Increasing Export Diversification in Commodity-Exporting Countries
A Theoretical Analysis
  • 1 0000000404811396 Monetary Fund

This paper draws on the neoclassical theory of international trade to examine the limits of efficient export diversification in low-income, commodity-exporting countries. It demonstrates the fundamental importance of relative factor endowments among countries for determining the commodity composition of international trade flows under both certain and uncertain economic conditions. In recognition of the importance of international financial markets for risk-spreading and allocating resources efficiently under uncertainty, the paper emphasizes the importance of “open” economic policies towards international trade in financial assets as well as goods.


This paper draws on the neoclassical theory of international trade to examine the limits of efficient export diversification in low-income, commodity-exporting countries. It demonstrates the fundamental importance of relative factor endowments among countries for determining the commodity composition of international trade flows under both certain and uncertain economic conditions. In recognition of the importance of international financial markets for risk-spreading and allocating resources efficiently under uncertainty, the paper emphasizes the importance of “open” economic policies towards international trade in financial assets as well as goods.

I. Introduction

Most low-income developing countries rely heavily on exports of primary commodities for their foreign exchange earnings. International prices for primary commodities, however, tend to be highly variable, introducing considerable uncertainty about the export proceeds these countries earn from one year to the next. In recent years, against the background of the large and widespread decline of relative commodity prices in world markets since the mid-1980s, this uncertainty has rekindled interest in national policies and possible multilateral initiatives to increase export diversification in “commodity-dependent” low-income countries as a means of improving their external positions and, more generally, their overall economic performance and welfare.

That export diversification can contribute to reducing the instability of export earnings is widely accepted in principle. If a country can judiciously foster a particular composition of exports with the result that the variability of earnings from one subset of exports is wholly or largely offset by that from another subset of exports, then the country will tend to face less uncertainty in connection with its ability to finance a given, or higher, level of imports. Thus, policymakers in commodity-exporting developing countries have frequently focused on trade, industrial, and other structural policies for fostering greater exports of nontraditional goods, especially manufactures.

Attempting to achieve greater stability of export earnings through the adoption of specific policy measures, however, raises some fundamental economic questions. One important set of questions relates to the comparative advantage of commodity-dependent countries. Namely, how do factor endowments influence the range of practical possibilities for diversifying the exports of these countries? Another set of difficult questions concerns why the “invisible hand” of the private market might fail to promote greater export diversification without special government incentives to support trade or other sectoral policy objectives.

Consider, for example, the situation of many Sub-Saharan African countries. In most cases, these countries are relatively highly endowed with one or more natural resources, but considerably less so with physical and social infrastructure, private capital, and skilled labor. The scope for efficiently introducing the production and export of new goods in these countries, especially modern manufactures requiring adequate infrastructure and capital, experienced management, and a trained work force, may be very limited. In other words, while uncertainty about economic conditions surrounding international trade in primary commodities may pose important problems for many low-income countries, the comparative advantage of these countries may still predominantly lie in primary commodities, or closely related products, so long as national investment to increase the relative stock of productive capital—including human capital—remains low.

A number of economic studies have sought to measure differences in the extent of export diversification among countries and to relate these differences to aspects of economic performance, especially the stability of export earnings and economic growth. 1/ The findings of these studies support many of the concerns raised frequently about the development problems facing low-income countries. For example, they confirm that dependence on commodity exports tends to be associated with lower growth rates of exports and domestic output. At the same time, however, they frequently find that neither export concentration, nor instability of export earnings, per se, are widely associated with poor export performance or low economic growth. 2/

Few studies, however, have investigated the relationship between export diversification and comparative advantage. With the notable exception of a seminal paper by Brainard and Cooper (1970), much of the recent economic literature on uncertainty and international trade has focused on the narrow, albeit important, issue of the implications of completeness of markets for securities and other risk-spreading financial instruments for the basic theorems of the pure theory of international trade. But, as noted by Pomery (1984), this literature largely passes over specific consideration of the particular economic circumstances of commodity-exporting developing countries, including especially the often highly skewed patterns of relative factor endowments observed in these countries.

This paper draws on the neoclassical theory of international trade under both deterministic conditions and uncertainty to explore the limits for low-income countries of what is termed here “efficient export diversification,” emphasizing the implications of factor endowments for determining the factor content, as well as the expected commodity composition, of trade. 3/ Also, the paper also considers briefly the appropriate role of national and multilateral policies towards export diversification. In this regard, it examines especially whether administered trade or other structural policies for promoting export diversification are appropriate in the modern world of increasingly sophisticated and widespread financial markets for risk-spreading.

The remainder of the paper is divided into three sections. Section II sets forth a simple model of international trade in which the relationships among factor endowments, comparative advantage, and diversification of production and exports in a deterministic world may be analyzed. This framework is broadened in Section III to consider the implications of uncertainty, using the modern framework pioneered by Arrow (1964) and others for incorporating uncertainty and markets for trading risks into general equilibrium models. Finally, Section IV summarizes the main conclusions of the paper and emphasizes the importance of the adoption of favorable public policies in low-income commodity-exporting countries towards the development of domestic markets for risk-spreading and their integration with world financial markets.

II. Export Diversification in a Deterministic Model

The neoclassical theory of international trade presents a number of paradigms in which differences in demand and supply conditions between countries give rise to specialization and trade of different commodities and goods. A variant of the familiar Heckscher-Ohlin-Samuelson (HOS) model of comparative advantage is employed here to consider the issue of export diversification. Among the assumptions of this model is that no uncertainty exists about the state of the world, including about conditions influencing the production of primary commodities in countries that are abundantly endowed with natural resources. 4/

The basic HOS model explores the economic implications of a 2-good, 2-factor, 2-country world. Trade and its commodity composition are determined predominantly by differences in endowments of primary resources between countries. The well-known prediction of the model is that, given similar consumption preferences and access to the same variable-proportions technologies in all countries, a country will tend to export the commodity that employs most intensively in production the resource with which the country is relatively abundantly endowed. In order to consider the issue of export diversification, an analytical framework of higher dimension is necessary, namely, one admitting the possibility that a country may export more than a single commodity. Although a number of different dimensional orderings of goods, factors and countries could be specified, the analysis here employs an ordering in which the number of traded goods is greater than the number of factors, namely, a simple 3-good, 2-factor model of international trade. 5/

1. Production and trade with factor price equalization

Following Dixit and Norman (1980) and Helpman and Krugman (1985), the conditions for attaining a so-called integrated world equilibrium in the 3-good, 2-factor HOS model may be summarized as follows:

Pi - ci(r1/r2) (i = 1,2,3)

Σi aki(r1/r2) Xi - Vk (k = 1,2)

αi(P1, P2, P3) - PiXii PiXi) (i - 1,2,3)


Pi = price of good i;

Xi = total world output of good i;

rk = price of the services of primary factor k;

Vk = total world supply of factor k;

ci () = unit cost function for producing good i;

aki () = input of factor k required per unit output of good i;

αi () = share of good i in total expenditures. 6/

These conditions imply that, although resources are immobile between countries, international trade occurs in an equilibrium consistent with that which would be found in a borderless, fully integrated world economy. Thus, in addition to familiar marginal conditions for the optimization of economic welfare and efficiency, factor price equalization (FPE) between countries occurs. In the model, trade in goods substitutes for international factor mobility, allowing the existence of an equilibrium in which relative factor rewards are identical to those that would occur if factors, in addition to goods, were mobile between countries. More specifically, for a given set of all factor endowments in the world economy, the equilibrium conditions above can be used to derive the set of possible distributions of the factor endowments between trading countries such that the integrated world equilibrium is attained and, by implication, world output is maximized. Notably, because the same factor prices, as well as goods prices, obtain in all countries in the integrated equilibrium, the same production methods will also be optimal in every country.

The integrated equilibrium in the 3-good, 2-factor, 2-country model is illustrated in Figure 1. The figure depicts an Edgeworth box diagram in which the dimensions of the box are given by the total endowment of primary resources in the world economy. The hexagonal area, OaGHOBG’H’OA, within the box consists of the set of possible distributions of factor endowments between the two countries, A and B, such that trade in the three goods results in both factor price equalization and the maximization of total world output at a unique set of relative output prices. The points forming the boundary of the FPE-area are equilibrium points along the world contract curve for the services of the two factors, say (for the purposes here), physical and human capital (K) and natural resources (R). 7/ The slope of each segment of the boundary indicates the optimal factor intensity of production for each good in the integrated world equilibrium, whereas the length of each segment measures the total factor requirements necessary to produce each good in equilibrium. 8/ Thus, the figure depicts an equilibrium in which total world production of each good employs differing amounts of capital and natural resources, and in which the optimal proportions in which the two factors are combined to produce the three goods in both countries also differ appreciably. With regard to factor proportions, the figure indicates that the first good (corresponding to the line segment OAG) is produced by relatively natural resource-intensive means, the second good (line segment GH) by more capital-intensive means, and the third good (line segment HOB) by the most capital-intensive means.

The alternative endowment points within the FPE-area correspond to incomplete specialization of production in country A and country B. That is, they correspond to production by both countries of positive amounts of at least two of the three traded goods. Points along the boundary of the shaded area also correspond to factor price equilization, but at points along the boundary complete specialization in the production of only one good by one of the countries is likely to occur, especially when there is a considerable disparity between the distribution of the endowments between the two countries (such as a point along the segment OAG in the figure). Outside of the boundary, factor price equalization with balanced trade is not possible because the level of full-employment production of some goods is greater than that which would be supported by total world demand in an integrated world equilibrium.

The commodity patterns of production (and trade) corresponding to endowment points within the FPE-area are indeterminate. For example, given the distribution of endowments represented by point Q in Figure 1, full employment of resources in country A using the optimal production methods depicted for the integrated equilibrium might be accomplished via a number of different production plans, including the three plans illustrated in the figure: OAFQ, OAIJO, and OAI’Q. The indeterminacy of the commodity pattern of trade is apparent upon introducing the consumption point C. This point lies on the budget constraint determined by the endowment point Q and the equilibrium ratio of factor rewards (rR/rK). It also necessarily lies on the diagonal of the Edgeworth box, OAOB, because of the assumption that preferences for goods (and hence for factor services) are identical, and linear homogeneous, in both countries. Thus, the three traded goods are consumed by both countries in similar proportions to total disposable income. In country A, the commodity pattern of consumption is given by OASTC in Figure 1. Then, depending upon the alternative production plans assumed, the commodity patterns of trade in Table 1 will be observed.

Table 1.

Country A: Possible Commodity Patterns of Foreign Trade

article image
Source: Figure 1. For each good, the direction of trade is given by the sign of the length of the line segment reported in parentheses.

Table 1 illustrates that in the 3-good, 2-factor version of the HOS model, the commodity composition of production and trade may take alternative, equally efficient, forms. Nonetheless, an important element of the familiar “2x2” HOS model is not contradicted by the equilibrium depicted in Figure 1. Specifically, the so-called net factor content of trade (Vanek, 1968), which refers to the embodiment of factor services in a country’s consumption versus production of goods, is invariant to the commodity pattern of production and trade. Essentially, this element, which is depicted in Figure 1 by the difference (QC) between the vectors OAQ and OAC, indicates the net exchange, through trade, of the services of natural resources for those of capital. Thus, in terms of the factor content of trade, the Law of Comparative Advantage is preserved in the “3x2” version of the HOS model.

2. Export diversification

Country A is assumed to be a low-income, resource-abundant country that predominantly produces and exports resource-intensive goods within an integrated world equilibrium that determines unique relative output prices, relative factor rewards, and the optimal methods for producing three goods. The low income level of the country is represented by the assumption that the country is endowed with only a small share of the world’s total supply of capital and natural resources. Finally, the “dependence” of the country on trade in primary commodities is represented by the assumption that the country’s endowment of resources is strongly “skewed” towards natural resources. Specifically, the country’s abundance of natural resources, relative to capital, is assumed to be greater than the relative factor intensities required in equilibrium to produce all but the third (most resource-intensive) good. 9/

Based on the conditions of the integrated equilibrium, the locus of efficient possibilities for the diversification of output in the resource-abundant country can be defined. Specifically, the “factor exhaustion” conditions of the integrated equilibrium can be employed to determine the implications for country A of undertaking a deliberate policy to promote greater production and exports of the more capital-intensive goods (i.e., goods 2 and 3) using “neutral” fiscal measures such as lump-sum taxes and subsidies. 10/

The factor exhaustion conditions for the commodity-dependent country can be expressed as


where A is a 2x3 matrix of equilibrium requirements for capital and natural resources per unit of output in the three goods sectors, q is a 3x1 vector of the quantity of goods produced in each sector, and v is a 2x1 vector of the country’s capital and natural resource endowments. There are fewer factor exhaustion conditions than unknowns (the elements of the output vector q). Nonetheless, the implications of promoting greater output of the more capital-intensive goods, without violating the FPE conditions, can be determined. Specifically, the linear system in equation (1) can be solved for equilibrium changes—that is, changes along an efficient diversification frontier—in the output of, say, goods 1 and 2 for similar changes in the output of the most capital-intensive good, good 3:



D = a1Ka2R - a1Ra2K < 0;

where aiK and aiR are the integrated equilibrium requirements for capital and natural resource inputs, respectively, per unit of output in each sector i, and where K and R denote the endowments of capital and natural resources in country A. The signs of the changes in output along the efficient diversification frontier are based explicitly on the assumption that country A is strongly abundant in natural resources relative to capital, as depicted in Figure 2. In algebraic terms, the inequality signs in equation (2) are derived from the following assumptions about the relative factor endowments of country A (i.e., K/R) versus the optimal relative factor requirements in the three output sectors:


As equation (2) indicates, along the efficient diversification frontier expansion of output in the third sector can only be achieved if output contracts in the second sector and expands in the first sector. 11/ In other words, greater output of more capital-intensive goods in country A can only be achieved if, simultaneously, output of the most resource-intensive good expands. This outcome demonstrates that efficient diversification is something of a zero-sum game. To preserve factor price equalization and the other elements of the integrated world equilibrium, the net factor content of trade between country A and the rest of the world must be maintained, namely, at the level indicated by the difference in Figure 2 between the endowment vector OAQ and the consumption vector OAC. 12/

Greater output diversification, however, does not necessarily imply greater export diversification. Indeed, along the efficient diversification frontier of country A, few points are likely to correspond to an appreciable level of exports of either of the two more capital-intensive goods. 13/ This occurs because all countries consume the three goods in similar proportions to their incomes. Thus, especially in low-income resource-abundant countries, the increased production of more capital-intensive goods is likely to be consumed for the most part domestically. Moreover, given the requisite expansion of production of the natural resource-intensive good along the efficient diversification frontier, the extent of export diversification, as measured by export concentration ratios or other indicies of product specialization, is unlikely to be substantially increased.

These considerations suggest that in a low-income country, such as country A, an export diversification program will “succeed” only if the domestic production plan occurs outside of the bounds of the integrated world equilibrium. This point is illustrated in Figure 3. Although a production plan such as OAWQ is optimal from the perspective of maximizing domestic output and economic welfare, it does not yield an appreciable level of exports of either of the two relatively capital-intensive goods, goods 2 and 3. Thus, an “active” policy to increase export diversification significantly would seek to expand the total production of, say, good 3 beyond OAW. 14/ As illustrated in Figure 3, however, production at a level such as OAV would result in unemployed natural resources at the integrated equilibrium ratio of factor rewards. Thus, if full employment of natural resources as well as capital is to be maintained under the export diversification policy, relative factor rewards must adjust in country A, inducing the adoption of more natural resource-intensive technologies in all sectors of the country’s economy, including the relatively capital-intensive sector. Full employment of all primary resources would then occur in the country at a lower ratio of factor rewards, (rR/rK)’, than in the integrated world equilibrium and would involve a production plan such as OAV’Q.

In this equilibrium, the country achieves substantial exports of good 3 and lower exports of good 1, thus realizing the principal objective of the export diversification policy. Less desirable aspects of the new equilibrium are apparent, however. Whereas along the efficient diversification frontier only momentary fiscal measures were necessary to induce changes in production, the active export diversification policy requires sustained fiscal measures to maintain production of the targeted good (good 3) at a sufficiently high level to achieve appreciable exports. Moreover, the decline in relative factor rewards amounts to a deterioration in the terms of trade at which the country exchanges the services of its relatively abundant natural resources, and, in effect, this decline finances the subsidization of output in the third sector. This outcome also results in a lower level of economic welfare in country A; as illustrated in Figure 3, consumption falls to point C’ from its maximum under the conditions of factor price equalization at point C. 15/

Thus, an active trade policy to diversify exports in a low-income country with a sharply skewed relative endowment of natural resources can involve appreciable economic costs in terms of foregone output and economic welfare. The outcome of this policy is strikingly similar to the outcome of import substitution. Although an export diversification policy does not involve direct discrimination in consumption between domestic and foreign goods, both policies tend to reduce imports of goods of a particular factor intensity, causing the domestic relative reward of the factor used more intensively in the “protected” sector to rise, and economic output and welfare to fall.

III. Export Diversification under Uncertainty

The foregoing analysis of export diversification in relationship to factor endowments and comparative advantage was based on a deterministic model of production, consumption and trade. It fails, however, to account for an essential aspect of the issues raised in connection with export diversification, namely, the greater uncertainty alleged to surround economic conditions in world markets for primary commodities versus other goods, especially manufactures. 16/ In order to be able to draw more reliable conclusions about the appropriateness of trade and other policy measures to diversify exports, uncertainty needs to be incorporated formally into the HOS framework.

1. Uncertainty in international trade models

Uncertainty has been represented in different ways in economic models. In an early, and particularly insightful, analysis of the implications of uncertainty for the production and trade of developing countries, Brainard and Cooper (1970) associated uncertainty principally with fluctuations in international prices. Subsequent analyses attempted to treat uncertainty more explicitly, often within HOS-based models, as originating in natural phenomena, especially periodic changes in weather conditions influencing agricultural production, and to a lesser extent in possible stochastic demand conditions. 17/ Basic portfolio choice theory is employed in most of these early analyses to explain how risk-averse producers of primary commodities marginally reduce their output, causing resources to be released to alternate economic activities that face less uncertain conditions. In the simplest of these models, uncertainty tends to reduce the volume of trade between countries because the optimal response of producers, especially in natural resource-abundant countries, is to reduce the extent of their specialization in the production, and hence exports, of primary commodities and to increase output of goods that require intensive inputs of other primary factors of production, such as physical or human capital. In versions of these models that consider a wide array of possible traded goods, portfolio choice theory suggests the adoption of production plans that would offset variations in prices of one commodity against another, thereby lowering the uncertainty surrounding both total earnings from production and foreign exchange earnings from exports. These models, however, frequently fail to distinguish explictily the limitations that factor endowments and comparative advantage place on efficient production plans in an integrated world economy.

More recently, the very general (and elegant) approach of Debreu (1959) and Arrow (1964) to uncertainty in general equilibrium models has been incorporated into neoclassical models of international trade. 18/ While models based on this approach continue to associate uncertainty principally with unanticipated supply-side developments, these models are more comprehensive in their treatment of the elements of an integrated equilibrium for the world economy, including more explicit consideration of the relative endowments of primary factors of production across countries. They also focus attention on the efficiency of financial markets for assessing economic risks and directing resources to productive uses, both within and between countries.

2. Integrated world equilibrium under uncertainty

Following the Debreu-Arrow approach, uncertainty is incorporated here into the 3x2 HOS model by introducing the notion of different possible “states” of the world. Each state corresponds to a possible outcome of nature in which either demand or supply conditions in one or more trading countries are affected. Conceptually, the range of possible states, both favorable and unfavorable, is very large and corresponds to a wide diversity of possible patterns of production and trade between countries. Indeed, in some states the familiar precepts of the deterministic HOS model may be completely negated. For instance, the productivity of a vital resource in a relatively resource-abundant country may be so impaired in a particular realized state that the country must export manufactures and import food or raw materials. Thus, in addition to the deterministic influence of natural resource and capital inputs to production, stochastic factors are assumed to influence ex ante decisions to allocate resources to the production of different traded goods, yielding an integrated world equilibrium in which risks are incorporated into economic decisions before the state of nature is made known to economic agents. Within this framework, extensive markets for securities, as well as goods, are assumed to exist, enabling individuals and firms to trade in “risks” in addition to goods. Firms supply financial markets with securities consisting of their expected profits for alternative states of the world in exchange for financing (to hire factor inputs) from individuals who must allocate their wealth—ownership of primary resources and initial equity shares in firms—to maximize their expected net worth, subject to their individual assessments of the probability of the occurance of different states of the world and their attitudes towards risk-taking.

“Complete” financial markets provide an essential facility for investors, with either wide or narrow differences in preferences for risk-taking, to assess and assume the risks associated with each possible state of nature, and, in so doing, to establish market-clearing prices for financial securities. In effect, these markets determine the socially optimal extent of risk-taking in the world economy and, accordingly, allocate optimal amounts of resources to firms in each sector according to the “present value” of firms’ profits:


where β(α) is the price of an elementary security paying one unit of numeraire value if state of nature α occurs and zero otherwise; pi(α) is the spot price in state α of good i; Fi(Ki, Ri,α) is the output in state α of the representative firm using inputs of physical and human capital (Ki) and natural resource (Ri); and, as before, prices for the two primary factors of production are denoted by, respectively, rK and rR.

This framework involves a distinct phasing between assessing the social value of different production activities and contracting for factor inputs to production, on the one hand, and the actual outcome of the state of nature, production, consumption and trade, on the other hand. International financial markets enable firms to contract for factor services before the actual state of the world is revealed. Then, when the state of the world is revealed, “spot” prices for commodities and goods are determined by the optimal consumption and production plans for the state of the world realized. Although ex post spot prices and patterns of consumption and production will vary, sometimes considerably, from one state to another, possible adversities arising in different states will have been “offset” substantially by the insurance against uncertainties provided by the initial trading in risks among firms and individuals.

Because the determination of factor prices and the allocation of resources between sectors are not dependent upon the actual outcome of the alternative states of the world, the integrated world equilibrium under uncertainty can be defined analogous to the integrated world equilibrium in the deterministic HOS model. In this equilibrium, international factor price equalization will hold, and firms in the same goods-producing sector of different countries will employ primary factors of production in the same proportions. As before, the precise production plan of a given economy remains indeterminate in the 3-good, 2-factor model. Also as before, the integrated world equilibrium and its relationship to the relative endowments of primary factors of production between countries can be considered with reference to a simple Edgeworth box diagram, but one in which the boundaries of the FPE set and equilibrium relative factor prices themselves are determined by financial market expectations regarding the likelihood of different possible states of nature as well as by economic factors known with certainty.

The precise implications of uncertainty for the integrated world equilibrium are revealed better formally than diagramatically. The Stopler-Samuelson theorem, which, under deterministic conditions, states that a change in relative output prices will increase the reward of the factor used most intensively in the production of the good whose relative price rises, provides an appropriate analytical framework for illuminating the implications of uncertainty in the 3x2 HOS model, as well as other HOS-based models. 19/

The conditions for competitive allocation of resources in the integrated equilibrium dictate that within each country factor rewards must be the same for any two industries, say, the natural resource-intensive industry producing good 1 and the more capital-intensive industry producing good 2:


where, for each industry i (i - 1,2), ki is the equilibrium capital-natural resource intensity ratio; f’i(ki,α) is the marginal product of capital in state α; and f(ki,α)-kif’i(ki,α) is the marginal product of natural resources in state α. This system of equations can be differentiated totally and solved for dk1 and dk2 (so long as k1 and k2 are not equal). Upon substituting back into equations (5) and (6) and combining the resulting expressions for drR and drK, one obtains the generalized Stopler-Samuelson equation:



Z = [Σαβ(α)p2(α)f2(α)]2/rK2

and where it is understood that the differential on the right-hand side of equation (7) is computed for constant input ratios (and therefore for constant values of the marginal factor productivities). This version of the Stopler-Samuelson equation, which relates relative goods prices to relative factor prices under uncertainty, states that, under the maintained assumption that k2>k1, changes in state-dependent prices for either traded goods or securities that increase the present value of output in the second sector relative to the first sector will cause the relative reward to capital, the factor used more intensively in the second sector, to rise.

Equation (7) reveals that uncertainty surrounding production of, say, the natural resource-intensive good (good 1) will tend to reduce the present value of output in the sector producing the good, assuming investors are predominantly risk-averse. Thus, efficient securities and other financial markets will tend to direct greater resources to those firms producing goods that are less affected by uncertainty, raising the relative rewards of the factors used most intensively in producing the latter goods.

3. Implications for export diversification

The implications of this result for export diversification in low-income, natural resource-abundant countries are remarkably similar to those reported in Section II. Given that the uncertainty surrounding the profitability of producing primary commodities will tend to reduce the relative reward to natural resources, commodity-dependent countries will find that their economic welfare is lower than if no uncertainty existed, in particular, because the terms of trade at which they are able to exchange the services of natural resources for the services of physical and human capital are lower than elsewise.

The uncertainty surrounding production of natural resource-intensive goods will tend to encourage less specialization in such goods. As before, however, the extent of efficient diversification of output and exports in the integrated world equilibrium remains indeterminate. Efficiency considerations will nonetheless continue to place important bounds on diversification, especially in respect to the expected factor content of the country’s trade. In other words, the initial relative endowment of primary factors of production, together with the assumed preference of low-income countries for consuming a more “balanced” menu of factor services than available domestically, will continue to dictate the efficient possibilities for diversification of production and exports. Moreover, policy-induced or other attempts to exceed the boundaries of efficient diversification will result in the consequences outlined before. If low-income countries endowed predominantly with natural resources adopt policies to promote export diversification beyond its efficient bounds, they will experience a deterioration in their effective factor terms of trade—and, by implication, in their economic welfare—additional to that engendered by the uncertainty surrounding economic conditions in the natural resource-intensive sectors.

IV. Conclusion

Low-income countries in the world economy today are mainly commodity-exporting countries, typically specializing in the export of only a small number of primary commodities. As such, they tend to face highly variable international terms of trade, primarily because of year-to-year changes in natural, and sometimes political economy, factors that often importantly affect global production and trade in natural resource-based commodities and related goods. For policymakers in many of these countries the uncertainty that accompanies these external conditions raises serious economic problems. Traditionally, the instability of export earnings has been regarded as a particularly difficult problem, requiring periodic macroeconomic adjustments to adverse external developments when international reserves or the availability of foreign financing were inadequate to maintain desired levels of aggregate consumption. More recently, with persistent balance of payments problems and mounting international debt obligations in many low-income countries, uncertain export prospects continue to pose difficult problems for these countries.

Export diversification has frequently been recommended as a means of effectively stabilizing the export earnings of commodity-dependent countries. While it is intuitively appealing, this recommendation raises basic questions about the adequacy of the resource endowments of many low-income countries to support, on a competitive and efficient basis, expanded production and export of goods different from those they presently produce and trade abroad. In addition, it should be asked why greater export diversification does not occur naturally or with more alacrity in these economies, obviating the basis for proposals for the establishment of special national, or even multilateral, facilities to promote greater diversification. 20/ To investigate these issues, this paper has sought to examine the relationship between comparative advantage and export diversification in a simple Heckscher-Ohlin-Samuelson model of international trade that admits exports of more than one good by a country and incorporates uncertainty surrounding the production of natural resource-based goods.

Within the familiar framework of an integrated world equilibrium, the analysis reveals the fundamental importance of relative endowments of different primary factors of production between countries, with similar consumption preferences, for determining the commodity composition of trade flows, and, in particular, how international trade allows countries to exchange the services of internationally immobile factors of production according to the relative abundance of the primary factors of production between countries. The analysis also demonstrates that a frontier of efficient possibilities for diversifying production can be defined for a small country unable to exercise appreciable power over the determination of relative prices in world markets, providing some margin for public policy to encourage diversification.

This frontier, however, is importantly circumscribed by the requisite factor content of the country’s trade, given the country’s particular resource endowments, to maximize output and economic welfare. In particular, a low-income, commodity-exporting country that wishes to expand its output of more capital-intensive goods, must also typically expand its production of natural resource-intensive goods to preserve full employment of its domestic resources and to ensure that the optimal factor content of its trade is achieved. The requirements to diversify exports are more stringent yet, because the production of nontraditional goods must first exceed the domestic demand for these products. Thus, administered trade or other commercial policies to promote export diversification pose the danger of expanding output beyond the locus of efficient possibilities for diversification with the result, remarkably similar to the outcome of an import-substitution policy, that full employment of relatively abundant natural resources in the economy can only be maintained by reducing the relative factor reward of these resources and thereby reducing the economic welfare of the country in general.

This conclusion is not substantially altered by more explicit consideration of the importance of uncertainty surrounding economic conditions in world markets for primary commodities. Indeed, incorporating uncertainty into the basic Heckscher-Ohlin-Samuelson model, following the modern approach to representing uncertainty in general equilibrium models, demonstrates that uncertainty leads risk-averse investors and competitive producers to direct resources, at the margin, away from specialization in economic activities facing considerable risks, and this, in turn, affects the relative rewards of primary factors of production in an integrated world economy. Explicitly taking into account uncertainty, however, also points to the important role domestic and international financial markets can play in risk-spreading and allocating resources in the world economy.

A “complete” system of efficient international financial markets would be expected to assess economic uncertainties in the world with considerable foresight and rationality and, accordingly, to direct resources in countries to their most productive uses. Although financial markets in the world economy today fall short of their depiction in economic models, they are becoming increasingly complete in many countries, including advanced developing countries that have pursued relatively liberal economic policies. Moreover, financial liberalization has deepened in many industrial and other countries, and the financial markets themselves have responded to new opportunities to trade in marketable risks both within and across national borders. 21/

This experience points to the large discrepancy in policies today towards financial markets and international transactions typically observed in industrial and other advanced countries, versus low-income commodity exporting countries. Whereas the former countries have increasingly allowed financial markets to play a larger role in allocating resources in their economies, the latter countries have tended to continue to enforce highly restrictive policies towards foreign investment and other international financial transactions, and towards the establishment (or expansion) of domestic markets for risk-spreading. In this light, the establishment of special domestic or multilateral facilities to promote and finance greater export diversification in low-income commodity-exporting countries should be regarded as a second-best prescription for dealing with uncertainty and the related problems that surround the external balance and general economic condition of these countries. Greater benefit would be derived from adopting more liberal economic policies towards domestic and international financial markets, in particular, encouraging the integration of these countries with the world economy in markets for risk as well as traded goods.


  • Arrow, Kenneth J.,The Role of Securities in the Optimal Allocation of Risk-Bearing,” Review of Economic Studies. Vol. 31, No. 2 (April 1964), pp. 9196.

    • Search Google Scholar
    • Export Citation
  • Batra, Raveendra, The Pure Theory of International Trade under Uncertainty (New York: Halstead Press, 1975).

  • Behrman, Jere R.,Commodity Price Instability and Economic Goal Attainment in Developing Countries,” World Development. Vol. 5, No. 5 (May 1987), pp. 559574.

    • Search Google Scholar
    • Export Citation
  • Bond, Marian, and Elizabeth Milne,Export Diversification in Developing Countries: Recent Trends and Policy Impact,” Staff Studies for the World Economic Outlook (Washington, D.C.: International Monetary Fund, 1987).

    • Search Google Scholar
    • Export Citation
  • Brainard, William C., and Richard N. Cooper,Uncertainty and Diversification in International Trade,” Economic Growth Center Paper No. 145 (New Haven: Yale University, 1970).

    • Search Google Scholar
    • Export Citation
  • Debreu, Gerard, Theory of Value (New Haven: Yale University Press, 1959).

  • DeRosa, Dean A. (1990), “Commodity Dependence and Export Diversification: Some Empirical Evidence Bearing on Proposals to Diversify LDC Exports,” Eastern Africa Economic Review, forthcoming.

    • Search Google Scholar
    • Export Citation
  • Dixit, Avinash K., and Victor Norman, Theory of International Trade (Cambridge: Cambridge University Press, 1980).

  • Dumas, Bernard,The Theorems of International Trade under Generalized Uncertainty,” Journal of International Economics. Vol. 10, No. 4 (November 1980), pp. 481498.

    • Search Google Scholar
    • Export Citation
  • Eaton, Jonathan, and Gene M. Grossman,Tariffs as Insurance: Optimal Commercial Policy When Domestic Markets are Incomplete,” Canadian Journal of Economics. Vol. 18, No. 2 (May 1985), pp. 258272.

    • Search Google Scholar
    • Export Citation
  • Ethier, Wilfred J.,Higher Dimensional Issues in Trade Theory,” in Ronald W. Jones and Peter B. Kenen, eds., Handbook of International Economics (Amsterdam: Elsevier Science Publishers, 1984).

    • Search Google Scholar
    • Export Citation
  • Grinols, Earl L., Uncertainty and the Theory of International Trade (Chur, Switzerland: Harwood Academic Publishers, 1987).

  • Helpman, Elhanan, and Paul R. Krugman, Market Structure and Foreign Trade (Cambridge, Massachusetts: The MIT Press, 1986).

  • Helpman, Elhanan, and Assaf Razin, A Theory of International Trade under Uncertainty (New York: Academic Press, 1978).

  • International Monetary Fund, International Capital Markets: Developments and Prospects (Washington, D.C.: International Monetary Fund, May 1991).

    • Search Google Scholar
    • Export Citation
  • Jones, Ronald W.,A Three-Factor Model in Theory, Trade, and History,” in J.N. Bhagwati, R.W. Jones, et al., eds., Trade. Balance of Payments. and Growth (Amsterdam: North Holland, 1971).

    • Search Google Scholar
    • Export Citation
  • Kemp, Murray C., Three Topics in the Theory of International Trade: Distribution. Welfare and Uncertainty (Amsterdam: North Holland, 1976).

    • Search Google Scholar
    • Export Citation
  • Knudsen, Odin, and Andrew Parnes, Trade and Economic Development (Lexington, Massachusetts: D.C. Heath and Company, 1975).

  • Krueger, Anne O., Growth. Factor Market Distortions, and Patterns of Trade among Many Countries. Princeton Studies in International Finance, No. 40 (Princeton, New Jersey: Princeton University Press, 1977).

    • Search Google Scholar
    • Export Citation
  • Lancaster, Kevin,The Heckscher-Ohlin Trade Model: A Geometric Treatment,” Economica. Vol. 24, No. 93 (February 1957), pp. 1939.

    • Search Google Scholar
    • Export Citation
  • Love, James,Concentration, Diversification and Earnings Instability: Some Evidence on Developing Countries’ Exports of Manufactures and Primary Products,” World Development. Vol. 11, No. 9 (September 1983), pp. 78793.

    • Search Google Scholar
    • Export Citation
  • Massell, Benton F.,Export Concentration and Fluctuations in Export Earnings: A Cross Section Analysis,” American Economic Review. Vol. 54, No. 1 (March 1964), pp. 4763.

    • Search Google Scholar
    • Export Citation
  • Melvin, James R.,Production and Trade with Two Factor and Three Goods,” American Economic Review. Vol. 58, No. 5 (December 1968), pp. 12491268.

    • Search Google Scholar
    • Export Citation
  • Newbery, David M. G., and Joseph E. Stiglitz, The Theory of Commodity Price Stabilization: A Study in the Economics of Risk (Oxford: Oxford University Press, 1981).

    • Search Google Scholar
    • Export Citation
  • Pomery, John,Uncertainty in Trade Models,” in Ronald W. Jones and Peter B. Kenen, eds., Handbook of International Economics (Amsterdam: Elsevier Science Publishers, 1984).

    • Search Google Scholar
    • Export Citation
  • Ruffin, Roy J. (1974a), “Comparative Advantage under Uncertainty,” Journal of International Economics. Vol. 4, No. 3 (August 1974), pp. 261273.

    • Search Google Scholar
    • Export Citation
  • Ruffin, Roy J. (1974b), “International Trade under Uncertainty,” Journal of International Economics. Vol. 4, No. 3 (August 1974), pp. 243259.

    • Search Google Scholar
    • Export Citation
  • Rybczynski, T. M.,Factor Econdowment and Relative Commodity Prices,” Economica. Vol. 22, No. 84 (November 1955), pp. 336341.

  • Satterthwaite, M. A.,On the Scope of Stockholder Unanimity Theorems,” International Economic Review. Vol. 22, No. 1 (February 1981), pp. 119133.

    • Search Google Scholar
    • Export Citation
  • Travis, William P., The Theory of Trade and Protection (Cambridge, Massachusetts: Harvard University Press, 1964).

  • United Nations Conference on Trade and Development (UNCTAD), “Recent Developments in Existing Compensatory Financing Facilities and Various Institutional Options,” TD/B/AC.43/8 (January 1989).

    • Search Google Scholar
    • Export Citation
  • Vanek, Jaroslav,The Factor Proportion Theory: The N-Factor Case,” Kyklos. Vol. 21, No. 4 (December 1968), pp. 749756.


The author wishes to acknowledge helpful comments received from John Karlik, Mohsin Khan, Jonathan Ostry, and Assaf Razin on an earlier version of the paper. The author, however, bears sole responsibility for any errors or omissions.


See, for example, Michaely (1962), Massell (1964), Love (1983), Bond and Milne (1987), and DeRosa (1990).


The factor content of trade refers to the embodiment of the services of different primary factors of production in goods traded between countries. Thus, for instance, the factor content of a country’s exports might be predominantly comprised of labor services, in which case the country would typically exchange these services for those of other factors—in relatively limited supply in the country—embodied in its imports.


The analysis is developed without extensive mathematics, relying instead upon diagramatic demonstration of basic concepts and, as necessary, appeal to the fundamental theorems of international trade theory. In addition to the seminal contributions of Heckscher, Ohlin and Samuelson discussed in standard economic textbooks on international trade, see Dixit and Norman (1980), Helpman and Krugman (1985), and Ethier (1984).


Regarding the relevance of this “3x2” model versus other “even” and “uneven” HOS-based models, see Ethier (1984). Closely related variants of the model employed here should be expected to yield similar analytical results. Among these variants, a particularly important one is the so-called specific factor model developed by Jones (1971) and applied extensively by Krueger (1977) and others. The specific factor model is sometimes viewed as providing a more apt description of the economic circumstances of commodity-exporting countries because, in specifying a wide range of factors by country, it is viewed as providing analytical assurance that countries can produce an equally wide variety of goods. As discussed in Dixit and Norman (1980) and Ethier (1984), however, concern for the ability of countries to produce a wide range of goods using a limited number of primary factors is often founded on some confusion between the partial versus general equilibrium determination of efficient possibilities for producing different goods in uneven HOS-based models. In the 3x2 version of the HOS model specified here, the integrated world equilibrium does not preclude efficient production of any good by any trading country, including relatively resource-abundant countries.


In addition, it is assumed that the two primary factors of production are inelastic in supply, production occurs under conditions of (quasi-concave) constant returns to scale, preferences are well-behaved and homothetic, and perfect competition prevails.


The boundary is determined by application of the so-called theorem of corresponding points to separate Edgeworth box diagrams for the two countries, given their actual endowments and the existence of similar preferences in both countries. See, Lancaster (1957), Travis (1964) and Melvin (1968).


Note that the upper boundary of the shaded area is the reflection of the lower boundary, thus enabling the measurement of factor requirements and intensities to produce each good from either country origin, OA or OB.


This assumption is given an explicit algebraic representation further below in the text. In Figure 1, the assumption is represented by the fact that the slope of a line segment drawn from OA to Q is less than the slope of either OAH’ or GH, and greater than the slope of OAG. Notably, this assumption also implies that the location of the endowment distribution point Q is not so “skewed” as to lie outside of the FPE set.


Along the efficient diversification frontier, fiscal incentives may be viewed as “momentary” measures, financed by nondiscriminatory lump-sum taxes on all firms. These taxes are used to promote different patterns of final goods production in the economy. As discussed further below, however, “neutral” fiscal measures to promote output of particular goods beyond the efficient diversification frontier are less benign. In particular, because they must be maintained indefinitely, these measures will cause relative factor rewards (and prices) in the home country to differ from those in the integrated world economy, and thereby will cause the welfare of the home country to fall.


Analogous results are obtained if output is initially expanded in sector 2 rather than sector 3. Thus, along the efficient diversification frontier, expansion of production of either of the two more capital-intensive goods will require an accompanying expansion of good 1, the most resource-intensive good.


This result contrasts sharply with that resulting from the familiar alternative development strategy of seeking through time to expand the relative stock of both physical and human capital, either through domestic saving or net foreign borrowing and investment. For a given ratio of factor rewards (and a given level of output of one of the three goods), an expansion of the stock of capital in country A will result in a greater than proportional expansion of output of the more capital-intensive good (good 2 or good 3) and a decline in output of the more resource-intensive good (good 1 or good 2). This “magnification” effect, first derived by Rybczynski (1955), occurs because expansion of the relatively scarce factor, capital, enables the country to diversify its output with reference to the requisite change in its net factor content of trade. More specifically, the requisite factor content of the country’s trade changes with the evolution of the country’s effective ratio of factor endowments, and hence it becomes efficient for the country to expand its output of the more capital-intensive good.


In the absense of international capital flows, a country can export at most only two goods in the simple 3-good HOS model considered here.


At the same time, production of good 3 would not be expanded to OAY, because at OAY the country’s entire stock of capital would be devoted to producing good 3. Also at OAY, an excess supply of natural resources would emerge at the equilibrium ratio of factor rewards, rR/rK.


Broadly speaking, the distance C’C measures the efficiency loss in aggregate output owing to the “nondistortionary” fiscal measures used to implement the export diversification policy.


Throughout this paper, a maintained assumption is that, from a global perpective, greater uncertainty surrounds the production of primary commodities than other traded goods.


See, among others, Ruffin (1974a, 1974b), Batra (1975) and Kemp (1976).


See in particular, Helpman and Razin (1977), Dumas (1980), and Grinds (1987).


The discussion that follows is based on the extension of the Stopler-Samuelson theorem to the case of uncertainty presented by Dumas (1980). Notably, Dumas uses a 2-good, rather than 3-good, model of international trade. However, his results are applicable, without modification, to 2-factor models of higher orders of (goods) dimensionality. On this point, see Ethier (1984).


See, for instance, UNCTAD (1989).


See, for instance, International Monetary Fund (1991). These developments support the contention of Satterthwaite (1981), as discussed in Pomery (1984), that where unexploited gains from trade in risks exist, missing financial markets will be created by market-making arbitrage. In a related vein, it should be noted that some theoretical studies, e.g., Newbery and Stiglitz (1981) and Eaton and Grossman (1985), find that incomplete financial markets can result in equilibria in which unrestricted trade in goods reduces economic welfare. While these studies have spawned arguments in favor of limiting the integration of world goods markets, namely, as a form of social insurance, they do not consider the longer-run implications for international trade and economic welfare of expanding domestic and international financial markets in response to the market-making incentives referred to by Satterthwaite.