Journal Issue
Share
Article

An Econometric Study of Primary Commodity Exports from Developing Country Regions to the World

Author(s):
International Monetary Fund. Research Dept.
Published Date:
January 1987
Share
  • ShareShare
Show Summary Details

THE pressing need for highly indebted countries to accelerate the growth of their exports has drawn increased attention to the export performance of developing countries in recent years. This paper focuses on that performance with regard to primary commodities. Data on the flow of primary commodity (henceforth, for simplicity, “commodity”) exports from selected regional groups of non-oil developing countries 1 are presented and analyzed. Such data can be useful both for projecting exports and for formulating exchange rate and trade policies.

Foreign trade data point to substantial changes in the commodity structure of developing country exports over the past two decades. First, the data show a decline from 1965 through 1980 in the share of goods from all developing countries in the commodity imports of industrial countries. Second, the data show that over the same period the Asian countries were the most successful in maintaining their market shares. Both demand-side and supply-side reasons for these changes are suggested. On the demand side, the paper discusses the role played by commodity composition, proximity to markets, and industrial country policies. On the supply side, the paper examines factors such as relative prices, domestic resource use, population growth, and the local endowment of natural resources, as well as the influence of domestic policies.

The export data are then analyzed econometrically to distinguish the relative effects of the world economic environment (demand) and of the domestic environment (supply) on the volume of exports. The countries studied are categorized into five geographical regions to highlight the interregional differences in demand and supply elasticities. The estimated price and income elasticities thus obtained are compared with estimated elasticities for individual commodities and for groups of commodities obtained from other studies.

It was decided to base the study on groups of commodities and on regions, rather than on individual commodities and countries, to permit an analysis of broad economic trends while at the same time allowing enough aggregation so that differences in these trends among commodity groups and regions could be distinguished. With developing countries aggregated into regions, the supply equations for each commodity group can be estimated on the assumption that the commodity group is differentiated by its source of supply and that importing countries distinguish the imported commodity originating with a regional group from the commodity produced domestically. The supply of commodity exports from a particular region is also assumed to increase as the price of the exported commodity rises relative to domestic prices in the region.

The possibility of using the imperfect substitutes model arises from the aggregation of countries into regions. Whereas an individual country may be able to change its supply of exports of certain primary products without any price change taking place, groups of developing countries can change supply conditions only by evoking a change in the price. Thus for groups of countries the imperfect substitutes model is the appropriate one to use. The elasticity estimates resulting from this model provide important information on the extent to which specific types of exports of particular groups of countries respond to world growth, world prices, and domestic regional prices.

The plan of the paper is as follows. Section I looks at the trends of commodity exports for developing countries as a whole and for the five selected regions; such trends reflect structural developments that have taken place over the past two decades. Section II presents a model that incorporates both the demand and supply determinants of commodity exports. Section III discusses the results derived from the model. Section IV presents a survey of empirical results from other studies and compares them with the results from this study. Section V summarizes the conclusions.

I. Trends in Commodity Exports of Developing Countries

This section of the paper discusses the trends in developing country export trade that took place over the years 1965 through 1980.2 The first subsection describes the growth in trade of commodity exports from developing countries in comparison with trade of similar exports from industrial countries; it describes how market shares of developing and industrial countries adjusted to changed economic conditions during the period examined.3 The second subsection describes the growth in trade of commodity exports, distinguishing among regions of developing countries, identifying the relatively more successful regions, and discussing the reasons for the success of these regions. The third and concluding subsection analyzes the causes of the trends identified, from both demand and supply perspectives, with a focus on the respective roles of the international economic environment, industrial country policies, and domestic policies of the developing countries.4

Trends in Developing and Industrial Country Exports

Trade in commodity exports, especially in agricultural commodities, is declining in relative importance for developing countries.5 Indeed, from 1965 through 1980 developing countries lost market shares in world commodity imports. During this period commodity export earnings (in U.S. dollars) for industrial countries grew by 15.0 percent, compared with 11.7 percent for developing countries (Table 1). To some extent this difference might be explained by the fact that over the same period manufactured exports from developing countries grew faster (18.6 percent a year) than did such exports from the industrial countries (16.9 percent a year) (Table 1). These figures reflect the growing trend in developing countries of exporting processed food and raw materials formerly exported without processing. Despite this gain in exports of manufactures from developing countries, total exports (commodities plus manufactures) of the industrial countries grew at an annual average rate that was 1.5 percent faster than the growth rate of the developing countries’ exports (Table 1).

Table 1.Exports of Commodity Groups from Developing Country Regions, Industrial Countries, and OPEC to the World, 1965 and 1980(In billions of U.S. dollars)
FoodBeverages and TobaccoAgri-cultural Raw MaterialsMineralsEnergyAll Commod-itiesManu-facturesTotal
Region1965198019651980196519801965198019651980196519801965198019651980
Africa1.123.790.722.020.861.820.984.180.025.253.7017.062.2323.555.9340.61
(8.5)(7.1)(5.1)(10.1)(45.0)(10.7)(17.0)(13.7)
Asia1.7010.670.611801.858.530.714.540.077.424.9432.952.9560.307.8993.20
(13.0)(7.5)(10.7)(13.2)(36.5)(13.5)(22.3)(17.9)
Europe0.593.360.240.520.321.550.150.950.000.001.316.392.7425.104.0431.49
(12.3)(5.3)(11.1)(13.1)(0.0)(11.1)(15.9)(14.6)
Middle East0.290.690.000.000.450.690.030.310.032.360.804.060.698.751.4912.80
(5.9)(0.0)(2.9)(16.8)(33.8)(11.4)(18.4)(15.4)
Western2.6713.121.627.741.032.691.366.650.162.256.8332.442.0319.878.8752.31
Hemisphere(11.2)(10.9)(6.6)(11.2)(19.3)(10.9)(16.4)(12.5)
Total6.3631.633.1912.084.5115.283.2316.630.2817.2817.5892.9010.64137.5728.21230.47
developing

countries
(11.3)(9.3)(8.5)(11.5)(31.6)(11.7)(18.6)(15.0)
Industrial13.97104.940.583.476.7235.195.1339.831.4342.3827.83225.8296.18995.62124.001221.44
countries(28.2)(12.7)(11.7)(14.6)(25.3)(15.0)(16.9)(16.5)
OPECa0.560.980.180.900.193.130.251.657.76186.758.94193.402.4218.6311.36212.03
(3.8)(11.3)(20.5)(13.4)(23.6)(22.7)(14.6)(21.5)
World20.89137.553.9516.4511.4253.608.6158.119.47246.4154.35512.12109.241151.82163.591663.95
average(13.4)(9.9)(10.8)(13.6)(24.3)(16.1)(17.0)(16.7)
Source: World Bank Trade System data base.Note: Numbers in parentheses are average annual growth rates.

Organization of Petroleum Exporting Countries.

Source: World Bank Trade System data base.Note: Numbers in parentheses are average annual growth rates.

Organization of Petroleum Exporting Countries.

Although the developing countries increased their share of total manufactured exports in world markets—from 9.7 percent in 1965 to 11.9 percent in 1980—their share of total commodity exports fell from 32.4 percent in 1965 to 18.1 percent in 1980 (Table 2). This decline was particularly marked for agricultural commodities: the share of developing countries in total food exports fell from 30.4 percent to 23.0 percent; in total exports of beverages and tobacco, from 80.7 percent to 73.4 percent; and in total exports of agricultural raw materials, from 39.5 percent to 28.5 percent. These declines can in part be explained by the increased processing capacity of developing countries, but this explanation does not complete the picture. In both food and agricultural output, developing countries experienced higher growth rates than industrial countries between 1960 and 1980 (Table 3), but much of the increased agricultural output was consumed domestically rather than exported, since the population of these countries grew during the 1960s and 1970s at much higher rates than that of the industrial countries.

Table 2.Exports of Commodity Groups from Developing Country Regions, Industrial Countries, and OPEC to the World, 1965 and 1980(As proportion of world exports, in percent)
Agri-
BeveragesculturalAll
andRawCommod-Manu-
FoodTobaccoMaterialsMineralsEnergyitiesfacturesTotal
Region1965198019651980196519801965198019651980196519801965198019651980
Africa5.32.818.212.37.63.411.47.20.22.16.83.32.02.03.62.4
Asia8.17.815.410.916.215.98.37.80.73.09.16.42.75.24.85.6
Europe2.82.46.13.22.82.91.81.60.00.02.41.32.52.22.51.9
Middle East1.40.50.00.03.91.30.30.60.31.01.50.80.60.80.90.8
Western Hemisphere12.39.541.047.09.05.015.811.41.70.912.66.31.91.75.43.2
Developing countries30.423.080.773.439.528.537.628.62.97.032.418.19.711.917.213.9
Industrial countries66.976.314.721.158.965.759.568.615.117.251.244.188.186.575.873.4
OPEC2.70.74.65.51.65.82.92.882.075.816.437.82.21.67.012.7
World average100.0100.0100.0100.0100.0100.0100.0100.0100.0100.0100.0100.0100.0100.0100.0100.0
Source: World Bank Trade System data base.
Source: World Bank Trade System data base.
Table 3.Growth Rates of Agricultural and Food Output by Major World Regions, Excluding China, 1960–80
Agricultural outputFood output
TotalPer capitaTotalPer capita
Region1960–701970–801960–701970–801960–701970–801960–701970–80
Developing countries2.82.70.30.32.92.80.40.4
Low-income2.52.10.2-0.42.62.20.2-0.3
Middle-income2.93.10.40.73.23.30.70.9
Africa2.71.30.2-1.42.61.60.1-1.1
Middle East2.52.70.00.02.62.90.10.2
Latin America2.93.00.10.63.63.30.10.6
Southeast Asia2.93.80.31.42.83.80.31.4
South Asia2.52.20.10.02.62.20.10.0
Southern Europe3.13.51.81.93.23.51.81.9
Industrial market
economies2.12.01.11.22.32.01.31.1
World total2.62.20.70.42.72.30.80.5
Source: World Bank (1983).Note: Production data are weighted by world export unit prices. Decade growth rates are based on midpoints of five-year averages, except that 1980 is the average for 1969–71.
Source: World Bank (1983).Note: Production data are weighted by world export unit prices. Decade growth rates are based on midpoints of five-year averages, except that 1980 is the average for 1969–71.

Although the developing countries’ share of world exports of energy and manufactures rose, this increase was by no means sufficient to offset the decline in their share of trade in primary products; consequently, developing countries’ share of total world exports fell from 17.2 percent in 1965 to 13.9 percent in 1980 (Table 2). Among the developing countries only the Asian countries resisted this trend; their combined share in total world exports rose from 4.8 percent in 1965 to 5.6 percent in 1980. By contrast, industrial countries’ share of total world exports fell very little during the same period; this loss of shares for industrial countries was due to the small fall in the share of manufactured exports.

Trends in Commodity Exports Among Different Developing Country Regions

The aggregates presented in the preceding subsection conceal substantial regional differences. Among regional groups of developing countries, Asia has experienced the fastest average rate of growth (13.5 percent) of commodity export earnings (Table 1, “All Commodities”). By contrast, the slowest rates of growth took place in Africa and the Western Hemisphere (10.7 percent and 10.9 percent, respectively). Asia also experienced the fastest growth rate in manufactured exports, which grew at an annual rate of 22.3 percent, compared with an average of 18.6 percent for the developing countries as a whole. Total export earnings for the Asian countries thus grew at an annual average rate of 17.9 percent, compared with a rate of 12.5 percent for the Western Hemisphere and of 13.7 percent for Africa. Because Asia also experienced one of the lowest growth rates in export unit values—7.5 percent a year (Table 4)—the growth in the volume of its exports (over 10 percent a year) was by far the largest such increase of any developing country region.

Table 4.Exports of Commodities from Developing Country Regions, Industrial Countries, and OPEC to the World, 1965 and 1980(Commodity prices and export unit values; 1980 = 100)
Agri-Esport
BeveragesculturalUnit
andRawValue
FoodTobaccoMaterialsMineralsOil(Total)
Region196519801965198019651980196519801965198019651980
Africa19.55100.020.65100.026.03100.044.56100.04.64100.029.34100.0
(11.5)(11.1)(9.4)(5.5)(22.7)(8.5)
Asia31.12100.039.77100.026.58100.032.07100.04.64100.033.73100.0
(8.1)(6.3)(9.2)(7.9)(22.7)(7.5)
Europe35.40100.041.59100.021.95100.034.66100.04.64100.034.84100.0
(7.2)(6.0)(10.6)(7.2,)(22.7)(7.3)
Middle East32.96100.036.82100.032.75100.030.20100.04.64100.0
(non-oil)(7.7)(6.9)(7.7)(8.3)(22.7)
Western26.97100.025.07100.029.01100.047.17100.04.64100.024.14100.0
Hemisphere(9.1)(9.7)(8.6)(5.3)(22.7)(9.9)
Total indebted29.59100.0
developing(8.5)
countries
Industrial29.38100.0
countries(8.5)
OPEC5.34100.0
(21.6)
World total28.61100.025.29100.036.04100.041.39100.04.64100.023.46100.0
(8.7)(9.6)(7.0)(6.1)(22.7)(10.1)
Source: International Monetary Fund.Note: Numbers in parentheses are average annual growth rates.
Source: International Monetary Fund.Note: Numbers in parentheses are average annual growth rates.

The decline in the relative importance of exports of food, beverages and tobacco, and agricultural raw materials for the developing countries as a whole (Table 5) conceals substantial differences between regions, particularly as regards food exports. In Asia, Europe, and the Western Hemisphere there was an increase in the annual per capita growth rates of agricultural output over the period and therefore very little decline in the relative importance of food exports. By contrast, the fall in the share of agricultural exports from Africa is a reflection both of the decline in the annual growth rate of African agricultural output, which felt from 2.7 percent in the 1960s to 1.3 percent in the 1970s (Table 3), and of the acceleration in the rate of population growth in that region. The implications of this decline are especially serious for the many African countries that remain highly dependent on agricultural exports as a source of foreign exchange earnings.

Not surprisingly, Africa also experienced the largest decline in agricultural exports as a proportion of world exports of agriculture (Table 2). Africa’s share in world imports of food fell from 5.3 percent in 1965 to 2.8 percent in 1980; in beverages and tobacco, from 18.2 percent to 12.3 percent; and in agricultural raw materials, from 7.6 percent to 3.4 percent. At the other end of the spectrum, the Asian countries lost very little in world market shares of agricultural exports between 1965 and 1980.

Demand and Supply Perspectives of Commodity Export Trends in Developing Countries

This subsection examines demand-side and supply-side explanations for changes in the commodity structure of developing country exports during the past two decades.

Table 5.Exports of Commodity Groups from Developing Country Regions, Industrial Countries, and OPEC to the World, 1965 and 1989(As proportion of all commodity exports, in percent)
Agri-
BeveragesculturalAllTotal Exports
andRawCommod-Commod-Manu-
FoodTobaccoMaterialsMineralsEnergyitiesitiesfactures
Region1965198019651980196519801965198019651980196519801965198019651980
Africa30.222.219.511.823.410.726.424.50.530.8100.0100.062.442.037.658.0
Asia34.332.412.45.437.525.914.413.81.422.5100.0100.062.635.337.464.7
Europe45.252.618.48.224.624.311.614.90.20.0100.0100.032.420.367.679.7
Middle East
(non-oil)36.517.00.40.055.817.03.47.93.958.1100.0100.053.731.746.368.3
Western
Hemisphere39.140.423.723.915.08.319.920.52.36.9100.0100.077.162.022.938.0
Total developing
countries36.234.018.113.025.716.518.417.91.618.6100.0100.062.340.337.759.7
Industrial
countries50.246.52.11.524.115.618.417.65.218.8100.0100.022.418.577.681.5
OPEC6.30.52.00.52.11.62.80.886.896.6100.0100.078.791.221.38.8
World average38.526.87.33.221.010.515.811.417.448.1100.0100.033.230.866.869.2
Source: World Bank Trade System data base.
Source: World Bank Trade System data base.

Demand for Commodity Exports of Developing Countries

Although activity in the industrial countries is by far the most important determinant of demand for developing country exports, commodity composition, commodity prices, geographical location, and industrial country policies also play a significant role.

In the first place, for the developing countries much success or the lack of success in exporting can depend on the types of commodities exported and how world demand for each commodity group moves over time. Exports of fuel and manufactures have increased in importance in world trade over the years. The elasticity of demand tends to be higher for these goods than for nonfuel primary commodities, and manufacturing prices tend to be more stable than prices for food and raw materials. Minerals and metals tend to have an income elasticity of demand greater than that for foodstuffs. In addition, the continual shift of mineral processing plants from industrial to developing countries has tended to raise industrial country demand for developing country mineral exports. Agricultural commodities have fallen as a share of developing country exports despite the accelerating demand for world foodstuffs and beverages that has taken place since 1973; this rise in demand has reflected the rapid increase in demand for food from the Organization of Petroleum Exporting Countries (OPEC), newly industrializing countries, and non-market economies that resulted from growth in these countries, rather than high demand elasticities. The rising demand for high-value products such as meat, poultry, dairy produce, fruit, and vegetables has been mainly met by industrial countries, with the biggest relative increase coming from the European Communities (EC). The performance for developing countries in these growing agricultural export markets has been quite mixed and has varied widely across regions and countries. The successes mainly took place in exports of nontraditional crops, such as soybeans, whereas in more traditional exports inelastic demand limited the increase. Therefore the failure of agricultural product exporters to diversify has led to a shrinking of their share of world trade.

In the second place, the growth of exports can depend in part on the location of the exporter. For example, most of the markets for Africa’s commodity exports are in the EC because of Africa’s proximity to Europe. With European agriculture expanding in the last decade, Africa’s exports to this region declined. By comparison, most of the markets for Asian exports have been rapidly growing. Trade in rice between developing countries of South and East Asia has been growing rapidly, as have Japan’s imports from the Asian countries.

In the third place, trade and agricultural policies of industrial countries influence the options for exporting that are open to developing countries. Agriculture in some industrial countries is becoming heavily protected and subsidized. The EC, for example, has become increasingly self-sufficient in a broad range of agricultural commodities, thanks to various protectionist measures. Therefore, in certain products developing countries have been prevented from expanding into traditional industrial country markets.

Supply of Commodity Exports from Developing Countries

Weather conditions, resource endowments, relative prices, technology, domestic market growth, and population growth all determine the supply of a developing country’s exports. Apart from the weather and resource endowment, all the other determinants of supply are affected by domestic policies.

First, pricing policies have to allow producer prices to reflect market prices if a country is to have an optimal production mix for domestic markets and exports. In some low-income African countries during the period examined, poor incentives to farmers, inefficient marketing systems, high inflation, and the maintenance of an overvalued nominal exchange rate led to an appreciation of real exchange rates, as relatively high rates of domestic inflation were not fully offset by a fall in nominal exchange rates. Analogously, pricing policies, particularly in agriculture, failed in many countries to display adequate flexibility in the face of domestic inflation; for example, in several African countries real producer prices were lower in 1980 than in 1970.

Second, policies that led to higher investment are highly correlated with export growth rates. The spread of new technology in the rural sector tends to raise agricultural capacity and provide increased export earnings, which can be used to import the capital and raw materials necessary to begin expansion in the manufacturing sector. New technology and technical schooling have also played an important role in the rapid diffusion of high-yielding crop varieties.

Third, population policy and export growth are related. Many countries have shown that effective measures can be taken to slow population growth. With the slowing of population growth, more domestic production can be exported—thus providing the means to purchase capital inputs so necessary for technological improvements.

II. Volume and Price Equations for Commodity Exports of Developing Countries

In the empirical equations developed in this section, five developing country regions and five commodity groups are distinguished. The regions are Africa, Asia, Europe, the Middle East, and the Western Hemisphere, as defined in the International Monetary Fund’s World Economic Outlook (April 1986, p. 174); they are enumerated in Appendix I. They were chosen because of the availability of data and because of certain similarities that exist among the countries within each region, such as closeness to industrial country markets and transport costs.

The five commodity groups are (1) food, (2) beverages and tobacco, (3) agricultural raw materials, (4) minerals, and (5) energy. This disaggregation is made because an analysis based on aggregate relationships covering all commodities could produce misleading results, owing to differences in the degree of sensitivity to price and income changes among types of export commodities. In addition, the specification of demand and supply equations may differ among commodity types. In particular, conditions in the energy market require estimation procedures that are different from those used for the other four commodity groups.

Demand and Supply Equations for Four Commodities

The demand and supply equations for five regions and the first four commodity groupings are based on the equations used by Goldstein and Khan (1978). Adjustment in export demand to changing market conditions occurs within a period of one year; adjustment in export supply allows for the possibility of delayed adjustment beyond one year. In the demand equation, commodity k is differentiated by its regional source of supply.6

Export Demand

The world demand for exports of commodity k from developing country region R is specified in log-linear form as follows:

where the variables are

All data are expressed in U.S. dollars.

Because equation (1) is specified in logarithms, a1 is the elasticity of world demand for regionR’s exports of the k th commodity with respect to the divergence between region R’s export price of the k th commodity and the average world price, and a2 is the elasticity of export demand for commodity k with respect to global real income. It is expected that a1 will be negative and that a2 will be positive.

Export Supply

The supply of exports of commodity k from region R is specified as a log-linear function of current and lagged ratios of the export price of commodity k to domestic price levels in producing countries in region R, an index of productive capacity in region R, and supply shocks:

where

All data, except domestic prices, are expressed in U.S. dollars.

Equation (2) embodies the notion that exporters increase their supply of exports as the price of exports rises relative to domestic prices. The lagged price variable allows for the possibility of delayed supply adjustment beyond the period of one year. Exports of commodity k are also expected to increase as productive capacity in region R increases. SSR reflects other factors that influence exports from region R; β1 and β2 are region R’s price elasticity and lagged price elasticity of export supply, respectively, and β3 is the elasticity with respect to productive capacity. It is expected that the sign of elasticities β1, β23 and β3 will be positive. Normalizing the equation for the price of exports in region R yields the following equation:

The normalized coefficients are related to the structural parameters in the following way:

Since β1, β2 and β3 are positive, it is expected that b1 > 0, b2 > 0, b3 < 0, and b4 < 0.

Equations (1) and (3) can be estimated simultaneously to obtain the estimates of the structural parameters.

Demand and Supply Equations for Energy

The equations for the energy market require a treatment different from that in the equations for the other four commodities because the links between energy prices, economic growth, and energy demand are complex and because energy supply is influenced by both economic and noneconomic factors. The equations for energy exports will be specified in a two-stage process. First, the export demand for world energy will be specified, with export supply of world energy assumed to be exogenous. Second, export demand for energy will be allocated across exporting developing country regions with reference to a trend term.

Export Demand for World Energy

The world demand for energy is assumed to be determined by the world price of energy relative to the world price level and to world income; this demand can be written as

where

Export Supply of World Energy

The world supply of energy (XSWen) is treated as exogenously determined because over the period of estimation (1963 through 1982) many production decisions were made by a partial cartel dominated by OPEC:

Export Demand and Supply for Energy from Five Developing Country Regions

World energy demand is allocated across regions with reference to each region’s share in 1980 in the world energy market and a trend variable:

where

By substituting equation (4) into equation (6), the following estimating equation, specified in log-linear form, was obtained for the demand for energy exports from region R:

The supply of energy from each developing country region is also assumed to be determined exogenously. These regions do not include major OPEC members, so that production decisions in these countries were not dominated by OPEC. In many of these countries, however, production decisions were dominated by national governments that provided capital toward initial investments in the oil sector. Thus noneconomic criteria also influenced the oil sectors of many of these countries:

III. Estimation of the Equations

Equations (1) and (3) were estimated using annual data from 1963 through 1982 for the five developing country regions and the first four commodity groups listed in Section II.7 Nonlinear least-squares estimation procedures were used to obtain the estimates using the MINDIS routine, which performs minimum distance estimation on a multiple equation model. Equation (7) was estimated for the energy commodity group and four developing country regions 8 using ordinary least-squares estimation procedures. All the Fund member countries were taken to represent the world,9 Definitions of the data used in the estimation and the sources of these data are given in Appendix II. The estimated coefficients and their respective t-statistics (in parameters), together with the coefficient of determination, R2, and the standard error of the estimate, SEE, are presented in Tables 6 and 7. The meaning of these last two statistics, however, is ambiguous in simultaneous equation models.10

Table 6.Estimates of Demand Equation for Food, Beverage and Tobacco. Agricultural Raw Material, Mineral, and Energy Exports of Five Developing Country Regions, 1963–82
Demand Equation
Trend and
Region anddummy
Commodityaoa1a2variablesR2SEE
Africa
Food296-0.321.010.740.15
(3.97)(6.59)(6.18)
Beverages and1.86-0.311.34-0.09t10.700.23
tobacco(1.41)(4.16)(3.37)(4.30)
Agricultural218-3.280.540.710.16
raw materials(0.77)(2.30)(0.91)
Minerals5.26-0.853.850.710.38
(2.64)(3.67)(5.07)
Energy-17.44-0.065.10-0.18t10.930.13
(7.24)(0.54)(8.45)(5.57)
Asia
Food-0.51-0.331.140.840.19
(0.46)(2.56)(4.62)
Beverages and4.420.08-0.140.01D10.340.12
tobacco(13.69)(2.63)(1.27)(1.87)
Agricultural-2.50-0.340.460.790.06
raw materials(7.77)(2.27)(6.59)
Minerals-0.94-0.401.190.860.12
(1.23)(1.51)(7.06)
Energy-10.21-0.363.560.29D10.980.08
(14.19)(7.06)(17.98)(3.14)
Europe
Food-6.26-0.141.120.760.18
(1.05)(1.81)(8.05)
Beverages and6.31-0.26-0.380.170.20
tobacco(6.03)(1.79)(1.61)
Agricultural1.39-0.211.150.560.15
raw materials(2.15)(3.98)(5.04)
Minerals-4.37-0.482.910.870.18
(4.76)(4.01)(7.42)
Middle East
Food-1.71-0.461.540.22D770.700.17
(4.00)(10.96)(9.63)(3.41)
Agricultural-7.56-0.090.410.670.16
raw materials(17.32)(3.73)(4.03)
Minerals-1.700.600.260.890.26
(1.34)(4.44)(0.54)
Energy-5.61-0.282.520.610.39
(1.51)(1.17)(2.48)
Western Hemisphere
Food0.35-0.111.320.760.16
(0.48)(1.58)(4.78)
Beverages and2.23-0.330.510.750.07
tobacco(4.43)(3.79)(4.46)
Agricultural5.78-0.140.040.370.21
raw materials(18.99)(3.53)(0.19)
Minerals-0.18-0.381.740.540.24
(0.20)(3.38)(4.69)
Energy0.750.390.810.60D1-1.69D10.920.16
(0.66)(4.14)(2.56)(3.36) (11.46)
Note: For an explanation of time-trend and dummy variables, see Section III; R2 is the coefficient of determination; SEE is the standard error of estimate; numbers in parentheses are t-statistics.
Note: For an explanation of time-trend and dummy variables, see Section III; R2 is the coefficient of determination; SEE is the standard error of estimate; numbers in parentheses are t-statistics.
Table 7.Estimates of Supply Equation for Food, Beverage and Tobacco, Agricultural Raw Material, and Mineral Exports of Five Developing Country Regions, 1963–82
Region

and
Supply Equation
Commoditybob1b2b3b4bsb6R2SEE
Africa
Food3.66-0.780.740.250.30D1-0.05t10.960.16
(1.68)(2.50)(7.86)(2.37)(3.21)(3.74)
Beverages and0.30-0.201.29-0.060.48D20.950.21
tobacco(0.28)(1.16)(12.77)(0.45)(4.32)
Agricultural-8.211.431.41-0.13-0.10D10.980.08
raw materials(2.29)(2.79)(7.44)(0.33)(2.30)
Minerals-2.36-0.530.040.051.940.02t10.950.11
(2.64)(3.01)(0.56)(1.12)(5.85)(5.36)
Asia
Food-2.610.821.420.20-0.880.53D10.990.06
(5.67)(2.15)(4.67)(1.53)(1.73)(5.63)
Beverages and-18.224.391.400.11-0.390.04D10.990.04
tobacco(2.50)(2.51)(1.59)(3.46)(1.29)(1.79)
Agricultural-25.075.894.62-0.12-4.01-0.08D10.980.10
raw materials(2.18)(1.70)(4.40)(0.27)(2.64)(0.66)
Minerals-9.344.183.19-0.08-4.19-0.07D10.990.05
(2.66)(1.95)(4.74)(0.26)(2.20)(1.02)
Europe
Food-8.095.231.050.77-0.420.980.07
(2.26)(2.13)(3.97)(2.78)(1.99)
Beverages and-4.080.810.25-0.090.920.970.06
tobacco(1.90)(1.96)(1.81)(0.55)(10.86)
Agricultural6.22-2.550.07-0.172.34-0.05D10.0lt10.980.09
raw materials(2.60)(4.21)(1.04)(3.11)(9.50)(1.31)(3.42)
Minerals-1.040.040.710.380.31D10.08t30.980.07
(1.31)(0.58)(5.08)(3.11)(6.10)(4.88)
Middle East
Food0.61-0.371.050.200.01t10.900.19
(0.46)(2.49)(14.98)(1.32)(1.02)
Agricultural20.15-3.690.310.39-0.06t10.990.08
raw materials(4.97)(4.64)(1.76)(4.15)(2.34)
Minerals-2.730.140.870.540.43D1-0.0lt30.950.15
(3.44)(0.54)(2.40)(3.61)(3.23)(0.86)
Western Hemisphere
Food1.54-2.490.220.182.800.21D10.970.10
(0.60)(1.97)(1.26)(1.87)(3.13)(3.98)
Beverages and-10.501.841.100.361.260.910.25
tobacco(2.21)(1.62)(4.85)(1.13)(3.90)
Agricultural18.99-3.351.76-0.100.01D10.10D20.990.07
raw materials(4.39)(3.68)(3.76)(0.26)(0.46)(4.31)
Minerals2.61-1.32-0.220.110.89-0.01t30.970.07
(2.37)(3.66)(3.24)(0.27)(7.39)(2.09)
Note: For an explanation of tune-trend and dummy variables, see Section III; R2 is the coefficient of determination; SEE is the standard error of estimate; numbers in parentheses are t-statistics.
Note: For an explanation of tune-trend and dummy variables, see Section III; R2 is the coefficient of determination; SEE is the standard error of estimate; numbers in parentheses are t-statistics.

From the results it appears that the model performs quite well in terms of yielding parameter estimates that are both of the expected sign and size and that are statistically significant. The estimated price coefficients in the export demand equation (1) carry the expected negative sign for most of the commodity groups and are significantly different from zero at the 5 percent level of significance in 14 of the 23 equations estimated. In all but one of the equations the estimated price elasticity is less than unity, a finding that implies a fairly limited short-term response of demand for exports to changes in relative prices. The average estimated price elasticities of demand for commodity groups have been computed as follows: food, -0.22; beverages and tobacco, -0.33; agricultural raw materials, -0.62; minerals, -0.51; and energy, -0.21 (Table 8). These estimated price elasticities therefore differ significantly across commodities; the lowest are for food and energy, and the highest are for agricultural raw materials. There is much less variation in the price elasticities calculated across regions than across commodities (Table 8, last column). The range is -0.21 to -0.33 if the high estimated elasticity for Africa is excluded. The price elasticity of demand for all commodities for all developing countries is calculated to be -0.35.

Table 8.Estimated Price Elasticities of Demand (a1) by Commodity and Region
BeveragesAgricultural
andRaw
RegionFoodTobaccoMaterialsMineralsEnergyTotala
Africa-0.32**-0.31**-3.28**-0.85*-0.06-0.68
Asia-0.33**-0.34**-0.40*-0.36**-0.33
Europe-0.14*-0.26**-0.21*-0.48*-0.22
Middle East-0.46**-0.09**-0.28-0.25
Western
Hemisphere-0.11*-0.33**-0.14**-0.38*-0.21
Totalb-0.22-0.33-0.62-0.51-0.21-0.35
Note: * indicates significance at the 90 percent level; ** indicates significance at the 95 percent level.

The weights of each commodity export in 1980 as a percentage of total exports for each region were used to obtain this total elasticity.

The weight of each region’s exports in 1980 as a percentage of total exports for each commodity were used to obtain this elasticity.

Note: * indicates significance at the 90 percent level; ** indicates significance at the 95 percent level.

The weights of each commodity export in 1980 as a percentage of total exports for each region were used to obtain this total elasticity.

The weight of each region’s exports in 1980 as a percentage of total exports for each commodity were used to obtain this elasticity.

The estimated income elasticities shown in Table 9 have the expected positive signs and are significantly different from zero at the 5 percent level in 17 of 23 equations. The average income elasticities are as follows: food, 1.20; beverages and tobacco, 0.68; agricultural raw materials, 0.56; minerals, 2.16; and energy, 3.53. These results support the view that exports of agricultural products (food, beverages and tobacco, and agricultural raw materials) are less sensitive to short-term fluctuations in world demand than other exports, such as minerals. Both price and income elasticities will be compared with ones obtained from similar studies in the next section of the paper.

Table 9.Estimated Income Elasticities of Demand (a2) by Commodity and Region
BeveragesAgricultural
andRaw
RegionFoodTobaccoMaterialsMineralsEnergy
Africa1.01**1.34**0.543.85**5.10**
Asia1.14**0.46**1.19**3.56**
Europe1.12**1.15**2.91**
Middle East1.54**0.262.52**
Western
Hemisphere1.32**0.51**1 74**0.81**
Totala1.200.680.562.163.53
Note: **indicates significance at the 95 percent level.

The weight of each region’s exports in 1980 as a percentage of total exports for each commodity were used to obtain this elasticity.

Note: **indicates significance at the 95 percent level.

The weight of each region’s exports in 1980 as a percentage of total exports for each commodity were used to obtain this elasticity.

The estimated coefficients in the export supply equation (3) also yield useful information, but in general the performance of the supply equation is poor. The estimate of the elasticity of supply with respect to the price can be derived from the estimated version of equation (3) by calculating (b1)_1, and the elasticity of supply with respect to lagged prices, by calculating -b3(b1)_1 These computed elasticities are given in Tables 10 and 11.

Table 10.Estimated Export Price Elasticities of Supply (β1) by Commodity and Region
RegionFoodBeverages

and

Tobacco
Agricultural

Raw

Materials
Minerals
Africa-1.28**0.70**-1.89**
Asia1.21**0.23**0.17*0.24*
Europe0.19**1.23**-0.39**
Middle East-2.70**0.27**
Western Hemisphere-0.40**0.54**-0.30**-0.76*
Total a0.700.660.430.24
Note: * indicates significance at the 90 percent level; ** indicates significance at the 95 percent level.

Only those coefficients with the “right” sign were used to obtain the mean total elasticity for each commodity group.

Note: * indicates significance at the 90 percent level; ** indicates significance at the 95 percent level.

Only those coefficients with the “right” sign were used to obtain the mean total elasticity for each commodity group.

Table 11.Estimated One-Year Lagged Export Price Elasticities of Supply (β2)by Commodity and Region
RegionFoodBeverages

and

Tobacco
Agricultural

Raw

Materials
Minerals
Africa0.32**0.09*0.09*
Asia-0.24*-0.02**0.020.02
Europe-0.15**0.11-0.06**
Middle East0.54*0.10*
Western Hemisphere0.07*0.030.08
Note: * indicates significance at the 90 percent level; ** indicates significance at the 95 percent level.
Note: * indicates significance at the 90 percent level; ** indicates significance at the 95 percent level.

The estimated coefficients of exports β^1 are positive and significantly different from zero at the 10 percent level in 6 of the 19 equations estimated for the first four commodity groups, implying a positively sloped supply function for exports for these six commodities and regions. In terms of the estimates of β^1 for the geographical regions, the equations for Asia perform the best; these results reflect the policies of Asian countries to allow the producer price to reflect export prices as a way to encourage export production. By contrast, the estimates of β^1 for Africa for some commodities are implausibly large and have the wrong sign. In Africa producer pricing policies that allowed the real producer price for food crops to fall during the late 1960s and 1970s, despite increases in export prices received by the authorities, led to less food production. Indeed, in many African countries during this period producer prices often moved in a different direction from the export price. Less food production, combined with increasing population growth, led to less food available for export. Under these conditions, the normal relationship between export prices and commodities produced for export was distorted.

In terms of the estimates of β^1 for the commodity groups, the equations for beverages and tobacco perform the best. The equations for minerals have the poorest performance. Modeling supply equations for minerals is a complicated process, and the supply equations here are probably too highly aggregated. Furthermore, many mineral products are subject to export quotas, and the export price may bear little relationship to the amount exported through quotas.

The equations perform poorly with regard to estimated coefficients of β^2, the elasticity with respect to lagged export prices. Only 6 of the 23 equations carry the expected positive sign and are significant at the 10 percent level. The lagged supply price elasticity for the food and agricultural raw materials groups performed the best. For most other groups and regions the data series are not sufficiently long, and the lag process is not well specified enough, to capture the long lag structure that exists for certain commodities between prices and export production. There are great differences in the lag structures for different commodities, and the estimated equations are unlikely to be disaggregated enough to capture the sophisticated lag structures that exist in commodity markets.

The capacity variable Y¯ was included in the supply equations of the developing countries to capture the effects of domestic capacity on exports. One problem with including this variable is that Y¯ is not independent of real world income, of exports in the supplying countries, or of the trend term; its inclusion therefore makes it more difficult to interpret the values or significance of the regression coefficients accurately. This problem was particularly severe for most of the equations for Africa and the Middle East, and consequently the capacity variable was dropped from these equations. Nevertheless, the capacity variable Y¯ is in accordance with standard theory and has the expected positive sign for 6 of the 19 equations.

The time trend was removed from most of the equations because it was highly collinear with the income and capacity variables; however, time trends were included for particular periods in regions where important structural changes took place over the estimation period. For example, the negative coefficients on the time trend for food exports from Africa probably reflect the unfavorable incentives for agriculture that had been created. The trends t1, t2, and t3 were used to measure these periods of structural change. Dummy variables D1 and D2 were included in the equations to measure the effects of the two oil crises that took place over the estimation period. On the demand side, the oil crisis led to an increase in the demand for some commodities, possibly because some countries faced with higher production costs produced less domestically.

On the supply side, the oil crisis led to both increases and decreases in exported commodities. For those countries and commodity groups where a fall in exported commodities was experienced, the oil crisis probably led to increased production costs that reduced output. For countries and commodity groups where exported commodities rose, production costs probably rose less, and these countries were able to expand supply of the relevant commodity to meet the rise in demand.

IV. Survey of Commodity Demand and Supply Elasticities

This section presents a comprehensive list of demand and supply elasticities of commodities estimated in other studies and compares these with the elasticities estimated in this study. On the demand side, the estimates for the five groups presented in this paper give good information for the purpose of comparison. On the supply side, the estimates for the four groups—excluding energy—are relatively poor, and more reliance may need to be placed on estimates from other studies.

The demand and supply elasticities presented in Table 12 are collected from a number of sources. The demand elasticities are obtained from studies by Behrman (1977) and UNCTAD (1974); these two studies calculate median demand elasticities from estimates gathered from about 200 studies. The supply estimates were obtained from the survey paper by Askari and Cummings (1977); for each individual commodity the mean elasticity is calculated from the hundreds of elasticity estimates presented in their study. In collecting the elasticities, zero elasticities and wrong-sign elasticities were excluded from the individual commodity group prior to its component items being summed. The elasticity estimates of Askari and Cummings were used as the main source for the supply elasticities. Where gaps occurred, these were supplemented by estimates from Behrman.

Table 12.Range of Estimates of Commodity Demand and Supply Elasticities
Price

Elasticity
Price

Elasticity

of Supply
CommodityIncome

Elasticity
of

Demand
Short-

run
Long-

run
Source
Food0.98-0.58Goldstein and Khan (1984)
Averagea0.50-0.420.430.80Author’s calculations
Barley0.420.79Askari and Cummings (1977)
Cereals0.520.73Askari and Cummings (1977)
-0.35UNCTAD (1974)
Dairy products0.181.01Askari and Cummings (1977)
Fats and oils0.491.06Askari and Cummings (1977)
-0.5UNCTAD (1974)
Fruit0.310.73Askari and Cummings (1977)
Maize0.440.57Askari and Cummings (1977)
-0.45UNCTAD (1974)
Meat0.410.80Askari and Cummings (1977)
0.3-0.4Behrman (1977)
Rice0.270.44Askari and Cummings (1977)
0.3-0.3Behrman (1977)
Soybeans1.141.16Askari and Cummings (1977)
Sugar0.490.87Askari and Cummings (1977)
1.1-1.1Behrman (1977)
Vegetables0.250.92Askari and Cummings (1977)
Wheat0.300.56Askari and Cummings (1977)
0.3-0.3Behrman (1977)
Beverages
and tobacco0.98-0.58Goldstein and Khan (1984)
Averagea0.35-0.450.270.46Author’s calculations
Cocoa0.380.79Askari and Cummings (1977)
0.4-0.4Behrman (1977)
Coffee0.370.53Askari and Cummings (1977)
0.5-0.6Behrman (1977)
Tea0.040.13Askari and Cummings (1977)
0.5-0.3Behrman (1977)
Tobacco0.290.41Askari and Cummings (1977)
-0.5UNCTAD (1974)
Agricultural
raw materials0.85-0.67Goldstein and Khan (1984)
Averagea0.8-0.440.330.51Author’s calculations
Cotton0.430.89Askari and Cummings (1977)
0.8-0.3Behrman (1977)
Jute0.530.74Askari and Cummings (1977)
-0.5UNCTAD (1974)
Rubber0.180.31Askari and Cummings (1977)
-0.8Behrman (1974)
Sisal0.460.33Askari and Cummings (1977)
Wood-0.40.30.5UNCTAD (1974)
Wool-0.20.10.3Behrman (1977)
Minerals-0.3UNCTAD (1974)
Averagea2.8-1.10.00.27Author’s calculations
Bauxite2.3-1.30.00.4Behrman (1977)
Copper1.0-0.20.00.2Behrman (1977)
Iron ore-0.70.00.3Behrman (1977)
Lead-0.2UNCTAD (1974)
Magnesium-0.1UNCTAD (1974)
Tin5.0-5.00.00.2Behrman (1977)
Zinc-0.1Behrman (1977)
Energy1.22-0.54Goldstein and Khan (1984)

The average is calculated as an arithmetic mean of the elasticities for individual commodities shown in this table.

The average is calculated as an arithmetic mean of the elasticities for individual commodities shown in this table.

These estimates need to be treated with caution for several reasons. First, there is wide variation in the quality of the studies from which these estimates were taken, especially with respect to supply responsiveness. Second, there is a wide range of quantitative estimates among the studies for single commodities because of the differences with regard to the price variables, the time periods, and the quality of data used.11 Third, some of the equations in the individual studies may be incorrectly specified. For example, a demand or supply equation may be specified separately when both would have been more appropriate. Fourth, lag structures are notoriously difficult to specify; modeling the adjustment process for producers’ price expectations may depend on several factors, such as changes in the weather, changes in output, and other exogenous economic events that can never be adequately captured by lag structures.

When making comparisons between these estimates it is also important to focus on how the results were obtained. In particular, it is necessary to distinguish between (1) the price elasticity of response for an individual commodity to a change in the relative price for that commodity, and (2) the price elasticity of response for a group of commodities, such as food, to a change in the group price. The former elasticity is likely to be larger than the latter because the substitution possibilities are much greater for an individual commodity than for a group of commodities. Similarly, the price elasticity for one region is likely to be somewhat larger than the price elasticity for all regions together because of the greater substitution possibilities between regions. This point should be borne in mind when comparing summed elasticities.

The income and price elasticities contained in Table 12 were summed and averaged across individual commodities to obtain estimates for each of the five commodity groups presented earlier in this paper. These elasticities were then compared with the mean estimates of all regions obtained in Section III of the paper and with mean demand elasticity estimates 12 obtained by Goldstein and Khan (1984). These three sets of elasticities are presented in Table 13.

Table 13.Aggregate Estimates of Commodity Demand and Supply Elasticities, All Regions
Commodity GroupGoldstein and Khan(1984)Average of

Individual

Commodities
This

Study
Income Elasticities
Food0.980.501.20
Beverages and tobacco0.980.350.68
Agricultural raw materials0.850.800.56
Minerals2.802.16
Energy1.223.53
Price Elasticity of Demand
Food-0.58-0.42-0.22
Beverages and tobacco-0.58-0.45-0.33
Agricultural raw materials-0.67-0.44-0.62
Minerals-1.1-0.51
Energy-0.54-0.21
Total-0.35
Short-runLong-run
Price Elasticity of Supply
Food0.430.800.70
Beverages and tobacco0.270.460.66
Agricultural raw materials0.330.510.43
Minerals0.270.24

One broad conclusion, gained from examining the estimates shown in Table 13, is that income elasticities of demand for developing country commodity exports fall in the range of 0.3 to 3.5. A further conclusion is that the demand for agricultural products is income inelastic. For the beverages and tobacco group the results from this study tend to be midway between the other two sets of results. The estimates for food tend to be higher in this study than in others. This finding could reflect a higher income elasticity for developing country exports than for world exports in general.

A further broad conclusion is that the demand for commodity exports is not very sensitive to short-run price changes. The estimates range from −0.2 to −1.1. Again, these elasticities tend to be lower for agricultural products than for minerals and energy, with the lowest for food exports. The estimated price elasticities of demand in this study are lower than the mean price elasticities of demand from other studies because this study does not include the substitution possibilities that are available in individual studies. The mean estimates from Goldstein and Khan (1984) are somewhat higher; these estimates include exports from industrial countries as well as developing countries, and it is quite likely that this is the reason for the higher price elasticity.

These results also confirm the conclusion that the supply of commodity exports is more sensitive to prices in the long run than in the short run, and that in the short run price elasticities of supply are in general lower than the corresponding price elasticities of demand. The short-run price elasticities of supply in this study are higher than those from other studies, whereas one might expect them to be lower in view of the lack of substitution possibilities. In this case more reliance might be placed on the lower estimates—except, perhaps, for the estimates obtained in this study for Asia, for which the equations perform quite well.

One of the uses for this information is to help prepare forecasts, projections, and simulations for a single country. Once the price and income elasticities are determined for those commodities relevant to a country, the elasticities can be summed over all commodities to determine the effect of changes in domestic prices, foreign prices, and foreign income on that country’s exports.

V. Conclusions

In many ways trade in commodity exports is no longer as important for developing countries as it once was. These countries lost considerable market shares to industrial countries between the years 1965 and 1980, in part because of the growing processing capacity in developing countries and in part because of the accelerating growth of these countries’ populations over the same period. Developing countries did increase their share in world exports of manufactures over this period, but this increase was by no means large enough to offset the decline that took place in commodity exports. If developing countries wish to solve their debt problems and to improve their long-run growth prospects, they need to recapture their export shares in world commodity markets. This means that protectionism in world commodity markets must be reduced and that developing countries have to maintain a domestic relative price structure that will ensure a sound commodity base, as well as to encourage domestic production in other ways.

The empirical results obtained in this study are very much in line with those obtained from other studies. The results demonstrate the inelastic nature of price responses in the demand for exported commodities and the inelastic income responses in the demand for food, beverage and tobacco, and agricultural raw material exports from developing countries. The evidence also shows that price elasticities of supply are in general lower than the corresponding price elasticities of demand in the short run, but that in the longer run the supply of commodity exports from developing countries is more sensitive to prices than to demand.

The results presented in this paper also provide further evidence of, and support for, the usefulness of pricing policy. Export supply in developing countries does indeed respond to improved price incentives. This evidence lends support to a developing country’s use of the exchange rate as a policy tool to improve the trade balance through both an increase in the demand for and an increase in the supply of commodity exports.

Furthermore, the grouping of developing countries allows more broadly based policy questions to be answered than would be possible for an individual developing country. For example, the empirical evidence suggests that exchange rate or producer pricing policies conducted simultaneously in a group of countries can give rise to an increase in exports of commodities from each of the countries within the group. At the same time the interregional differences that exist show that responses to exchange rate policies will differ among regions. On the demand side, however, there is in general much less difference between regions than between commodities. In general, though, price elasticities of demand tend to be larger for Africa and Asia than for the other regions; they also tend to be slightly larger for agricultural raw materials and minerals than for other commodities.

There remains a considerable amount of work to be done in this area. In particular, more work is needed to derive a set of estimates for the supply equation that are completely satisfactory and that allow one to give unequivocal support to the results obtained. Further work on the demand for and supply of exports of manufactures from developing countries would also prove fruitful and would add to our knowledge of how countries can use policies to change the mix of their exports and thereby improve their external positions.

APPENDIX I

Classification of Countries

The classification of countries adopted in this paper is the same as that adopted by the International Monetary Fund in its World Economic Outlook (Washington, April 1986, p. 174):

African countries
AngolaGambia, TheRwanda
BeninGhanaSt. Helena
BotswanaGuineaSao Tome and Principe
Burkina FasoGuinea-BissauSenegal
BurundiKenyaSeychelles
CameroonLesothoSierra Leone
Cape VerdeLiberiaSomalia
Central Africa RepublicMadagascarSudan
ChadMalawiSwaziland
ComorosMaliTanzania
CongoMauritaniaTogo
Côte d’IvoireMauritiusTunisia
DjiboutiMoroccoUganda
Equatorial GuineaMozambiqueZaire
EthiopiaNigerZambia
GabonReunionZimbabwe
Asian countries
AfghanistanIndiaPakistan
American SamoaKampuchea, Dem.Papua New Guinea
BangladeshKiribatiPhilippines
BhutanKoreaSingapore
BruneiLao P.D. RepublicSri Lanka
BurmaMacaoThailand
China, People’s Republic ofMalaysiaTongo
FijiMaldivesVanuatu
French PolynesiaNauruViet Nam
GuamNepalWestern Samoa
Hong KongNew Caledonia
European countries
CyprusHungaryRomania
Faeroe IslandsMaltaTurkey
GibraltarPortugalYugoslavia
Greece
Middle Eastern countries
BahrainJordanYemen Arab Republic
EgyptLebanonYemen, P.D. Republic
IsraelSyrian Arab Republic
Western Hemisphere countries
Antigua and BarbudaEcuadorMonserrat
ArgentinaEl SalvadorNetherlands Antilles
BahamasFalkland IslandsNicaragua
BarbadosGreenlandPanama
BelizeGrenadaParaguay
BermudaGuadeloupePeru
BoliviaGuatemalaSt. Christopher and Nevis
BrazilGuiana, FrenchSt. Lucia
Cayman IslandsGuyanaSt. Pierre and Miquelon
ChileHaitiSt. Vincent
ColumbiaHondurasSuriname
Costa RicaJamaicaTrinidad and Tobago
DominicaMartiniqueUruguay
Dominican RepublicMexicoU.S. Virgin Islands
APPENDIX II

Definitions and Data Sources

All data are annual. Definitions of the variables in the model are as follows:

  • PRER = an index of the consumer price index in producing countries in region R, in U.S. dollars (1980 = 100)13

  • PXRk = an index of the export price of commodity k, from region R, in U.S. dollars (1980 = 100)

  • PWen = an index of the average world price of energy, in U.S. dollars (1980 = 100)

  • PWk = an index of price of commodity k, in international markets, in U.S. dollars (1980 = 100)

  • PW = an index of consumer price index for the world, in U.S. dollars (1980 = 100)

  • XRk = an index of the volume of exports of commodity k from region R, in billions of U.S. dollars (1980 prices)

  • XWRen = an index of the volume of energy exports from region R, in billions of U.S. dollars (1980 prices)

  • Y¯R = the trend of real output in region R in index form

  • YW = an index of real world income, in U.S. dollars (1980 = 100).

Data sources were the World Bank Trade System data base, the Fund’s International Financial Statistics, and data files of the Commodities Division and the Current Studies Division of the Fund’s Research Department.

REFERENCES

    AskariHossein and John T.CummingsAgricultural Supply Response: A Survey of the Econometric Evidence (New York: Praeger1976).

    AskariHossein and John T.Cummings“Estimating Agricultural Supply Response with the Nerlove Model: A Survey,”International Economic Review (Osaka) Vol. 18 (June1977) pp. 25792.

    BehrmanJere R.International Commodity Agreements: An Evaluation of the UNCTAD Integrated Commodity Programme (Washington: Overseas Development CouncilOctober1977).

    ChuKe-young and Thomas K.Morrison“The 1981–82 Recession and Non-Oil Primary Commodity Prices,”Staff PapersInternational Monetary Fund (Washington) Vol. 31 (March1984) pp. 93140.

    ChuKe-young and Thomas K.Morrison“World Non-Oil Primary Commodity Markets: A Medium-Term Framework of Analysis,”Staff PapersInternational Monetary Fund (Washington) Vol. 33 (March1986) pp. 13984.

    GoldsteinMorris and Mohsin S.Khan“The Supply and Demand for Exports: A Simultaneous Approach,”Review of Economics and Statistics (Cambridge, Massachusetts) Vol. 60 (May1978) pp. 27586.

    GoldsteinMorris and Mohsin S.Khan“Income and Price Effects in Foreign Trade,”in Handbook of International Economicsed. by Ronald W.Jones and Peter B.Kenen (Amsterdam: North-Holland1984).

    International Monetary FundWorld Economic Outlook: A Survey by the Staff of the International Monetary FundWorld Economic Financial Surveys (WashingtonApril1986).

    UNCTADSurvey of Commodity Demand and Supply Elasticities Research Memorandum 48 UNCTAD/RD/70 (Geneva: United Nations Conference on Trade and Development, Research DivisionMarch 1974).

    World BankWorld Bank Development Report (New York: Oxford University Press1983).

Ms. Bond, Senior Economist in the Research Department at the time this paper was written, is now Senior Economist in the Asian Department of the Fund. She is a graduate of the University of Essex and the London School of Economics and Political Science. She has been a faculty member of the University of Reading, England.

Non-oil developing countries are defined in this paper as in the International Monetary Fund’s World Economic Outlook (Washington: April 1986, p. 174) and are listed in Appendix I.

Although data for 1981 and 1982 were used in the empirical estimation, they were excluded from the data used for this discussion on trends because they contain cyclical effects.

A more detailed discussion of how these market shares changed can be found in World Bank (1983).

A wider focus on demand and supply factors in world non-oil primary producing commodity markets was given by Chu and Morrison (1984, 1986).

The data presented below exclude data for centrally planned economies that are not members of the Fund.

Commodity exports of different regions are treated as imperfect substitutes in this paper.

The country classification is given in Appendix I of the paper.

Energy exports from Europe were very small over the estimation period, and therefore energy equations were not estimated for this region. For similar reasons, equations were not estimated for beverages and tobacco for the Middle East.

Fund data exclude nonmember centrally planned economies.

The R2 statistic is not banded (0,1) but (-∞, 1), so that the small values are not an indication of a “poor” fit.

Calculating the mean elasticity for individual commodities smooths out these differences.

The mean elasticities are an average of the activity estimates taken from import equations from a variety of studies and presented in Table 4 of Goldstein and Khan (1984). These elasticities are thus income elasticities of demand for imports rather than for exports.

This series was calculated in the following way:

where αj is the weight of country j’s GDP in region R in U.S. dollars.

Other Resources Citing This Publication