The Macroeconomic Effects of Changes in Barriers to Trade and Capital Flows: A Simulation Analysis

Mohsin S. Khan; Roberto Zahler

doi:10.5089/9781451946895.024.A001

Jump to Content Jump to Main Navigation

The Macroeconomic Effects of Changes in Barriers to Trade and Capital Flows: A Simulation Analysis

Author: Mr. Mohsin S. Khan and Mr. Roberto Zahler¹

¹ 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

Publication Date:: 01 Jan 1983

eISBN:: 9781451946895

ISBN:: 9781451946895

Language:: English

Keywords:: SP; rate of inflation; rate of interest; trade balance; price level; monetary policy; Tariffs; Capital account; Trade balance

Download PDF (2.9 MB)

The issue of opening up domestic economies to the world economy through the liberalization of trade and capital movements—either across the board or selectively—has been of perennial concern to policymakers in the developing world. Recently there has been a renewal of interest in this subject in these countries, including many in Latin America. While the reasons for this heightened interest are varied, three are of principal importance.

Abstract

First, there is the “demonstration effect” imparted by the economic performance of a select group of developing countries, particularly in Southeast Asia, where the growth of trade has played a major role. A number of recently completed studies of individual-country experiences (i.e., Bhagwati and Srinivasan (1979), Keesing (1979), Balassa (1980), Krueger and others (1981)) have shown that, at the broadest level, the countries adopting outward-looking development strategies have fared far better in terms of economic growth, employment, economic efficiency, and adjustment to external shocks than those that have engaged in more inward-looking strategies. The outward-oriented strategies have been typically characterized, inter alia, by the provision of incentives for export production, the encouragement of import competition for most domestically produced goods, and the use of the nominal exchange rate for the maintenance of realistic real exchange rates. At the same time, various developing countries have experienced a relative loss of dynamism in their industrialization processes that are based on import substitution; thus, they have had to consider designing new strategies for the external sector much along the lines adopted by the outward-looking developing economies.

Second, in certain countries there has been a move toward greater stress on the role of market forces in the functioning of the economic system, and this has led to a revival of the propositions associated with the well-known theory of the gains from trade. Briefly, according to this theory, international trade is believed to contribute to the development process in the following ways: trade allows a country to follow the route indicated by the theory of comparative advantage; it offers greater opportunities to exploit economies of scale; it increases the supply capacity of the economy through imports of capital goods, raw materials, and other inputs in the production process; and finally, by providing competition for tradable goods, it is a source of both stimulus and pressure for domestic production and, depending on the exchange rate policy being pursued, can set limits to the domestic inflation rate. In a similar vein, insofar as liberalization of capital movements is concerned, proponents argue that capital flows can increase the supply of financial savings, augment the stock of capital, and induce competition and efficiency in the domestic financial system.

Third, the generally increasing integration of the world economy, in both goods and capital markets, has meant that countries have been drawn into closer international relationships, whether expressly desired or not. This growing interdependence has also contributed to a reorientation of thinking and policies in these countries so that they can adjust to international developments.¹

It can clearly be argued that a combination of these theoretical and empirical factors prompted several Latin American countries to move in the direction of opening up their economies to international competition. The most dramatic shifts in policy occurred in the Southern Cone countries—Argentina, Chile, and Uruguay—although other countries in the region also followed this path, albeit in a more restrained and selective fashion. Uruguay started the process about 1974, and the liberalization reforms were instituted in Argentina and Chile during the years 1975–77. Venezuela moved to rationalize its protection system in 1978–79, and Colombia, Mexico, and Peru also took steps toward freer trade and the liberalizing of capital movements. Moreover, it appears that other countries, both within and outside Latin America, are closely watching and judging these experiments.

Nevertheless, a majority of developing countries are reluctant to embark on a program of liberalizing their trade and exchange systems. If such policies seem so attractive on economic grounds, then it is certainly relevant to ask what the reasons are that underlie this supposed reluctance. The answers, of course, are not clear cut by any means, but several factors bear on this issue.² At the outset it should be recognized that many government policies, including controls on trade and capital flows, may have been implemented with full knowledge of their likely adverse effects on resource allocation and efficiency, as measured by market prices. These policies are, however, regarded as having strong implications from the point of view of equity so that welfare considerations can play a fundamental, if not overriding, role in any decision to alter existing distortions and controls in the foreign sector. The arguments based on externalities that are created by offering incentives to produce domestically rather than to import, and achieving a minimum stage of industrialization, are of no less importance, particularly if it is not known where a country’s dynamic comparative advantage lies. Broadly speaking, there may be both long-run economic and social costs that have to be balanced against the long-run economic benefits to be expected from opening up.

The longer-term factors just discussed can clearly inhibit the opening up of an economy, but perhaps equally important in this context are the short-run and medium-run effects that occur when such a strategy is adopted. Simple casual observation shows that there are serious, even if transitory, costs as a country moves from a relatively closed economy to a more open one. These costs can include losses in output and employment, current account deficits, a fall in fiscal revenues, and increases in external debt. Whatever the long-term effects are in net terms, these short-run costs can provide an effective barrier to a country’s desire to open up its economy. Some consideration has to be given to these potential shorter-term problems as well. The identification of such costs, at both macroeconomic and microeconomic levels, and to determine if the costs are general or country specific, is therefore of utmost importance; the available evidence on this is extremely scanty. Furthermore, if these costs are general in nature, and thus are to be expected whenever a policy of liberalization is adopted, do they depend on the specific type of liberalization strategy chosen? Opening up has not, of course, been uniform with respect to the trade and capital accounts, nor has the speed with which these reforms have been implemented in various countries. For example, Argentina and Uruguay eliminated capital controls first and then proceeded slowly toward goods-market liberalization. On the other hand, Chile chose the opposite sequence of moving first toward the more rapid removal of trade restrictions and then allowing a greater degree of capital mobility. Mexico and Venezuela had had no capital controls to speak of for some time and, therefore, their concentration was exclusively on the trade side. Finally, can the negative aspects associated with opening up be mitigated or minimized by an appropriate mix of different policies? To our knowledge, these issues have not been dealt with in the literature in any generalized and systematic fashion.

Analyzing this latter set of questions—namely, relating to the shorter-term aspects of liberalization policy—is the focus of this study. Such an examination would make it possible to shed some light on whether these short-run effects are significant enough to warrant concern. It should be stressed that no explicit attempt is made here to consider the long-run aspects of liberalization but simply to determine the process of adjustment of the economic system after the new policy has been initiated. While this type of approach does exclude a number of interesting and important issues, it is believed that the exercise does provide useful insights in an area where information is lacking. Furthermore, restricting the analysis to the short term enables one to stay clear of difficult welfare-related issues, which naturally arise when one looks into the long-term consequences of opening up an economy. It was also the express intention to concentrate on the effects of changes in policy on such variables as overall economic activity, prices, the balance of payments, external indebtedness, and interest rates. As such, the exercise is exclusively macroeconomic in nature. It is considered that studying the simultaneous interaction of the main macroeconomic variables and their initial response to policy stimuli generated in the context of removal of restrictions on trade and capital flows is a necessary first step in properly ascertaining the response of the economic system.

The various short-term aspects of opening up are handled within the framework of a dynamic general-equilibrium model constructed for this specific purpose. This approach is more systematic than analyzing the liberalization experiences of individual countries, as, for example, has been done by Wonsewer and Saráchaga (1980), Zahler (1980), Diaz-Alejandro (1981), Gaba (1981), and Mezzera (1981). This model has the capability of explaining the more important macroeconomic variables, and the structure is based on well-grounded economic theory. It has its roots in a variety of theoretical and empirical general-equilibrium models that have been formulated to deal with issues similar to the ones that are of interest here;³ the model also has strong links to monetary-oriented models that have been proposed to analyze short-term stabilization policies.⁴ The model does go well beyond the standard monetary models and contains an explicit, albeit simplified, treatment of the real sector and the interaction of this sector with the monetary sector. Nevertheless, the crucial role played by monetary disequilibrium in the behavior of major macroeconomic variables is stressed in the model, and in principle it remains consistent with the basic long-run tenets of the monetary approach to the balance of payments.⁵ In fact, even though attention is focused exclusively on the short-run path of the economy, considerable care is taken throughout to ensure that long-term conditions that are consistent with general-equilibrium theory are satisfied. However, the dynamic aspects are considered crucial, since the trajectories of the variables are of great interest to policymakers, and it is precisely in this area that there is a serious gap in the literature.⁶

The model is taken to be representative of a typical Latin American economy, and thus does not pretend to be exactly applicable to any single country. For this reason, its structure is quite general in form and aims at covering only the essential features of a broad group of countries. It does not seek to incorporate institutional or other characteristics that would necessarily have to be taken into account to apply the model to any particular country. The economy represented by this model is assumed to be small in relation to the world and to operate under a fixed exchange rate regime.⁷ It is also assumed that the domestic markets are relatively free of direct government control. This last assumption implies that prices are essentially market determined and that financial reforms to liberalize the domestic financial system have already been implemented.

Nevertheless, even though highly aggregated, the model is sufficiently rich and complex to yield the necessary answers to the relevant questions. It is also, as mentioned previously, dynamic in that it allows for delayed response of variables and can trace out the path of adjustment from one equilibrium position to another. It is obvious that questions of transition definitionally imply the study of the time path of the variables after there has been some type of shock to the system. The complex and dynamic structure of the model precludes its solution in any simple analytical fashion. Numerical values for the parameters are required, and these can be either obtained through econometric estimation or imposed from outside the system. The latter approach was taken here for two reasons. First, econometric estimation of such a model is quite difficult; further, it was not the intention to apply this model in its present stage to any single country or group of countries. Second, a considerable amount of information was already available in the literature for a number of relevant parameters, and this could be readily utilized. As such, the model is properly described as a “simulation” model, and naturally all the results are conditional on the values of the parameters chosen.⁸ Some sensitivity analysis through the variation of certain key parameters was performed to determine if the basic results were significantly different.

The initial equilibrium—namely, the closed economy—was defined as one in which there was a uniform tariff on imports,⁹ and capital inflows and outflows were completely restricted. Trade liberalization was defined then as a reduction in the tariff rate to zero, and, correspondingly, capital account liberalization as the complete removal of the existing restrictions. With the properties of the initial equilibrium defined, and using a ceteris paribus assumption regarding the international picture and with respect to other domestic policies,¹⁰ the following simulation experiments were conducted: (a) gradual and sudden reduction in tariffs; (b) gradual and sudden removal of restrictions on capital flows; (c) simultaneous gradual removal of restrictions on trade and capital flows; and (d) sequential gradual removal of restrictions on trade and capital flows. The results from these simulations allow one to evaluate the effects of “gradual versus shock” types of policy, as well as the effects of a combination of alternative policies implemented both simultaneously and with differing speeds. The last simulation is particularly interesting, since there has been a lively discussion in the literature on whether a country should liberalize its trade account first or its capital account.¹¹ Together these four sets of simulations are reasonably comprehensive and are believed to cover most of the actual types of liberalization strategy.

While the model is utilized here only for the study of the effects of opening up, it can easily be extended to analyze a variety of other issues. For example, it can be used to simulate the effects of other domestic policy actions, as well as changes in foreign variables. Furthermore, within the model it is possible to design compensatory stabilization-type policies that operate in conjunction with the opening-up policies. One such experiment is undertaken here to demonstrate this particular capability of the model. Also, even though growth is not explicitly allowed for in the present version of the model, it can easily be introduced exogenously without radically altering the basic structure. What the model is not able to do is handle growth endogenously so that it cannot deal with the longer-term aspects of opening up. Capital formation, labor force participation, transfer of factors, wealth accumulation, etc., have to be introduced if such long-term effects are to be covered adequately. To do so, however, was not the intention in this present study.

The remainder of the paper proceeds as follows. Section I describes the basic structure of the model. Section II considers the various simulation experiments, and Section III contains the principal conclusions of the study.

I. Specification of the Model

The model consists of 37 equations, of which 17 are behavioral, while the remaining are definitions or identities. In a sense, the size of this model represents a compromise between the much larger computational general-equilibrium models developed by Feltenstein (1980) and Derviş, and others (1981) among others, on the one hand, and the more aggregated models of Clements (1980), Blejer and Fernandez (1980), and Khan and Knight (1981), on the other hand. There are basically two reasons for the level of aggregation chosen. First, as it is sought to evaluate the behavior of the main macroeconomic variables of a general rather than a specific economy, it is necessary to restrict the complexity of the model and to limit it to its essential features. In other words, country-specific institutional details are excluded from consideration. Second, since this is a simulation exercise, numerical values are required for the parameters; the more the model is expanded, the greater is the degree of arbitrariness in the selection of parameters, thereby possibly increasing the danger of making the results less general.

Essentially, this model describes an economy that is “small” in the sense that it takes prices and interest rates in the world markets as given parameters. It assumes a fixed nominal exchange rate or, more strictly, one that is not subject to market forces but that can be altered by the authorities in whatever manner they choose. Foreign trade, while permitted, is subject to a relatively high tariff on imports that creates a wedge between domestic prices and the domestic-currency price of importable goods. Capital movements, however, are restricted completely. In contrast to the restrictions on trade and capital flows, domestic prices and interest rates are assumed to respond to market forces, implying that reforms affecting the operation of the domestic money and credit markets are assumed to have been carried out.¹²

Potential output, which can be identified as the “transformation curve” of the economy, is assumed to be determined exogenously, that is to say, it is independent of the endogenous evolution of the variables in the system. This assumption means that the productive capacity of the economy is unaffected by the opening up, even though it could, in principle, be allowed to increase over time. This is a serious assumption and warrants comment, since it implies that—in the short run—capital formation, technical progress, etc., are unaffected by any liberalization policy. While one can legitimately assume the growth of the labor force to be exogenous, determined, for example, by population growth, the assumption of no increase in net investment is not as easy to justify, since it necessarily implies zero net additional national savings. However, it is well known that policies of opening up are usually accompanied to a greater or lesser extent by current account deficits, which constitute external savings. To assume that net savings (and therefore net investment) are constant, it is necessary to impose the condition that changes in external savings are exactly offset by changes in domestic savings. While this is possibly an extreme assumption, some empirical evidence does point to partial offsetting in developing countries, in the sense that part of external savings goes into financing consumption.¹³ Basically, this assumption of a given transformation curve was made to avoid consideration of sectoral production functions, which would have greatly complicated the structure of the model.

Actual real output is, of course, endogenously determined so that the model attempts to capture the short-term deviations of output from potential output. The gap between potential and actual output measures unemployment of resources, which, with limitations and in an indirect manner, can be interpreted as unemployment of labor.¹⁴ The long-run equilibrium for output is therefore defined as a point on the transformation curve—namely, where actual and potential output are equal, allowing for some small level of permanent unemployment of resources.

In line with the current theory of international trade, the model contains three types of goods—importables, exportables, and nontradables. These categories are based essentially on the relative degree of substitution between domestic and foreign goods in consumption and production (reflected basically though price differentials, including transport costs, tariffs and other trade distortions, and any other adjustments). This classification, which is useful for the present purpose (directed as it is toward analysis of the foreign sector), does not incorporate any distinction between consumption goods and investment goods. This omission does constitute a further limitation, since one of the most frequently discussed issues in the process of opening up is the impact of such a policy on the rate of capital formation.¹⁵ Neither, for that matter, is there any consideration of intermediate inputs so that all goods are “final” goods.

Finally, the model has three important general-equilibrium characteristics. First, the quantities produced of each good are limited by the aggregate transformation curve, the position and shape of which is in turn determined by the resource endowment and technology of the economy; in the long term, the vector of quantities produced satisfy this restriction. Second, the demand equations for these goods satisfy the theoretical conditions of homogeneity, symmetry of substitution effects, and additivity. The last general characteristic of the model is the explicit introduction of budget constraints, both for the government and for the economy as a whole.¹⁶ The government budget constraint makes it possible to link the balance of payments and the fiscal and monetary sectors to domestic expenditures and income. These links are modeled explicitly and, together with the first two elements referred to earlier, ensure the global and sectoral consistency of the model.

The model contains the following six sectors: (1) production and supply; (2) expenditures; (3) prices and unemployment; (4) money and credit; (5) the balance of payments; and (6) the government.

Production and Supply

The system of equations determining “desired” aggregate supply for the three types of goods—importables (I^s*), exportables (X^s*), and nontradables (N^s*)—is derived in a manner outlined by Clements (1980) in the framework of a multiproduct supply model. This involves the maximization of the value of the national product subject to the restriction represented by the transformation curve and the respective prices of the three goods—Pi, Px, and Pn. The shape of the transformation curve describes the technological possibilities of transformation of one good into another, and the distance of the transformation surface from the origin represents the available resource endowment, given by the potential real output (y^*). Assuming that the technology is characterized by a quadratic transformation function, the producers’ problem can be formally described as

Maximize P′Z
Subject to Z′ΛZ = y^*2

where P is a price vector [Pi, Px, Pn], Z is the quantity vector [I^s*, X^s*, N^s*], and Λ = diagonal [γ₁, γ₂, γ₃].¹⁷ The γ₁, γ₂, and γ₃ are the price parameters of supply of importables, exportables, and nontradables, respectively. Also, y^* is a positive scalar that determines the distance of the transformation surface from the origin and represents the total endowment of resources. For the present study, it has been assumed to be constant, that is,

\begin{matrix} y^{*} = {\bar{y}}^{*} & (1) \end{matrix}

The solution to the maximization problem is given by

\begin{matrix} Z = \frac{y^{*}}{{(P' Λ^{- 1_{P}})}^{\frac{1}{2}}} Λ^{- 1_{P}} & (2) \end{matrix}

so that the desired supplies of importable, exportable, and non-tradable goods depend exclusively on the relative prices of the three goods, the technical conditions of transformation of one good into another, and the resource endowment. The system of equations (Z) is homogeneous of degree zero in prices, the cross-price effects are symmetrical, and the weighted sum of the price effects across equations is zero.¹⁸ The specific desired supply equations for the three goods are as follows:

\begin{matrix} I_{t}^{s *} = {[\frac{γ_{2} γ_{3} y_{t}^{* 2} P i_{t}^{2}}{D}]}^{\frac{1}{2}} & (2') \end{matrix}

\begin{matrix} X_{t}^{s *} = {[\frac{(γ_{1}^{2} γ_{3} / γ_{2}) y_{t}^{* 2} P x_{t}^{2}}{D}]}^{\frac{1}{2}} & (2'') \end{matrix}

and

\begin{matrix} N_{t}^{s *} = {[\frac{(γ_{1}^{2} γ_{2} / γ_{3}) y_{t}^{* 2} P n_{t}^{2}}{D}]}^{\frac{1}{2}} & (2''') \end{matrix}

where

D = γ_{1} γ_{2} γ_{3} {P i}_{t}^{2} + γ_{1}^{2} γ_{3} P x_{t}^{2} + γ_{1}^{2} γ_{2} P n_{t}^{2}

Dynamics in the supply of tradable goods are introduced by allowing for the actual supplies of importables and exportables to respond gradually to any changes in relative prices or the resource endowment. The dynamics are modeled in a simple way by specifying partial-adjustment mechanisms, whereby actual supplies adjust to the difference between current desired supply and the supply in the previous period, in the following manner:

\begin{matrix} \begin{matrix} Δ I_{t}^{s} = λ_{1} [I_{t}^{s *} - I_{t - 1}^{s}] & 0 \leq λ_{1} \leq 1 \end{matrix} & (3') \end{matrix}

\begin{matrix} \begin{matrix} Δ X_{t}^{s} = λ_{2} [X_{t}^{s *} - X_{t - 1}^{s}] & 0 \leq λ_{2} \leq 1 \end{matrix} & (3'') \end{matrix}

The λ₁ and λ₂ are coefficients of adjustment, and Δ is a first-difference operator, $Δ I_{t}^{s} = I_{t}^{s} - I_{t - 1}^{s} .$ . Substituting from equations (2′) and (2″), one obtains the actual supply equations for importable and exportable goods

\begin{matrix} I_{t}^{s} = λ_{1} {[\frac{γ_{2} γ_{3} y_{t}^{* 2} P i_{t}^{2}}{D}]}^{\frac{1}{2}} + (1 - λ_{1}) I_{t - 1}^{s} & (4') \end{matrix}

\begin{matrix} X_{t}^{s} = λ_{2} {[\frac{(γ_{1}^{2} γ_{3} / γ_{2}) y_{t}^{* 2} P x_{t}^{2}}{D}]}^{\frac{1}{2}} + (1 - λ_{2}) X_{t - 1}^{s} & (4'') \end{matrix}

The supply of nontradable goods is determined in a somewhat different manner, on the grounds that disequilibrium in this market leads to changes in both prices and quantities.¹⁹ The specification is as follows:

\begin{matrix} \log N_{t}^{s} = \log N_{t}^{s *} + λ_{3} [\log N_{t}^{d} - \log N_{t}^{s *}] & (5) \end{matrix}

Equation (5) states that supply will equal the desired supply only if the demand for nontradables (N^d) is equal to the desired supply. If at the prevailing price there is excess demand (supply) with respect to desired supply, the supply of nontradables will be larger (smaller) than desired supply. The second term, therefore, can be viewed as representing variations in inventories. The purpose of formulating this type of equation was to have a direct link between the demand for and the supply of nontradable goods, other than through variations in the price of such goods.²⁰ With λ₃ = 0, the aggregate supply of nontradable goods is totally independent of aggregate demand in quantity terms, and with λ₃ = 1, the adjustment of supply takes place along the demand function.²¹

Given the values of I^s, X^s, and N^s, one can obtain real output (y) by the equation²²

\begin{matrix} y_{t} = {[γ_{1} I_{t}^{s 2} + γ_{2} X_{t}^{s 2} + γ_{3} N_{t}^{s 2}]}^{\frac{1}{2}} & (6) \end{matrix}

The level of nominal income in turn is given by the identity

\begin{matrix} Y_{t} = {P i}_{t} I_{t}^{s} + P x_{t} X_{t}^{s} + P n_{t} N_{t}^{s} + \in_{t} \cdot τ_{t} \cdot P f_{t} \cdot I_{t} & (7) \end{matrix}

where Y is national income, ϵ is the exchange rate (defined in terms of units of domestic currency per unit of foreign currency), τ is the uniform tariff rate on imports, and I is the level of imports in foreign currency. In nominal income, one must add tariff revenues, because this corresponds to income generated, although it accrues in this case to the government rather than to producers. Disposable nominal income (YD) is calculated simply by deducting tax revenues (T) and tariff revenues from equation (7), that is,

\begin{matrix} {Y D}_{t} = Y_{t} - T_{t} - ϵ_{t} \cdot τ_{t} \cdot P f_{t} \cdot I_{t} & (8) \end{matrix}

Expenditures

The desired level of private nominal expenditures (including consumption and investment expenditures) are specified following the approach outlined by Dornbusch and Mussa (1975) with respect to savings. Here private expenditures are related to disposable nominal income, the nominal excess supply of money, and the rate of interest²³

\begin{matrix} \log {E P R D}_{t}^{d} = γ_{4} \log {Y D}_{t} + γ_{5} (\log M_{t} - \log M_{t}^{d}) - γ_{6} r d_{t} & (9) \end{matrix}

where EPRD^d is the desired level of private expenditures, YD is disposable income, M and M^d are the stock of money and the nominal demand for money, respectively, and rd is the domestic interest rate.

Residents of the country are assumed to be able to spend more or less than their disposable income depending on whether they are running down or accumulating cash balances. This latter term represents, in other words, a “hoarding” effect that is related to the wealth effect on private expenditures. A rise in the domestic interest rate is assumed to have a depressing effect on private nominal expenditures through, presumably, the reduction of the investment component of such expenditures.²⁴

Actual private expenditures are assumed to adjust to the difference between the current desired level and the actual level in the previous period

\begin{matrix} Δ \log {E P R D}_{t} = λ_{4} [\log {E P R D}_{t}^{d} - \log {E P R D}_{t - 1}] \\ 0 \leq λ_{4} \leq 1 & (10) \end{matrix}

Substituting for EPRD^d from equation (9), one obtains the dynamic version of the nominal private expenditures equation

\begin{matrix} \log E P R D_{t} = λ_{4} [γ_{4} \log Y D_{t} + γ_{5} (\log M_{t} - \log M_{t}^{d}) \\ - γ_{6} r d_{t}] + (1 - λ_{4}) \log E P R D_{t - 1} & (11) \end{matrix}

Equation (11) represents total private expenditures (including spending on both goods and nonfinancial services) and expenditures associated with interest payments on foreign debt. Private expenditures on goods alone require the latter to be subtracted from EPRD, that is,

\begin{matrix} {E P}_{t} = {E P R D}_{t} - r d_{t} \cdot {B f}_{t} & (12) \end{matrix}

where EP is private expenditures on goods and Bf is the stock of foreign debt. Obviously, in an economy with total restrictions on capital movements in the past, private foreign debt would be zero, but this variable is bound to become quantitatively important as the opening-up policy is instituted.²⁵

Total expenditures (E) in the economy are the sum of private expenditures (EP) and government expenditures (G)

\begin{matrix} E_{t} = {E P}_{t} + G_{t} & (13) \end{matrix}

Once total expenditures are given, the distribution between importable, exportable, and nontradable goods is determined by a process of maximization, subject to the budget constraint represented by total nominal expenditures. Such an approach assumes strict separability, implying a unidirectional causal relationship running from total expenditures toward its components.²⁶ Since one is dealing with total expenditures, one is also (for the sake of simplicity) making the explicit assumption that government expenditures are allocated among the three goods on the same basis as are private expenditures.

Consequently, the problem consists of maximizing the utility function f(Q) subject to the constraint that

P' Q = E

where P = [Pi, Px, Pn] and Q is the vector of quantities demanded of the three goods [I^d, X^d, N^d]. E corresponds to the nominal expenditure on goods (E = PiI^d + PxX^d + PnN^d), and f(Q) represents a general utility function.²⁷ The solution to the maximization problem yields the demand equations for importable, exportable, and nontradable goods, respectively²⁸

\begin{matrix} Δ \log {(P i I^{d})}_{t} & = & Δ \log E_{t} + Δ \log P i_{t} + (1 / w_{i}^{d}) [- γ_{7} Δ \log P i_{t} \\ + (w_{n}^{d} + (γ_{7} + γ_{8} - γ_{9} - 1) / 2) Δ \log P x_{t} \\ + (w_{x}^{d} + (γ_{7} - γ_{8} + γ_{9} - 1) / 2) Δ \log P n_{t}] & (14') \end{matrix}

\begin{matrix} Δ \log {(P x X^{d})}_{t} & = & Δ \log E_{t} + Δ \log P x_{t} + (1 / w_{x}^{d}) \\ [w_{n}^{d} + (γ_{7} + γ_{8} - γ_{9} - 1) / 2 Δ \log P i_{t} \\ - γ_{8} Δ \log P x_{t} + (w_{n}^{d} + (- γ_{7} + γ_{8} + γ_{9} - 1) / 2) \\ Δ \log P n_{t}] & (14 ″) \end{matrix}

\begin{matrix} Δ \log {(P n N^{d})}_{t} & = & Δ \log E_{t} + Δ \log P n_{t} + (1 / w_{n}^{d}) \\ [w_{x}^{d} + (γ_{7} - γ_{8} + γ_{9} - 1) / 2) Δ \log P i_{t} \\ + (w_{i}^{d} + (- γ_{7} + γ_{8} + γ_{9} - 1) / 2) \\ Δ \log P x_{t} - γ_{9} Δ \log P n_{t}] & (14 ‴) \end{matrix}

where γ₇, γ₈ and γ₉ are the respective price parameters of demand for importable, exportable, and nontradable goods, and the proportion of total expenditures on each type of good is given by the (variable) weights $w_{i}^{d}, w_{x}^{d}, and w_{n}^{d} .$ . The corresponding price elasticities, therefore, are $γ_{7} / w_{i}^{d}, γ_{8} / w_{x}^{d}, and γ_{9} w_{n}^{d} .$ . The demand equations are homogeneous of degree zero in prices, the matrix (∂Q/∂P′) is symmetrical, and the property of additivity is satisfied.

Prices and Unemployment

In equilibrium, the price of importable goods is defined as equal to an index of foreign prices (Pf),²⁹ adjusted by the exchange rate (ϵ) and the level of tariff protection (τ).³⁰ In log-linear terms, this can be expressed as

\begin{matrix} \log \bar{P} i_{t} = \log P f_{t} + \log ϵ_{t} + \log (1 + τ_{t}) & (15) \end{matrix}

where the $\bar{P} i$ represents the equilibrium price of importables. Since it is often observed that the actual price of importables does not adjust immediately to changes in the variables on the right-hand side,³¹ a partial-adjustment function is specified for the actual price of importables

\begin{matrix} \begin{matrix} Δ \log P i_{t} = λ_{5} [\log \bar{P} i_{t} - \log P i_{t - 1}] & 0 \leq λ_{5} \leq 1 \end{matrix} & (16) \end{matrix}

Substituting equation (15) in equation (16) and solving for the price of importable goods yields

\begin{matrix} \log P i_{t} = λ_{5} [\log P f_{t} + \log ϵ_{t} + \log (1 + τ_{t})] \\ + \log (1 - λ_{5}) \log P i_{t - 1} & (17) \end{matrix}

The price of exportable goods is definitionally equal to the product of the foreign price level and the exchange rate

\begin{matrix} \log P x_{t} = \log P f_{t} + \log ϵ_{t} & (18) \end{matrix}

Prices of nontradable goods are assumed to be essentially demand determined, and therefore the specification abstracts from all types of cost factors. In other words, these prices respond to excess real demand for nontradable goods and, in the absence of any disequilibrium, will be changing according to the variation in the price of tradable (importable) goods. The influence of the prices of tradable goods can be rationalized on two basic grounds. First, it can be argued that this term captures expectations of future inflation, and, second, for long-run consistency it is necessary to ensure that all prices move in line in the steady state.³² This yields the equation

\begin{matrix} Δ \log P n_{t} = λ_{6} [\log N_{t}^{d} - \log N_{t}^{s}] + λ_{7} Δ \log {P i}_{t} & (19) \end{matrix}

One would generally expect both λ₆ and λ₇ to be positive. A small value of λ₆ would imply slow clearance of the nontradable goods market, and continuous equilibrium would require λ₆ → ∞.³³ The movements in quantity and prices of nontradable goods when there is disequilibrium in this market would obviously depend on the relative values of λ₆ and λ₃, the parameter in the nontradable goods supply equation. Also, in the long run, one would expect the parameter λ₇ to tend to unity.

The general price index is specified as a Divisia index, where the percentage change in prices is a weighted average of the percentage changes in the prices of importable, exportable, and nontradable goods, with the weights being the shares of the expenditures on each of the three goods

\begin{matrix} Δ \log P_{t} = w_{i}^{d} Δ \log P i_{t} + w_{x}^{d} Δ \log P x_{t} + w_{n}^{d} Δ \log P n_{t} & (20) \end{matrix}

As a measure of expected inflation is required, the adaptive-expectations model of Cagan (1956) is used to generate it. In this context, expectations are revised proportionally according to the difference between the actual rate of inflation and the rate that was expected in the previous period

\begin{matrix} Δ Π_{t} = λ_{8} [Δ \log P_{t} - Π_{t - 1}] & (\begin{matrix} 21 \end{matrix}) \end{matrix}

where Π is the expected rate of inflation, and λ₈ measures the extent to which the revision of expectations responds to the error, 0 ≤ λ₈ ≤ 1.

The unemployment of resources is modeled as a simple function of the difference between potential and actual output—namely, the so-called output gap

\begin{matrix} U_{t} = U_{0} + γ_{10} (\log y_{t}^{*} - \log y_{t}) & (22) \end{matrix}

where U is the level of unemployment and U₀ represents some level of “normal” underutilization of resources. If the relation between the labor force and the level of aggregate output is stable, then equation (22) can be interpreted as an equation for labor unemployment.³⁴ In general, however, one has to exercise caution when dealing with distinct sectors, which may, for example, have differing factor intensities, to make such an assumption.

Money and Credit

The monetary sector, which is crucial to the operation of the model, is formulated in a fairly straightforward manner. The nominal demand for money is specified as a function of a scale variable, in this case nominal income, and the opportunity costs of holding financial assets in monetary form. Since the public can hold real assets (goods) as well as financial assets, these opportunity costs are the expected rate of inflation (II) and the domestic interest rate (rd). Formally, the function for nominal demand for broad money can be written in log-linear terms as³⁵

\begin{matrix} \log M^{d} = α_{1} + γ_{11} \log Y_{t} - γ_{12} Π_{t} - γ_{13} {r d}_{t} & (23) \end{matrix}

Obviously, more general formulations can be considered that could include, among other variables, the “own” rate of interest or, if residents hold wealth in the form of foreign financial assets, the foreign interest rate and the expected change in the exchange rate. For the moment, it is assumed that money pays no interest³⁶ and that residents hold only domestic financial assets.

The supply of money (equal to the actual stock)—broadly defined to include currency, demand deposits, and time and savings deposits—is equal to the stock of international reserves (in domestic currency terms) and the level of domestic credit extended by the banking system. This definition allows changes in the money supply to be brought about by variations in the balance of payments; this phenomenon is the central element of the monetary approach to the balance of payments (Frenkel and Johnson (1976) and IMF (1977)). Domestic credit, which is the basic monetary tool, is made up of credit to the government and credit to the private sector. In the absence of sterilization, monetary policy in this framework is essentially passive. Using these distinctions, the identity for the money supply can be expressed as

\begin{matrix} M_{t} = {C R G}_{t} + {C R P}_{t} + R_{t} & (24) \end{matrix}

where CRG and CRP are credit to the government and to the private sector, respectively, and R is the stock of net international reserves (in terms of domestic currency).

Since it has been assumed that the domestic financial system has already been liberalized, the domestic interest is free to adjust to market forces. The formulation chosen here relates the changes in the interest rate to monetary disequilibrium as follows:

\begin{matrix} Δ {r d}_{t} = γ_{14} [\log M_{t}^{d} - \log M_{t}] & (25) \end{matrix}

An excess demand for money can be expected to raise the interest rate, and vice versa, so that the parameter γ₁₄ would be positive. Its size would naturally determine the speed at which the interest rate moves to equilibrate the money market.³⁷ The nominal rate, as determined from equation (25), feeds back into the rest of the model, although the real interest rate, however defined, does not.

Balance of Payments

Imports (I), valued in foreign currency terms, are defined as the difference between domestic demand and domestic supply for importables³⁸

\begin{matrix} I_{t} = P f_{t} [I_{t}^{d} - I_{t}^{s}] & (26) \end{matrix}

Similarly, exports (X) are equal to domestic excess supply of exportables

\begin{matrix} X_{t} = P f_{t} [X_{t}^{s} - X_{t}^{d}] & (27) \end{matrix}

and the current account (in domestic currency terms) is equal to the difference between exports and imports less the interest payments on foreign debt

\begin{matrix} {C A}_{t} = ϵ_{t} [X_{t} - I_{t}] - r d_{t} \cdot B f_{t} & (28) \end{matrix}

The equation for capital movements is derived on the basis that, apart from some autonomous components, capital flows respond to the differential between domestic and foreign interest rates, adjusted for expected exchange rate changes and other factors, such as country risk and differences in bank reserve requirements. In the absence of controls, the function determining capital movements could be specified as

\begin{matrix} {D K}_{t} = λ_{9} ϵ_{t} + γ_{15} (r d_{t} - r f_{t} - ρ_{t} - Δ \log ϵ_{t}) & (29) \end{matrix}

where DK is the flow of capital, rf is the foreign interest rate, and ρ is the risk premium.³⁹ The first term simply gives the domestic currency value of autonomous capital flows, and it is assumed that the expected change in the exchange rate can be represented by the percentage change in the actual rate, Δlog ϵ_t.

The value of γ₁₅ measures the degree of response of capital flows to interest rate differentials, and to the extent that there are controls or restrictions on capital mobility, this parameter will be smaller. Equation (29) is redefined in the following way to take into account various degrees of opening up of the capital account

\begin{matrix} {D K}_{t} = λ_{9} ϵ_{t} + β [γ_{15} ({r d}_{t} - {r f}_{t} - ρ_{t} - Δ \log ϵ_{t})] & (29') \end{matrix}

In this formulation, by varying β one can control the degree of restrictions on movements of capital. If β = 0, then the economy is totally closed to international capital flows, whereas β = 1 implies complete liberalization. Values between zero and unity do not necessarily reflect varying degrees of restriction but mainly whether the response of capital flows is slow or rapid. Gradual opening up of the capital account can be represented in two alternative ways. First, β can be allowed to go from zero to unity gradually, or, second, it can be fixed at some positive value. Both methods, although not strictly equivalent, yield broadly similar results in the simulations performed. For the purpose of the exercise here, it was decided to work with the second of the alternatives.⁴⁰

The overall balance of payments (BP) is given by the identity

\begin{matrix} {B P}_{t} = {C A}_{t} + {D K}_{t} & (30) \end{matrix}

and the stock of international reserves (in domestic currency terms) by

\begin{matrix} R_{t} = R_{t - 1} + {B P}_{t} & (31) \end{matrix}

It is worth noting the way in which foreign indebtedness is incorporated in the model. It is assumed that foreign residents acquire (sell), in the home country, domestic financial assets issued only by the private sector and that there are no government bonds.⁴¹ There is an initial stock of private bonds, and no new issues are assumed to take place.⁴² External debt is therefore given by

\begin{matrix} {B f}_{t} = Σ_{j = 0}^{\infty} {D K}_{t - j} & (32) \end{matrix}

The risk premium is assumed to have a constant component (ρ₀) and a variable component related to the ratio of foreign debt to total income, that is,

\begin{matrix} ρ_{t} = ρ_{0} + γ_{16} {(B f / Y)}_{t} & (33) \end{matrix}

This function basically assumes that the supply curve of international capital to the country is not infinity elastic, and the relevant cost—that is, the foreign interest rate adjusted for exchange rate changes and the risk premium—rises as more external debt is contracted. It is arbitrarily assumed here that when the ratio of foreign debt to income reaches a certain level, domestic and foreign interest rates (including the now larger risk premium) will be equalized.⁴³ As foreign debt rises beyond this point, capital outflows will be generated and the domestic interest rate will have to rise to compensate for the increase in the risk premium.

Government Sector

The model incorporates the fiscal sector in a rudimentary fashion, since the purpose is solely to introduce an explicit government budget constraint into the framework. Government expenditure in nominal terms (G) is defined as the sum of taxes, tariff revenues, and the public sector deficit

\begin{matrix} G_{t} = T_{t} + ϵ_{t} \cdot τ_{t} \cdot P f_{t} \cdot I_{t} + {G D}_{t} & (34) \end{matrix}

where GD is the government deficit.

Taxes are related in a linear fashion to the level of nominal income

\begin{matrix} T_{t} = t_{0} + γ_{17} Y_{t} & (35) \end{matrix}

where γ₁₇ is the marginal tax rate.

Finally, it is assumed that all public sector deficits are financed by the issuance of money, namely, through variations in credit to the government, that is,

\begin{matrix} Δ {C R G}_{t} = {G D}_{t} & (36) \end{matrix}

The linkage between fiscal policy, as represented by GD, and monetary policy is immediate and stems from the fact that no other forms of financing public expenditure, in particular, borrowing from the nonbank public, have been modeled.⁴⁴

For purposes of convenience, the complete model, including the behavioral equations as well as identities and definitional equations, are presented in Appendix I.

II. Simulation Experiments

Utilizing the values for the parameters (shown in Appendix II), the complete model can be simulated for various types of change related to the opening-up process. To begin with, following conventional analysis, the economy was assumed to be in initial equilibrium; this may appear to be a rather arbitrary starting point, given that the implementation of a liberalization strategy generally presupposes the existence of some disequilibria in the economy. Nevertheless, any attempt to begin the analysis from a disequilibrium position would pose three problems. First, the main features of the disequilibrium cannot be chosen arbitrarily, since this would probably violate the internal consistency of the model. In other words, if one starts from a position of a particular current account deficit, this implies specific values for other variables in the system so that all of these must be taken into consideration. Second, if the analysis were begun from a position of disequilibrium, it would be quite difficult to distinguish those changes that occurred as a result of discretionary policy actions from those that would have occurred in any case, owing to the automatic processes that tend to adjust, for example, the balance of payments and the domestic inflation rate in a small open economy. Finally, the time path of the variables during the transition period, and indeed the transition period itself, is not independent of the initial conditions. Consequently, it would be hard to determine whether the behavior of a variable during transition was due to the policy or simply to the position from which it started. For these reasons, the simulations are begun in full equilibrium, not with any intention of realism but to be exactly aware of how the adjustment process operates within the context of the model.

The initial equilibrium was defined as one in which the economy is protected by a uniform tariff of 100 percent (τ = 1.0) on all imports and is closed to international capital movements (β = 0). The economy is assumed to be on its transformation curve (assuming 5 percent normal unemployment) so that actual and potential output are the same and all desired quantities are equal to their actual levels. Prices are assumed to be in equilibrium for a fixed exchange rate (ϵ = 1.0); the current account, capital account, and overall balance of payments are in balance; and the level of international reserves and the stock of money are constant. In the context of the simulations, trade liberalization corresponds to a lowering of the tariff rate (τ) to zero, and opening up the capital account means increasing the value of β.⁴⁵

Given these initial conditions, and the further assumptions of an unchanging international environment⁴⁶ and no variation in other policies,⁴⁷ the following simulation experiments were conducted: (1) gradual and sudden reduction in tariffs; (2) gradual and sudden removal of restrictions on capital movements; (3) simultaneous gradual removal of restrictions on trade and capital flows; and (4) sequential gradual removal of restrictions on trade and capital flows. These four sets of experiments are believed to cover most of the liberalization scenarios that one has observed. In addition, for purely illustrative purposes, a “compensatory” policy simulation was conducted in which monetary policy was used to keep the current account from deteriorating when tariffs were lowered. The purpose of this particular simulation is to demonstrate the capability of the model to handle this type of question and further to highlight the kind of trade-offs that emerge when it is desired to stabilize certain variables in the context of opening up.

The effects of the policy changes were naturally traced out for all the endogenous variables, although the results are reported and discussed for only a select few that are regarded as of central interest.

Gradual and Sudden Reduction in Tariffs

Assume that the level of restrictions on capital flows is maintained (β = 0) but the tariff rate is lowered from 100 percent to zero in two specific ways. In the first case the reduction takes place gradually over four periods starting in period 3, and in the second case in a sudden fashion, also in period 3.⁴⁸ The effects of these two types of policy change on the level of prices (in logarithms), the domestic interest rate, the current account and the level of international reserves, and the gap between potential output and actual output (unemployment of resources), are shown in Chart 1, A–D.

Considering the gradual tariff reductions first, one can observe that the overall price index falls (Chart 1-A). This fall is a consequence of both the direct effect of the drop in the price of importable goods and the effect of the decline in the price of nontradables. The price of nontradable goods faces downward pressure on two counts. First, the percentage change in the price of importables enters the equation for the percentage change in the price of nontradable goods, measured by the parameter λ₇; second, the change in relative prices leads to a transfer of resources toward the production of nontradable (and exportable) goods and, at the same time, diverts demand from these two sectors toward the market for importables.⁴⁹ As a result of this transfer, an excess supply of nontradable goods is created, which causes a decline in the price of nontradables and the overall price level.

The fall in prices, by lowering the nominal demand for money, creates an excess supply of money that manifests itself both in the financial sector and in expenditures. The rate of interest drops temporarily (the “liquidity effect”) for about six periods after the policy of tariff reduction is initiated⁵⁰ and then starts to rise as the monetary disequilibrium is steadily eliminated (Chart 1-B). After approximately 14 periods, the rate of interest stabilizes at its original level.

The combination of the excess supply of money and the change in relative prices results in a current account deficit that persists for 14 periods or so (Chart 1-C). There is a steady loss of international reserves until they reach about one third of their original level. In fact, it is this loss of reserves, together with the initial decline in the rate of interest, that eventually brings about equilibrium in the money market.

Finally, as adjustment in the tradable goods market is not instantaneous, and, more specifically, importable goods are assumed to adjust more rapidly than exportable goods (λ₁ > λ₂), the resources released by the importables sector are not absorbed by the other sectors, thereby resulting in unemployment (Chart 1-D).⁵¹ The gap between potential and actual real output rises by more than 3 percentage points and then declines to the original 5 percent level of “normal” unemployment. It can be observed that the unemployment created by the policy of tariff reduction tends to persist for a considerable period.⁵² Both the extent and duration of this unemployment depend crucially on the relative values of the adjustment parameters in the supply equations (λ₁, λ₂, and λ₃). When adjustment is instantaneous in both the importable and exportable goods markets, and λ₃ = 0, one would observe that trade liberalization has basically the same effects on the time paths of the other variables in the system but that the resource gap turns out to be zero. However, instantaneous adjustment is clearly an extreme assumption, and our choice of parameters appears more realistic.

When tariffs are reduced in a “shock” fashion, that is, they are immediately reduced to zero in the first period, there is essentially no qualitative difference in the results; see Chart 1, A–D. Clearly, the effects of such a policy result in a more pronounced movement in the initial periods and the transition path is generally less “smooth.” Given the different time paths of the excess supply of money in this simulation, the initial fall in prices and the interest rate is sharper, but it can be seen that the latter variable starts to rise back to its original level earlier. It is interesting that the accumulated current account deficit (loss of international reserves) does not differ according to whether the opening up is gradual or sudden. What is evident is that when the policy is implemented suddenly these deficits are larger originally and then smaller later (Chart 1-C).

A similar pattern is evident in the behavior of the gap between potential and actual output, but to properly compare the areas under the curves in Chart 1-D, one would have to calculate the present values using some type of social rate of discount. The difference observed between the cases of the sudden and the gradual opening up is related not so much to the time during which output remains below “full” employment, but rather to the fact that in the former the peak of unemployment is higher and the distribution of the resource gap is more asymmetrical than in the scenario for the gradual opening up.

Gradual and Sudden Removal of Restrictions on Capital Flows

These experiments start with the assumption that there is a 100 percent tariff on imports and capital movements are completely restricted. Maintaining the tariff and eliminating the restrictions on capital flows both gradually and suddenly—that is, by varying β—gives a picture of what would be expected if only the capital account were liberalized. The results of this experiment on the various important macroeconomic variables are shown in Chart 2, A–E.⁵³

The immediate effect of removing restrictions on the capital account is an inflow of capital as the domestic interest rate is above the corresponding foreign interest rate. This positive differential in favor of the domestic economy is a phenomenon typically observed in a number of developing countries that have liberalized their domestic financial markets and removed capital controls. (See Mathieson (1979; 1980) and Zahler (1980).) The inflow of capital creates an excess supply of money that, in turn, has an expansionary effect on aggregate demand, which is reflected in an initial rise in prices (Chart 2-A) and a worsening of the current account (Chart 2-C). The excess supply of money, as expected, also lowers the domestic interest rate (Chart 2-B). In theory, the interest rate move should persist until interest parity is established, which in this specific case implies that the rate should fall initially and then rise to its equilibrium level.⁵⁴ How fast this occurs depends on the size of the response of capital flows to the interest rate differential (γ₁₅) and the effect on the rate of interest of monetary disequilibrium, measured by γ₁₄. The values chosen for these parameters imply fairly slow adjustment of the interest rate so that even though it initially falls, it nevertheless starts to rise slowly as the stock of external debt increases and pushes up the risk premium. This slow adjustment would seem to accord with actual experience, and it has been hypothesized that such factors as market segmentation, nontradable assets, and other rigidities prevent the emergence of instantaneous interest parity.⁵⁵ While these factors have not been modeled explicitly, the choice of parameters here does yield a rough approximation to this type of behavior.

Moreover, as long as equality between interest rates does not exist, foreign debt will continue to increase (Chart 2-D); this means that the economy must generate a trade surplus in order to cover the rising interest payments.⁵⁶ In this process, there is a slight increase in the share of tradable goods and a corresponding fall in the proportion of nontradable goods in total output because of the need to depress aggregate demand so as to generate the resources for paying interest on external debt. This fall causes a small decline in the price of nontradable goods, and the general price level, after the temporary rise that occurred owing to the initial excess money supply (Chart 2-A).

Despite the surplus on the trade account, a sustained deficit in the current account can be noted, although its size decreases gradually as the economy moves toward equilibrium. Given the values of the parameters, the current account deficit is, however, more than compensated for by the inflow of capital (Chart 2-C) so that there is an increase in international reserves in the beginning. The higher initial level of reserves reflects the increased demand for money that results from the decline in the interest rate; of course, at the same time, the stock of foreign debt is also larger (Chart 2-D).

In contrast to the simulations related to trade liberalization, the opening up of the capital account has only a small impact on the resource gap. After a slight increase in output owing to the initial excess money supply, the gap between potential and actual output reaches a peak that is only about 0.4 percentage point above the equilibrium level of full employment (Chart 2-E).

The main differences between the gradual and sudden policies with respect to the opening up of the capital account lie in the distribution of the variables in question. The time path of the variables in the scenario for the gradual opening up are generally smoother than when β is set to unity in one period. In the shock case, external indebtedness toward the end is larger, and, while the initial decline in the resource gap is greater, the eventual rise in unemployment is somewhat higher.

In short, capital account liberalization, unlike trade liberalization, does not significantly affect relative prices, the level of domestic prices, or resource unemployment,⁵⁷ nor is there a significant loss of international reserves, but rather an initial gain. There is associated with this policy, however, a process of growing external indebtedness that involves a continuing current account deficit that is due to the need to make interest payments on the foreign debt. Finally, there is a temporary decline in both the nominal and the real rates of interest.

Simultaneous Gradual Removal of Restrictions on Trade and Capital Flows

Assume that the authorities undertake to liberalize the trade and capital accounts simultaneously, rather than separately. In this case, the tariff rate (τ) is reduced to zero and the coefficient measuring the degree of restrictions (β) is raised to one half, and both policies are implemented gradually.⁵⁸ That is, τ is reduced to zero in four periods and β is set at equal to 0.5 in the third period. The results of this simulation experiment are shown in Chart 3, A–E.

In the first place, the combined effect of such policies is to bring about a greater degree of monetary disequilibrium in the economy, compared with the consideration of the two policies separately. The nominal demand for money falls because of the drop in the price level resulting from the tariff reduction and the money supply increases owing to the inflow of international capital. In net terms, there would appear to be a larger excess supply of money initially; this causes the price level and the interest rate to fall relatively more than was observed in the two previous simulations (Charts 3-A and 3-B). The interest rate falls rather sharply in the beginning; even when it begins to rise again, it remains below the time path followed when only the capital account was opened up.

The initial excess money supply creates a current account deficit that is also larger than when the two policies were undertaken individually. The deterioration caused by the tariff reduction is reinforced by the further small worsening that results from opening up the capital account (Chart 3-C). The time path of the trade balance lies between the corresponding paths of the two previous simulations, initially following (for about 10 periods) the direction of the trade balance picture emerging from trade liberalization but later beginning to mimic the behavior of the trade balance that resulted from the capital account liberalization. This is so because a surplus must be generated continually to finance the interest payments on the growing external debt.

The final level of international reserves is about the same as was observed with the trade liberalization, although during the transition the paths of reserves do deviate from one another (Chart 3-D). It is readily apparent that the capital inflows generated by the policy of removing the relevant restrictions are not sufficient to cover the current account deficits so that the country will lose reserves. The stock of external debt has a somewhat different path from the simulation experiments related to capital account liberalization (Chart 3-D). There is an initial rise but then a fall for a few periods and later a smooth increase. This gyration is a direct consequence of the cyclical way in which the interest rate behaves in this particular simulation.

Finally, the structure of production and the resource gap behave in a way that is similar to the trade liberalization, which can be explained by the small impact of capital account liberalization on relative prices (Chart 3-E). Even though there is initially a greater excess money supply, this appears to be reflected in a current account deficit rather than affecting the real sector through a change in relative prices. Unemployment rises by a little more than 3 percentage points above its long-run equilibrium level.

To sum up, the simultaneous application of the two types of opening-up policy is not simply the same as the sum of each of them considered separately. Although the structure of production and resource unemployment, together with prices, tends to broadly reproduce the situation observed for the reduction in tariffs, the financial and foreign sector variables behave in a different manner from that resulting from the two policies considered individually. This shows up, in particular, in a lower level of external indebtedness, a smaller surplus in the trade balance (but with larger imports than in the initial situation), and a different time path of the interest rate in comparison with capital account liberalization alone.

Sequential Gradual Removal of Restrictions on Trade and Capital Flows

Much of the recent discussion regarding opening-up strategies has focused on the sequence in which these reforms ought to be implemented, namely, whether the trade account should be liberalized first and then the capital account, or vice versa. (See McKinnon (1982) and Frenkel (1982).) Since theory provides limited guidance on the issue of sequencing, the arguments have been based on essentially casual empiricism. Without going into the discussion of which sequence is better or more likely to be successful, one can, in the context of this model, outline the consequences of two alternative types of strategy. First, the trade account is liberalized (gradually) in period 3 and then, after the subsequent four periods, the restrictions on capital account are removed (also gradually).⁵⁹ In the second stage, the sequence is reversed, with the capital restrictions removed first and then, again after four periods have elapsed, the tariff rate is gradually lowered to zero. The results of this particular experiment are reported in Chart 4, A–F.

It appears to be a matter of indifference as to the sequence of policies adopted insofar as the price level is concerned (Chart 4-A). Prices decline by approximately the same amount in the two cases, except that when the trade account is liberalized first, the effect occurs earlier. Since it has been shown in the previous experiments that the effect on prices of the opening up of the capital account is negligible, this result is not too surprising.

The effect on the interest rate is slightly different, depending on the chosen sequence. When the capital account is liberalized first, the rate of interest initially declines more slowly, but then the fall is accelerated as tariffs are removed (Chart 4-B). This is so because the excess supply of money generated by the fall in prices owing to the reduction in tariffs is greater than the excess supply of money created by the capital inflow. The overall decline in the interest rate also turns out to be somewhat greater in this particular sequence of reforms.

The trade balance and current account pictures are, however, quite different. If the capital account is liberalized initially, for a few periods there is a slight worsening of the trade balance (Chart 4-D). The tariff adjustment pushes the trade balance further into deficit, but the cumulative deficit turns out to be smaller than if tariff reductions had been introduced first. The payment of interest on foreign debt also makes the path of the current account somewhat different (Chart 4-C). For this variable, the deficit is smaller in the beginning when the capital account is liberalized first, but later the deficit turns out to be larger than when the opposite sequence is implemented.

The behavior of international reserves and the foreign debt is also interesting. As capital restrictions are removed, there is an immediate inflow of capital, and both reserves and the stock of foreign debt rise (Chart 4-E). However, as the tariff reduction lowers the demand for money, which further reduces the interest rate, this process is temporarily reversed and both the stock of foreign reserves and debt fall. Eventually, as the system stabilizes, the level of international reserves is approximately the same in both scenarios, and the level of external indebtedness is slightly higher when the capital account is opened up initially.

As the liberalization of the capital account has a negligible effect on the output gap, there is no noticeable difference when the different sequences of policies are adopted (Chart 4-F). All that one observes is that the unemployment is generated later when the opening up of the capital account is instituted first. Other than that, the time paths of unemployment are exactly the same.

In summary, as with the simultaneous liberalization experiments, little difference is found in the effects on prices and the real sector. The only question that can be raised in relation to the latter is whether unemployment is preferred now or later. The main differences arise in the financial and foreign sectors, where the decline in the interest rates, the deterioration in the current account, and the loss of international reserves is initially less if capital flows are liberalized before any trade reforms are initiated than if the opposite sequence is followed. At the same time, however, the stock of external debt will be smaller in the long run if the trade account is liberalized first. Therefore, these results do not indicate clear-cut support for the propositions made by McKinnon (1982) and Frenkel (1982) that tariff reforms should necessarily take place prior to policy changes that affect capital movements. Explicit trade-offs are evident, and the issue cannot be resolved on theoretical grounds.

Compensatory Policies

It is evident from the analysis of the various simulations that opening-up policies are generally accompanied by transitory effects (of varying duration) that may be considered undesirable by the policymakers. The effects can include current account deficits, loss of international reserves and/or greater external indebtedness, increases in the real interest rate, and resource unemployment. Each of the strategies tends to yield some combination of these “costs.” Of course, the authorities could in principle use compensatory demand management policies, such as monetary and fiscal or exchange rate policy, to minimize some of these costs. The present model allows one to design such policies; purely for illustrative purposes, one such experiment is described here.

In the context of trade liberalization alone, it is clear from all these experiments that the current account deteriorates in the short run. Suppose that the authorities wish to prevent this and are prepared to use monetary policy to this end. To avoid a current account deficit and, given that capital movements are not allowed, a consequent loss of international reserves, it would be necessary to implement a restrictive monetary policy consisting of reducing domestic credit—in this case, credit to the private sector (CRP)—so as not to generate the excess money supply that typically emerges when tariffs are reduced.⁶⁰ Chart 5-A describes the magnitude and the time path that domestic credit must follow in the attempt to secure permanent equilibrium in the monetary sector when trade is opened up gradually. For purposes of comparison, the (constant) path of domestic credit has been plotted when trade was liberalized gradually without any change in monetary policy. It can be seen that a strong contraction in domestic credit would be necessary, being more pronounced in the initial periods and subsequently less marked.

The paths of the international reserves and current account as a consequence of this restrictive monetary policy are shown in Charts 5-B and 5-C, where again the results have been plotted in the absence of changes in monetary policy. International reserves fall slightly; they do not remain unchanged because guaranteeing equilibrium in the monetary sector alone is not sufficient for this. In addition, as the adjustment in the exportable goods market is slower than in the importable goods market, the change in relative prices creates an asymmetrical response in the two markets, thereby generating a current account deficit in the first few periods—even though there is no excess aggregate demand.

While the level of international reserves can be stabilized by a suitable tightening of monetary policy, this policy can itself have obvious adverse effects on other variables in the system. For example, Chart 5-D shows what happens to private expenditures as a consequence of such a policy. For several periods, private expenditures are lower than the rate witnessed when tariffs were reduced without any variation in monetary policy. Although not shown here, a relatively larger rise in the real interest rate and some further lowering of inflation and growth were also observed.

The basic exercise here has pointed out that designing an “optimal” mix of policies necessarily involves trade-offs, and it is up to the authorities to decide the weights that they assign to the various effects. It appears quite evident that there is no easy escape from having to choose among alternative combinations of transitory effects.

III. Conclusions

This study, while focusing on the narrow issue of the characteristics of the short-run and medium-run adjustment path of some main macroeconomic variables as a consequence of opening up the economy to the free flow of goods and financial capital, has yielded, nonetheless, some important insights that can now be summarized. The main results are presented for convenience in Table 1. The simulation exercises produced a number of interesting and plausible results, based—it should be stressed—on a representative set of parameters. Certain combinations of parameters, of course, could produce somewhat counterintuitive results, but it seems that the values needed for the coefficients would be fairly unrealistic and of only possible academic interest.

Table 1.

Summary of Short-Run Effects of Liberalization of Trade and Capital Movements on Selected Economic Variables

^¹Capital controls retained.

^²Tariff barriers retained.

^³This simulation corresponds to the gradual removal of tariff barriers and then the elimination of capital controls, and vice versa. While the overall effects of the two sequences are similar, as reported in this table, the time paths of the variables do differ.

Table 1.

Summary of Short-Run Effects of Liberalization of Trade and Capital Movements on Selected Economic Variables

Policy	Price Level	Current Account	Inter- national Reserves	Real Output	External Debt
Trade liberalization¹	Lowered	Deficit	Lowered	Lowered	—
Capital account liberalization²	Unchanged	Deficit	Initially increased	Unchanged	Increased
Simultaneous liberalization	Lowered	Deficit	Lowered	Lowered	Increased
Sequential liberalization³	Lowered	Deficit	Lowered	Lowered	Increased

^¹Capital controls retained.

^²Tariff barriers retained.

Table 1.

Summary of Short-Run Effects of Liberalization of Trade and Capital Movements on Selected Economic Variables

Policy	Price Level	Current Account	Inter- national Reserves	Real Output	External Debt
Trade liberalization¹	Lowered	Deficit	Lowered	Lowered	—
Capital account liberalization²	Unchanged	Deficit	Initially increased	Unchanged	Increased
Simultaneous liberalization	Lowered	Deficit	Lowered	Lowered	Increased
Sequential liberalization³	Lowered	Deficit	Lowered	Lowered	Increased

^¹Capital controls retained.

^²Tariff barriers retained.

Briefly, the principal results are as follows:

(1) It was demonstrated that the effects of trade liberalization are quite different from those resulting from the opening up of the capital account. The former type of policy results in significant changes in resource allocation and in both relative and absolute prices, in current account deficits, in loss in international reserves, and in a relatively large fall in aggregate production. The impact of capital account liberalization, on the other hand, results in greater external indebtedness and current account deficits, although international reserves rise initially and then fall. The effects on the structure of production, prices, and the resource gap, however, are small.

(2) Another interesting result is that the speed with which reforms are instituted also matters. As would be expected, a shock-type approach, whether in the trade sector or on the capital account side, has a more pronounced impact at the beginning, but the adjustment to equilibrium is faster. If the policy is implemented more gradually, the time paths of the variables are smoother, but naturally the adjustment process is more delayed. While, on the face of it, this would seem to suggest that a more gradualist approach would reduce the cost of the opening-up policy, it should be noted that making an adequate comparison of the effects of alternative strategies requires the introduction of an appropriate social rate of discount. This is something that is not done here, since it would involve a fair degree of speculation.

(3) If a choice must be made as to which type of liberalization policy to implement first, it is clear that different policy sequences yield different results from the point of view of the time paths of the principal macroeconomic variables, and it is not immediately obvious which pattern would be preferable. Also, it appears that the effects of the simultaneous liberalization of trade and the capital account is not simply a linear combination of the policies implemented separately.

It may, therefore, be concluded from the simulations that the economic authorities cannot be indifferent about sequences, magnitudes, and relative speeds of alternative opening-up strategies.

The basic time paths of the variables in the system, which are in a sense the focus of the exercise, depend on a variety of factors; it may be useful to discuss some of these factors:

(1) In general terms, the study indicates that even when there is rapid opening up, it takes a number of periods to secure general macroeconomic equilibrium. In particular, even for a sudden liberalization of the capital account, it takes a while for the domestic interest rate to converge to the international rate, appropriately defined. The sensitivity of the speed of convergence of domestic and international interest rates depends on the interest elasticity of the flow of international capital to the country, the proportion of the monetary disequilibrium that spills into the domestic financial market, and the effect on the country-risk premium of changes in the external debt situation of the country.

(2) A further element that plays an important role in determining the time path of the various macroeconomic variables is the adjustment of prices and quantities in the nontradable goods market. It is observed that the slower the adjustment of this market and the more important the direct role of effective demand in the supply of such goods, the costlier the process of trade liberalization tends to be in terms of production losses. Likewise, the speed with which domestic and foreign prices tend to equalize when there is removal of trade barriers depends strongly on the importance of the production of nontradable goods within total supply and on the speed of adjustment of prices to disequilibrium in the nontradables market.

(3) A final element that stands out in defining the time path of the macroeconomic variables in a process of opening up is that associated with the speed of response of the production of exportable and importable goods to changes in relative prices. The greater the intersectoral mobility of resources and the smaller the differences between the lags in the desired and effective supplies of exportable and importable goods, the smaller will be the transition costs related to unemployment of resources arising from the opening up.

Some other specific interesting results that emerge from the exercise are as follows:

(1) When the domestic economy is opened up to the world economy, the global balance of payments position may improve, even though the current account deteriorates and external indebtedness grows.

(2) Real output temporarily falls when tariffs are reduced. The duration of the resource unemployment is independent of the speed of trade liberalization, but, as mentioned, the resource gap is greater at the beginning for “shock” policy, although later its magnitude is less than that corresponding to the “gradual” scenario. This result is sensitive to the relative speed of adjustment of quantity and prices in the nontradable goods market.

(3) As expected, the rate of inflation falls when tariffs are reduced, and the domestic interest rate moves toward the value of the international rate, appropriately defined, when the economy is opened up to capital flows. However, these two variables do not converge instantaneously toward their international levels, and the time that it takes depends, to some extent, on the speed of the respective processes of opening up.

(4) Although the production of importable goods falls with tariff reductions as its price falls, the production of exportable goods rises. The proportion of nontradable goods in the final equilibrium, compared with the initial equilibrium, depends on the own and cross-price elasticities of substitution of production and expenditure. In general, there is a change in the structure of production even though the level may remain constant.

(5) If initially the domestic nominal interest rate is above the relevant (adjusted) foreign rate, it will tend to fall when the economy is initially opened up, but the real rate may rise substantially above its initial value for some time. This case is frequently observed in sudden opening-up strategies in the trade sector and is explained by the (relatively) sharp drop in the domestic (actual and expected) rate of inflation, which results from the reduction in the price of importable goods.

(6) Compensatory policies can be designed that tend to reduce some of the undesirable transitory effects of the process of opening up. The model makes it possible to devise such policies and to evaluate the trade-offs implicit in their application.

Before ending, it is important to list certain caveats. The analysis was conducted in a nongrowth framework, and it was assumed that potential output is exogenous. In the opening-up process, when a current account deficit occurs, to preclude net capital accumulation (a necessary condition for maintaining real potential output as exogenous), it has been assumed that foreign savings are a perfect substitute for domestic savings. Furthermore, the assumption of exogenous real potential output implies that the model does not incorporate the possibility that, in the process of opening up, some part of the existing capital stock can become economically (but not necessarily technically) obsolete, and that there may be an improvement in efficiency owing to inputs of new modern capital goods.

Some further caveats, which do not necessarily limit the relevance of this study, should also be pointed out. For example, even though the results here have obvious implications regarding the behavior of factor markets and changes in the distribution of income, property, or wealth, which may be quite important in some experiences of economic opening up, no explicit analysis of such phenomena has been conducted. Also, all goods in the model are final goods, and, as such, there is no allowance for intermediate inputs; therefore, there is no direct analysis of costs of production, the level and composition of which may change in the process of opening up. The model also makes no distinction between consumption and investment goods. This omission may be important, particularly as one of the most frequently discussed topics in processes of opening up is that of their impact on the rate of accumulation. Finally, expectations have been modeled in a fairly simplified fashion, and it is possible that the introduction of more sophisticated expectations-generating schemes could alter some of the results. In this context, one should also be aware that a model such as that developed here is vulnerable to the so-called Lucas critique. In other words, it is quite likely that certain structural parameters that have been assumed to be constant, particularly those relating to expectations, could be altered as a consequence of the change in the policy regime. Unfortunately, to our knowledge, there is no obvious way to take account of such a criticism.

Also, note that the various simulations upon which these conclusions are based were conducted using the strong assumption that other domestic policies, as well as international factors, remained unchanged. It is obvious that the results could be significantly altered if, for example, the government ran large fiscal deficits, engaged in excessive monetary expansion, or maintained an unrealistic exchange rate at the same time that it was engaged in a liberalization process. Furthermore, changes in the international environment, as have been witnessed recently—namely, the high foreign real interest rates, declining growth in the industrial world, rising protectionism in the export markets of developing countries, and sharp changes in commodity prices that have generally worsened the terms of trade of these countries—would also be expected to change the outcome of the liberalization experiments. Although such factors have not been dealt with here explicitly, the model is quite capable of handling them.

Finally, further development of this model will make it possible to facilitate the analysis of real cases. The next stage would undoubtedly involve, apart from some theoretical refinements, the actual estimation of structural parameters, lags, and coefficients of adjustment, either for individual countries or using crosscountry data. The expectation is that the basic model designed and studied here will serve as a foundation on which more detailed structures can be built, taking due care to incorporate institutional and other characteristics of the particular case at hand. Such models would help to evaluate more precisely the economic policy options related to the liberalization process. In addition, it would be quite informative to compare the short-term results from this model with the longer-term results emerging from static general-equilibrium models that incorporate some type of welfare function. However, in conclusion, it should be reiterated that even in its current stage, the model has provided systematic general information on the macroeconomic effects associated with the liberalizing of trade and capital flows.

APPENDICES

I. Simulation Model

A. Equations

Production and Supply

Factor endowment

\begin{matrix} y_{t}^{*} = {\bar{y}}_{t}^{*} & (1) \end{matrix}

Supply of importable goods

\begin{matrix} I_{t}^{s} = λ_{1} {[\frac{γ_{2} γ_{3} y_{t}^{* 2} P i_{t}^{2}}{γ_{1} γ_{2} γ_{3} P i_{t}^{2} + γ_{1}^{2} γ_{3} P x_{t}^{2} + γ_{1}^{2} γ_{2} P n_{t}^{2}}]}^{\frac{1}{2}} + (1 - λ_{1}) I_{t - 1}^{s} & (2) \end{matrix}

Supply of exportable goods

\begin{matrix} X_{t}^{s} = λ_{2} {[\frac{(γ_{1}^{2} γ_{2}) y_{t}^{* 2} P x_{t}^{2}}{γ_{1} γ_{2} γ_{3} P i_{t}^{2} + γ_{1}^{2} γ_{3} P x_{t}^{2} + γ_{1}^{2} γ_{2} P n_{t}^{2}}]}^{\frac{1}{2}} + (1 - λ_{2}) X_{t - 1}^{s} & (3) \end{matrix}

Desired supply of nontradable goods

\begin{matrix} N_{t}^{s *} = {[\frac{(γ_{1}^{2} γ_{2} / γ_{3}) y_{t}^{* 2} P n_{t}^{2}}{γ_{1} γ_{2} γ_{3} P i_{t}^{2} + γ_{1}^{2} γ_{3} P x_{t}^{2} + γ_{1}^{2} γ_{2} P n_{t}^{2}}]}^{\frac{1}{2}} & (4) \end{matrix}

Supply of nontradable goods

\begin{matrix} \log N_{t}^{s} = \log N_{t}^{s *} + λ_{3} [\log N_{t}^{d} - \log N_{t}^{s *}] & (5) \end{matrix}

Real output

\begin{matrix} y_{t} = {[γ_{1} I_{t}^{s 2} + γ_{2} X_{t}^{s 2} + γ_{3} N_{t}^{s 2}]}^{\frac{1}{2}} & (6) \end{matrix}

Nominal income

\begin{matrix} Y_{t} = P i_{t} I_{t}^{s} + P x_{t} X_{t}^{s} + P n_{t} N_{t}^{s} + ϵ_{t} \cdot τ_{t} \cdot I_{t} & (7) \end{matrix}

Disposable income

\begin{matrix} {Y D}_{t} = Y_{t} - T_{t} - ϵ_{t} \cdot τ_{t} \cdot {P f}_{t} \cdot I_{t} & (8) \end{matrix}

Expenditures

Total private expenditures

\begin{matrix} \log E P R D_{t} = λ_{4} [γ_{4} \log Y D_{t} + γ_{5} (\log M_{t} - \log M_{t}^{d}) - γ_{6} r d_{t}] \\ + (1 - λ_{4}) \log E P R D_{t - 1} & (9) \end{matrix}

Private expenditures on goods

\begin{matrix} {E P}_{t} = {E P R D}_{t} - r d_{t} \cdot {B f}_{t} & (10) \end{matrix}

Total expenditures

\begin{matrix} E_{t} = {E P}_{t} + G_{t} & (11) \end{matrix}

Total expenditures on importable goods

\begin{matrix} Δ \log {(P i I^{d})}_{t} & = & Δ \log E_{t} + Δ \log P i_{t} + (1 / w_{i}^{d}) [- γ_{7} Δ \log P i_{t} \\ + (w_{x}^{d} + (γ_{7} + γ_{8} - γ_{9} - 1) / 2) Δ \log P x_{t} \\ + (w_{n}^{d} + (γ_{7} - γ_{8} + γ_{9} - 1) / 2) Δ \log P n_{t}] & (12) \end{matrix}

Total expenditures on exportable goods

\begin{matrix} Δ \log {(P x X^{d})}_{t} & = & Δ \log E_{t} + Δ \log P x_{t} \\ + (1 / w_{x}^{d}) [w_{n}^{d} + (γ_{7} + γ_{8} - γ_{9} - 1) / 2) Δ \log P i_{t} - γ_{8} Δ \log P x_{t} \\ + (w_{i}^{d} + (- γ_{7} + γ_{8} + γ_{9} - 1) / 2) Δ \log P n_{t}] & (13) \end{matrix}

Total expenditures on nontradable goods

\begin{matrix} Δ \log {(P n N^{d})}_{t} = Δ \log E_{t} + Δ \log P N_{t} \\ + (1 / w_{n}^{d}) [(w_{x}^{d} + (γ_{7} - γ_{8} + γ_{9} - 1) / 2) Δ \log P i_{t} \\ \begin{matrix} + (w_{i}^{d} + (- γ_{7} + γ_{8} + γ_{9} - 1) / 2) Δ \log P n_{t} \end{matrix}] & (14) \end{matrix}

Prices and Unemployment

Prices of importable goods

\begin{matrix} \log {P i}_{t} = λ_{5} [\log {P f}_{t} + \log ϵ_{t} + \log (1 + τ_{t})] + (1 - λ_{5}) \log P i_{t - 1} & (15) \end{matrix}

Prices of exportable goods

\begin{matrix} \log P x_{t} = \log P f_{t} + \log ϵ_{t} & (16) \end{matrix}

Prices of nontradable goods

\begin{matrix} Δ \log P n_{t} = λ_{6} [\log N_{t}^{d} - \log N_{t}^{s}] + λ_{7} Δ \log P i_{t} & (17) \end{matrix}

General price index

\begin{matrix} Δ \log P_{t} = w_{i}^{d} Δ \log {P i}_{t} + w_{x}^{d} Δ \log {P x}_{t} + w_{n}^{d} Δ \log {P n}_{t} & (18) \end{matrix}

Expected inflation

\begin{matrix} Δ Π_{t} = λ_{8} [Δ \log P_{t} - Π_{t - 1}] & (19) \end{matrix}

Resource unemployment

\begin{matrix} U_{t} = U_{0} + γ_{10} (\log y_{t}^{*} - \log y_{t}) & (20) \end{matrix}

Money and Credit

Nominal demand for money

\begin{matrix} \log M^{d} = α_{1} + γ_{11} \log Y_{t} - γ_{12} Π_{t} - γ_{13} {r d}_{t} & (21) \end{matrix}

Nominal money supply

\begin{matrix} M_{t} = {C R G}_{t} + {C R P}_{t} + R_{t} & (22) \end{matrix}

Domestic interest rate

\begin{matrix} Δ r d_{t} = γ_{14} [\log M_{t}^{d} - \log M_{t}] & (23) \end{matrix}

Balance of Payments

Imports (in foreign currency)

\begin{matrix} I_{t} = P f_{t} [I_{t}^{d} - I_{t}^{s}] & (24) \end{matrix}

Exports (in foreign currency)

\begin{matrix} X_{t} = P f_{t} [X_{t}^{s} - X_{t}^{d}] & (25) \end{matrix}

Current account

\begin{matrix} {C A}_{t} = ϵ_{t} [X_{t} - I_{t}] - r d_{t} \cdot B f_{t} & (26) \end{matrix}

Capital flows

\begin{matrix} {D K}_{t} = λ_{9} ϵ_{t} + β [γ_{15} ({r d}_{t} - r f_{t} - ρ_{t} - Δ \log ϵ_{t})] & (27) \end{matrix}

Balance of payments

\begin{matrix} {B P}_{t} = {C A}_{t} + {D K}_{t} & (28) \end{matrix}

Stock of international reserves

\begin{matrix} R_{t} = R_{t - 1} + {B P}_{t} & (29) \end{matrix}

Foreign debt

\begin{matrix} {B f}_{t} = {B f}_{t - 1} + {D K}_{t} & (30) \end{matrix}

Risk premium

\begin{matrix} ρ_{t} = ρ_{0} + γ_{16} {(B f / Y)}_{t} & (31) \end{matrix}

Government Sector

Government expenditures

\begin{matrix} G_{t} = T_{t} + ϵ_{t} \cdot τ_{t} \cdot P f_{t} \cdot I_{t} + {G D}_{t} & (32) \end{matrix}

Taxes

\begin{matrix} T_{t} = t_{0} + γ_{17} Y_{t} & (33) \end{matrix}

Credit to government

\begin{matrix} Δ {C R G}_{t} = {G D}_{t} & (34) \end{matrix}

Definitional Equations

Proportion of expenditures on importable goods

\begin{matrix} w_{i t}^{d} - {({P i I}^{d})}_{t} / E_{t} & (35) \end{matrix}

Proportion of expenditures on exportable goods

\begin{matrix} w_{x t}^{d} - {({P x X}^{d})}_{t} / E_{t} & (36) \end{matrix}

Proportion of expenditures on nontradable goods

\begin{matrix} w_{n t}^{d} - {({P n N}^{d})}_{t} / E_{t} & (37) \end{matrix}

B. Exogenous Variables

Ῡ ^* = potential real output
ϵ = exchange rate (index of units of domestic currency per unit of foreign currency), set equal to unity
τ = uniform nominal tariff on imports
Pf = index of foreign prices, equals 100
rf = foreign rate of interest, set equal to 0.1
β = index of restrictions on capital movements
CRP = credit to private sector
GD = government fiscal deficit

II. Values of Parameters Used in Simulations

Despite the seemingly large size of the model, it contains only 17 structural and 9 adjustment parameters and 4 constants. In choosing the values for these various coefficients, we were guided by two basic principles. First, the parameters should be consistent, in that the various theoretical restrictions implicit in the model, particularly relating to the supply and demand elasticities, be satisfied. Second, the combination of parameters should be such as to ensure that the model be dynamically stable and that it settle down to a steady state, which may or may not necessarily be equal to the original equilibrium. The specific values of the parameters used in the simulations reported in the text are given in Table 2.

Table 2.

Values of Parameters

Table 2.

Values of Parameters

		Parameter Values
Equation	Variable	Structural	Adjustment	Constant
(2‴)	N^s* = desired supply of nontradable goods	γ₃ = 0.3591	—	—
(4′)	I^s = supply of importable goods	γ₁ = 2.0	λ₁ = 0.8	—
(4″)	X^s = supply of exportable goods	γ₂ = 1-0	λ₂ = 0.4	—
(5)	N^s = supply of nontradable goods	—	λ₃ = 0	—
(11)	EPRD = private expenditures	γ₄ = 1.0	λ₄ = 1.0	—
		γ₅ = 0.3	—	—
		γ₆ = 0.5	—	—
(14′)	I^d = demand for importable goods	γ₇ = 0.4721	—	—
(14″)	X^d = demand for exportable goods	γ₈ = 0.0833	—	—
(14‴)	N^d— demand for nontradable goods	γ₉ = 0.4446	—	—
(17)	Pi = price of importable goods	—	λ₅ = 1.0	—
(19)	Pn = price of nontradable goods	—	λ₆ = 0.6	—
		—	λ₇ = 0.7	—
(21)	Π = expected inflation	—	λ₈ = 0.5	—
(22)	U = resource unemployment	γ₁₀ = 1.0	—	U₀ = 0.05
(23)	M^d = nominal demand for money	γ₁₁ = 1.0	—	α₁ = 0.2924
		γ₁₂ = 1.0	—	—
		γ₁₃ = 1.0	—	—
(25)	rd = domestic interest rate	γ₁₄ = 0.5	—	—
(29′)	DK = capital flows	γ₁₅ = 100.0	λ₉ = 0.0	—
(32)	ρ = risk premium	γ₁₆ = 0.2	—	γ₀ = 0.05
(35)	T = taxes	γ₁₇ = 0.1222	—	t₀ = 0.0

Table 2.

Values of Parameters

		Parameter Values
Equation	Variable	Structural	Adjustment	Constant
(2‴)	N^s* = desired supply of nontradable goods	γ₃ = 0.3591	—	—
(4′)	I^s = supply of importable goods	γ₁ = 2.0	λ₁ = 0.8	—
(4″)	X^s = supply of exportable goods	γ₂ = 1-0	λ₂ = 0.4	—
(5)	N^s = supply of nontradable goods	—	λ₃ = 0	—
(11)	EPRD = private expenditures	γ₄ = 1.0	λ₄ = 1.0	—
		γ₅ = 0.3	—	—
		γ₆ = 0.5	—	—
(14′)	I^d = demand for importable goods	γ₇ = 0.4721	—	—
(14″)	X^d = demand for exportable goods	γ₈ = 0.0833	—	—
(14‴)	N^d— demand for nontradable goods	γ₉ = 0.4446	—	—
(17)	Pi = price of importable goods	—	λ₅ = 1.0	—
(19)	Pn = price of nontradable goods	—	λ₆ = 0.6	—
		—	λ₇ = 0.7	—
(21)	Π = expected inflation	—	λ₈ = 0.5	—
(22)	U = resource unemployment	γ₁₀ = 1.0	—	U₀ = 0.05
(23)	M^d = nominal demand for money	γ₁₁ = 1.0	—	α₁ = 0.2924
		γ₁₂ = 1.0	—	—
		γ₁₃ = 1.0	—	—
(25)	rd = domestic interest rate	γ₁₄ = 0.5	—	—
(29′)	DK = capital flows	γ₁₅ = 100.0	λ₉ = 0.0	—
(32)	ρ = risk premium	γ₁₆ = 0.2	—	γ₀ = 0.05
(35)	T = taxes	γ₁₇ = 0.1222	—	t₀ = 0.0

Structural parameters

Little information is available on the price elasticities of the demand for and supply of importables, exportables, and nontradable goods, other than those for the United States contained in the study by Clements (1980). Other general-equilibrium models either use arbitrary values themselves or do not deal with sectors at the level of aggregation used in this study. For the supply side, therefore, we arbitrarily chose a value for the price parameter of importable goods (2.0) that was twice as large as the corresponding price coefficient of the supply of exportable goods. Developing countries in general, because of the type of goods that they export (primary and semimanufactured), are characterized as having a relatively low supply response of export goods, while the supply response of importables is considered to be larger. After these two coefficients were determined, the price coefficient of the supply of nontradable goods was readily obtained from the restriction that the weighted sum of the three must equal zero. The relevant weights themselves were calculated from national accounts data for six Latin American countries—Argentina, Brazil, Chile, Colombia, Mexico, and Uruguay—using the methodology outlined by Clements (1977) and Goldstein and Officer (1979). The production and expenditure weights obtained are shown in Table 3.

Table 3.

Six Latin American Countries: Weights of Importable, Exportable, and Nontradable Goods in Total Income and Expenditure, 1970–79¹

^¹Calculated as averages of the data for Argentina, Brazil, Chile, Colombia, Mexico, and Uruguay.

Table 3.

Six Latin American Countries: Weights of Importable, Exportable, and Nontradable Goods in Total Income and Expenditure, 1970–79¹

	Proportion of
Goods	Income	Expenditure
Importables	0.35	0.47
Exportables	0.17	0.08
Nontradables	0.48	0.45

^¹Calculated as averages of the data for Argentina, Brazil, Chile, Colombia, Mexico, and Uruguay.

Table 3.

Six Latin American Countries: Weights of Importable, Exportable, and Nontradable Goods in Total Income and Expenditure, 1970–79¹

	Proportion of
Goods	Income	Expenditure
Importables	0.35	0.47
Exportables	0.17	0.08
Nontradables	0.48	0.45

^¹Calculated as averages of the data for Argentina, Brazil, Chile, Colombia, Mexico, and Uruguay.

The price elasticities of demand for importables (γ₇) and exportables (γ₈) were calculated as an average of the price elasticities of demand for imports and exports, respectively, as reported by Khan (1974). Again, the use of the weights for expenditures (Table 3) yielded the estimate of the price elasticity of demand for nontradable goods.

The parameters in the private expenditures equation are based on the empirical results of Aghevli and Khan (1980) and Knight and Mathieson (1983), and the money demand coefficients correspond to those obtained for developing countries by Khan (1980), Mathieson (1981; 1983), and others. The risk premium parameter (γ₁₆) is selected so that when the ratio of the stock of external debt to nominal income is 0.25, and the constant risk premium is 0.05, the total risk premium should be 0.1. The tax parameter (γ₁₇) is also set so as to maintain equality between tax revenues and nominal income, assuming the constant to be zero. The remaining parameters were imposed to guarantee a consistent initial equilibrium.

Adjustment parameters

The adjustment parameters were selected for the most part to ensure non-instantaneous adjustment to equilibrium, in keeping with the intention of tracing out the transition path of certain important variables. The main distinctions were made in the supply sector and the determination of the price of nontradable goods. Basically, it was assumed that the supply of importables would generally respond faster than the supply of exportable goods, implying thus that λ₁ > λ₂. On the nontradable goods side, λ₃ was set equal to zero in the simulations reported in the text, meaning that adjustment of the quantity of nontradables responded to changes in demand through variations in prices rather than directly. The parameter λ₃ was, however, allowed to vary so as to permit a degree of direct demand effects on the behavior of nontradable goods. The price of nontradable goods was assumed to adjust slowly to excess demand because of “stickiness,” and in the long run was expected to grow at a rate somewhat less than the rate of increase of the price of importables.⁶¹ For the present analysis it was assumed (although it is not necessary to do so) that the price of importables adjusted immediately to a change in tariffs, that is, that λ₅ = 1. The coefficient of expectations (λ₈) is an approximate average of the values obtained by Khan (1980) in the context of a study of the demand for money function for a group of developing countries. No autonomous capital flows were permitted so that λ₉ was initially set to zero.

To allow for some permanent or “normal” level of underemployment, U₀ was fixed at 5 percent. The constant in the money demand equation was chosen simply to ensure equilibrium in the money market, given the initial values of the variables involved and the relevant parameters. The constant risk premium is assumed to be 5 percent, and the average tax rate zero.

While there is no doubt that the choice of parameters for the simulation experiments is, in the final analysis, essentially arbitrary, the exercise was repeated by varying certain key parameters. The sensitivity analysis, while yielding different transition patterns for some of the variables, did not qualitatively change the results or the main conclusions of the study.⁶²

REFERENCES

Aghevli, Bijan B., and Mohsin S. Khan (1978), “Government Deficits and the Inflationary Process in Developing Countries,” Staff Papers, Vol. 25 (September 1978), pp. 383–416.
- Search Google Scholar
- Export Citation
Aghevli, Bijan B., and Mohsin S. Khan (1980), “Credit Policy and the Balance of Payments in Developing Countries,” in Money and Monetary Policy in Less Developed Countries: A Survey of Issues and Evidence, ed. by Warren L. Coats, Jr., and Deena R. Khatkhate (Oxford, 1980), pp. 685–711.
- Search Google Scholar
- Export Citation
Alexander, Sidney S., “Effects of Devaluation on a Trade Balance,” Staff Papers, Vol. 2 (April 1952), pp. 263–78.
- Search Google Scholar
- Export Citation
Ayza, Juan, Gerard Fichet, and Norberto Gonzalez, América Latina: Integración económica y sustitución de importaciones, Comisión económica para América Latina (Mexico City, 1975).
- Search Google Scholar
- Export Citation
Balassa, Bela, The Process of Industrial Development and Alternative Development Strategies, Essays in International Finance, No. 141, International Finance Section, Princeton University (December 1980).
- Search Google Scholar
- Export Citation
Bhagwati, Jagdish N., and T.N. Srinivasan, “Trade Policy and Development,” in International Economic Policy: Theory and Evidence, ed. by Rudiger Dornbusch and Jacob A. Frenkel (Johns Hopkins University Press, 1979), pp. 1–35.
- Search Google Scholar
- Export Citation
Blejer, Mario I. (1977), “The Short-Run Dynamics of Prices and the Balance of Payments,” American Economic Review, Vol. 67 (June 1977), pp. 419–28.
- Search Google Scholar
- Export Citation
Blejer, Mario I. (1982), “Interest Rate Differentials and Exchange Risk: Recent Argentine Experience,” Staff Papers, Vol. 29 (June 1982), pp. 270–79.
- Search Google Scholar
- Export Citation
Blejer, Mario I., and Roque B. Fernandez, “The Effects of Unanticipated Money Growth on Prices and on Output and Its Composition in a Fixed-Exchange-Rate Open Economy,” Canadian Journal of Economics, Vol. 13 (February 1980), pp. 82–95.
- Search Google Scholar
- Export Citation
Brown, Alan, and Angus Deaton, “Surveys in Applied Economics: Models of Consumer Behavior,” Economic Journal, Vol. 82 (September 1972), pp. 1145–236.
- Search Google Scholar
- Export Citation
Cagan, Phillip, “The Monetary Dynamics of Hyperinflation,” in Studies in the Quantity Theory of Money, ed. by Milton Friedman (University of Chicago Press, 1956), pp. 25–117.
- Search Google Scholar
- Export Citation
Clements, Kenneth W. (1977), “The Trade Balance in Monetary General Equilibrium” (unpublished doctoral dissertation, University of Chicago, June 1977).
- Search Google Scholar
- Export Citation
Clements, Kenneth W. (1980), “A General Equilibrium Econometric Model of the Open Economy,” International Economic Review, Vol. 21 (June 1980), pp. 469–88.
- Search Google Scholar
- Export Citation
Derviş, Kemal, Jaime de Melo, and Sherman Robinson, General Equilibrium Models for Development Policy (Cambridge University Press, 1981).
- Search Google Scholar
- Export Citation
Diaz-Alejandro, Carlos F., “Southern Cone Stabilization Plans,” in Economic Stabilization in Developing Countries, ed. by William R. Cline and Sidney Weintraub, The Brookings Institution (Washington, 1981), pp. 119–41.
- Search Google Scholar
- Export Citation
Dornbusch, Rudiger, and Michael Mussa, “Consumption, Real Balances and the Hoarding Function,” International Economic Review, Vol. 16 (June 1975), pp. 415–21.
- Search Google Scholar
- Export Citation
Feltenstein, Andrew, “A General Equilibrium Approach to the Analysis of Trade Restrictions, with an Application to Argentina,” Staff Papers, Vol. 27 (December 1980), pp. 749–84.
- Search Google Scholar
- Export Citation
Ffrench-Davis, Ricardo, Economía internacional: Teorías y políticas para el desarrollo, Fondo de Cultura Económica (Mexico, 1979).
- Search Google Scholar
- Export Citation
Frenkel, Jacob A., “The Order of Economic Liberalization: Lessons from Chile and Argentina—A Comment,” in Economic Policy in a World of Change, ed. by Karl Brunner and Allan H. Meltzer, Carnegie-Rochester Conference Series on Public Policy, Vol. 17 (Amsterdam, 1982), pp. 199–201.
- Search Google Scholar
- Export Citation
Frenkel, Jacob A., and Harry G. Johnson (eds.), The Monetary Approach to the Balanceof Payments (London, 1976).
- Search Google Scholar
- Export Citation
Fry, Maxwell (1980), “Saving, Investment, Growth and the Cost of Financial Repression,” World Development, Vol. 8 (April 1980), pp. 317–27.
- Search Google Scholar
- Export Citation
Fry, Maxwell (1981), “Interest Rates in Asia” (unpublished, International Monetary Fund, June 25, 1981).
- Search Google Scholar
- Export Citation
Fry, Maxwell (1982), “Models of Financially Repressed Developing Economies,” World Development, Vol. 10 (September 1982), pp. 731–50.
- Search Google Scholar
- Export Citation
Gaba, Ernesto, “The Argentine Financial Reform—Lessons from an Experience,” paper presented at the XVIII Central Bank Technicians Meeting, held in the Dominican Republic (November 1981).
- Search Google Scholar
- Export Citation
Goldstein, Morris, and Lawrence H. Officer, “New Measures of Prices and Productivity for Tradable and Nontradable Goods,” Review of Income and Wealth, Series 25 (December 1979), pp. 413–27.
- Search Google Scholar
- Export Citation
Goldstein, Morris, and Mohsin S. Khan (1982 a), Effects of Slowdown in Industrial Countries on Growth in Non-Oil Developing Countries, Occasional Paper No. 12, International Monetary Fund (Washington, August 1982).
- Search Google Scholar
- Export Citation
Goldstein, Morris, and Mohsin S. Khan (1982 b), “Income and Price Effects in Foreign Trade” (unpublished, International Monetary Fund, September 20, 1982).
- Search Google Scholar
- Export Citation
International Monetary Fund, The Monetary Approach to the Balance of Payments (Washington, 1977).
- Search Google Scholar
- Export Citation
Keesing, Donald B., “Trade Policy for Developing Countries,” World Bank Staff Working Paper No. 353 (August 1979).
- Search Google Scholar
- Export Citation
Khan, Mohsin S. (1974), “Import and Export Demand in Developing Countries,” Staff Papers, Vol. 21 (November 1974), pp. 678–93.
- Search Google Scholar
- Export Citation
Khan, Mohsin S. (1980), “Monetary Shocks and the Dynamics of Inflation,” Staff Papers, Vol. 27 (June 1980), pp. 250–84.
- Search Google Scholar
- Export Citation
Khan, Mohsin S., and Malcolm D. Knight, “Stabilization Programs in Developing Countries: A Formal Framework,” Staff Papers, Vol. 28 (March 1981), pp. 1–53.
- Search Google Scholar
- Export Citation
Knight, Malcolm D., and Clifford R. Wymer, “A Macroeconomic Model of the United Kingdom,” Staff Papers, Vol. 25 (December 1978), pp. 742–78.
- Search Google Scholar
- Export Citation
Knight, Malcolm D., and Donald J. Mathieson, “Economic Change and Policy Response in Canada Under Fixed and Flexible Exchange Rates,” Ch. 18 in Economic Interdependence and Flexible Exchange Rates, ed. by Jagdeep S. Bhandari and Bluford H. Putnam (MIT Press, 1983), pp. 500–29.
- Search Google Scholar
- Export Citation
Krueger, Anne O., and others (eds.), Trade and Employment in Developing Countries: Vol. 1—Individual Studies (University of Chicago Press, 1981).
- Search Google Scholar
- Export Citation
Massad, Carlos, “El enfoque monetario de la balanza de pagos y el enfoque de absorción: ¿Resultados contradictorios?” Estudios de Economía, No. 15 (Primer Semestre, 1980), pp. 97–128.
- Search Google Scholar
- Export Citation
Mathieson, Donald J. (1979), “Financial Reform and Capital Flows in a Developing Economy,” Staff Papers, Vol. 26 (September 1979), pp. 450–89.
- Search Google Scholar
- Export Citation
Mathieson, Donald J. (1980), “Financial Reform and Stabilization Policy in a Developing Economy,” Journal of Development Economics, Vol. 7 (September 1980), pp. 359–95.
- Search Google Scholar
- Export Citation
Mathieson, Donald J. (1981), “Inflation, Interest Rates, and the Balance of Payments During a Financial Reform: The Case of Argentina” (unpublished, International Monetary Fund, June 18, 1981).
- Search Google Scholar
- Export Citation
Mathieson, Donald J. (1983), “Estimating Models of Financial Market Behavior During Periods of Extensive Structural Reform: The Experience of Chile,” Staff Papers, Vol. 30 (June 1983), pp. 350–93.
- Search Google Scholar
- Export Citation
McKinnon, Ronald I., “The Order of Economic Liberalization: Lessons from Chile and Argentina,” in Economic Policy in a World of Change, ed. by Karl Brunner and Allan H. Meltzer, Carnegie-Rochester Conference Series on Public Policy, Vol. 17 (Amsterdam, 1982), pp. 159–86.
- Search Google Scholar
- Export Citation
Mezzera, Jaime, “El proceso de apertura uruguayo en la esfera real” (unpublished, Economic Commission for Latin America, January 1981).
- Search Google Scholar
- Export Citation
Mikesell, Raymond F., and James E. Zinser, “The Nature of the Savings Function in Developing Countries: A Survey of the Theoretical and Empirical Literature,” Journal of Economic Literature, Vol. 11 (March 1973), pp. 1–26.
- Search Google Scholar
- Export Citation
Stone, Richard, “Linear Expenditure Systems and Demand Analysis: An Application to the Pattern of British Demand,” Economic Journal, Vol. 64 (September 1954), pp. 511–27.
- Search Google Scholar
- Export Citation
Villanueva, Javier, “La industrialización latinoamericana y las nuevas condiciones de la economía internacional,” Integración Latinoamericana (No. 31, December 1978), pp. 3–31.
- Search Google Scholar
- Export Citation
Weisskopf, Thomas E., “The Impact of Foreign Capital Inflow on Domestic Savings in Underdeveloped Countries,” Journal of International Economics, Vol. 2 (February 1972), pp. 25–38.
- Search Google Scholar
- Export Citation
Wonsewer, Israel, and Darío Saráchaga, “La apertura financiera” (unpublished, Comisión Económica para América Latina, Montevideo, October 1980).
- Search Google Scholar
- Export Citation
Zahler, Roberto, “The Monetary and Real Effects of the Financial Opening Up of National Economies to the Exterior: The Case of Chile, 1975–1978,” CEPAL Review (No. 10, April 1980), pp. 127–54.
- Search Google Scholar
- Export Citation

Mr. Khan, Advisor in the Research Department, is a graduate of Columbia University and the London School of Economics and Political Science.

Mr. Zahler, Economist, Economic Commission for Latin America, is a graduate of the University of Chile and the University of Chicago. He was formerly a member of the Research Department of the Central Bank of Chile and Professor of the Graduate Program of Latin American Economic Studies (escolatina).

This study is the output of a joint project undertaken by the Economic Commission for Latin America (CEPAL) and the International Monetary Fund. The CEPAL participation was conducted within the framework of UNDP Project RLA/77/021, Implications for Latin America of the Situation of the International Monetary System, directed by Carlos Massad. The authors are grateful to Ricardo Arriazu, Reynaldo E. Bajraj, Andres Bianchi, Kenneth W. Clements, Jacob A. Frenkel, Ernesto Gaba, Norberto Gonzalez, Carlos Massad, and Daniel Schydlowsky, as well as some of Mr. Khan’s colleagues in the Fund, whose many comments and suggestions at various stages of the work were extremely valuable. Nicolas Eyzaguirre not only provided excellent research assistance but also contributed to the development of the model.

The United Nations has published this study under the title, Efectos Macroeconómicos de Cambios en las Barreras al Comercio y al Movimiento de Capitales: un Modelo de Simulación (Sales No. S.83.II.G.12, Santiago de Chile, December 1982).

For a detailed study on the effects of the changing international environment on developing countries, see Goldstein and Khan (1982 a).

Some of these factors have been covered in the studies by Ayza and others (1975), Villanueva (1978), and Ffrench-Davis (1979).

See, for example, the models of Clements (1977; 1980) and Feltenstein (1980).

These models include those developed by Blejer (1977), Knight and Wymer (1978), Aghevli and Khan (1980), Blejer and Fernandez (1980), and Khan and Knight (1981).

See Frenkel and Johnson (1976) and IMF (1977). For a more recent discussion of the relationship between the monetary and the absorption approaches, see Massad (1980).

Both Clements (1980) and Feltenstein (1980), for example, deal with models in continuous equilibrium.

The choices could include pegging the nominal exchange rate, crawling peg regimes, and preannounced exchange rate schedules.

This is the type of approach taken in the larger computational general-equilibrium models. See Feltenstein (1980).

The assumption of a uniform tariff does not involve any loss of generality and has been made only for simplicity.

This assumption was made solely for the simulation experiments. The model is able to simultaneously handle both domestic policy changes and alternative international scenarios.

See McKinnon (1982) and Frenkel (1982) for a discussion of some of these issues.

Such financial reforms, taking into account the complex interactions between the money and credit markets, distinctions between lending and borrowing rates, and differential reserve requirements, etc., have been analyzed in detail for various countries by Mathieson (1979; 1980), Wonsewer and Saráchaga (1980), Zahler (1980), Fry (1981; 1982), and Gaba (1981).

See, for example, Weisskopf (1972), Mikesell and Zinser (1973), and Fry (1980).

This interpretation requires assuming that, at the aggregate level, a reasonably stable relationship exists between unemployment and total real output. However, as the productivity of labor in different sectors is usually heterogeneous, changes in the level and structure of production may be quite different with respect to their impact on employment.

This discussion usually involves considerations that are not easy to handle in a model such as the one used here. These considerations include the uneven structure of the protection given to consumption and investment goods, possible initial situations of disequilibrium with respect to stocks of consumption goods (usually associated with an excess demand for such goods), the impact of opening up on the demand for domestic credit and, in a more general sense, on the evolution of the capital market.

Namely, that government spending is constrained by fiscal revenues and borrowing, and the current account of the balance of payments is equal to the excess of absorption over income (Alexander (1952)).

In general, A is a positive definite symmetrical matrix of parameters, but for simplicity it has been assumed to be diagonal.

Gross substitutability has been assumed.

For tradable goods, quantity adjustments take place via the foreign sector.

The link via prices is discussed later in this paper.

In a sense, allowing for both price and quantity adjustments is a generalization of the Clements (1980) model in which prices clear instantaneously to keep supply of and demand for nontradable goods in continuous equilibrium.

The assumption of a quadratic form of the transformation function yields the equation for real output in this form.

See Aghevli and Khan (1980), Clements (1980), and Knight and Mathieson (1983) for similar formulations.

In fact, consumption expenditures could very well decline as well. However, here, as mentioned before, no distinction is made between the two types of expenditure.

It does not make any important difference to assume either a zero or a positive initial stock of foreign debt.

This assumption is standard in studies of demand systems. See, for example, the survey by Brown and Deaton (1972).

The use of this function was considered to be more general than the specific forms associated with the linear-expenditure system (Stone (1954)). This was possible because one is not actually estimating the system and does not have to be overly concerned with the complexity of the form.

Because one is working with a general utility function, unlike the supply equations, the variables are in rates of change rather than in levels.

It is assumed that a single index of foreign prices is relevant to importables and exportables alike. This assumption does not introduce any loss of generality, since, if it were desired to study, for example, the impact of variations in the terms of trade, two different foreign price indices could be employed.

Actually, the variable τ is assumed to include all types of imperfections and distortions (such as transport costs) that keep the domestic price of importables different from the foreign price index adjusted for the exchange rate.

Indeed, it is frequently argued that opening up by lowering tariffs does not lead to an immediate fall in Pi, thus allowing importers to obtain as profits the revenues that previously accrued to the government.

For a rationale of this formulation, see Khan and Knight (1981), Appendix II.

A large value for λ₆ yields instantaneous equilibrium, as in the model of Clements (1980).

It should be stressed that resources engaged in reallocation between sectors are also viewed here as unemployment.

As long as γ₁₁ is unity, this function is exactly the same as an equation specified in real terms.

Or, alternatively, the net effect of a change in, say, deposit interest rates on broad money is zero, as some components of the money stock rise while others fall. For a discussion of this issue, see Mathieson (1981; 1983).

Strictly speaking, it is the combination of γ₁₄ and the interest elasticity of the demand for money, γ₁₃, that determines how quickly the interest rate moves to eliminate disequilibrium in the monetary sector.

This formulation corresponds to the so-called perfect-substitutes model of import behavior. See Goldstein and Khan (1982 b).

Equation (29) corresponds to what has been termed in the literature the stock specification for international capital flows.

To be precise, gradual opening up in the simulation experiments is defined as changing β from zero to one half.

For the present, the issue of government bonds has not been taken into consideration because of the complexities involved in the possible “net wealth effects,” the effects that this would have on the government budget constraint, and the problems involved in modeling market segmentation of private and government bonds.

Relaxing these rather restrictive assumptions greatly complicates the structure of the model. To properly handle these issues, one would have to formulate a complete portfolio model.

This approach has been adopted because the equilibrium properties of the model related to the fixed transformation curve require that the rate of interest return to its original level eventually. Therefore, any capital account liberalization will lower the domestic interest rate, but as p rises, this rate is pushed back up. Note that in the original equilibrium, one has

$r d > r f + ρ + Δ \log ϵ$

but in the final equilibrium, because p has increased, one has interest parity

$r d = r f + ρ + Δ \log ϵ$

For a model of this type, see Aghevli and Khan (1978).

See Section I.

This amounts to assuming that the foreign price level and foreign interest rate are constant.

This is a strong simplifying assumption, since in fact most liberalization policies have typically been accompanied by stabilization programs. See, for example, Diaz-Alejandro (1981).

Although the unit of observation is not defined explicitly, the parameters utilized are either independent of time or correspond to yearly data. The first two periods in all the simulations correspond to the initial equilibrium so that all policy changes are assumed to occur in the third period.

These effects indicate that the reduction in tariffs would lead to an increase in imports and exports.

While the nominal interest rate declines, as the decline in expected inflation is greater, the real interest rate actually rises during the process.

One is working with the assumption that the demand for nontradable goods does not have a direct impact on the supply (λ₃ = 0). If λ₃ were positive, then, as there is initially an excess supply in the nontradables market, the magnitude of the unemployment of resources would be greater.

The creation and duration of unemployment is one of the main aspects of a trade liberalization strategy that has been the subject of criticism.

Recall that the change in policy is always assumed to take place in the third period.

That is, the foreign interest rate plus the increased risk premium.

See, for example, Blejer (1982).

It is assumed that interest payments on foreign debt begin immediately and that international reserves earn no interest so that one can abstract from the concept of “net” interest payments.

This neutrality no doubt reflects the relative speeds at which the goods and financial markets clear.

Since the results for the shock scenario are similar, only the results for the gradual opening up are reported here.

The four periods correspond to the complete implementation of the gradual tariff reduction.

For experiments of a similar nature, see Khan and Knight (1981).

As the period of analysis does not correspond strictly to a full long-run situation, it was not deemed necessary to set λ₇ = 1.

The results of these additional exercises are available from Mr. Khan, whose address is Research Department, International Monetary Fund, Washington, D.C. 20431.

Other Publishers

Save
Cite
Email this content

Share Link

Copy this link, or click below to email it to a friend
Email this content
or copy the link directly:

https://www.elibrary.imf.org/view/journals/024/1983/002/article-A001-en.xml
The link was not copied. Your current browser may not support copying via this button.

Link copied successfully

IMF Staff papers: Volume 30 No. 2

Author: International Monetary Fund. Research Dept.

Volume/Issue:: Volume 1983: Issue 002

Publisher:: International Monetary Fund

ISBN:: 9781451946895

ISSN:: 1020-7635

Pages:: 242

DOI:: https://doi.org/10.5089/9781451946895.024

Topics

Countries

Series

About

Resources

Abstract

I. Specification of the Model

Production and Supply

Expenditures

Prices and Unemployment

Money and Credit

Balance of Payments

Government Sector

II. Simulation Experiments

Gradual and Sudden Reduction in Tariffs

Gradual and Sudden Removal of Restrictions on Capital Flows

Simultaneous Gradual Removal of Restrictions on Trade and Capital Flows

Sequential Gradual Removal of Restrictions on Trade and Capital Flows

Compensatory Policies

III. Conclusions

APPENDICES

I. Simulation Model

A. Equations

Production and Supply

Expenditures

Prices and Unemployment

Money and Credit

Balance of Payments

Government Sector

Definitional Equations

B. Exogenous Variables

II. Values of Parameters Used in Simulations

Structural parameters

Adjustment parameters

REFERENCES

Same Series

Other IMF Content

Other Publishers

Asian Development Bank

Inter-American Development Bank

Nordic Council of Ministers

The World Bank

Share Link

Table of Contents