects on developing countries’ debt service burdens, even though there is no change in the real interest rate. When nominal interest rates rise in line with inflation because of floating rate loans, the higher interest payments include a component to compensate the lender for the erosion of the real value of loans. Therefore, higher inflation rates produce larger real debt repayments in the near future and lower real debt repayments near the end of the loan repayment schedule. In this way, higher nominal interest rates can result in debt-servicing problems, particularly if new lending is limited.

A depreciation of the U.S. dollar tends to reduce both the debt service and external debt ratios of developing countries. At the end of 1985, approximately 80 percent of the external debt of the capital importing countries was denominated in dollars. Since this is larger than the share of U.S. production and consumption in world markets for those goods and services exported by developing countries, a depreciation of the dollar raises the dollar value of developing country exports by more than the dollar value of their external debt. For example, calculations based on the 1985 values for exports of goods and services and external debt indicate that a 10 percent depreciation of the U.S. dollar would reduce the debt ratio of the capital importing developing countries from 162 percent to 155 percent.40

Rationing International Credit

Commercial banks make decisions not only regarding the distribution of their assets and interest rates but also on which customers to make loans to and to what extent. There has been considerable discussion in recent years about the relationship between credit rationing and default risk. Credit rationing exists when banks are unwilling to lend to a borrower even when the latter is willing to pay a higher interest rate. The availability of funds to developing countries is determined to a large extent by the judgments of commercial banks regarding the creditworthiness of countries. The major determinants of creditworthiness include the borrowers’ export performance, soundness of domestic financial policies, rate of economic growth, international reserve holdings, debt service ratio, and the existing exposure of banks. Consequently, when export earnings decline because of recession in the industrial world or when international interest rates rise, the actual and perceived capacity of developing countries to service their debt is diminished. As a result, the impact of such developments on these countries’ growth prospects can be magnified by a sudden reduction in the availability of external financing.

Several explanations have been offered to explain the phenomenon of credit rationing. It has been observed that a bank’s profit rate could actually decline if interest rates charged to borrowers were raised from “adverse selection effects” as those borrowers who are discouraged are likely to be those who intended to invest in relatively safe projects.41 The lower loan profitability might also be caused by the “incentive effects” of higher interest rates; at higher interest rates, borrowers choose riskier projects. In these circumstances, the resulting equilibrium may be one in which there is an excess demand for loanable funds at a particular interest rate, but banks do not have an incentive to raise the interest rate to eliminate the excess demand. In attempting to maximize their profits, banks ration credit to particular borrowers because the gains from higher interest rates are more than offset by the increase in the expected loss.

Credit rationing will also exist when there are doubts regarding the capacity or willingness of a borrowing country to meet its future debt service payments.42 Although a country that defaults on its debt incurs costs in terms of exclusion from new lending and interruption of trade flows, the threat of default may still limit the amount that any country can be lent since at some level of debt the benefits of default—the real income advantage of not having to service the debt—will outweigh the costs of default—the inability to secure new credit to reduce the variability of consumption or to finance profitable investment projects. An empirical analysis along these lines by Eaton and Gersovitz (1981) estimated that around 80 percent of the developing countries covered in their sample had experienced credit rationing.

Recent trends in net borrowing from private creditors by major groups of developing countries illustrate the changing perceptions of the creditworthiness of developing countries and the consequent availability of funds. The size of private foreign capital inflows expressed as a ratio to developing countries’ GDP or to their imports of goods and services has varied considerably over time and across groups (Table 64). Although these overall trends mask some notable contrasts among individual developing countries, several interesting observations can be made from the aggregate data. For the capital importing developing countries, net external borrowing from private creditors rose from an average of 2 percent of GDP in 1973–77 to almost 3 percent in 1981–82 before declining to less than 1 percent in 1983–85.43 Expressed as a ratio of imports of goods and services, net external borrowing from private creditors was 9 percent in 1973–77, 11 percent in 1978–80, but less than 3 percent in 1983–85. The impact of the cutback in private capital flows on the subgroup of countries with debt-servicing problems and the fifteen heavily indebted developing countries has been particularly striking. For countries with debt-servicing problems, net external borrowing from private creditors as a ratio to GDP averaged 4 percent in 1978–80 but declined sharply to minus 1/10 of 1 percent in 1983–85; for the fifteen heavily indebted countries, the decline in the ratio was from 5 percent to 1/10 of 1 percent. In contrast, for countries without debt-servicing problems, this ratio averaged only 1.6 percent in 1978–80 and showed a slight decrease in 1983–85.

Table 64.

Capital Importing Developing Countries: Average Saving, Investment, and External Borrowing, 1973–85

(In percent of GDP)

article image
Note: For classification of countries in groups shown here, see Appendix 1.

Residual. GDP less private and government consumption.

Gross capital formation.

Including official transfers.

Residual. Except for minor discrepancies in coverage, amounts shown reflect almost exclusively net external borrowing from private creditors.

Estimates of net disbursements by official creditors (other than monetary institutions).