V A Scenario and Forecast Adjustment Model for Developing Countries
Author:
Mr. Charles Adams https://isni.org/isni/0000000404811396 International Monetary Fund

Search for other papers by Mr. Charles Adams in
Current site
Google Scholar Close
and
Claire Hughes Adams
Search for other papers by Claire Hughes Adams in
Current site
Google Scholar Close
Language:
English
Print Publication Date:
01 Jan 1989
Keywords:
WEFS; export volume; import volume; price level; market distortion; resource allocation; export price; Merchandise import values; import price; world export price; reserve-import ratio; Imports; Debt service ratios; Europe
Download PDF (3.8 MB)

Abstract

For several years, the discussion of policy issues in the World Economic Outlook has been based on alternative medium-term scenarios. The scenarios help to identify and determine the implications of tensions and imbalances in the medium-term projections for key economic indicators, such as output, prices, and current accounts. In addition, they provide a vehicle for examining the sensitivity of developments in various groups of countries to changes in the global environment caused, for example, by shifts in economic policies in larger countries.

For several years, the discussion of policy issues in the World Economic Outlook has been based on alternative medium-term scenarios. The scenarios help to identify and determine the implications of tensions and imbalances in the medium-term projections for key economic indicators, such as output, prices, and current accounts. In addition, they provide a vehicle for examining the sensitivity of developments in various groups of countries to changes in the global environment caused, for example, by shifts in economic policies in larger countries.

Because of the increasing importance which is now attached to scenarios, the Research Department’s modeling effort has been increased considerably. A major aspect of this effort has been the expansion of the Department’s main simulation model, MULTIMOD, to incorporate more detail on the interactions among the seven major industrial countries as well as between those countries and the developing countries.1 The expanded MULTIMOD system, while including developing countries, currently incorporates distinctions between developing countries only according to their net external asset position. However, the developing countries differ in many other respects, including the structure of their external trade, their degree of openness, and the level of their external debt. Economic developments, including responses to external disturbances may be greatly influenced by these factors. In order to enrich our understanding of the responses of different groups of countries to external disturbances, the Research Department has developed a disaggregated model system for those developing countries that are classified as net debtors.2 The new system, which includes models for each of these countries,3 was first used in the April 1998 World Economic Outlook and has been expanded considerably since then. In addition to generating alternative scenarios, these models also play a major role in the continuous updating of the staffs World Economic Outlook projections, for example, following changes in international interest rates or other external environmental variables.

This paper outlines the structure of the developing country model system and describes its major properties. The paper is organized as follows. It first discusses the key features of the individual country models including the linkages to industrial countries; it also describes the specifications of the equations for exports, external financing, and domestic sectors and their significance for the models’ properties. It then describes the structure of the model in detail, presents parameter estimates, and discusses the constraints that were imposed in the estimation process. Finally, it considers the simulation properties of the model and the channels through which external shocks are transmitted to different groups of developing countries.

Developing Country Model System

Main Features of the Model

The individual country models describe the behavior of broad aggregates, such as GNP, exports, imports, the terms of trade, and capital flows and are intended primarily to be used for forward-looking simulations rather than tracking historical data. A common model structure is applied for each individual net debtor country, and country results are aggregated to generate the various regional and functional groups used for analytical purposes in the World Economic Outlook. Differences in economic behavior among countries are reflected in country-specific behavioral and structural parameters.

A number of considerations were important in specifying the structure of the model. While it was recognized that model specifications tailored to individual countries would best capture major differences across countries, given the objective of focusing on aggregates for groups of countries, such an approach was viewed as unnecessary. At the same time it was essential to choose a specification that would be applicable to countries with very different economic structures and which could reflect those differences in a parsimonious way.

There are several types of model structure that could have been used for the system.4 One possibility would have been to adapt the structure typically used in many models of industrial countries (Bryant, Henderson. Holtham, Hooper, and Symansky (1988)).5 These models, which assume that each country is specialized in production, capture major linkages between countries through their trade in goods and financial assets. Furthermore, they can be adapted to allow for a variety of assumptions about the degree of substitution among commodities and financial assets and can include many structural differences across countries. Unfortunately, however, the highly aggregative structure of such models means that they are not well suited to analyze the resource allocation and supply problems faced by developing countries. Alternatively, a highly disaggregated multi-sector model could have been specified. Such models assign an important role to resource allocation and relative price effects, and can easily be extended to incorporate the types of supply considerations that are important for developing countries (Taylor (1983)). These models, however, are very intensive in their data requirements; the information base available for many developing countries is frequently insufficient. Moreover, multi-sector models tend to be better suited to microeconomic issues than to those of macroeconomic interdependence.6

The approach adopted in the developing country models presented here is a compromise between the highly aggregated structure of typical industrial country macroeconomic models and the disaggregated general equilibrium approach often used for microeconomic analysis. The approach is based on a disaggregation of each country’s output into two classes of goods: internationally tradable goods and nontradable goods. Structures of this kind have been used extensively in analytical work on developing countries; a similar setup is used for the net debtor developing country block of MULTIMOD.7

In addition to the breakdown between tradable and nontradable goods, the specification adopted here includes equations for both export volumes and prices, which allows for differences in the relative importance of manufactures, non-oil primary goods, and fuel. Allowance also is made for differences in price behavior among broad groups of non-oil primary commodities through separate equations for the prices of foodstuffs, beverages, raw materials, and metals and minerals. Moreover, the equations for import volumes and prices depend directly on the financing available to each country; import price behavior also reflects the commodity composition of imports.

On the domestic side, the country models include equations for both output and prices of nontradable goods. In addition, total demand is disaggregated into consumption and investment, and investment is assumed to augment capacity in the tradable and nontradable sectors.8 The models allow the prices of nontraded goods to be predetermined in the short run, with output determined by demand; their evolution is assumed to depend on relative demand pressure in the nontradable goods sector. Finally, the structure includes equations that allocate the capital stock between tradables and nontradables and thereby determine the long-run supply of each category of output.

As regards the balance of payments, the models include equations for official and private transfers, investment income payments and receipts, and exports and imports of non-factor services. The capital account is disaggregated into net lending (or borrowing), non-debt creating capital flows, and changes in international reserves. The models also include identities that relate the net international asset (investment) position to balance of payments flows and exchange rate valuation effects.

As noted above, differentiation among countries is introduced through the models’ behavioral and structural parameters. The behavioral parameters include income and price elasticities of export demand, marginal propensities to spend, and measures of the sensitivity of capital flows to interest rates. Behavioral parameters are estimated using time series data for each country or are obtained from existing studies. The structural parameters include the shares of various goods in countries’ exports and imports, and export market shares. Such parameters capture many of the differences in economic structure among countries. These structural differences are also captured in the procedures used to aggregate country results, where differences in countries’ relative sizes as well as in their degrees of openness are reflected directly in group totals.

Given the questionable quality of data for many smaller developing countries, and the limited number of data points available,9 several of the models’ behavioral parameters were estimated subject to constraints determined by the structural differences between countries. This procedure helps to ensure that the models’ results reflect such factors as the commodity composition of exports, the geographical distribution of exports, and the degree of indebtedness, rather than differences stemming from poorly determined behavioral parameters. The approach is consistent with the emphasis placed in the models on differences in responses to external disturbances among countries. Given the time period over which most equations were estimated (1973—88), and the economic difficulties which many countries encountered during this period, historical relationships may not provide much guidance to future behavior. This is an additional reason for imposing constraints on the behavioral parameters. Particularly in the case of countries that encountered debt-servicing difficulties, a number of parameters may well change in the period ahead as a result of the ongoing efforts to improve economic policies. As such, parameters were obtained more with a view to generating reasonable simulation results rather than to track history closely.

The international transmission mechanisms incorporated in the model can be briefly summarized. Disturbances in industrial countries are transmitted to developing countries through economic activity, export and import prices, interest rates, and capital flows.10 Changes in economic activity in the industrial countries have direct effects on developing country exports that differ according to the commodity and geographic composition of trade and domestic supply conditions in each country. Variations in industrial country growth also may lead to changes in net transfer payments, in official development assistance, and in flows of non-factor services. The effects of changes in export and import prices depend on the structure of trade as well as the behavioral characteristics of each country model. In the current specification, changes in exchange rates influence mainly the valuation of external debt according to shares in dollar and non-dollar currencies.11

Increases in interest rates directly affect net factor payments, the size of the effects depending on the level and maturity structure of each country’s external debt. Interest rates may also influence capital flows of countries that are creditworthy and are able to borrow freely on international capital markets.

Finally, the models allow for linkages with industrial countries through direct investment and capital flows. The willingness of foreign investors to lend, or the willingness of developing countries to borrow, influences countries’ overall availability of external funds. The amount of external financing in turn affects the level of demand that can be sustained in developing countries, and also plays an important role in determining the amount of domestic investment these countries are able to undertake. Together with export earnings, the amount of financing also determines countries’ imports from (and hence the exports of) the rest of the world.

Specifications for Exports, External Financing, and Domestic Sectors

Specifications for Exports

Exports provide a major link between developing and industrial countries. The approach adopted here assumes that non-oil export volumes12 (primary products plus manufactures) are a function of demand conditions in partner countries and supply conditions at home. The variables representing external demand conditions are income in partner countries and external competitiveness. The export supply function includes the real exchange rate and the capital stock in the tradable goods sector.

Estimation is based on a reduced form representation of the export demand and supply functions. The use of the reduced form addresses the problem that the structural export equations applicable to exporters of primary products and manufactures are likely to be quite different, and also addresses the difficulties in determining whether structural parameters capture demand or supply factors.

The nature of this problem can be illustrated by considering the very different market conditions that are likely to be faced by exporters of primary products and manufactures in the short term (Figure 1) and how their structural export equations might differ.

Figure 1.
Figure 1.

Examples of Different Market Conditions Faced by Developing Country Exporters

Differences in market conditions faced by exporters can be considered in terms of two polar cases: a country exporting homogeneous primary goods and one that exports differentiated manufactured goods. In the former case (shown in the top panel of Figure 1), the country is usually a price taker in world markets and probably faces an infinitely elastic (horizontal) demand curve. Furthermore, given the lags that are normally involved in expanding the production of many primary products, the short-run supply curve may be relatively steep.13 Under these conditions, shifts in international demand would be reflected predominantly in export prices rather than quantities and would trace out a short-run supply curve for exports.

At the other extreme (shown in the bottom panel of Figure 1) is the case of a country producing differentiated manufactured products and facing a relatively inelastic (steep) demand curve. In contrast to the primary product exporter described above, the lags in expanding supply could be fairly short and the short-run supply curve relatively flat. In this case, shifts in international demand would be reflected in the quantity of exports rather than their price.

In practice, of course, countries differ markedly in the composition of their trade between manufactured and primary product exports, which suggests that market conditions may also be quite different. This is evident from Table 1, which shows the commodity composition of merchandise trade across regions. Because of these differences, export equations relevant to specific market conditions could be estimated only if countries can be categorized into those for whom short-run exports are mainly determined by demand or supply conditions. This problem is avoided by estimating reduced-form equations and then determining elasticities of demand or supply, depending on the category of the country. The reduced-form function is applicable regardless of whether a country is a price setter or taker in export markets.14

Table 1.

Net Debtor Developing Countries: Commodity Composition of Merchandise Trade, 19851

(Percent shares in total trade)

article image
Source: World Bank Trade System; and Fund staff estimates.

Group totals are derived as weighted averages of the trade composition of individual countries; weights are based on 1985 values of merchandise exports and imports.

External Financing Specifications

The terms on which external financing can be obtained are important for the linkages between developing and industrial countries, and for countries’ responses to external developments. Developing countries may seek to obtain external financing for a variety of reasons, the most important of which is the desire to supplement domestic resources in order to finance investment in pursuit of economic development and growth. In addition, a high level of external financing can be used to smooth demand when a country experiences shortfalls (or windfall gains) in export earnings as a result, for example, of sharp movements in the terms of trade.

In recent years, the flow of external financing to developing countries has depended to a large extent on their ability to borrow in international capital markets. Many countries have been faced with increasing debt burdens but have lacked the necessary resources to meet their debt-servicing obligations. In response to this deterioration in countries’ debt-servicing capacity and, hence, creditworthiness, foreign creditors have sharply curtailed the availability of additional loans (Table 2). Consequently, for countries that have been facing severe debt-servicing problems, the amount of financing has been determined by foreign creditors. On the other hand, countries that have pursued sound macroeconomic performances and maintained creditworthiness have faced these constraints to a much smaller extent.

Table 2.

Net Debtor Developing Countries: External Financing and Debt, 1973—88

article image
article image
Source: World Economic Outlook, April 1989.

In percent of exports of goods and services.

Ratio of new lending over last period’s stock of debt.

Balance of goods and services, with signs reversed.

For the purpose of modeling the external financing behavior of developing countries, countries are divided into two regimes: an unconstrained regime in which external financing is determined by demand in the home country; and a constrained regime in which foreign creditors limit the amount of external funding that a country obtains.15

(1) Specification for countries that are not finance constrained

Countries that are not finance constrained are assumed to be able to borrow as much as they desire at prevailing interest rates. Several variables are likely to be important determinants of the demand for lending. These include the cost of borrowing (real interest rates), the growth and variability of exports, the import needs (represented by the ratio of imports to GDP), and the level of output.16 Owing to the limited availability and doubtful quality of some of the data, the determinants of the demand for external financing have been limited to the real interest rate, changes in a country’s terms of trade, and the level of output. These factors and their significance for external borrowing are discussed below.

(2) Specification for finance-constrained countries

When a country is finance constrained, the amount it can borrow is assumed to be determined by the amount of lending foreign creditors are willing to supply.17 This in turn is dependent on creditors’ perceptions of a country’s creditworthiness. The major determinants of a country’s creditworthiness include its export performance, the soundness of its domestic policies, its rate of economic growth, its debt-service ratio, and its existing exposure to banks.

In the present model, a persistently high ratio of debt to exports is assumed to provide an indication of a country’s creditworthiness and is used to determine whether the country is likely to be subject to a finance constraint.18 Taking the above financing characteristics into account, the model imposes an exogenous finance constraint on countries that have been showing high debt-export ratios over the preceding three years.19 Specifically, if a country’s debt-export ratio over the last three years has averaged 200 percent or more, the amount of financing it can obtain becomes limited. This amount is computed exogenously and is assumed to correspond to the average level of financing that the country in question received from 1982 to 1988 (during which period new financing fell considerably).20 Given this supply constraint together with the demand for financing equation, the models allow for shifts in borrowing behavior if a country’s debt-export ratio over the last three years declines below 200 percent. At that point, the external finance equation switches back to the unconstrained regime and financing becomes a function of the determinants of demand for external funds as specified above.

Specifications of Domestic Sector

The linkages from external financing to the rate of growth of output in the nontradable sector are of key importance for much of the analysis of trends in developing countries in the World Economic Outlook. Following sharp reductions in external lending in the early 1980s, output growth in many countries fell sharply and has not recovered to rates achieved in the 1970s. Clearly, the model must be able to account for such effects as well as for the implications of possible future improvements in financing conditions.

The linkages from external financing to the nontradable sector operate through their influence on the demand and supply of nontradable goods and services. On the demand side, a decline in external Financing will induce a reduction of imports: the effects on activity will depend on the channels through which imports are reduced.21 If imports are reduced by curtailing spending, the demand for non-tradables would be likely to fall which would result in a reduction in their relative price. If prices are slow to adjust, either because they are administered or because of wage contracts, the result would be a temporary decline in growth. These effects are captured by including external borrowing in the demand function for nontradables and by allowing for a Phillips curve type of response of nontradable prices. If the coefficient on net lending is positive, a drop in lending will reduce the demand for nontradables and temporarily lower output.

On the supply side, reductions in net lending from abroad may affect growth through two channels.22 First, reduced lending may lower investment in the nontradable sector. Such an effect works with significant lags but would eventually tend to lower potential and actual output. Second, if accompanied by reductions of intermediate imports, a reduction of lending might directly limit productive capacity.23 The significance of this effect, discussed by Goldstein and Khan (1982), can be illustrated by noting that if there were no substitutability between intermediate imports and home goods, imports clearly would become a binding constraint on growth. In the more general case of limited substitution possibilities, lower imports could temporarily reduce supply until the production of substitutes could he enhanced.

Detailed Specifications and Parameter Estimates

The parameters for the behavioral equations were obtained from econometric estimation and from other empirical studies carried out within and outside the Fund. In general, the equations were estimated over the period 1973–1988 using ordinary least squares. However, as discussed above, constraints were imposed on the range of allowable parameter values so as to produce stable models24 and to ensure a degree of consistency with MULTIMOD. Under these conditions, the models do not necessarily track history closely but are intended to provide broad quantitative guidance of the impact of various exogenous shocks that may occur over the projection period. Parameter estimates are presented for country groups classified either by region or by principal export commodity, rather than for individual countries even though it is the individual country parameters that appear in the models. Unless indicated otherwise, the parameter presented for each group is a weighted average of the parameters for the countries in the group.25 The values of the dependent variable for countries in the group are used as weights.26

The presentation of the equations is organized into four sectors: trade and current account flows; capital account and reserve flows, including net external lending; debt and reserve stocks, including interest payments and valuation effects; and domestic sectors, including price determination.

Trade and Current Account Flows

The current account balance is disaggregated into the exports and imports of goods and non-factor services, net transfer receipts, and net investment income flows:

Current account balance =

Exports of goods and non-factor services less

Imports of goods and non-factor services plus

Net private transfers plus net official transfers plus

Net investment income receipts

Exports of Goods and Non-Factor Services

Exports of goods and non-factor services are disaggregated into oil, non-oil, and non-factor services. The specifications for non-oil exports apply to exporters of manufactures and to exporters of non-oil primary commodities.27 The structural export equations are as follows: export demand is assumed to depend on export prices relative to world prices (competitiveness), and on income in partner countries (income variable); the supply of exports is assumed to depend on export prices relative to the prices of nontradable goods (the real exchange rate), and on the capital stock in the tradable goods sector.

The structural non-oil export demand and supply functions are given by:

Non-oil export demand = F(Export price/ world export price, world income). ( 1 )
Non-oil export supply = F(Export price/ domestic price of nontradables, capital stock). ( 2 )

The equations include two variables capturing structural features of a country’s exports: the world export price is a dollar index of the prices of manufactured and non-oil primary commodities, with weights reflecting their shares in each country’s exports;28 world income is an index of real GNP/GDP in partner countries, with weights reflecting the market shares shown in Table 3.

Table 3.

Shares of Non-Oil Export Markets, 19851

(In percent)

article image
Source: World Economic Outlook, April 1989.

Weighted by 1985 value of non-oil exports.

Includes developing countries that are net external creditors.

Improvements in competitiveness or increases in real output in partner countries are expected to raise the demand for a country’s exports. A rise in the domestic capital stock, or an increase in the prices of exportables relative to domestic goods, would be expected to increase export supply.

For reasons discussed above, it was decided to estimate reduced-form equations for export volumes and prices rather than structural equations. Summary results from the estimation of the reduced forms implied by equations (1) and (2) are shown in Tables 4 and 5.29

Table 4.

Parameter Estimates for Non-Oil Export Volumes

article image
Table 5.

Parameter Estimates for Non-Oil Export Prices

article image

Four features of the estimates recorded in Tables 4 and 5 stand out.30 First, income in partner countries has a large impact on non-oil export volumes in all groups. The average income elasticity is 1.6 and it ranges from 2 ½ in Asia to ½ in Africa. Moreover, the elasticity is higher for exporters of manufactures than for other exporters. The estimates are comparable to those in other studies (see Goldstein and Khan (1982) and Goldsbrough and Zaidi (1986)) and confirm that developing country non-oil export volumes are quite sensitive to income in partner countries. Conversely, a strong effect of foreign income on dollar export prices was not found for any group mainly because foreign demand conditions are largely subsumed into the world price variable, particularly for commodities. The average elasticity of non-oil export prices with respect to partner country income is only 0.04 (Table 5).

Second, for most groups the estimates suggest that the capital stock influences non-oil export volumes. For primary product exporters, the elasticity is 0.6, which suggests that a 10 percent increase in the capital stock would raise export volumes by 6 percent. Regarding the impact of capacity on prices, the estimates point to a relatively small elasticity of -0.06 which implies that a 10 percent increase in the capital stock would only reduce non-oil export prices by about ½ of t percent. A small elasticity is expected for price-taking exporters of primary products, but is surprising for manufactured exporters. In the latter case, an expansion of capacity would be expected to have large implications for competitiveness.31

Third, the average elasticity of export volumes with respect to relative prices averages –0.25, implying that a 10 percent rise in relative prices would reduce export volumes by 2.5 percent. Finally, the average elasticity of export prices with respect to world prices is estimated at 0.8; as was to be expected, the elasticity is higher for primary product exporters than for exporters of manufactures.

The modeling of oil exports presents particular difficulties. In accordance with the approach adopted in MULTIMOD, it is assumed that in OPEC members set the world price of oil in real terms and allow the quantity of their exports to be determined by demand.32 Oil export volumes of the net debtor countries are assumed to be exogenous, and are allocated across countries according to historical relationships.

Oil export volume = F(Total oil export of net debtor countries). ( 3 )

Given oil export volumes, the dollar export price of oil in each individual country is assumed to follow the world price of oil:

Oil export price = F(World oil price). ( 4 )

The parameters for equations (3) and (4) are shown in Tables 6 and 7. As can be seen, each group’s oil export volumes move closely with total oil exports across all groups, A close relationship also holds for export prices. The estimates suggest that on average 90 percent of any increase in world oil prices will be passed into higher export prices within a year. By the second year, the pass-through is complete.

Table 6.

Parameter Estimates for Oil Export Volumes

article image
Table 7.

Parameter Estimates for Oil Export Prices

article image

The long-run (two-year) elasticity is equal to unity.

Exports of non-factor services comprise receipts from tourism, banking, and other services. These receipts, measured in dollars, are assumed to depend on the level of dollar GDP/GNP in industrial countries as given by equation (5). The real value of non-factor services is determined using the dollar GNP deflator for industrial countries.33

Exports of F(Industrial country non-factor = current dollar services GNP/GDP). ( 5 )

The parameters for equation (5) suggest an average elasticity of non-factor service exports with respect to industrial country GNP of 1 (Table 8). The elasticity implies that a 10 percent increase in industrial country GNP would raise exports of non-factor services by 10 percent. The elasticity ranges from 1.3 in Asia to 0.7 in Africa.

Table 8.

Parameter Estimates for Exports of Non-Factor Services

article image

As noted, the models include behavioral equations for non-oil export prices which depend on developments in world prices of manufactures and of non-oil primary commodities. For each country the weights of the export price indices reflect the shares of manufactures and non-oil primary products in 1985 exports. The non-oil primary commodity export price depends on developments in the prices of food, beverages, raw materials, and metals and minerals. The world price of manufactures is exogenous but the models include equations for the prices of food, beverages, raw materials, and metals and minerals: these are related to economic activity in industrial countries and to an exogenous world price of non-oil primary commodities.

Imports of Goods and Non-Factor Services

Imports of goods and non-factor services are largely determined by the amount of financing available to each country, where foreign exchange availability is derived from export earnings, transfer receipts, net capital flows, and changes in reserves. The specification of imports is based on the assumption that when there are changes in external financing conditions countries will respond by changing their imports.34 This is consistent with the experience of countries that have become finance constrained and have dramatically and quickly reduced their imports.35

The specifications are given by equations (6) through (8) which describe the determination of total imports of goods and non-factor services in terms of available financing (equation (6)): the division of imports between merchandise trade and non-factor services (equation (7)); and the breakdown of merchandise imports between oil and non-oil products (equation (8)).

Imports of goods and non-factor services = Exports of goods and non-factor services p ⁢ l ⁢ u ⁢ s Net transferreceipts p ⁢ l ⁢ u ⁢ s Net investment income receipts p ⁢ l ⁢ u ⁢ s Net capital inflows l ⁢ e ⁢ s ⁢ s Accumulation of international reserves. ( 6 )
Merchandise imports = F(Imports goods and non-factor services). Non-factor service Imports = F( Imports goods and non-factor services). ( 7 )
Non-oil imports = F(Merchandise imports). Oil imports = F(Merchandise imports). ( 8 )

Finally, the models include equations for merchandise import prices which are used to deflate merchandise import values to obtain real magnitudes. Imports of non-factor services are deflated using the dollar GNP deflator for industrial countries. Equation (9) specifies that the dollar prices of non-oil imports depend on indices of the world prices of manufactures and non-oil primary commodities, with weights reflecting 1985 import shares. Equation (10) relates oil import prices to the world oil price.

Non-oil import price = F(World price of mnufac- tures, world price of non- oil primary commodu- ties). ( 9 )
Oil import price = F(World oil price). ( 10 )

The parameters for equation (10) suggest that, on average, almost 90 percent of any increase in “world”‘ oil prices is passed into higher import prices within a year (Table 9).

Table 9.

Parameter Estimates for Oil Import Prices

article image

The long-run (two-year) elasticities equal unity.

Private and Official Transfers

The importance of private and official transfers differs considerably among developing countries (Goldsbrough and Zaidi (1986)). Private transfers mainly comprise remittances from industrial countries (in Europe) and the high income oil exporters (in the Middle East).36 These remittances are likely to be closely linked to some measure of economic activity in the host country. Official transfers include bilateral and multilateral development assistance. The specifications adopted here relate private transfers to Middle East oil export earnings and to GDP/GNP in the industrial countries of Europe. Official transfers are assumed to be a function of official development assistance.37

Private transfers =   F(Export earnings of Middle East oil exporters, exporters, European GNP/GDP) ( 11 )
Official transfers = F(Official development assistance). (12)

The parameters for equations (11) and (12) are shown in Tables 10 and 11. Estimates suggest that on average increases in Middle East export earnings of $1,000 would raise private transfers to each net debtor country by $0.81; a comparable increase in European GNP/GDP would raise private transfers by $0.07. The effect varies across groups: transfers to the non-oil Middle East, Europe, and Asia are most closely linked to Middle East oil export earnings: service and remittance countries receive the largest in crease in transfers as a result of higher growth in Europe. As regards official transfers, the estimates suggest that an increase in official development assistance of $1,000 would raise official transfers to each net debtor country by $4.80. On a regional basis, the largest recipient of the increased aid would be the non-oil Middle East.

Table 10.

Parameter Estimates for Private Transfers1

article image

Coefficients show the dollar impact on transfers of a $1,000 increase in a given variable.

Table 11.

Parameter Estimates for Official Transfers, Net

article image

The coefficients show the dollar impact of a $1,000 increase in ODA.

Net Investment Income

Investment income credits depend on a country’s stock of foreign assets and their rate of return. The equations relate investment income credits to stocks of international reserves multiplied by the LIBOR on six-month dollar deposits; however, they are allowed to vary more than proportionately either to reflect the earning of income at interest rates different from the dollar LIBOR or assets in addition to official reserves.38

Total investment income debits as defined in the WEO data base comprise payments on direct foreign investments and “other” investment income debits. In the absence of information on past direct investments, foreign direct investment income debits are assumed to be a function of domestic output, “Other” investment income debits comprise (scheduled) interest payments on external debt and are assumed to depend on debt-service payments as given by the debt block of the model. However, allowance is made for recorded payments to differ from those given in the debt block owing to data problems and arrears accumulation.

The investment income equations are as follows:

Investment income credits = F(LIBOR * international reserves). ( 13 )
Investment income debits = F(Domestic GNP in dollars ). ( 14 )
“ ⁢ Other ” ⁢ i nvestment income debits = F(Debt-service payments). ( 15 )

The parameters for these equations are shown in Tables 12 through 14. The use of reserves in the investment income credits equation does not lead to a biased representation of the behavior of investment income credits: on average, a $1 increase in interest earnings on reserves generates increases in investment income credits of $1.14. In the case of the non-oil Middle East, however, the effect is much larger, reflecting the region’s accumulated external assets. The investment income debits parameters suggest that an increase in domestic GDP of $1,000 on average is associated with an outflow of $16, with the figure for Asia being higher than average. Finally, the estimates for “other” investment income debits, which provide a link with the interest payments calculated in the debt block of the model (see below), suggest that there is a close relationship between the two measures, at least at the margin.39

Table 12.

Parameter Estimates for Investment Income Credits

article image

Coefficients show the impact of a $1 increase in interest earnings on reserves.

Capital Account and Reserve Flows

The capital account is disaggregated into three components: non-debt creating capital flows, net external borrowing, defined as new borrowing less amortization, and the accumulation of international reserves.40

Table 13.

Parameter Estimates for Investment Income Debits

article image

Coefficients show the impact of a $1,000 increase in domestic economic activity.

Table 14.

Parameter Estimates for Other Investment Income Debits

article image

Coefficients show the impact of a $1 increase in interest payments.

Non-Debt Creating Capital Flows

Non-debt creating capital flows comprise direct foreign investments, new allocations of SDRs valuation adjustments, and a balancing item necessary to reconcile the inevitable discrepancies that exist between the current and capital accounts of the balance of payments. For simplicity, these flows are modeled as if they consist exclusively of direct investments and are assumed to depend on the factors that would influence such investments; these include changes in income in the recipient country and real interest rates. However, since it proved extremely difficult to identify a role for the interest rate, this term was excluded from the equation. The specification is given by:

Non-debt creating = F(Change in domestic capital flows GDP). (16)

The parameters for this equation are shown in Table 15 and suggest that on average an increase in domestic GDP of $1,000 would be associated with net non-debt creating capital outflows of $30.

Table 15.

Parameter Estimates for Non-Debt Creating Capital Flows

article image

Coefficients show the impact of a $1,000 increase in domestic output.

Net External Borrowing

Net borrowing is the difference between new borrowing and amortization. As discussed above, new borrowing is assumed to be demand determined when a country is not finance constrained, and to be determined by lenders when it is. There are accordingly two net lending equations, one for each regime.

When a country is not finance constrained new borrowing is given by equation (17).

Unconstrained borrowing = F(Real LIBOR, terms of trade changes, domes- tic GNP). ( 17 )

When a country is finance constrained new borrowing is given by equation (18)

Constrained borrowing = Average amount of dollar financing received over the period 1982 − 88. 41 ( 18 )

The unconstrained financing equation was estimated over the period 1973–81 for those countries that are assumed to have become finance constrained in the 1980s; and it was estimated up to 1988 in the case of those which did not become finance constrained.42 Results from the estimation are shown in Table 16. Notwithstanding the large coefficient on the real interest rate,43 which reflects differences in units of measurement, the impact of interest rates on borrowing is relatively small. Instead, an important role is played by changes in the terms of trade (multiplied by baseline imports): the estimates suggest that on average a $1 decline in the purchasing value of exports, owing to a change in the terms of trade, will lead to increased borrowing of $0.36. Conversely, an increase in domestic GDP of $1,000 is found to reduce borrowing by $8. Parameters vary across groups with the terms of trade effect being higher than average for the Western Hemisphere. Of course, to the extent that a country has become finance constrained in the 1980s, new borrowing will be determined by foreign creditors and the unconstrained financing equation will not apply.44

Table 16.

Parameter Estimates for External Borrowing

article image

Coefficients show the impact of a $1,000 increase in domestic income.

Amortization payments depend on the maturity profile of external debt.45 In the absence of direct information on maturities, the specification adopted here simply relates amortization payments to the external debt, lagged one period:

Amortization due = F(Debt stock at the end of previous year). ( 19 )

The reciprocal of the parameter obtained for this equation measures the average maturity structure of external debt, adapted for slippages in countries’ payments schedules.46 Estimates of the equation are given in Table 17 and suggest an implicit (average) maturity structure of debt for the net debtor developing countries of l6½ years; the implicit maturity structures range from 10 years in Asia to 34 years in the Western Hemisphere. These estimates should obviously be treated with caution since they reflect both the maturity structure of debt and the countries’ past record in repaying principal. A country that has accumulated arrears in principal payments will appear incorrectly to have a long maturity structure.

Table 17.

Parameter Estimates for Amortization Due

article image

Reserve Accumulation

The specification of changes in international reserves is based on an assumption that countries seek to maintain a constant long-run ratio of reserves to imports. As given by equation (20), deviations from the long-run or average reserve ratio are assumed to be reduced gradually over time.

Reserve-import ratio = F(Gap between actual and tar- get reserve import ratio). 47 ( 20 )

The parameters for this equation are shown in Table 18 and suggest a relatively slow adjustment of actual reserves to their long-run level across all groups.48

Table 18.

Parameter Estimates for Reserve-Import Ratio

article image

Debt and Reserve Stocks

Changes in stocks of external debt and international reserves are driven by flows in the balance of payments together with exchange rate valuation effects. The specifications assume that the stock of external debt evolves according to equation (21).

Debt, = Last period ’ s stock of debt adjusted for valuation effects, plus new external bor- rowing = ( 1 + e ) * ( 1 − v ) * debt t − 1 + new borrowing t ⁢ . ⁢ ( 21 )

Valuation effects are given by ((1 + e)∗(l – v))∗ debtt-1, where e is the proportional change in the dollar MERM exchange rate from time t – 1 to t, measured so that an increase in e means a faster rate of dollar depreciation; v is the share of debt in non-dollar currencies (see Table 19).

Table 19.

Structure of External Debt

article image
Source: World Economic Outlook data base.

Debt from official creditors as a share of total 1987 value.

Country shares weighted by the 1987 value of debt.

According to equation (21), the stock of debt valued in dollars at time t is equal to new borrowing plus the effects on existing debt of changes in the dollar exchange rate. Other things equal, a depreciation of the dollar raises the dollar value of debt denominated in non-dollar currencies (Table 19).49

For simplicity, all reserves are assumed to be in dollars, giving rise to the simple form for the reserves equation (Equation 22).

Reserves = Reserves last period p ⁢ l ⁢ u ⁢ s reserve flows. ( 22 )

Finally, interest payments on the outstanding debt are disaggregated between payments on debt contracted at variable versus fixed interest rates. A temporary (one year) increase in interest rates is assumed to raise interest payments only on debt at variable interest rates; a permanent increase is assumed to eventually raise interest payments on all debt, as maturing debt is refinanced (or rescheduled) at higher interest rates. Short-run effects are captured by multiplying the share of debt at variable interest rates by the dollar LIBOR. Longer-term effects are captured by multiplying a weighted average of interest rates by the share of debt at fixed interest rates. The formulation is given by equation (23) where v is the share of debt at variable interest rates, m is the average maturity structure, and w(L) is a polynomial in the lag operator reflecting the maturity structure. The coefficients in w(L) sum to unity implying that an increase in interest rates will eventually raise interest payments on all debt.50

Interest payments duet t = v * LIBOR, + (1 − v )* w ⁡ ( L ) * LIBOR t − m + constant. ( 23 )

An implication of equation (23) is that interest payments are related directly to the dollar LIBOR. The possibility of a risk premium is reflected in a constant term in the estimated equation.

Domestic Sectors and Price Determination

The demand and supply of nontradables are influenced by developments in the domestic sectors. The specifications are based on the assumption that the (nominal) price of nontradables is predetermined in the short term with supply responding to meet demand. Prices are assumed to adjust in response to any difference between the demand and supply of nontradables.

The demand for nontradables comprises consumption and investment, and is modeled as a function of real income, the relative price of tradable and nontradable goods, and external borrowing.51 The demand function is given by equation (24) where real income is equal to real GNP (GDP less net factor income payments abroad) adjusted for the terms of trade; real external borrowing is the growth of real debt; and the real exchange rate is the relative price of tradables and nontradables.

Nontradables F(Real income, demand = real borrowing, real exchange rate). (24)

The supply specifications are given by equations (25) and (26) which describe, respectively, the pricing behavior of nontradables and the long-run supply of these goods.

Change in nontradables = F(Excess demand for prices nontradables). ⁢ (25)
Long-run supply of = F(Real exchange rate, nontradable goods capital stock). ⁢ (26)

These equations are based on the assumption that the long-run supply of nontradables will fall as the relative price of nontradables decreases and rise as the capital stock in the nontradable sector is augmented. Estimation is based on the substitution of equation (26) into equation (25)—giving a price equation—and the estimation of equation (24). The estimated parameters in equation (24) suggest that the marginal propensity to spend on home goods averages ½ and that real external borrowing has a positive impact on the demand for home goods (Table 20).52

Table 20.

Parameter Estimates for Real Spending on Home Goods

article image

Numerous difficulties were encountered in determining an appropriate price equation; the parameters for the equation included in the current model are shown in Table 21. They suggest a relatively weak link between domestic and foreign prices and a small effect of the capital stock on prices. In addition, the estimates do not imply a very large impact of demand on prices.53

Table 21.

Parameter Estimates for Inflation in Home Prices

article image

Finally, given output in the nontradable sector as determined by equation (24), output is allocated exogenously between consumption and investment spending. This formulation was adopted to allow the choice between consumption and investment to be specified in order to reflect alternative assumptions about intertemporal choices.54

In addition to the behavioral equations discussed in the foregoing, the models include a series of identities which describe components of the balance of payments and determine aggregate output from production of tradables and nontradables.55

Simulation Experiments

This section considers the sensitivity of the models to various exogenous disturbances and discusses the major channels through which disturbances affect different groups of developing countries. Two kinds of simulations are discussed. The first are partial equilibrium sensitivity analyses which examine the impact of changes in the models’ exogenous variables. These simulations do not take into account any relationships among exogenous variables, or feedbacks from developing to industrial countries. In the general equilibrium simulations, disturbances are generated using MULTIMOD and involve simultaneous changes in several variables that are exogenous.

For both sets of simulations, results for individual countries are aggregated to various regional and analytical groups. As is well known, there is no unambiguous way to aggregate country results to obtain group totals: different procedures can lead to different results. For example, each series could be weighted by its share in the group total for that series; or, it could be weighted by country shares in the group’s GNP. In the former case, large differences in economic structure between countries will influence group totals. Alternatively, if series are weighted by GDP, many structural differences are concealed and group totals are dominated by countries with large shares of the groups’ GNP.

The aggregate results reported below, which are expressed as deviations from the baseline, are based on weighting estimates for individual countries by shares in the group total for each series.56 This procedure, which is used in the World Economic Outlook, implies that structural differences among countries in each group are directly reflected in the aggregate results.

Partial Equilibrium Simulations

Four sensitivity analyses were carried out: (a) faster industrial country real GNP/GDP growth of 1 percentage point a year; (b) a rise in the LIBOR interest rate of 1 percentage point; (c) a 10 percent rise in oil prices; and (d) increases in non-oil primary product prices of 10 percent. In each case, simulations were run over the period 1989–94, and results are presented as deviations from the baseline projections reported in the April 1989 World Economic Outlook.

Faster Industrial Country Growth

In this simulation, a sustained rise in the rate of growth of GNP in the industrial countries of 1 percentage point a year leads to increased demand for developing country exports, a strengthening of import growth, and a significant increase in real GDP in the developing world (Table 22). On average, the growth of developing country exports would be raised by 1 ¾ percentage points a year and the growth rate of real GDP would rise by 2/3 of 1 percentage point. By 1994, these effects would cumulate to large increases in the levels of exports and GDP relative to the baseline. Across groups, exporters of manufactures would experience a particularly large increase in export volumes; the expansion for fuel exporters would be relatively small as suggested by the export elasticities listed in Table 4.

Table 22.

Medium-Term Implications of Faster Industrial Country Real GNP/GDP Growth of 1 Percentage Point a Year, 1989–94

(Difference from the reference scenario in percentage points)

article image
article image

Ratios to exports of goods and services.

Includes Taiwan Province of China.

Increases in real GDP in this simulation would reflect the expansion of export volumes and the responses of domestic spending to higher incomes. Group totals reflect the economic structure of each region, including the share of exports in GDP. Regions, such as the Western Hemisphere, which include some large but relatively closed economies, would experience a modest increase in real GDP even though these countries would also register a large increase in their export volumes.

Increase in LIBOR Interest Rate

An increase in the LIBOR of 1 percentage point would raise countries’ scheduled debt-service payments which, in the absence of new financing, would lead to an immediate compression of import volume (Table 23). The average effects would be relatively modest over the simulation period. Nevertheless, in the Western Hemisphere and Europe, where a large proportion of external debt is at variable interest rates, the assumed increase in debt-service payments would require immediate reductions in import volumes; there would also be downward pressure on output in these regions as lower real disposable incomes reduce the demand for home goods. In contrast to the situation in most of the highly indebted developing countries, the newly industrialized economies in Asia would be likely to raise their spending in response to higher interest earnings on their net external assets.57

Table 23.

Medium-Term Implications of an Increase in LIBOR by 1 Percentage Point a Year, 1989–94

(Difference from the reference scenario in percentage points)

article image
article image

Ratios to exports of goods and services.

Includes Taiwan Province of China.

Increase in Oil Prices

The effects of an increase in oil prices depend on whether countries are net oil exporters or importers (Table 24). Since most groups include both importers and exporters of oil the results from this simulation are somewhat complex (see Table 1). Fuel exporters would experience a substantial improvement in their terms of trade which would, in turn, generate higher domestic spending, imports, and real GDP. The exporters of manufactures and primary products, on the other hand, would tend to experience a terms of trade deterioration which would lead to a compression of imports. However, since both of these groups include some fuel exporters, there would be offsetting effects which would depend on the “average” economic structure of the countries included. Overall, real GDP would tend to rise relative to the baseline in those groups that are dominated by fuel exporters and would fall elsewhere. In the Western Hemisphere and Asia, for example, real GDP would tend to rise owing to the large weight of oil exporting countries in both regions.

Table 24.

Medium-Term Implications of a 10 Percent Rise in Oil Prices, 1989–94

(Difference from the reference scenario in percentage points)

article image
article image

Ratios to exports of goods and services.

Includes Taiwan Province of China.

Increase in Non-Oil Commodity Prices

A 10 percent increase in non-oil commodity prices would raise export earnings of a large number of developing countries. However, as a result of very large differences in the commodity composition of both exports and imports across countries, the net effects would vary significantly between groups (Table 25). Those countries that are classified as primary product exporters would enjoy a relatively large terms of trade gain which over time would stimulate investment in the tradable goods sector and thus raise their export volumes. In contrast, countries for which primary commodity exports constitute a small proportion of their total exports would tend to experience a deterioration in their terms of trade. On balance, real GDP tends to rise in most groups but by amounts depending on the magnitudes of the terms of trade and export supply effects.

Table 25.

Medium-Term Implications of an Increase in the World Price of Non-Oil Primary Commodities by 10 Percent, 1989–94

(Difference from the reference scenario in percentage points)

article image
article image

Ratios to exports of goods and services.

Includes Taiwan Province of China.

General Equilibrium Simulations Using MULTIMOD

For many purposes it is necessary to use the developing country model system in conjunction with MULTIMOD to determine the implications for developing countries of policy changes and other disturbances originating in industrial countries. Only two scenarios are discussed here, since detailed results for a number of simulations are included in various editions of the World Economic Outlook. In the first simulation, it is assumed that inflation in industrial countries turns out to be higher than envisaged in the baseline as a result of stronger industrial country demand, and an over-estimation of supply potential in the United States. In the second, it is assumed that the monetary authorities in industrial countries would react to the hypothetical pick-up in inflation by raising interest rates. This would result in a temporary slowing of growth in the industrial countries but would lead to an eventual reduction in their rates of inflation.

Higher Inflation in the Industrial Countries

A pick-up in inflation in industrial countries has very different effects across groups of developing countries (Table 26). Lower industrial country demand would weaken the growth of world trade. This is accompanied by higher cost increases for manufactured and fuel products which would weaken the terms of trade of most developing countries, tending to reduce real export earnings. Furthermore, to the extent that an increase in inflationary pressures in industrial countries would be accompanied by upward pressures on interest rates, countries with external debt at variable interest rates would experience a rise in debt-service payments. On balance, real GDP in most regions tends to fall as a result of weaker terms of trade and declines in export volumes. In all regions, import volumes would tend to decline over time as a consequence of weaker terms of trade and higher interest rates.

Table 26.

Medium-Term Implications of Higher Inflation in Industrial Countries, 1989–92

(Difference from the reference scenario in percentage points)

article image
article image

Ratios to exports of goods and services.

Includes Taiwan Province of China.

Higher Inflation in the Industrial Countries Combined with a Tightening of Monetary Policy

When higher inflation in industrial countries is combined with tightening of monetary policies, the effects on developing countries are unfavorable. The monetary tightening tends to lower industrial country growth in the short run. As a result, demand for developing country exports declines and their terms of trade eventually weaken. Furthermore, in contrast to the first simulation with MULTIMOD, there would be much stronger upward pressure on international interest rates; this would lead to larger increases in debt-service payments, principally in countries with external debt at variable interest rates. As can be seen from Table 27, the Western Hemisphere would be particularly adversely effected by lower exports and a sharp rise in debt-service payments. In the absence of additional financing, there would be sharp compression of spending and imports in this region. In other regions, the effects would also be negative even though some groups would experience a temporary improvement in their terms of trade, partly offsetting the effects of lower export volumes.

Table 27.

Medium-Term Implications of Higher Inflation in Industrial Countries with an Immediate Monetary Tightening, 1989–92

(Difference from the reference scenario in percentage points)

article image
article image

Ratios to exports of goods and services.

Includes Taiwan Province of China.

References

  • Berg, Andrew and Jeffry Sachs, “The Debt Crisis: Structural Explanations of Country Performance,” in NBER Working Paper 2607 (Cambridge, Massachusetts: National Bureau of Economic Research, June 1988).

    • Search Google Scholar
    • Export Citation

  • Bond, Marian E., “Export Demand and Supply for Groups of Non-Oil Developing Countries, “ International Monetary Fund, Staff Papers (Washington), Vol. 32 (March 1985), pp. 56–77.

    • Search Google Scholar
    • Export Citation

  • Bond, Marian E., “An Econometric Study of Primary Commodity Exports from Developing Country Regions to the World,” International Monetary Fund, Staff Papers (Washington), Vol. 34 (June 1987), pp. 191–227.

    • Search Google Scholar
    • Export Citation

  • Bryant, Ralph, Dale Henderson, Gerald Holtham, Peter Hooper, and Steven Symansky (eds.), “Empirical Macroeconomics for Interdependent Economies, “ (Washington: The Brookings Institution, 1988).

    • Search Google Scholar
    • Export Citation

  • Chu, Ke-young, and Thomas K. Morrison, “The 1981–82 Recession and Non-Oil Primary Commodity Prices,” International Monetary Fund, Staff Papers (Washington), Vol. 31 (March 1984), pp. 93–140.

    • Search Google Scholar
    • Export Citation

  • Chu, Ke-young, and Thomas K. Morrison, “World Non-Oil Primary Commodity Markets: A Medium-Term Framework of Analysis,” International Monetary Fund, Staff Papers (Washington), Vol. 33 (March 1986).

    • Search Google Scholar
    • Export Citation

  • Corden, Max W., and Michael P. Dooley, “Issues in the Debt Strategy: An Overview,” (unpublished, Washington: International Monetary Fund, June 6, 1988).

    • Search Google Scholar
    • Export Citation

  • Dornbusch, Rudiger, Open Economy Macroeconomics (New York: Basic Books, 1980).

  • Easterly, William R., “Fiscal Adjustment and Deficit Financing During the Debt Crisis,” Conference paper prepared for Workshop on Dealing with the Debt Crisis, Macroeconomic Adjustment and Growth Division, Country Economics Department, (Washington: World Bank, January 1988).

    • Search Google Scholar
    • Export Citation

  • Eaton, Jonathan, and Lance Taylor, “Developing Country Finance and Debt,” Journal of Development Economics (Amsterdam), Vol. 22 (October 1985), pp. 209–265.

    • Search Google Scholar
    • Export Citation

  • Frenkel, Jacob A., and Assaf Razin, “The Mundell-Fleming Model a Quarter Century Later: A Unified Exposition,” International Monetary Fund, Staff Papers (Washington), Vol. 34 (December 1987), pp. 567–620.

    • Search Google Scholar
    • Export Citation

  • Goldsbrough, David, and Iqbal Zaidi, “Transmission of Economic Influences from Industrial to Developing Countries,” Staff Studies for the World Economic Outlook (Washington: International Monetary Fund, July 1986), pp. 150–195.

    • Search Google Scholar
    • Export Citation

  • Goldstein, Morris, and Mohsin S. Khan, “Effects of Slowdown in Industrial Countries on Growth in Non-Oil Developing Countries,” IMF Occasional Paper 12, (Washington: International Monetary Fund, August 1982).

    • Search Google Scholar
    • Export Citation

  • Haque, Nadeem Ul, Peter Montiel, and Steven Symansky, “A Forward-Looking Macroeconomic Model for a Developing Country,” (unpublished, Washington: International Monetary Fund, 1989).

    • Search Google Scholar
    • Export Citation

  • Masson, Paul, Steven Symansky, Richard Haas, and Michael Dooley, “MULTIMOD: A Multi-Regional Econometric Model,” Staff Studies for the World Economic Outlook (Washington: International Monetary Fund, July 1988), pp. 50–103.

    • Search Google Scholar
    • Export Citation

  • Saunders, Paul, and Andrew Dean, “The International Debt Situation and Linkages Between Developing Countries and the OECD,” Organization of Economic Cooperation and Development, Economics and Statistics Department, (Paris), pp. 155–203.

    • Search Google Scholar
    • Export Citation

  • Suttle, Phil, “Debt Projection Models: A Survey,” Oxford Review of Economic Policy, Vol. 2, No. 1 (1986).

  • Taylor, Lance, Structuralist Macroeconomics, Applicable Models for the Third World (New York: Basic Books, 1983).

1

The structure of MULTIMOD is described in a recent staff study for the World Economic Outlook: Paul Masson, Steven Symansky, Richard Haas, and Michael Dooley, “MULTIMOD—a Multi-Region Econometric Model,” Staff Studies for the World Economic Outlook (Washington, International Monetary Fund, July 1988). The Group of Seven version is to be documented in a forthcoming working paper by Paul Masson, Steven Symansky, and Ralph Tryon, “MULTIMOD MARK II: The Group of Seven Version.”

2

The country groups referred to in this paper are those used in the April 1989 World Economic Outlook in which the previous classification of capital exporting and importing developing countries was replaced by the concepts of net creditor and net debtor countries (see Introduction to Statistical Appendix).

3

The system will be expanded to include the developing countries that are classified as net creditor countries.

4

Taylor (1983) discusses a number of developing country models. See Haque, Montiel, and Symansky (1989) for recent application within the Fund.

5

Such a structure is used for the industrial country block of MULTIMOD. Structures of this kind are often referred to as extensions of the Mundell-Fleming model (Frenkel and Razin (1987)).

6

Structures of this kind have been used extensively for studying resource allocation or sectoral problems in industrial countries. They have been applied to problems as diverse as the adjustment to higher oil prices in resource dependent economies, tax reform and its implications for different sectors of an economy, and the sectoral effects of the European Community’s Common Agricultural Policy.

7

For examples of analytical applications of the tradable/nontradable framework, see Dornbusch (1980).

8

As in MULTIMOD, the models do not al this stage distinguish between private and public expenditures. Subsequently, such a distinction is to be incorporated into the models.

9

Most equations were estimated on annual data for the period 1973 to 1988.

10

When the models are simulated in conjunction with MULTIMOD, feedbacks from developing to industrial countries are taken into account.

11

Subsequently, the models are to be extended to include a monetary side and the determination of exchange rates.

12

Fuel exports are treated separately and are modeled as depending on the total volume of the net debtpr countries’ fuel exports. (See below).

13

The slope of the supply curve is also affected by stocks of inventories. If these are relatively large and responsive to prices, the short-run supply elasticity can be higher.

14

Subsequently, attempts will be made to disaggregate merchandise trade into manufactures and non-oil primary commodities.

15

The sharp distinction between these regimes is blurred when countries resort to exceptional sources of financing such as arrears accumulation. The use of exceptional financing was incorporated in earlier versions of the models but was subsequently excluded because of difficulties in modeling such flows consistently across countries. Such financing has, of course, been of considerable importance since 1982, and its determinants are critical to an understanding of finance flows over this period.

16

The role of these factors is discussed by Eaton and Taylor (1985).

17

The existence of non-price rationing can be explained in a number of ways, including moral hazard and adverse selection considerations. See Eaton and Taylor (1985).

18

Alternatively, indicators such as the debt-service ratio could serve this purpose. In MULTIMOD, the interest payments ratio is used as a proxy for creditworthiness.

19

Alternatively, in forward-looking models such as MULTIMOD, expected future movements in the debt ratio can be allowed to influence current lending.

20

This threshold debt ratio was chosen in order to ensure that those countries that are classified as having experienced debt-servicing difficulties in the World Economic Outlook are defined as being finance constrained from 1982 to 1988.

21

When external financing was abruptly curtailed in 1982–83, countries relied on a wide range of measures to reduce their imports. These included demand contraction, import controls, two-tier foreign exchange markets and, in some cases, foreign exchange controls. Depending on the measures adopted, an incipient excess demand or supply of nontradables could be created. The text focuses on the more realistic case where an incipient excess supply is created.

22

Both effects below apply to the tradable sector and to the growth of exports.

23

Such an effect could be captured by specifying a gross output function including intermediate goods.

24

The small number of observations for each country precludes more sophisticated estimation techniques as does the quality of some of the data. Unfortunately, the small number of observations also implies that outlier observations can have a very large impact on the estimates.

25

Constant terms are not shown since weighted aggregates of such terms contain limited information.

26

Measures of goodness of statistical fit are not provided. This is because the models’ emphasis is on the simulation of future disturbances rather than on tracking history.

27

As discussed above, the market conditions faced by exporters of manufactures are likely to be quite different from those faced by primary product exporters, suggesting that it might be useful to adopt separate specifications for the exports of these goods. However, the data base for the World Economic Outlook does not currently allow such an approach.

28

The non-oil primary commodity price is an index of the prices of four commodity groups: food, beverages, raw materials, and metals and minerals. As discussed below, the weights in this index are country specific and reflect the importance of each commodity group for a country’s exports.

29

The reduced-form equations are found by setting the demand for non-oil exports equal to supply, and solving for equilibrium export prices and volumes.

30

On account of the log-linear form, coefficients can be interpreted as elasticities.

31

For no country do the elasticities with respect to capacity imply immiserizing effects whereby larger capacity would reduce export values on account of their implications for prices.

32

The volume of oil exports is then determined by the demand for oil in industrial and developing countries, with allowance for oil production by non-OPEC countries.

33

Use of the GNP deflator is necessitated by lack of a separate deflator for exports of non-factor services. A similar problem appears for imports of non-factor services.

34

The formulation gives rise to price elasticities of demand for imports that equal (minus) unity. This is because, when financing is provided, any rise in import prices must be offset by an equivalent proportional reduction in import volumes to hold import values constant.

35

Alternatively, arrears in debt-service payments or reserves can slow the process down.

36

Remittances from the United States and Canada will be incorporated into the models in the future to account for private transfers in the Western Hemisphere.

37

In addition, the models include an equation explaining official development assistance in dollar terms as a function of the level of dollar GNP/GDP in industrial countries.

38

The specifications assume that all reserves assets are short term and are held at variable interest rates. The use of the dollar LIBOR as the interest rate applicable to these assets is a simplification.

39

Estimated equations included a constant term, which in several cases was significantly different from zero.

40

Because of discrepancies between balance of payments flows and data on debt stocks, it was necessary to define non-debt creating capital residually to ensure that simulated increments to the various elements of the overall balance of payments sum to zero.

41

This specification was discussed above.

42

As discussed earlier, the constraint leads to those countries classified as having experienced debt-servicing difficulties being finance constrained after 1981.

43

The interest rate is deflated by export prices in the home country.

44

This assumes that borrowers will not run up arrears and force reschedulings.

45

Ideally, a distinction should be made between scheduled and actual amortization payments. At this stage, such a distinction can not reliably be made for many countries.

46

In this framework, arrears and reschedulings of amortization payments can be represented either by the incorporation of dummy variables in the specified equations or by changes in the parameter measuring the average maturity structure of debt.

47

Attempts were made to explain the long-run reserve import ratio in terms of a number of factors including: international interest rates, GDP, and the terms of trade. No systematic relationships were found that were robust across countries.

48

For many countries, reserves were difficult to model. In the case of the newly industrialized economies in Asia, the reserve import ratio appeared non-stationary and had to be constrained to converge to a long-run level. There is some evidence that countries use reserves to smooth demand in the face of temporary shortfalls in export earnings. The result, however, is not robust.

49

It is the intention to modify this equation to permit a disaggregation of non-dollar debt into alternative currencies. In addition, an exchange rate series will be constructed with weights reflecting shares of currencies in external debt. The use of the MERM exchange rate is a simplification that may mismeasure valuation effects to the extent to which it puts inadequate weight on currencies such as the French Franc which are important numeraires for external debt but have a relatively small weight in the index.

50

The values of the weight polynomial are related directly to the (average) maturity structure of a country’s debt. For example, if the average maturity of a country’s debts is three years the w coefficients have the pattern: 1/3,1/3,1/3.

51

As discussed above, external borrowing is included in the demand function to capture the implications of changes in external financing conditions.

52

Owing to data difficulties, it was necessary to estimate the marginal propensity to spend on home goods under very tight constraints. Specifically, it was necessary to constrain marginal propensities to spend to be below unity in several countries.

53

In specifying the price equation, it was necessary to purge the domestic price series of trend inflation which was at very high levels in a number of countries. Implicitly, the equation for domestic inflation includes a trend capturing the path for inflation.

54

An alternative assumption is that the choice between current and future consumption (investment) is determined endogenously by rate of return and cost considerations. It proved very difficult, however, to identify a stable investment function, and there was evidence that the function applicable to the finance-constrained environment of the 1980s was very different to that in the 1970s. The formulation that was eventually adopted, by exogenizing the choice between current and future consumption, allows the effects of alternative policy assumptions to be considered.

55

A listing of the complete model structure, including identities, can be obtained from the authors.

56

Percentage deviations from the baseline for exports, imports, and real GDP are weighted by shares in the respective total for each series. Terms of trade deviations are based on weighting export and import prices by respective export and import shares. Deviations from debt and current account balances are summed across countries.

57

The results for this group reflect the inclusion of Taiwan Province of China.

Within Same Series

Other IMF Content

Other Publishers Content

  • Collapse
  • Expand
Cover World Economic and Financial Surveys
World Economic and Financial Surveys
Staff Studies, 1989
ISBN:
9781451943726
ISSN:
0258-7440
Publication Date:
01 Jan 1989
Publisher:
International Monetary Fund
DOI:
https://doi.org/10.5089/9781451943726.083
Pages:
136
  • Figure 1.
    View in gallery
    Figure 1.

    Examples of Different Market Conditions Faced by Developing Country Exporters

Figure 1.

Examples of Different Market Conditions Faced by Developing Country Exporters