Effectiveness of Exchange Rate Policy for Trade Account Adjustment

ALFRED STEINHERR*

Abstract

ALFRED STEINHERR*

ALFRED STEINHERR*

During the last decade, frequent autonomous or induced changes in exchange rates that are in excess of inflation rate differentials have taken place. The magnitudes of these real exchange rate changes are often considerable, yet trade balances have tended to respond sluggishly and, in several cases, have even exhibited a “perverse” movement. For example, the major real appreciations of the yen and deutsche mark and depreciation of the U. S. dollar that occurred in the late 1960s and early 1970s did not lead for many years to any weakening of the current accounts of Japan and the Federal Republic of Germany or strengthening of the U. S. current account. 1 The length of time over which such responses have been observed makes it difficult to dismiss them as mere short-term J-curve phenomena. 2

The monetary approach to the balance of payments provides an explanation for the absence of any real effects of exchange rate adjustments in the steady state of an economy. This paper focuses on the intermediate run with the aim of explaining why exchange rate policy may sometimes lead to perverse results. For this purpose, a framework is developed that is consistent with the long-run predictions of the monetary approach but that highlights the adjustment of the trade account within a time frame shorter than the one for steady-state equilibrium. Throughout the paper, the exchange rate is treated as an exogenous variable.

The novel feature of the approach in this paper is a distinction, within a general equilibrium model, of industries that sell or buy at a given world market price from those that face a downward-sloping demand curve. The latter industries may represent oligopolies in the world market, or may correspond to a collection of domestic firms, each of which acts as a price taker but manufactures goods for which no perfect substitutes are being produced in the rest of the world. Motivated by the present focus on foreign trade, all these industries are called price setters (the S-sector), while industries in the first group are referred to as price takers (the T-sector). In the presence of nontraded goods, the simple two-sector structure may still be retained by aggregating nontraded goods with traded goods produced by price setters; in both cases the domestic industry, or even the firm, faces a downward-sloping demand curve. Among other things, the S-sector’s average propensity to export would then reflect the share of non-traded goods. As to imports, the assumption is made throughout this paper that the domestic economy is a price taker.3

What makes this framework worthy of interest and, from the present perspective, preferable to that based on the distinction between tradable and nontradable goods? 4 First of all, since the terms of trade become endogenous in this framework, it may be of greater empirical relevance to countries for which the small-country assumption stretches the imagination excessively: unlike the tradables/nontradables approach, with its infinite elasticity of substitution between traded goods produced at home and abroad, the present approach lends itself more readily to an elasticity analysis of the trade account without, however, sharing the partial-equilibrium shortcomings of the traditional elasticities approach. Second, most so-called nontradables are, in principle, tradable but trade does not take place because imperfections in domestic markets are likely to exist, and these goods fall naturally into the proposed category of S-goods. 5 The imperfections are caused by government monopoly, trade barriers, prohibitive transport and information costs, etc. 6

This framework, set forth and analyzed in Sections I and II, was applied to a study of the effects of revaluation of the deutsche mark on the structure of the economy of the Federal Republic of Germany (Morel and Steinherr (1978); Steinherr and Morel (1979 a; 1979 b)). Most interpretations of the long-time current account surplus of the Federal Republic of Germany—despite real revalutions of the deutsche mark—have invoked low elasticities of the demand for that country’s export products, and below-average income growth there, without focusing, however, on the relationship between relative prices and income growth. This interdependence is exerted through various channels that can be fully appreciated only in a general equilibrium model, such as the one proposed in this paper. Real revaluations that lead to terms of trade gains tend to raise income by more than what is indicated by the growth of gross domestic product. However, an increase in the terms of trade tends to lower, ceteris paribus, the volume of exports relative to the volume of imports and thereby exerts a downward pressure on the domestic employment level. Finally, in the Federal Republic of Germany, comparison of export unit values with domestic unit labor costs suggests that the appreciation of the deutsche mark combined with a downward rigidity of nominal wages had been translated into an increasing loss of profitability, and that exporters there had accepted lower profit margins to hold on to their export market shares. This fact squares with the substantial decline of the share of gross capital formation in gross domestic product in the 1970s. Thus, the slow growth of the economy of the Federal Republic of Germany in the 1970s is in part due to the slowdown of domestic investment, accompanied by a sharp rise in investment abroad, which itself is related to the gradual loss of competitiveness of the economy on world markets. While it is true that slow income growth has had a positive influence on the trade balance, this argument is not independent of relative price, and hence exchange rate, developments.

If income could somehow be kept fixed during a relative price change, attention could be concentrated on price elasticities in foreign trade. Most studies concur in finding low values for these elasticities. For example, in the world trade model estimated by Deppler and Ripley (1978), for 2 out of 5 countries the long-run import price elasticities are below unity, and the long-run export price elasticities are below unity for 7 out of 11 countries. These are estimates for trade in manufacturing, and elasticities for total trade or for the current account can be expected to be much lower. Such low values for price elasticities imply a substantial degree of product differentiation in foreign trade that should become an integral part of trade account modeling. This would be necessary in order to capture the effects of changes in market structure initiated, for example, by a change in relative prices. Moreover, owing to the empirical problem of aggregation over markets and commodities, a serious bias could taint the estimates of foreign trade elasticities so that it might still be desirable to search for explanations not relying exclusively on elasticity pessimism.

In Morel and Steinherr (1978), an attempt has been made to incorporate market structure into the analysis of the trade account of the Federal Republic of Germany. The industries of its manufacturing sector are classified on the basis of their pricing behavior on international markets as either price setters or price takers. In 1978 about one half of the manufacturing exports originated in industries with some market power, and those industries’ average propensity to export (40 per cent) substantially exceeded that of the price-taking sector (27 per cent).

The disaggregation of manufacturing into price-taking and price-setting sectors obviously has implications for the effects of a revaluation, since the relative price between these goods now plays a key role. Consider a revaluation. By definition, the domestic currency price of T-goods falls by the full rate of revaluation, while the price of S-goods is likely to decrease less than proportionately, owing to cost-plus pricing practices or to excess foreign demand, etc. In most cases, a higher relative price of S-goods would lead to a reduction of both domestic and foreign demand, and, in the absence of compensatory fiscal demand, to a downward adjustment of production and, hence, higher unemployment. However, in many instances, and in the Federal Republic of Germany, fiscal policy tended to react countercyclically to stabilize employment. As a consequence of higher relative prices for S-goods, in effect sustained by fiscal policy, Steinherr and Morel (1979 a) have verified that productive resources over time had moved into the S-sector. Since productivity in the S-sector had been growing at a lower rate than in the T-sector, the economy’s average productivity growth had declined as a consequence of the shift in resources. Thus, after an increase in the country’s prices relative to world prices, the change in the value of exports and imports certainly depended on the elasticity of world demand, but the rise of imports was dampened by a slower growth of employment (reinforced by a shift from private to fiscal expenditure with the latter falling predominantly on domestic goods) and lower long-run income growth owing to slower average productivity growth. Clearly, these income effects on imports were direct consequences of higher prices in the Federal Republic of Germany relative to those in the rest of the world.

Within this framework of noncompetitive pricing, another reason for inelastic export demand (and, hence, perverse exchange rate effects) that may be particularly relevant for some sub-periods is consumer rationing on export markets. Although rationing cannot characterize a state of long-run equilibrium, it arises in the short run when domestic producers are unable to step up production enough to satisfy growing world demand and fail to adjust prices sufficiently for market clearing. International market penetration as portrayed by the product-cycle theory, together with both the higher costs of foreign as opposed to domestic market exploration and the additional risks faced abroad, suggests that rationing of foreign demand may occur more frequently than rationing of domestic demand. Another theory consistent with the rationing of foreign demand is represented by the “export-led growth” hypothesis that is often applied to the growth of the Federal Republic of Germany: foreign demand is rationed as domestic production lags behind growing demand, while prices are set according to a long-run notion of market equilibrium and not so as to clear the market in the short run. 7 In this case, revaluation produces a shift of resources into the sector for exportables as long as export prices remain below their long-run equilibrium levels, facilitating an increase in the volume of exports despite higher prices in foreign currency. The “observed” elasticity of world demand for domestic products would then be positive, and an increase in the trade surplus seems to be a likely result of revaluation.

Thus, Steinherr and Morel (1979 a) argue that the trade surplus of the Federal Republic of Germany was increasing, at least for a substantial period, as a result of the real appreciation of the deutsche mark. This perverse effect derived, certainly, from low trade elasticities but also from a variety of other factors. In particular, lower productivity growth and lower investment, partly as a result of revaluation, tended to slow down income growth and hence the growth of import demand. While world demand for exports from the Federal Republic of Germany was found to be inelastic, export performance was reinforced after revaluation by the sectoral resource reallocation toward more export-intensive industries with enhanced priceset-ting power.

The model developed in Section I was used in Steinherr and Morel (1979 b) to simulate revaluation of the deutsche mark. The simulated change in the trade account depends mainly on whether fiscal or monetary policy is passive or responds to revaluation, on whether nominal wages are rigid in nominal or real terms, and on a number of other factors. In all situations, perverse effects are possible outcomes.

Section I of this paper develops the framework of the model. Section II discusses the results derived from the model. The analysis is grouped into three parts according to its three focuses: in the first part, the adjustment process for the asset accumulation approach is derived. In the second part, a fiscal sector is incorporated into the analysis, followed by a comparison of the effects of exchange rate policy with those of fiscal policies. Wage rigidities and their effects on the results of the model are presented in the third part. While not all results of this section are entirely novel, some are; and an additional merit might be seen in the presentation of a model able to embrace in a consistent manner various approaches to balance of trade analysis and apt to indicate the restrictions required to obtain those results that are familiar from the elasticity, absorption, and asset theories.

I. The Model

For the theoretical analysis that follows, the framework discussed earlier has been simplified. In particular, all dynamic aspects regarding investment and productivity growth are not considered. On the basis of previous discussion, the economy is dichotomized. For all industries in the first group, the market price is a parameter. Industries in the second group face a downward-sloping demand curve. For simplicity of structure as well as empirical relevance, import-competing industries are associated with the first group, while industries producing exportables are linked to the second category. Thus, the economy is considered to be small enough not to affect import prices and to possess specialized production knowledge for some of its exports. 8

The economy then produces quantity XS at unit price PS for local consumption and export (“exportables”), and quantity XT at unit price PT for local consumption as a perfect substitute for imports (“importables”). 9 The price of importables is determined by the world market so that

PT=P˜Te(1)

where P˜T denotes the world market price and e the exogenous exchange rate, expressed as the local currency price of one unit of foreign exchange.

Since P˜T is exogenous and can be considered fixed, PT is set equal to one, and relative prices, equal to the terms of trade, are defined as

q=Ps/e(2)

Determination of the price of exportables is described later.

At any time, factor supplies are given. With only two factors of production, labor (L) and capital (K), the production possibilities of XS and XT (measured as value added) are obtained from the transformation function

τ(Xs,XT;K,L)=0(3)

The production levels of both goods are functions of the relative price q, with the restrictions on the partial derivatives ∂XS/∂q ≡ XSq as indicated:

XS=Xs(q);XSq0XT=XT(q);XTq0(4)

The real value of domestic production (value added) is defined as

X=(PsXs+eXT)/P(5)

with the expenditure price index (P) equal to

P=αPs+(1α)e(6)

where α is the share of S-goods, and (1 - α) the share of T-goods in overall domestic expenditure. To obtain real disposable income, tax payments are subtracted from equation (5). To separate fiscal and monetary policies, the balanced budget condition is imposed so that the government budget constraint becomes

R=PsGs+eGT(7)

where R is taxes and GS and GT are fiscal expenditure on exportables and importables, respectively. Real disposable income is then equal to

Y=XR/P(8)

For analyzing changes in relative prices, it is convenient to multiply equation (8) by P/e, on substitution from equations (2) and (5)

YP/e=qXs+XTR/e(8)

For small changes in the exchange rate, the change in real disposable income is obtained by differentiating equation (8'). Noting that qXSq + XTq = 0 at any point of tangency to the transformation curve, and using initial values of P, PS, and e equal to one, the change in real income is equal to

dY=[(XsαY)qe+R]de(8)

Since α is the share of S-goods in total domestic expenditure, it is immediately verified that Xs – αY ≥ 0 if the country is a net exporter of S-goods, but the change in disposable income is ambiguous.

The private sector is assumed to hold two kinds of asset, foreign (M1) and domestic (M2) currency, considered as perfect substitutes under fixed exchange rates and in the absence of expected exchange rate changes. 10 The real value of private wealth (A) is therefore 11

A=(M1e+M2)/P(9)

The excess demand for importables, QT, is equal to actual imports. This is obtained from the assumption of price taking for T-goods, and can be written as

QT=ET(q,Y,A)+GTXT(q)(10)

where ET and GT are domestic private and government demand for T-goods, respectively, and XT is domestic production of T-goods. In other words, imports are a residual in a competitive market. By contrast, the demand for domestic exports, Ws, is a function of the terms of trade

Ws=Ws(q)(11)

Excess demand for exportables, QS, is then

Qs=Ws(q)+Es(q,Y,A)+GsXs(q)=0(12)

where ES and GS are domestic private and government demand for S-goods, respectively, and XS is domestic production of S-goods. The current account surplus can now be derived in a number of ways: as the excess demand for cash balances, as the excess of income over domestic expenditure, or as the surplus of exports over imports. The second and third definitions can be verified to be identical, and both can be made equivalent to the first by a suitable choice of the period of analysis, or by using the appropriate speed of adjustment in the money market. Retaining the third definition, one obtains from equations (10) and (11)

B/e=qWsQT(13)

where B is the current account surplus in domestic currency, and qWs and QT are exports and imports, respectively, both measured in units of T-goods.

II. Exchange Rate Policy

A change in a nominal variable—say, the exchange rate—can produce lasting effects on real variables, such as the current account surplus, only if some rigidities exist in the economy. Some rigidities may be durable, accounting then for permanent effects, or may diminish over time together with any real effects of exchange rate policy.

In what follows, several rigidities of practical importance are considered. The first flows naturally from the distinction between the two types of industry described earlier. Firms in the group of price setters may fail to adjust their prices to clear the market in the short run, either because they set prices according to some notion of long-run equilibrium or because they follow cost-plus rules. Instead, a quantity adjustment may take place. Whether this is a short-run phenomenon or not depends, of course, on the structure of the particular market. A second type, related to the rigidity of some prices, is a structural feature that pertains to the composition of domestic wealth between domestic and foreign assets. It will be shown that this composition, in relation to that of domestic expenditure between exportables and importables, plays a role in the effects of an exchange rate change. A third type of rigidity derives from the government sector. Depending on whether the government determines its budget in real or nominal terms, exchange rate changes may or may not affect real variables. Finally, wage rigidity is considered in combination with the assumption of an active fiscal policy that is committed to maintaining full employment.

PRICE RIGIDITIES

Examining the effects of a change in the exchange rate, what is the result of revaluation (de < 0)? To simplify the analysis, the existence of the fiscal sector is ignored, initially. Equation (12) shows that the excess demand for exportables (QS) is a function of relative prices (q), income (Y), and wealth (A). By virtue of equation (8"), real disposable income, Y, is a function of relative prices only, if the fiscal sector is neglected. Hence, on substitution, QS = QS(q, A). Revaluation has the immediate effect of raising the real value of domestic currency assets and creating thereby excess demand for S-goods. To re-establish equilibrium in that market, the relative price of S-goods must increase so that PS cannot fall by as much as the exchange rate. As a consequence of this change in relative prices and of a shift in the real value of existing assets, domestic demand for importables increases while world demand for exportables declines as domestic production rises. 12 Making the assumption of an initially balanced current account, a deficit develops that leads to a decrease in the stock of domestic assets, thereby reversing the initial price and wealth effects of revaluation. In final equilibrium, relative prices and real wealth return to their initial values. 13

The preceding results depend heavily on the assumption of a high speed of price adjustment in the S-market. They are also at variance with the observation of J-curve effects of exchange rate policy. If one now assumes that export prices do not adjust immediately to clear markets (for example, and to mark a sharp contrast, that prices of exportables are fixed and production is adjusted to the level of demand), the simplicity and clarity of the effects of exchange rate policy are lost. Depending on whether price rigidity is only a short-run phenomenon or a structural characteristic of the industries involved, the results may be interpreted either as short-run effects, possibly accounting for the J-curve observation, or as longer-lasting effects. 14

From definition (13), and on substitution of equations (10) and (11), the effect of revaluation on the current account surplus is derived as 15

dBde=(1+ηw)WsηTET+εTXT+ωTdAde+μTdYde(14)

where, for i = T, S, ηi are demand elasticities, ϵi are supply elasticities, and μi and ωi are marginal propensities to spend out of income and wealth, respectively. The change in the value of domestically owned assets is discussed later, and from equation (8') the change in real disposable income is equal to

dYde=XsαY+εTXTdXsde(15)

where, from equation (12),

dXsde=ηsWs+ηsEsωsdAdeμsdYde(16)

In equation (14) the change in the current account surplus is seen to be the sum of three effects: first, the price effects of revaluation on exports and imports; second, the effect on real domestic wealth and hence, via the marginal propensity to spend, on imports; and third, the effect on real disposable income and thereby on demand for imports. The change in real disposable income, equation (15), itself is related to terms of trade effects (XS - αY), the fall of production in the T-sector (ϵTXT), and the change in the production level of the S-sector (dXS/de). From equation (16) the latter is seen to depend on the fall of demand for S-goods caused by an increase in their relative price plus any changes in demand resulting from different real wealth and income levels. Substitution of equation (16) into equation (15), of equation (15), and of the differential of equation (9) into equation (14) yields 16

dBde=1Δ[B(ωT(1μs)+ωsμT)(A(1α)M1)+(μsB+μTαY)+(1μsμT)((1+ηW*)WsηT*ET+εTXT)](17)

where ηi* are compensated demand elasticities, and

Δ=(1μS)(1+ωT)+ωSμT

The result is now ambiguous for essentially three reasons. First, with an initial surplus, there is a traditional valuation effect that is assumed to be negative here. 17

Second, with PS fixed, the real value of domestically held assets may increase or decrease, depending on the proportion of foreign assets in total domestically owned assets, compared with the weight of importables in the domestic price index. From equation (9), on substitution from equation (6), one obtains

dA=[M1(1α)A]de(18)

so that real domestic wealth decreases after revaluation (de < 0) if M1/A > (1 - α), that is, if the proportion of foreign assets in total assets exceeds the share of importables in domestic expenditure. Hence, with an ambiguous change in the value of real assets, the adjustment of relative prices also becomes ambiguous, as does the effect on the current account. Thus, the composition of domestic asset holdings as well as the structure of domestic expenditure is of direct relevance for exchange rate policy.

A third source of ambiguity derives from the various income effects induced by revaluation. While the positive terms of trade effect increases real income, price rigidity causes some reduction in employment and production that may more than offset the gain in the terms of trade. In this case (for precise conditions see equation (45), in the Appendix), lower domestic real income tends to reduce imports, thus counteracting the pure price effects that unambiguously reduce the surplus if the demand for exports is elastic. Pure price effects then become effective only when the marginal propensity to spend is less than unity (1 - μs - μT > 0), a result that is well known from the income approach. Collecting these various elements, sufficient conditions (albeit not necessary ones) for a decrease of the trade surplus are (i) elastic export demand combined with 1 - μS - μT > 0; (ii) a sufficiently low share of foreign assets in domestic wealth (αA - M1 > 0); and (iii) a low value for SB + μTαY), which is part of the income effect, requiring together with (ii) a low marginal propensity to consume importables T). A large initial surplus is also seen to augment the difficulty to adjust.

Expression (17) is of a general nature, consistent with various particular approaches to trade account analyses. The restrictions to be imposed on equation (17) to obtain the results for the income and asset approaches are obvious; but it may be useful to collapse equation (17) explicitly into the familiar Marshall-Lerner conditions. To this end, assume an initial trade equilibrium (B = 0); then set ωT = ωS = μT = 0. With these restrictions Δ = (1 - μS) and equation (17) becomes

dBde=(1+ηw*)WsηT*ET+εTXT(19)

Defining the (compensated) elasticity of import demand as ηI*QT=ηT*ETεTXT, where QT =ET - XT, with ηI*ηT*, and using again the condition of an initial trade equilibrium (Ws = QT) yields

dBde=(1+ηw*ηI*)Ws(19)

Thus, it clearly appears from equations (17) and (19') that the Marshall-Lerner condition is not particularly useful, since it is neither a necessary nor a sufficient condition for a revaluation to reduce (or a devaluation to increase) the trade surplus. In particular, it neglects the induced income and wealth effects of an exchange rate change and fails to explicitly incorporate the importance of the distribution of domestic expenditure. In other words, it neglects the importance of the structure of the economy for the effects of exchange rate policy.

REVALUATION AND FISCAL POLICY

The focus of the analysis is now directed to the fiscal sector. Since the considerations related to asset valuation and asset accumulation, and those built on income effects induced by passive fiscal policy, are additive, it is now assumed that either the real value of domestic assets is constant (which could be achieved by appropriately adjusting the domestic component of the money supply M2) or the value of assets is irrelevant for expenditure decisions. To cut off fiscal policy’s umbilical cord with asset accumulation, the fiscal budget is constrained to balance. It follows that, depending on whether the government fixes the level of expenditure or of taxes, the budget constraint makes either one or the other endogenous. Of some importance is whether the government determines its target in nominal or real terms. If the authorities fix the real value of expenditure, then the excess demand function for exportables, equation (12), becomes exclusively a function of relative prices and, if the initial equilibrium is globally stable, exchange rate policy will exert no effects on relative prices and hence will not produce any real effects. 18 Exchange rate policy can lead to real effects only if the budget of the public sector, and hence the tax burden, is defined in nominal terms. In this case, a revaluation increases the real taxation of the private sector so that private expenditure decreases, offsetting higher real expenditure by the government. If the entire increase in government expenditure falls on exportables, 19 then total demand for exportables increases and their relative price must rise to clear the market, that is, the price of exportables falls by less than the price of importables. 20 As a result, real disposable income falls despite improved terms of trade, thereby lowering the demand for importables. The overall effect on the current account surplus is

dBde=B+RΔ{(1μsμT)[(1+ηW*)WsηT*ET+εTXT]+μTαY}(20)

In equation (20) the change in the current account can again be related to influences opposite in sign. The valuation effect is negative if the initial situation is one of surplus, while the income effect is positive. Again, for price effects to reduce the current account surplus, a sufficient but not necessary condition is an elastic demand for exports. As before, the sum of price and income effects is ambiguous in sign.

In the present framework, monetary and fiscal policies have effects that are similar to those of exchange rate policy. An expansion of the money supply 21 unambiguously reduces the current account surplus, and, with initial equilibrium, the effects of a revaluation and of an expansion of the money supply on the current account are identical up to a linear transformation. 22 Perhaps less obvious is the fact that contractionary fiscal policy also has the same effects on the current account as a revaluation (compare equations (46) and (50), in the Appendix), apart from the valuation effect and up to a linear transformation. In particular, a reduction in government expenditure paralleled by a reduction in taxes produces results that depend on the structure of expenditure. If expenditure on importables is reduced, the current account improves, while this effect remains ambiguous when expenditure on exportables is decreased.

WAGE RIGIDITIES

One possible reason for slow price adjustment considered earlier is the rigidity of nominal wages. The effects of revaluation on the current account surplus are similar to those already obtained, although the interpretation changes slightly. Production levels now decrease even if all firms behave as short-run profit maximizers. In the following, the implications of rigid wages are sketched, under the assumption that active fiscal policy maintains full employment.

With a given capital stock in each sector, the full employment condition for labor can be written

Ls(wPs)+LT(we)=L(21)

where Ls and LT are demands for labor in each sector, expressed as negative functions of sectoral real wages. If nominal wages are rigid downward (a case empirically pertinent for revaluation), the real wage in the T-sector of the economy increases and the production of importables declines. To maintain full employment, the price of exportables must increase to lower the real wage in the S-sector and to stimulate demand for labor. Since the fiscal expenditure necessary to maintain full employment is financed through higher taxes, disposable income declines. With this effect, which compounds the one where nominal taxes are held constant, the likelihood of a perverse effect of revaluation on the current account is clearly enhanced. Alternatively, if fiscal expenditure were not raised to the level where PS is consistent with full employment, excess supply would arise in the goods markets. In these circumstances, PS would have to fall so as to clear the goods market, in the process creating some unemployment, which would affect disposable income in much the same way as increased taxation, and hence produce similar effects on the trade surplus. Thus, relative prices are now determined by equation (21) and no longer by the condition for market clearing in the S-sector. To make demand for S-goods consistent with production plans at full employment, government must soak up excess supply; that is, Gs becomes endogenous.

When capital is mobile across sectors and nominal wages are rigid, the price of exportables can be derived from the competitive profit conditions

aLTw¯+aKTr=PTaLSw¯+aKSr=PS(22)

where aij are the input/output coefficients, w¯ is the fixed nominal wage rate, and r is the rental on capital. 23 Logarithmic differentiation of equation (22) yields

θKTr^=P^TθKSr^=P^S(23)

where θKjS represent capital’s distributive share in each sector, and the circumflex denotes a percentage change. By assumption, the decrease in PT is equal to the rate of revaluation. Since θKT, θKS < 1, the return of capital must decrease by more than either PT or Ps. The decrease of PS is equal to (θKS/θKT)P^T and depends therefore on the capital intensities in both sectors. If exportables are more capital intensive than importables, then θKSKT > 1 and the price of exportables will decrease by more than the rate of revaluation. Since the major countries with a current account surplus export capital-intensive goods, this case is retained here. The relationship between changes in factor rewards and output is then

r^<P^S<P^T<w^(24)

implying dq < 0. Following revaluation, resources are shifted into industries producing importables while demand is displaced toward exportables. The volume of imports is reduced as domestic demand for importables falls while domestic production rises.

Excess demand for exportables must now be matched by a reduction in government expenditure on exportables. Thus, with capital-intensive exports, nominal wage rigidity, and fiscal policy committed to maintain full employment, a revaluation leads to results opposite to those obtained when full employment is maintained in the presence of wage flexibility: the terms of trade fall, price effects stimulate exports and diminish imports, but the income effect on imports tends to be positive. Accordingly, perverse responses are likely possibilities in at least one of the two cases, corresponding, respectively, to nominal wage flexibility and nominal wage rigidity. 24

By contrast, if real wages are rigid and capital is mobile across sectors, the terms of trade cannot change, as can be verified from equation (22), and revaluation has no effect on the trade surplus.

III. Summary and Conclusions

In view of the sometimes low effectiveness of real exchange rate adjustment on trade imbalances, disappointing both policy-makers and trade theorists, a framework has been proposed that elucidates why trade account reactions to exchange rate changes may be insignificant or even perverse. The model incorporates the spirit of the elasticity, income, and asset approaches in a general equilibrium framework. It is shown that over the short term and medium term a revaluation may or may not affect the trade surplus depending on the structural characteristics of the economy, and, if it does, may increase or lower it. The most important structural characteristics are the size and behavior of the nonprice-taking sector, the composition of domestic demand, the existence of wage and price rigidities, the composition of domestic wealth, and the objectives of demand policy.

It seems that exchange rate policy yields unambiguously conventional results only where the price mechanism results in rapid market clearing at unchanged unemployment, with government expenditure fixed in real terms. Relaxation of either one of those two conditions makes exchange rate policy dependent on the structural features mentioned earlier.

While the elasticities of export and import demand are important, the Marshall-Lerner conditions are neither necessary nor sufficient for a successful exchange rate policy. What the elasticity approach inappropriately assumes away is the induced income and wealth effects of exchange rate changes. The results in this paper can also be looked upon from the viewpoint of the absorption approach: the change in the trade surplus is equal to the change in net private savings when the government account is kept in balance. If revaluation augments the terms of trade and has a perverse effect in the intermediate run, the increase of the trade surplus is tantamount to an increase in net private savings—a result akin to the Laursen-Metzler effect. 25

The analysis in this paper carries some implications for policy. In economies for which the terms of trade are not exogenously given, the effects of a real exchange rate change on the trade account cannot be assessed easily and in abstracto. As was shown in this paper, they depend on a number of structural features of the economy and on the extent and nature of other policies that are being applied. For a proper assessment of the effects on the trade balance, the information requirements concerning economic structure and the behavior of wage earners and firms seem to be quite high, possibly exceeding the degree of information at the disposal of policymakers. In practice it may therefore be rather difficult to predict in the short run and the intermediate run—be it only in sign—the effects on the trade account of a change in the exchange rate in excess of inflation differentials.

APPENDIX

article image
article image

Operators

xy(x)=yx

dx = differential of x

dxx=x^

determinant A = Δ (where A is a matrix)

The Model

From the description in the main text, the following model is obtained where initial values for P, Ps, e, and q are all set equal to unity.

B/eqWS(q)+XT(q)ET(q,Y,A)GT(25)
XqXS(q)+XT(q)(26)
YP/eXR/e(27)
R/eqGS+GT(28)
PαPS+(1α)e(29)
qPS/e(30)
APM1e+M2(31)
QS=WS(q)+ES(q,Y,A)+GSXS(q)(32)
dM1=dBBde(33)

The restrictions on the derivatives of the excess demand functions are as follows:

dQT=(ηTETεTXT)dq+μTdY+ωTdA>0

dQS=(ηW+ηTESεSXS)dq+μsdY+μSdA0

where

ηi = qEiq/Ei is the elasticity of the domestic demand in sector i

ηW = qWSq/WS is the elasticity of foreign demand for domestic exportables

ηi* = ηi + μiES is the elasticity of the compensated demand function in sector i; ηT*0,ηS*0,ηT*ET/ηS*ES=q

ωi = EiA is the marginal propensity to spend on good i, out of wealth; 0 ≤ ωS, ωT ≤ 1

μi = EiY is the marginal propensity to spend on good i, out of income; 0 < qμS, μT, qμS + μT ≤ 1

Exchange rate policy: no fiscal sector

Setting R, Gs, and GT equal to zero and differentiating yields the following linear approximation to an (assumed) initial equilibrium point:

[1ωTΦ00ωSγ011αA00011][dBdAdqdPS]=de[B0A+BM11]+dM2[0010](34)

where

Φ=(1+ηW)WS+εTXTηT*ET+μTESμT(XSαX)(35)

Sufficient conditions for Φ < 0 are as follows:

(i) 1 + ηW < 0

(ii) ηT*ET+μTES<0, the normal goods assumption

(iii) XS - αX ≥ 0

which is always true for a net exporter of S-goods.

γ=ηWWS+ηS*ES+μS(XSESαX)ɛSXS<0(36)

It is immediately verified that the system (34) is locally stable for empirically reasonable values of the parameters. The value of the determinant, denoted Δ, is

Δ=ωs(αAΦ)γ(1+ωT)>0

The following comparative static results can be derived:

dBde=BΔ(ωSαAγ)(A+BM1)Δ(ωSΦ+ωTγ)>0(37)
dAde=(AM1)γ/Δ<0(38)
dqde=(A+BM1)Δ<0(39)
dBdM2=(ωSφ+ωTγ)/Δ<0(40)
dAdM2=γ/Δ>0(41)
dqdm2=ωs/Δ>0(42)

Suppose that dPS = 0, so that dq = -de. Equation (32) is now replaced by equation (32’), which describes the adjustment of output to demand

dXs=(ηWWS+ηSES)de+ωSdA+μSdX(32)

and the change in real income is

dY=(XSαY+εTXT)de+dXS27

System (34) then becomes

[1ωTμT0110000110ωSμS1][dBdAdYdXS]=de[ζ1ζ2ζ3ζ4](34)

where

Δ=(1μS)(1+ωT)+ωSμT>0

ζ1=(1+ηW)WS+εTXTηTETB0

ζ2=(1α)AM10

ζ3=XSαY+εTXT0

ζ4=ηSWS+ηSES<0

dBde=(1μS)BΔ(1μSμT)Δ[(1+ηW*)WSηT*+εTXT]μTαYΔ+((1α)AM1)Δ(ωT(1μS)+ωSμT)(43)
dade=(1μS)Δ(ζ2+ζ1)μΔ(ζ3+ωSζ4)(44)
dYde=(1ωT)Δ(ζ3+ζ4)ωSΔ(ζ1+ζ2)(45)

Exchange rate policy: no wealth considerations

Setting dA = 0 but retaining equations (27) and (28) yields a system where the effects of an exchange rate policy depend on what is fixed by fiscal policy. Three cases might be considered: (i) dR/de = 0; (ii) dGS = 0 with dGT = 0; and (iii) dR = 0. In cases (i) and (ii), Qs in equation (32) becomes a function of q alone. Hence, q does not change and a revaluation has no effect on the current account. In case (iii), we have

[1μTΦ00μsγ001(XsαY)00011][dBdYdqdPS]=de[BRR1](46)

where

Φ=(1+ηW)WS+εTXTηT*ET+μTES0

γ=ηWWS+ηS*ESεSXSGSμSES<0

Δ=μS(XSαY)γ>0 (assumed)

Again it can be verified that system (46) satisfies necessary and sufficient conditions for local stability for “reasonable” values of the parameters. One then obtains

dBde=BRΔ{(1μSμT)[(1+ηW)WSηT*ET+εTXT]+μTαY}0(47)
dYde=RΔ[(1+ηW)WS+(1μS+ηS*)ESεSXSαY]>0(48)
dqde=RΔ(1μS)<0(49)

Fiscal policy

With the exchange rate fixed, dq = dPS and the last row and column of equation (46) can be deleted. Imposing a balanced budget

dR=GSdq+dGS+dGT

leads to the following results where the determinant Δ is minus one times the determinant of equation (46) and where – (XS - αY) in (46) becomes – (XS - GS- αY).

dBdGS=(1μSμT)Δ[(1+ηW)WSηT*ET+εTXT]μTαYΔ0(50)
dYdGS=(q+XSαYGS)Δ<0(51)
dqdGS=(1μS)Δ>0(52)
dBdGT=(1μSμT)Δ(ηWWSηT*ET+ɛTXT)+μSαYΔ<0(53)
dYdGT=γ/Δ<0(54)
dqdGT=μS/Δ<0(55)

BIBLIOGRAPHY

  • Aukrust, Odd, PRIM 1: A Model of the Price and Income Distribution Mechanism of an Open Economy,” Review of Income and Wealth, Vol. 16 (March 1970), pp. 5178.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bruno, Michael, The Two-Sector Open Economy and the Real Exchange Rate,” American Economic Review, Vol. 66 (September 1976), pp. 56677.

    • Search Google Scholar
    • Export Citation
  • Cassing, James, International Trade in the Presence of Pure Monopoly in the Non-Traded Goods Sector,” Economic Journal, Vol. 87 (September 1977), pp. 52332.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Deppler, Michael C., and Duncan M. Ripley,The World Trade Model: Merchandise Trade,” Staff Papers, Vol. 25 (March 1978), pp. 147206.

  • Dornbusch, Rudiger (1974), “Real and Monetary Aspects of the Effects of Exchange Rate Changes,” in National Monetary Policies and the International Financial System, ed. by Robert Z. Aliber (University of Chicago Press, 1974), pp. 6481.

    • Search Google Scholar
    • Export Citation
  • Dornbusch, Rudiger (1975), “Exchange Rates and Fiscal Policy in a Popular Model of International Trade,” American Economic Review, Vol. 65 (December 1975), pp. 85971.

    • Search Google Scholar
    • Export Citation
  • Hahn, Frank H., The Monetary Approach to the Balance of Payments,” Journal of International Economics, Vol. 7 (August 1977), pp. 23149.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Jones, Ronald W., and W. Max Corden,Devaluation, Non-Flexible Prices, and the Trade Balance for a Small Country,” Canadian Journal of Economics, Vol. 9 (February 1976), pp. 15061.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kravis, Irving B., and Robert E. Lipsey,Price Behavior in the Light of Balance of Payments Theories,” Journal of International Economics, Vol. 8 (May 1978), pp. 193246.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Laursen, Svend, and Lloyd Appleton Metzler, Flexible Exchange Rates and the Theory of Employment,” Review of Economics and Statistics, Vol. 32 (November 1950), pp. 28199.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Malinvaud, Edmond, The Theory of Unemployment Reconsidered (Oxford, 1976).

  • Miles, Marc A., The Effects of Devaluation on the Trade Balance and the Balance of Payments: Some New Results,” Journal of Political Economy, Vol. 87 (June 1979), pp. 60020.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Morel, Colette, and Alfred Steinherr,An Empirical Study of World Market Influences on Price Formation in West Germany,” Empirical Economics, Vol. 3 (No. 3, 1978), pp. 183201.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Nordhaus, William D., and Wynne A. H. Godley, Pricing in the Trade Cycle,” Economic Journal, Vol. 82 (September 1972), pp. 85382.

  • Steinherr, Alfred, and Colette Morel (1979 a), “The Reaction of Prices and of the Balance of Payments to Revaluation of the Deutsche Mark,” Weltwirt-schaftliches Archiv, Vol. 115 (No. 3, 1979), pp. 42549.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Steinherr, Alfred, and Colette Morel (1979 b), “Une analyse des effets de réévaluation en Allemagne Fédérate,” Recherches Economiques de Louvain, Vol. 45 (March 1979), pp. 95108.

    • Search Google Scholar
    • Export Citation
  • Vernon, Raymond, International Investment and International Trade in the Product Cycle,” Quarterly Journal of Economics, Vol. 80 (May 1966), pp. 190207.

    • Crossref
    • Search Google Scholar
    • Export Citation
*

Mr. Steinherr, Consultant in the External Adjustment Division of the Fund’s Research Department when this paper was prepared, is Associate Professor of Economics at the Catholic University of Louvain, in Belgium. He holds degrees from the University of Lausanne, the University of Florence, McMaster University, and Cornell University.

The author is grateful to a number of colleagues in the Fund for their contributions, but, as customary, they should not be held responsible for misconceptions or imperfections.

1

Important reversals of current account imbalances occurred, however, in 1974-75 (except for the Federal Republic of Germany) and again at the end of the 1970s.

2

This casual observation is supported by a recent study on the effects of devaluation that concludes that although devaluation usually improves the overall balance of payments, in most cases it does not improve the trade balance (Miles (1979)).

3

The assumption that most countries face horizontal supply curves for their imports seems to fit empirical observation. The assumption is also useful to restrict the explanation of possible difficulties with adjusting the current account to domestic factors and to eliminate such arguments as inelastic import supply.

4

Dornbusch (1975) has reinterpreted the elasticity approach within a general equilibrium model that makes the distinction between tradables and non-tradables.

5

Cassing (1977) appropriately analyzes the case of pure monopoly in the nontraded goods sector; in the “Scandinavian model” (Aukrust (1970)) cost-plus pricing is assumed to prevail in that sector.

6

The view that trade among developed economies is not predominantly characterized by industries behaving as price takers can be supported by both empirical evidence and theory. For example, the product-cycle theory as well as the Linder theory is built on the observation that specialization in production leads to imperfect substitutability on international markets. The principal exponent of the product-cycle theory is Vernon (1966). For evidence, see, for example, Kravis and Lipsey (1978).

7

A good example is provided by Mercedes-Benz automobiles. For at least a decade, customers have accepted waiting periods of up to three years. The company neither stepped up production nor raised the price enough to eliminate excess demand. Revaluations of the deutsche mark, however, have been regularly passed on to foreign currency prices.

8

In reality, many export industries are, of course, price takers on world markets. This fact is reflected by the elasticity of aggregate export demand. Furthermore, industries producing nontradable goods are included in the aggregation of exportable industries, since, by definition, they face a downward-sloping demand curve. The Hicksian conditions for industry aggregation are assumed to be satisfied, so that the analysis can be confined to a two-sector case.

9

A complete list of the variables used can be found in the Appendix.

10

An alternative interpretation would allow the private sector to hold only domestic currency. In this case, part of the domestic money supply would be backed by the monetary authority’s holdings of foreign currency that the private sector reckons to own indirectly. All money in this model is of the outside kind.

11

The value of private wealth depends, of course, also on the evaluation of future earnings of capital and labor. Although of considerable importance for the effects of exchange rate changes, these issues raise subtle questions better analyzed in a different framework and are therefore not addressed in this paper.

12

This interpretation corresponds to results (37) to (39), in the Appendix.

13

Uniqueness of equilibrium is assumed here as elsewhere in the literature. As Hahn (1977) points out, there is no strong justification for such an assumption. Also, if one starts out from an initial current account surplus, the interpretation has to be changed as follows: “In final equilibrium relative prices and real wealth are independent of the exchange rate.”

14

Other interpretations are, of course, consistent with this framework. For example, a permanent increase in the relative price could also be achieved if monetary policy were used to fix the general price level. From equation (6), a revaluation then leads to an increase in the relative price level.

15

With the exchange rate defined in units of domestic currency, dB/de represents the effect of a devaluation; for a revaluation, one has -dB/de = dB’/de.

16

See equation (43), in the Appendix.

17

If exports and imports are priced in domestic currency, the valuation effect is zero in terms of domestic currency and amounts to a gain equal to the surplus times the rate of revaluation in terms of foreign currency. If all trade is billed in foreign currency, the valuation loss in domestic currency is equal to the trade surplus times the rate of revaluation. Most industrialized countries tend to trade in domestic currency a higher share of exports than of imports. A revaluation then leads to smaller valuation losses in domestic currency than the one indicated in the derivation of equation (17).

18

Here it is again assumed that the speed of adjustment is infinite.

19

This assumption is, of course, important for the results obtained; with any other assumption the probability of a perverse effect is reduced.

20

It is postulated here that an increase in government expenditure does not enter into the utility functions of private agents. Relaxation of this assumption leads, in the present case, to a direct effect of substituting public for private expenditure to satisfy individual utility and to an indirect effect via a lower relative price of exportables. Both effects point to lower imports and increased exports, thus strengthening the likelihood of a perverse effect.

21

Since there is only outside money in the model, the increase in the stock of money would have to be matched by a fiscal deficit whose effect is neglected in this argument.

22

Compare equations (37) and (40), in the Appendix.

23

Factor-proportion theory for an analysis of devaluation is used by Jones and Corden (1976).

24

Notable among the shortcomings of the present analysis is that investment has been left exogenous, as no convincing theory of investment is available for an essentially static framework. To mask this void it may be worthwhile to sketch some relevant arguments. Unemployment is certain to discourage investment. Alternatively, with full employment but higher real wages (e.g., when nominal wages are rigid), the reduction of the return on capital is likely to stimulate capital deepening but discourage capital widening with an ambiguous net result. Hence, it is difficult to predict the variation in investment, but the possibility of a reduction in investment, reinforcing the decrease in net absorption, cannot be dismissed.