Quantitative Controls and Unofficial Markets in Foreign Exchange: A Theoretical Framework

The purpose of this paper is to examine the effects of monetary, fiscal, and exchange rate policy in a dual exchange rate regime that is characterized by an official market and by a parallel market that has arisen in response to the imposition of exchange controls. Results from the study suggest that the standard model of an open economy provides a misleading guide for policy under such a dual regime.

Abstract

The purpose of this paper is to examine the effects of monetary, fiscal, and exchange rate policy in a dual exchange rate regime that is characterized by an official market and by a parallel market that has arisen in response to the imposition of exchange controls. Results from the study suggest that the standard model of an open economy provides a misleading guide for policy under such a dual regime.

The purpose of this paper is to examine the effects of monetary, fiscal, and exchange rate policy in a dual exchange rate regime that is characterized by an official market and by a parallel market that has arisen in response to the imposition of exchange controls. Results from the study suggest that the standard model of an open economy provides a misleading guide for policy under such a dual regime.

Multiple exchange rate regimes have been a common feature of the international monetary system since the 1950s. Although the characteristics of these regimes have varied widely among countries and at different times, attention in the literature has usually been focused on the dual arrangements adopted by certain industrial countries. Conceived as a means of providing insulation from large and potentially volatile capital flows, this type of dual arrangement is characterized by an official market, in which current transactions are conducted at a fixed or managed exchange rate, and a financial market for capital transactions, in which the exchange rate is freely determined.1 Although exchange controls are necessary to segregate the two markets, quantitative restrictions on current payments in the official market are usually limited, if not altogether absent. Consequently, the official market in this type of regime clears through some combination of reserve intervention and exchange rate adjustment.

This paper focuses on a second, broad type of multiple exchange arrangement that has attracted rather less interest despite its prevalence, particularly among developing economies. It is distinguished from the first type by the presence of quantitative exchange and trade controls in the official market that are designed to protect depleted reserve positions and to avoid unpalatable official exchange rate adjustments in the face of balance of payments pressures. Under such a regime there exists no mechanism, such as reserve movements, to ensure that all prevailing demand for foreign exchange at the official exchange rate is met and that the official market clears. If the costs of engaging in illegal transactions are not prohibitive, the excess demand for foreign exchange in the official market is satisfied at a premium price in a secondary, or parallel, market. The domestic cost of imports will, therefore, reflect the parallel foreign exchange rate.

Several writers have recognized that the standard model of a restriction-free economy may not be applicable to an economy in which there is rationing in the official foreign exchange market.2 Little progress has been made, however, in incorporating quantitative restrictions into a general equilibrium framework for an open economy.3 In developing such a framework, this paper presents some results that differ significantly from those associated with the standard model. It is suggested, therefore, that reliance on the standard model may give rise to erroneous policy prescriptions.4

The analysis demonstrates that, under a reasonably plausible set of assumptions, an official devaluation of the domestic currency will have no inflationary effect; indeed, under certain conditions the price level may actually fall. Furthermore, for a given level of reserve intervention by the central bank, it is shown that a devaluation of the exchange rate will lead to an appreciation of the domestic currency in the parallel market, and that changes in domestic credit and in the official exchange rate may have permanent real effects.

The paper is structured as follows. Section I presents a partial equilibrium, pure-flow model of the foreign exchange market that incorporates smuggling. Section II extends the analysis to a general equilibrium framework that features currency substitution and rationally formed exchange rate expectations. Section III examines the implications of the model for exchange rate policy; and Section IV summarizes the paper’s conclusions.

I. The Foreign Exchange Market in a Partial Equilibrium Setting

The economy in question is small in the sense that it cannot influence its external terms of trade (in foreign currency terms). It comprises a private sector and an official sector. The private sector produces a nontraded good and an exported good with a variable factor, labor, which is geographically mobile and flexible in price. It also consumes two goods, the nontraded good and an imported good. All three goods are nonstorable.

The official sector comprises a government and a central bank. The central bank operates an official foreign exchange market in which all transactions are conducted at a fixed price. Both the government and the private sector participate in this market. Producers of the export good are legally obliged to sell their output to the government, which in turn surrenders the foreign exchange proceeds to the central bank. The central bank also sells foreign exchange for the purchase of the imported good. Exchange control regulations are in force, however, that restrict the availability of foreign exchange for the purchase of the imported good and prohibit entirely use of foreign exchange for capital transactions. The excess demand thus created is satisfied in a secondary or parallel market. The source of foreign exchange to this market consists of proceeds from smuggling the export good. The illegality of such transactions involves a cost that is partly reflected in a premium of the parallel market price over the official price.

It is assumed that all imports purchased in the official market are resold. Hence, although the imported good may be purchased at two different exchange rates, the price at which it is consumed is uniform, being determined by its (marginal) cost on the parallel market. It is further assumed that there are no risks or penalty costs associated with the purchase of foreign currency on the parallel market or with the resale of the imported good.

prohibitive smuggling costs

The consequence of quantitative import controls when prohibitive costs rule out smuggling are illustrated in Figure 1. Demand for the imported good, II0, and supply of the export good to the official market, XX0¯, are depicted as functions of the unofficial exchange rate, e (defined as the price of foreign currency in domestic currency). At the official exchange rate, e¯1 the prevailing demand for imports will be fully satisfied if official reserves, equal to AB, are sold each period.5 If the quantity of imports is restricted to OA through, say, a licensing system, then no reserve sales are necessary at e¯1. Consequently, at this rate there is an excess demand for imports, and the foreign exchange market does not clear. But imports purchased at e¯1 will be resold at the premium price of e1, the profits from this activity being given by the product of (e1 − ē1) and OA. Thus, although no foreign exchange is transacted outside the official market, e1 may be interpreted as a “shadow” parallel market exchange rate. A discriminatory arrangement that charged importers e1 but offered exporters e¯1 would have equivalent price and quantity effects, but the exchange profits would accrue to the central bank rather than to the private sector.6

Figure 1.
Figure 1.

Determination of Parallel Market Exchange Rate: Without Smuggling

Citation: IMF Staff Papers 1984, 002; 10.5089/9781451946918.024.A005

It is also clear from Figure 1 that, for a given level of official reserves, successive devaluations of the official rate will produce appreciations of the shadow parallel rate and, therefore, will cause a decline in the price of imports until the foreign exchange market clears at e2. This deflationary effect simply reflects an expansion in the supply of imports arising from the increased availability of foreign exchange.

nonprohibitive smuggling costs

We now turn to the more interesting case, in which the costs of smuggling are not prohibitive.7 The left-hand quadrant of Figure 2 depicts, as a function of the parallel market exchange rate e, the foreign currency value of sales of the export good to the official market, X, whereas the right-hand quadrant traces sales of the export good to the parallel market, also as a function of e. The schedules are drawn for given values of all other prices, including the official exchange rate e.

Figure 2.
Figure 2.

Determination of Parallel Market Exchange Rate: With Smuggling

Citation: IMF Staff Papers 1984, 002; 10.5089/9781451946918.024.A005

Schedule OA is drawn to show supply of the export good to the parallel market when the official exchange rate is set, for illustrative purposes, equal to zero. Its slope shows the expansion in production of the export good that occurs as the price of the export good rises relative to that of the nontraded good.8 As the official rate is devalued, it becomes profitable for those producers whose costs of smuggling are greatest to switch supply to the official market.9 Thus, if the official rate is, say, e¯1 and there is no parallel market, OB units will be sold on the official market. At e2, the parallel market premium (e21) is sufficiently great to divert all sales from the official to the parallel market. The official and parallel market supply curves corresponding to e¯1 are, therefore, BCe2 and e¯1 DA, respectively. Curve e¯1 D is flatter than OD because, as the parallel rate rises, sales of the export good are diverted from the official to the parallel market. An increase in the official rate, from e¯1 to e¯2, produces an outward shift of the official market supply curve from BCe2 to EFG.

The sale of official foreign exchange reserves is depicted by an outward shift in the official market supply schedule. Thus, the schedule HJKL traces the total supply of foreign exchange in the official market when the price in this market is e¯1 and reserves are being sold at a rate of BH per period.

To see how the parallel market rate is determined, consider an initial situation in which all of the export good is sold on the parallel market because the official rate is “too low” and in which there are no reserve sales by the central bank. The equilibrium parallel rate will be e0 when, for purely illustrative purposes, the official rate is set equal to zero. This rate corresponds to the intersection of the supply schedule OA and the demand schedule II0. If the official rate e is raised to e¯1, then the equilibrium parallel rate will be e1 because, at this price, total demand (e1M) equals supply from the official market (e1 N = PM, by construction) plus supply from the parallel market (e1 P). Successive increases in the official rate will lead to a decline in the parallel rate along a “unification path,” QPe¯2, until the two rates merge at e¯2 and the parallel market ceases to function. Note that at e¯2 the quantity of foreign exchange demanded (e¯2 R) is equal to the quantity supplied in the official market (Fe¯2). Both e1 (the parallel rate consistent with an official rate of e¯1) and e¯2 are equilibrium exchange rates in the sense that, at these prices, the demand for foreign exchange is equal to supply. Only e¯2, however, is consistent with the absence of binding restrictions on the availability of foreign exchange in the official market. In this regard, e¯2 may be referred to as a “restriction-free” equilibrium.

The inverse relationship between the official and the parallel market exchange rates conforms with results obtained by others (see Michaely (1954) and Sheikh (1976)). The reason for the relationship is straightforward. By diverting resources from production of the nontraded good to the export good, a devaluation of the official rate will, other things being equal, generate an expansion in the total supply of the export good even though sales to the parallel market contract. Because the parallel rate equilibrates the total demand and supply of foreign exchange in the economy, the parallel rate must appreciate not only relative to the official rate but also in absolute terms.10

By similar reasoning, an increase in official reserve sales will also lead to an appreciation of the parallel rate. This effect is also illustrated in Figure 2, where, at the official rate e¯1, reserves of BH are sold each period. Since, by construction, demand at this price (e¯1S) equals the amount supplied in the official market (e¯1J), the two markets are effectively unified. However, e¯1 can only be regarded as a temporary restriction-free equilibrium because the capacity of the central bank to sell foreign currency reserves over time is limited. For reserve sales of less than BH at e¯1, the parallel market rate will lie on the unification path e¯1P. It is entirely possible, therefore, that the parallel rate is more appreciated than the restriction-free equilibrium rate, e¯2. The implication here is that use of the parallel rate as an indicator of the restriction-free equilibrium exchange rate must take into account movements in net official reserves or, equivalently, the size of the officially recorded overall balance of payments position (as well as, of course, the incidence of quantitative exchange controls).

II. The Foreign Exchange Market in a General Equilibrium Framework

This section extends the framework of analysis from a partial to a general equilibrium framework, thereby permitting an examination of the macroeconomic effects of major policy shifts when there is rationing in the official exchange market. The analysis is confined to a descriptive exposition of the workings of the model; a mathematical treatment is provided in the Appendix.

structure of the economy

Real Sector

Domestic production consists of the output of the nontraded good and the exported good. By the assumption of a small open economy, demand for the export good is perfectly elastic at an exogenous foreign currency price, so that its output is supply determined. Furthermore, it is assumed initially that the only factor of production, labor, is fixed in supply. Aggregate output is therefore exogenous and equal in equilibrium to the sum of the demand for the nontraded good by the private sector and by the government, N( ) and G, respectively, and the supply of the export good to the official market and to the parallel market, X( ) and X( ), respectively:11

Y=N(e/pn,Y)+X¯(e¯/e,e¯/pn)e¯+X(e/e¯,e/pn)+G,N1,N2,X¯1,X¯2,X1,X2>0(1)

where e and e denote the official and parallel market exchange rates,12 respectively, and pn is the price of the nontraded good.

With reference to Figure 3, the IS curves show combinations of e and pn that satisfy the condition that the demand for the economy’s output be equal to the supply; they are upward sloping because an increase in pn produces an excess supply of domestic output that must be accompanied by a rise (depreciation) in e for equilibrium to be maintained. Point Z on IS0 represents the stage reached when the parallel market premium, e/e, has increased to a level at which all production of the export good is sold on the parallel market.13 IS has an elasticity greater than 1 to the southwest of point Z and an elasticity equal to 1 to the northeast of Z.14

Figure 3.
Figure 3.

Effects of an Increase in Domestic Credit and of a Devaluation

Citation: IMF Staff Papers 1984, 002; 10.5089/9781451946918.024.A005

Financial Sector

The private sector holds two financial assets, domestic currency and foreign currency. Neither is interest bearing. The demand for domestic currency in real terms is related positively to the level of output, negatively to the expected rate of return on holding foreign currency, but is independent of wealth. It is assumed that the sale of foreign exchange for capital transactions in the official market is prohibited; the return on foreign currency therefore corresponds to the rate of depreciation of the domestic money in the parallel market. The supply of domestic money is defined as the sum of domestic credit and the net foreign exchange reserves of the official sector valued in the local currency.

Because of the presence of quantitative controls on both imports and capital payments, the stock of reserves valued in foreign currency may be treated as exogenous or, equivalently, as a target variable. This is a relatively strong assumption, since it implies that any parametric shift in the supply function for foreign exchange in the official market will be accompanied by an intensification or relaxation of controls so as to maintain the reserve or balance of payments target.

These considerations imply the following equilibrium condition for domestic money:

R+D=L(Y,e+1*/e)p,L1>0;L2<0(2)

where R is the level of net official reserves valued in local currency, D is domestic credit, and e+1* is the expected value of e one period hence.

The aggregate price level, p, is defined as a geometrically weighted average price of the nontraded good and the imported good:

p=pnαe(1α),0<α<1.(3)

LM0 in Figure 3 depicts the locus of points along which the equilibrium condition for domestic money is satisfied. It is downward sloping because, for equilibrium, an excess demand for money caused by an increase in pn must be matched by a fall in e. While LM is stationary, the price level remains unchanged. Rightward shifts in the curve arising, for example, from an increase in the money supply imply an increase in the price level.

Assuming for the moment static exchange rate expectations,15 equations (1) through (3) can be used to solve for the endogenous variables e, pn, and p. The remainder of this section examines the effects of various policy instruments on these variables.

impact of policy shocks

Change in Domestic Credit

Consider first the effect of an increase in domestic credit to the private sector.16 With the net stock of foreign reserves given (and stationary), this produces an expansion of the money supply and a rightward shift of the LM curve in Figure 3 from LM0 to LM1. As a result, the parallel rate depreciates (that is, e rises), prices of all goods increase, and the economy moves from an initial equilibrium at point A to point B.17 Because the IS curve has an elasticity greater than 1 over this range, the depreciation of the parallel rate is proportionately greater than the increase in the price of the nontraded good. This shift in the structure of relative prices demonstrates the non-neutrality of money with respect to real variables when the availability of foreign exchange through official channels is restricted. The shift simply reflects the fact that e, the price received for selling the export good in the official market, is fixed while all other prices and costs are free to vary.

This result stands in sharp contrast to those derived from the standard models of fixed-flexible exchange rates (or their dual market variant). In the fixed rate model, an expansion in domestic credit is fully offset through the balance of payments, whereas in the flexible rate model it produces a proportionate and uniform rise in all prices, leaving their relative levels unchanged.

Money-Financed Government Spending

The consequences of a permanently higher level of money-financed government expenditure on the nontraded good are shown in Figure 4. In the current period there is a leftward shift of the IS curve and a rightward shift of the LM curve, such that the economy moves from point A to point B. Thus, although the overall effect is inflationary in the sense that the aggregate price level rises, the direction of change in the parallel rate is ambiguous. In subsequent periods, however, continued deficit financing generates successive outward shifts of the LM curve along IS1, raising all prices including the parallel rate.18 Eventually, the point Z is reached, at the intersection of LM2 and IS1 when supply of the export good to the official market is exhausted. Beyond this point the economy will be in a steady-state equilibrium with D, e, pn, and p all increasing at the same constant rate.

Figure 4.
Figure 4.

Effects of an Increase in Money-Financed Government Spending

Citation: IMF Staff Papers 1984, 002; 10.5089/9781451946918.024.A005

Devaluation Under Static Expectations and Fixed Output

In Section I it was demonstrated that, in a partial equilibrium setting, a devaluation of the official rate leads to an appreciation of the parallel rate when reserves remain constant. This result continues to hold within a general equilibrium framework. With reference again to Figure 3, a devaluation causes a leftward shift of the IS curve from IS0 to IS1. Equilibrium in the economy, therefore, shifts from point A to point C, where the parallel rate is more appreciated (that is, e is lower) and the price of the nontraded good is higher. Because of the assumptions of static exchange rate expectations and fixed aggregate output, however, the LM curve remains stationary, and the price level is therefore unchanged. The absence of any effect on the aggregate price level may appear a somewhat counterintuitive result, but it is clear from the condition for domestic money equilibrium in equation (2) that, with Y and e+1*/e fixed by assumption, the price level is determined solely by domestic credit and the stock of foreign exchange reserves, both of which are exogenously given. Only if the central bank decides, contemporaneously with the devaluation, to accumulate foreign exchange reserves on a net basis will the price level rise.

This example highlights a fundamental assumption of the model: the exogeneity of reserves. This condition implies that any shift in the supply of foreign exchange to the official market as a result, for example, of a devaluation must be accompanied by an offsetting shift in the incidence of controls in order to maintain an unchanged net reserve position. In practice, since a devaluation of the official rate is often supported by official sector borrowing, there will be a tendency for net reserves to decline. Thus, a devaluation may be associated with a deflationary, rather than inflationary, effect on prices.

Devaluation Under Rational Expectations and Flexible Output

A further implication of the model is that, when either the assumption of static expectations or that of fixed aggregate output is relaxed, a devaluation may have deflationary consequences. Consider first the role played by exchange rate expectations. If these are formed rationally, an unanticipated devaluation of the official rate in the current period will not affect the expected rate of change of the parallel rate, e+1*/e; both the numerator and the denominator of this term will decline proportionally. Consequently, as in the case of a devaluation under the assumption of static expectations, the price level will again be unaffected.

In contrast, an anticipated devaluation will bring about a temporary decline in the price level as follows.19 Under the assumption of rational expectations, the time paths for the parallel market rate and for the domestic price level associated with a preannounced devaluation of the official rate (see the Appendix for their mathematical derivation) will be as shown in Figure 5. At time t, the central bank announces that it will devalue the exchange rate at time t + θ from ē0 to e¯1. Because all market participants are presumed to know at time t that this devaluation will involve an appreciation of the parallel rate from e0 to e1, the actual return on holding foreign currency, e+1*/e, declines immediately. The resulting contraction in the demand for foreign currency holdings is matched by a rise in the demand for domestic currency, and, for equilibrium to be maintained, there must be a drop in both the parallel rate (that is, it appreciates) and the price level at time t. Thereafter, the parallel exchange rate continues to appreciate at an increasing rate (that is, e+1*/e declines further) until the devaluation actually takes place at time t + θ.20 At this point no further change in the parallel rate is anticipated, and the rate of return on holding foreign currency rises abruptly. The resulting substitution from domestic back into foreign currency thereby causes the price level to jump back to its original level, p0, at time t + θ.

Figure 5.
Figure 5.

Effects of a Preannounced Devaluation1

Citation: IMF Staff Papers 1984, 002; 10.5089/9781451946918.024.A005

1 An announcement is made at time t that the official exchange rate e0 will be devalued to e¯1 at time t + θ. The price level falls temporarily after the announcement but jumps back to its original level once the exchange rate is adjusted. The assumption in the text that p0 and e0 are set equal to 1 is relaxed here for illustrative purposes. In addition, the graph is drawn in continuous time.

Although the price level is therefore independent of the official exchange rate in the long run when output is fixed, an observer measuring the movement in prices at time t + θ may erroneously conclude that the effect of a devaluation is inflationary. Moreover, the devaluation may incorrectly appear to have only a minimal effect on the parallel rate. This example thus illustrates the importance of assumptions about expectations in any empirical analysis of price and exchange rate movements.

Once the assumption of exogenous aggregate output is relaxed (that is, once labor is no longer in fixed supply), the domestic price level ceases to be independent of changes in the official exchange rate in the long run. An unanticipated devaluation leads to an increase in total output that, by raising the demand for domestic money, exerts downward pressure on the domestic price level. The deflationary effect of the devaluation is reinforced by a fall in the price of the imported good as a result of the appreciation of the parallel market rate. A corollary of this result is that the imposition of controls leads to a permanent, once-and-for-all reduction in output and to an increase in the price level. Again, these predictions stand in sharp contrast to those obtained from the standard unified or dual market models, which imply that a devaluation will raise the money supply and all prices proportionately, thereby leaving relative prices and all real variables unchanged.

exchange rate unification

It should be clear from the analysis above that, once a parallel market emerges, no automatic forces operate to bring about a reunification of the exchange rate system. Indeed, in the case of an increase in money-financed government expenditures or in an inflationary environment, the official exchange rate will become increasingly overvalued. In turn, the parallel market premium will increase over time, and the volume of transactions conducted in the official market will contract. A merging of the two markets can only be achieved through a reduction in domestic credit, a devaluation, the sale of official foreign exchange reserves,21 or a relaxation of controls.

III. Some Implications for Exchange Rate Policy

An important implication of the model is that the imposition of quantitative controls as a substitute for a formal devaluation does not avoid the adverse repercussions that a devaluation has on the rate of inflation or real wages. The emergence of a parallel market in response to such controls and the depreciation of the exchange rate in this market have inflationary consequences similar to those of an official devaluation. Furthermore, the illegality of transacting in the parallel market gives rise to real resource costs that are absent in a unified exchange system. Such a dual market also provides an environment conducive to corrupt practices and permits economic rent to accrue to certain groups that are in a position to exploit the exchange rate differential between the two markets.

Under a fixed rate regime, the effect of expansionary policies on prices will usually be mitigated by a loss of foreign exchange reserves. This constraint on inflationary pressures is absent if quantitative controls are imposed as a means of averting reserve losses. Consequently, if and when the official exchange rate is devalued, there may be little if any direct inflationary effect because the increase in prices and the reduction in real wages normally associated with a devaluation will already have taken place through a depreciation of the currency in the parallel market. This point is worth iterating by way of the following, simple example.

Assume first that the domestic economy and the rest of the world are in stationary equilibrium, with no inflation and no growth in real or monetary variables. Also assume, for simplicity, that the net international reserves of the domestic economy are zero. Then, suppose that there is an increase in credit in the domestic economy. In the absence of foreign exchange restrictions, this credit expansion will be fully offset by a loss in reserves, and the domestic money supply and price level will remain unaffected. An unanticipated devaluation, intended to restore the stock of reserves to the level prevailing before the credit expansion, will produce a once-and-for-all increase in the money supply, with corresponding increases in the price of imports and in the aggregate price level.

Now consider, for the same economy, a similar increase in domestic credit—but one accompanied by exchange controls to avoid a loss in reserves. In such a situation a parallel foreign exchange market will develop, and the exchange rate in this market will depreciate. The rate of inflation will be equal to the rate of credit expansion, but the price of imports will rise more rapidly. A reunification of the exchange system brought about by a devaluation of the official rate and by the removal of all controls (such that the stock of reserves remains unchanged) will cause the parallel rate to appreciate and the price of imports to decline. But the devaluation will have no inflationary effect in the current or subsequent periods because, with reserves constant, the money supply will be determined only by movements in domestic credit. Hence, although the time paths for prices will be different under the two exchange regimes, the rate of inflation over the period as a whole will be the same. The movement in prices and reserves under the restrictive and restriction-free regimes is illustrated in Figure 6.

Figure 6.
Figure 6.

Comparison of Price and Reserve Movements Under Alternative Control Regimes1

Citation: IMF Staff Papers 1984, 002; 10.5089/9781451946918.024.A005

1 In this example, domestic credit expansion begins at time t, and an unanticipated devaluation of the official exchange rate takes place at t + θ. In the absence of controls, the rise in the price level occurs at the time of the devaluation, t + θ, whereas under the restrictive regime the rise takes place between t and t + θ. The actual inflation rate between t and t + θ is, however, identical under both regimes. For illustrative purposes, the graph is drawn in continuous time.

The preceding argument is also applicable to movements in the real wage rate under the two trade regimes. Although the model described in this paper assumes a flexible money wage rate, the implications for exchange rate policy remain qualitatively the same if the wage rate is sticky and lies above its market-clearing level. Under a restrictive trade regime, the real wage rate will fall as the parallel market rate depreciates, but it will be unaffected by a devaluation of the official rate.22 This is in contrast to the situation under a restriction-free regime, where a devaluation results in a decline in the real wage.

The assumption that there is only a single imported good and that its price reflects the parallel exchange rate is clearly a strong one. It can readily be shown, however, that relaxing this assumption does not affect the result that, under conditions of fixed output and static exchange rate expectations, the aggregate price level will be independent of changes in the official exchange rate. For example, consider the case in which there are two imported goods, the demand for one of which is fully satisfied in the official market. In these circumstances, a devaluation will lead to a rise in the local price of the import purchased entirely in the official market but, for the reasons described above, will produce a decline in the price of the import purchased in both markets. As a result, there will be a change in relative prices, but, in the absence of any shift in the supply or demand for domestic money, the aggregate price level will remain unchanged.

These considerations notwithstanding, it is not necessarily the aggregate price level that is of prime concern to policymakers. More often, attention is focused on the consumer price index, which will ordinarily be much narrower in scope than an aggregate price index. In particular, such an index may exclude certain imports purchased in the parallel market—either because these imports are not considered consumer goods or because prices transacted in illegal markets are not monitored. Consequently, a devaluation of the official exchange rate may have an inflationary effect (as measured by the official consumer price index) even though the aggregate price level (as measured, for example, by the domestic expenditure deflator) remains unchanged. By similar reasoning, if there is more than one imported good, real wages measured on the basis of an official consumer price index may show a fall following a devaluation.

IV. Conclusions

This paper has developed a simple macroeconomic model in which the official exchange rate is fixed and the availability of foreign exchange through official channels is restricted. Thus, unlike conventional models of fixed exchange rates (including the dual market variants) that abstract from the existence of quantitative controls, the stock of official foreign exchange reserves is autonomously given. The foreign exchange premium offered in the parallel market as a result of inconvertibility in the official market provides an incentive for exporters to sell their output illegally. Thus, although the parallel rate equilibrates supply and demand for foreign exchange in the parallel market, it responds directly to shocks in the official market such as a change in the official exchange rate or an intensification of import controls.

The analysis has demonstrated within both partial equilibrium and general equilibrium frameworks that the official and parallel market exchange rates are inversely related to each other. Furthermore, in the absence of foreign exchange sales by the central bank, the parallel rate will be more depreciated than the “restriction-free” equilibrium rate. The relation between these two rates will, however, be reversed beyond a given level of reserve sales by the central bank. Thus, when the central bank undertakes large reserve sales, some caution should be exercised in using the parallel market rate as an indicator of the appropriate magnitude of exchange rate adjustment necessary to attain balance of payments equilibrium.

The model presented in this paper generates several results that differ strikingly from those derived from conventional fixed-flexible exchange rate models and their dual market variants. Because not all prices are flexible, specifically prices received by exporters who sell on the official market, and because the stock of reserves is in effect fixed through the imposition of controls, changes in domestic credit and in the official exchange rate have real effects in the long run. In addition, the model gives rise to some interesting conclusions about the price effects of a devaluation of the official rate. Because the presence of controls prevents a deterioration in the balance of payments when inflationary financial policies are being pursued, the domestic currency will undergo an effective depreciation in the parallel market. If the official exchange rate is devalued, there will be no subsequent inflationary impact because the increase in prices and the reduction in real wages normally associated with a devaluation will already have taken place. Indeed, the analysis shows that deflationary forces are set in motion if the devaluation is preannounced or widely anticipated, or if it is associated with an expansion in domestic output.

The analysis also underscores the importance of expectations in any empirical study of price and exchange rate movements. In particular, it shows that failure to take expectations of official exchange rate changes adequately into account may erroneously lead to the conclusion that a devaluation of the official rate in the presence of quantitative restrictions is inflationary, and that its influence on the parallel market rate is minor.

Finally, it is shown that suppression of parallel market activities and a reunification of the exchange rate system can only be brought about through a tightening in domestic credit, a devaluation of the official rate, the sale of official reserves, or a relaxation of exchange and trade controls. No automatic forces operate to realign the two exchange rates. Reunification also has important welfare implications: it raises the surpluses enjoyed by consumers of the imported good and by producers of the export good, and it eliminates the economic rent accruing to traders who can obtain access to the official foreign exchange market.

APPENDIX: Formal Description of the Model

partial equilibrium analysis

This section of the Appendix provides a more thorough description of the model’s real sector than that contained in the text and presents a formal derivation of the partial equilibrium results obtained graphically in Section I.

Production of both the nontraded and export good is assumed to take place in a competitive environment in which each producer takes the product price as given. All producers of the export good face identical (variable) costs of production but not of distribution. Specifically, it is assumed that, whereas it is costless for producers to sell the export good through official channels, the illicit nature of sales to the parallel market involves a penalty cost. This cost, which is known with certainty (admittedly a simplification; uncertainty combined with risk neutrality on the part of producers will generate equivalent results), is strictly proportional to the distance between the production site and the frontier and to the actual volume sold. The purpose of these assumptions is to ensure that, for a range of parallel market premiums, the exported good is sold simultaneously through both official and illegal channels. Hence, the marginal cost of distribution is constant for each producer engaged in parallel market trade but varies across producers according to distance from the frontier. The marginal distribution costs facing producers operating at the frontier are taken to be greater than zero but negligible.

It follows that the individual producer’s decision of whether to produce the nontraded or the export good, and in what quantities, will depend on relative output prices and wage costs. Wage costs will be positively related to the price of the private sector’s consumption bundle because the supply of labor is assumed to vary with the real wage rate. Any decision to produce the export good will involve a secondary decision of which market to supply. The producer will not sell through official channels if the premium he can obtain in the parallel market exceeds his distribution (penalty) costs per unit.

These considerations imply a set of supply functions of the following form:

X¯=X¯(e¯/e,e¯/pn,e¯/p),X¯1,X¯2,X¯3>0(4)
X=X(e/e¯,e/pn,e/p),X1,X2,X3>0;X1=X¯1(5)
S=S(pn/e¯,pn/e,pn/p),S1,S2,S3>0(6)

where X and X represent the quantities of the export good sold on the official and parallel markets, respectively; S is production of the nontraded good; e and e are the prices of foreign currency on the official and parallel markets, respectively; pn is the price of the nontraded good; and p is the price of the private sector’s consumption bundle (or aggregate price level). The foreign currency prices of the export and the imported good are assumed to be fixed and have been set equal to 1.

Thus, an increase in e relative to e induces a switch of sales of the export good from the official to the parallel market, whereas an increase relative to pn results in an increase in production of the export good and a decline in production of the nontraded good. A rise in p produces a reduction in the supply of labor (through a decline in the real wage) and, hence, a decline in the output of both goods.

The supply functions (4) and (5) will be subject to certain discontinuities. For example, when the parallel market premium is sufficiently large (or zero), none (or all) of the export good will be sold on the official market. Additional parameter restrictions depend on the elasticity of the supply of labor with respect to the real wage. If, at one extreme, this elasticity is zero, the aggregate volume of production of domestically produced goods will be fixed in the sense that the economy will, at all times, be on a given production possibility frontier. This implies that X3, X3, and S3 = 0. Production decisions will, therefore, be independent of the real wage. If, at the other extreme, the supply of labor is perfectly elastic, then the aggregate volume of output will be variable, with production decisions independent of relative output prices. This implies that X2, X2, S1, and S2 = 0.

The private sector’s consumption of the nontraded and import good will depend on the relative prices of these goods and on income:

N=N(e/pn,Y),N1,N2>0(7)

and

I=I(e/pn,Y),I1<0,I2>0(8)

where N and I are the demands for the nontraded and imported good, respectively, and Y is the aggregate level of output measured as the sum of the production of the exported and the nontraded good, both valued in constant base-period prices.

Equilibrium in the parallel market is given by the condition that the demand for foreign exchange be equal to its supply:

II¯+ΔF=X(9)

where I denotes import purchases in the official market, F represents resident holdings of foreign currency, and δ is a first-difference operator. The feature that distinguishes this model from that of the standard dual market approach is the presence of binding quantitative controls in the official market, which renders the stock of reserves exogenous. The amount of official market imports is therefore determined by the available supply in that market:

I¯=X¯ΔR/e¯(10)

where R is the stock of official reserves valued in domestic currency. Equation (10) is the officially recorded balance of payments. Adding equations (9) and (10) yields

I+ΔF=X¯+XΔR/e¯.(10)

Ex ante, equation (11) may be used to determine the parallel market rate for given values of all other variables; ex post, it is the balance of payments identity.

Some of the results obtained in the text of the paper may now be readily verified algebraically. Holding F constant, substituting equations (4), (5), and (8) into equation (11), and differentiating with respect to time yields a solution for changes in the parallel market rate in terms of changes in the official rate and in reserves (with reserves held constant in the base period):

ΔeΔe¯=X¯2+X¯3I1X2X3<0

and

ΔeΔR=1e¯(I1X2X3)>0.

The result is therefore confirmed that, in a partial equilibrium setting, the parallel rate appreciates when the official rate is devalued and when the central bank sells reserves.

general equilibrium analysis

Fixed Output with Static Expectations

The impact multipliers for e and p in terms of the policy variables D, e, and G may be derived from equations (1) through (3) with the addition of a government budget constraint, δD = pnG (with G in the base period, for simplicity, assumed to be equal to zero). Letting a circumflex (ˆ) denote a proportionate rate of change and a prime (′) an elasticity, the solution is

[e^p^]=[θ1θ2αe¯X¯X¯2Y(θ1αG)θ2DR+D0DR+D][D^G^e¯^]

where

θ1=(NN1+XX2+e¯X¯X¯2)DR+D>0

and

θ2=[NN1+e¯X¯X¯2(1α)+XX2]>0

therefore

e^D^>p^D^>p^nD^>0e^e¯^<0;p^e¯^=0;p^ne¯^>0e^G^0;p^nG^>p^G^>0.
Flexible Output with Static Expectations

Under the assumption that the supply of labor is perfectly elastic, the partial derivatives X2, X2, S1, and S2, in equations (4) through (6) are set equal to zero. This renders output endogenous. Y, e, pn, and p may then be solved using equations (2) and (3), together with the modified clearing condition for domestic output:

Y=N(e/pn,Y)+X¯(e¯/e,e¯/p)e¯+X(e/e¯,e/p)+G(12)

and the clearing condition for the nontraded good:

S(pn/p)=N(e/pn,Y).(13)

The effect of a devaluation of the official exchange rate on the price level is

p^e¯^=e¯X¯X¯3[L1(1α)(N1+S3)+αN1]θ3<0(14)

where

θ3=(1α)S3(Ye¯X¯X¯3L1NN2)+N1[Y(1α)e¯X¯X¯3L1]+αXX3N2.

The numerator in equation (14) is unambiguously negative, whereas the denominator, θ3, must be positive if the tâtonnement process is to be stable.

Rational Expectations with Fixed Output

We now assume that exchange rate expectations are formed rationally, but return to the assumption of fixed output. As in the text, attention is confined to the effect over time that a devaluation has on the parallel rate and the aggregate price level. For convenience, the weighting of the price level is changed from a geometric to an arithmetic average:

p=αpn+(1α)e,0<α<1.(3a)

The reduced solutions for e+1* and p in terms of e and e are derived from equations (1), (2), and (3a). Linearizing around long-run equilibrium yields

et+1*=β1et+β2e¯t+β3Z(15)

and

pt=δ1et+δ2e¯t+δ3Z(16)

where Z denotes a vector of exogenous variables that are constant over time and where

β1=1[N1+X2+(1α)e¯X¯2]L2γ>1β2=αe¯X¯2L2γ>0δ1=N1+X2+(1α)e¯X¯2γ>0δ2=αe¯X¯2γ>0γ=N1+e¯X¯2+X2>0.

It may be noted that δ1β21 − 1 = δ2. Since the model is nonstochastic, e+1* = e+1, and equation (15) represents a first-order difference equation in e. However, because β1 > 1, it is necessary to solve equation (15) in a forward direction and to set the arbitrary constant equal to zero in order to ensure a stable time path for e and, therefore, for p.

The forward-looking solution for e is

et=(β3/1β1)Z(β2/β1)Σi=0(1/β1)ie¯t+i.(17)

Letting t be the current time period, t be the time of the announcement of the devaluation, t + θ be the time it actually takes place, and e¯1 be the value of the official exchange rate prior to t + θ, equation (17) may be written as

et¯=(β3/1β1)Z(β2/β11)e¯0(β2/β1)Σt+θt¯(1/β1)i(e¯e¯0)t¯+i,(18)

where

e¯=e¯0fort<t+θ.

From equation (16), the time path for p is given by:

pt¯=[δ1β3/(1β1)+δ3]Z+δ2(e¯e¯0)t¯(δ1β2/β1)Σt+θt¯(1/β1)i(e¯e¯0)t¯+i.(19)

The time paths, equations (18) and (19), are illustrated in Figure 5 of the text. It may be confirmed that when tt − 1 (that is, before the announcement is made) and when tt + θ (that is, after the exchange rate is changed)

pt¯=[(δ1β3/1β1)+δ3]Z.

Hence, in the long run, the aggregate price level is independent of the official exchange rate when output is fixed.

SUMMARIES

The Disequilibrium Real Wage Rate Hypothesis: An Empirical Evaluationjacques r. artus (pages 249–302)

Both the share of labor costs in value added and the unemployment rate rose in many industrial countries during the 1970s and early 1980s. These increases have led many observers to conclude that real wages are now too high and are a source of “classical” unemployment. Such conclusions are not necessarily valid. The increase in the labor share could be warranted by long-run changes in production techniques, in the price of energy, or in the relative availability of labor and capital. Moreover, the observed unemployment could be structural or Keynesian, rather than classical. In this paper a production function approach is used to examine these possibilities and to subject the disequilibrium real wage rate hypothesis to a more rigorous empirical test than has been conducted in the past. Both the capital stock and the exchange rate are taken as given, and the study focuses only on the manufacturing sector.

The results indicate that, for France, the Federal Republic of Germany, and the United Kingdom, there are indeed strong reasons to believe that real wages in the manufacturing sector are now too high, in the sense of being incompatible with “high” employment. In particular, the study did not find any evidence that a large part of the actual increase in the share of labor costs in value added is warranted by long-run changes in production techniques, in the price of energy, or in the relative availability of labor and capital. For Canada and the United States, however, the results indicate that there is no real wage problem for the manufacturing sector as a whole, although there may of course be a problem in specific industries. For Japan and Italy the findings are less conclusive. For Japan, the results indicate that the large increase in the share of labor costs in value added is not fully warranted by concomitant changes in the factors considered in the study. At the same time, the initial labor share was so small that this increase may be less of a problem than it is in France, the Federal Republic of Germany, and the United Kingdom. For Italy, the results suggest that there is no real wage problem, but poor data prevent any firm conclusion in this regard.

Relative Prices, Real Wages, and Macroeconomic Policies: Some Evidence from Manufacturing in Japan and the United Kingdomleslie lipschitz and susan m. schadler (pages 303–38)

Countries differed in their adjustment to the commodity price shocks of the 1970s. In some countries there was little wage adjustment, so that the brunt of the terms-of-trade loss was borne by profits; in others greater wage adjustment allowed the terms-of-trade loss to be distributed more evenly between labor and capital. The extent of wage adjustment is important to the current unemployment problem and to the debate on the efficacy of demand management. Where there has been little wage adjustment, poor output and employment performance is likely to be due to (“classical”) excessive wage costs rather than simply to a (“cyclical”) deficiency of demand.

This paper uses an econometric model to separate the cyclical and classical influences on manufacturing output and employment. It compares results for Japan and the United Kingdom, countries chosen because of significant differences between their labor markets. The comparison indicates that labor cost developments depressed output and employment growth in the United Kingdom during the last decade; in Japan, however, labor cost developments exerted a small positive influence on manufacturing output and employment during the same period.

Two measures of “warranted” changes in real wages are constructed from the estimated model. The first indicates wage movements consistent with fixed factor shares in value added, and the second, wage movements that would equilibrate labor supply and demand. The gap between actual and warranted wages provides an indication of the adequacy of wage adjustment and, in turn, of the sustainability of an output and employment response to an expansion of demand. During the last decade, and especially since the second oil price increase, the wage gap widened considerably more rapidly in the United Kingdom than in Japan. The wage gap estimates underscore the dangers of too rapid an expansion of demand, particularly in the United Kingdom.

Optimum Taxation and Tax Policynicholas h. stern (pages 339–78)

What types of goods should be taxed? How progressive should the income tax be? What should be the balance between the taxation of commodities and the taxation of income? These questions have concerned many leading economists of the last two centuries, and the past 15 years has seen a tremendous surge in the formal analysis of these problems. This more recent literature is often termed “optimum taxation.” The paper gives a nontechnical introduction to this topic in order to provide a broad understanding of the methods of approach, the type of arguments used, and the conclusions reached.

The study of the optimum taxation of commodities goes back to Ramsey (1927) and was developed by Samuelson and Boiteux and, more recently, by Diamond and Mirrlees (1971) and others. The Ramsey rule for the one-consumer economy is that, for small taxes, the proportional reduction in compensated demand (compared with demands in the absence of taxes) should be the same for all goods. In the many-consumer case, one also has to take into account the importance of the consumption of any good in the budget of deserving groups. The study of optimum income taxation was created by Mirrlees (1971), and the nonlinear and linear cases are presented, together with calculations for the latter. The optimum combination of direct and indirect taxes is shown to be sensitive to the sophistication of the tax tools involved (for example, whether the income tax is linear or nonlinear), to the way incomes are generated, and to the form of supply and demand functions.

The theory of reform (that is, the appraisal of small movements from a given initial condition) is introduced and compared with that of optimality. The data requirements and assumptions involved in such an analysis are more modest than those for optimality.

Certain robust and general conclusions are identified: for example, (1) efficiency considerations point to the taxation of inelastically demanded goods (with cross-elasticities treated properly, however), and (2) the interrelation between taxes is a major determinant of their optimum level.

Government Policy and Private Investment in Developing Countriesmario i. blejer and mohsin s. khan (pages 379–403)

The behavior of private investment in developing countries has obvious implications, both for long-term development and for the design of shorter-term stabilization programs. What determines the rate of private investment and how this rate responds to changes in government policies are questions of considerable importance to policymakers and academics alike. For example, would a tightening of monetary policy result in a fall in real private capital formation or in no change? Similarly, would an increase in government capital expenditures have a negative or positive effect on private investment? Clearly, any meaningful analysis of growth in developing countries must take into account questions of this nature.

Although private investment behavior in industrial countries has been studied extensively, there is as yet little systematic evidence on this subject for developing countries. In this paper a model of private investment in developing countries is formulated to analyze the empirical relations between private investment and some of its main determinants. The exercise focuses on the influences that variations in bank credit and in government capital formation have on the private sector’s investment decisions. The concentration on these two factors allows an explicit treatment of the issue of real and financial “crowding out,” a subject over which there is considerable controversy. Furthermore, the analysis attempts to make an empirical distinction between public investment related to the development of infrastructure, which is likely to be complementary with private investment, and other types of government investment, which may substitute for private capital formation.

The model is estimated for 24 developing countries by using a data set for private and public investment that was specially constructed for this exercise. Data on the public sector—properly defined to include the general government, autonomous institutions, and nonfinancial state enterprises—are not readily available and must be assembled from different sources. The empirical results indicate that it is possible to identify a fairly well-behaved private investment function for developing countries. The principal policy-related conclusions of the study are twofold. First, changes in bank credit to the private sector have a significant effect on private investment, such that a reduced flow of real credit to the private sector, for whatever reason, causes real private investment to decline. Second, an increase in the infrastructural component of government capital formation (represented by various empirical proxies) raises private investment, but similar increases in other kinds of public investment appear to result in some crowding out.

Quantitative Controls and Unofficial Markets in Foreign Exchange: A Theoretical Frameworkmichael nowak (pages 404–31)

This paper examines the effects of monetary, fiscal, and exchange rate policy in a dual exchange rate regime that has arisen in response to exchange controls in the official market. A key feature distinguishing this arrangement from other dual exchange systems is the absence of an automatic mechanism, such as reserve movements or exchange rate changes, to ensure that the official foreign exchange market clears. As a consequence, there are incentives for illegal foreign exchange transactions to take place at a premium price in a secondary or parallel market. Starting from a partial equilibrium setting and extending the analysis to a general equilibrium framework, the paper presents results that differ significantly from those of standard models of an open economy that allow the official foreign exchange market to clear.

The principal conclusions of the paper are as follows. First, under a plausible set of assumptions, a devaluation of the official exchange rate results in an unambiguous appreciation of the parallel market rate. Second, when the official foreign exchange market is not allowed to clear, changes in domestic credit and in the official exchange rate have permanent real effects. In addition, because the presence of controls prevents a deterioration in the balance of payments when inflationary financial policies are being pursued, the domestic currency will undergo an effective depreciation in the parallel market. Devaluing the official exchange rate under such conditions will have no inflationary effect because the increase in prices and the reduction in real wages normally associated with a devaluation will already have taken place. Indeed, the analysis shows that deflationary forces are set in motion if the devaluation is preannounced or widely anticipated. Third, the paper shows that, in a restrictive exchange regime, failure to take adequately into account expectations of changes in the official exchange rate may lead to the erroneous conclusion that a devaluation of the official rate is inflationary and that its influence on the parallel market rate is minor.

RESUMES

L’hypothèse du déséquilibre du taux de salaire réel : évaluation empiriquejacques r. artus (pages 249–302)

Au cours des années 70 et au début des années 80, la part des coûts de main-d’oeuvre dans la valeur ajoutée et le taux de chômage ont augmenté dans la plupart des pays industrialisés. De nombreux observateurs en ont conclu que les salaires réels sont maintenant trop élevés et provoquent un chômage de type classique. De telles conclusions ne sont pas nécessairement valables. Il se pourrait que cette augmentation de la part de la main-d’oeuvre soit justifiée par l’évolution à long terme de certains facteurs tels que les techniques de production, le coût de l’énergie, ou l’offre relative de main-d’oeuvre et de capital. En outre, le chômage observé pourrait être structurel ou de type keynésien et non classique. Dans la présente étude, on utilise une fonction de production pour examiner ces possibilités et soumettre l’hypothèse du déséquilibre du taux de salaire réel à un test empirique plus rigoureux que par le passé. Le stock de capital et le taux de change sont considérés comme donnés, et l’analyse porte uniquement sur le secteur manufacturier.

D’après les résultats obtenus, on a de bonnes raisons de croire que les salaires réels dans le secteur manufacturier sont maintenant trop élevés, en ce sens qu’ils sont incompatibles avec un haut niveau d’emploi, dans trois pays : France, République fédérale d’Allemagne et Royaume-Uni. En particulier, l’étude conclut que l’augmentation actuelle de la part des coûts de main-d’oeuvre dans la valeur ajoutée ne semble pas être due, pour l’essentiel, à l’évolution à long terme des techniques de production, du coût de l’énergie, ou de l’offre relative de main-d’oeuvre et de capital. Au Canada et aux Etats-Unis, par contre, il n’existe, d’après les résultats, pas de problème de salaires réels pour le secteur manufacturier dans son ensemble, mais un tel problème peut, évidemment, se poser dans certaines branches d’activité spécifiques. En ce qui concerne le Japon et l’Italie, les conclusions sont plus hésitantes. Au Japon, la forte progression de la part des coûts de main-d’oeuvre dans la valeur ajoutée n’est pas pleinement justifiée par l’évolution concomitante des facteurs envisagés dans l’étude. En même temps, la part de la main-d’oeuvre était initialement si faible que cette augmentation pose peut-être moins de problèmes qu’en France, en République fédérale d’Allemagne et au Royaume-Uni. Dans le cas de l’Italie, les résultats donnent à penser qu’il ne se pose pas de problème de salaires réels, mais la mauvaise qualité des données statistiques n’autorise pas de conclusion définitive à cet égard.

Prix relatifs, salaires réels et politiques macroéconomiques : résultats obtenus dans l’industrie manufacturière au Japon et au Royaume-Unileslie lipschitz et susan m. schadler (pages 303–38)

Les pays ont réagi différemment aux chocs causés par les prix des matières premières au cours des années 70. Dans certains d’entre eux, l’ajustement des salaires a été faible de sorte que l’essentiel de la perte des termes de l’échange a été supporté par les bénéfices des entreprises; dans d’autres, où l’ajustement des salaires a été plus fort, la perte des termes de l’échange a pu être répartie de façon plus équitable entre le travail et le capital. Le degré d’ajustement des salaires joue un rôle important lorsqu’on cherche à résoudre le problème actuel du chômage et lorsqu’on cherche à déterminer l’efficacité de la gestion de la demande. En cas de faible ajustement des salaires, des résultats médiocres en matière de production et d’emploi sont probablement imputables au niveau excessif des coûts salariaux (“influences classiques”) et pas simplement à l’insuffisance de la demande (“influences cycliques”).

Les auteurs du présent document utilisent un modèle économétrique afin de distinguer les influences cycliques et les influences classiques qui s’exercent sur la production manufacturière et sur l’emploi. Ils comparent les résultats obtenus pour le Japon et pour le Royaume-Uni, pays qui ont été choisis en raison des différences considérables que présentent leurs marchés du travail. La comparaison indique que l’évolution des coûts de main-d’oeuvre a déprimé la croissance de la production et de l’emploi au Royaume-Uni au cours de la dernière décennie; au Japon, toutefois, l’évolution des coûts de main-d’oeuvre a exercé une influence légèrement positive sur la production manufacturière et sur l’emploi pendant la même période.

Deux mesures des variations “justifiées” des salaires réels ont été établies à partir du modèle estimé. La première indique les variations de salaires compatibles avec des parts de facteurs fixes dans la valeur ajoutée, et la seconde, les variations de salaires qui permettraient d’équilibrer l’offre et la demande de main-d’oeuvre. L’écart entre les salaires effectifs et les salaires justifiés donne une indication du caractère adéquat de l’ajustement des salaires et, partant, de la durée de la réaction de la production et de l’emploi à une expansion de la demande. Au cours des années 70, et en particulier depuis la deuxième hausse des prix pétroliers, l’écart de salaires s’est creusé beaucoup plus rapidement au Royaume-Uni qu’au Japon. Les estimations des écarts de salaires mettent en évidence les dangers d’une expansion trop rapide de la demande, en particulier au Royaume-Uni.

Imposition optimale et politique fiscalenicholas h. stern (pages 339–78)

Quelles catégories de produits doivent-elles être imposées? Quelle doit être la progressivité de l’impôt sur le revenu? Quel doit être l’équilibre à établir entre l’imposition des biens et celle du revenu? Ces questions ont retenu l’attention de nombreux économistes parmi les plus importants depuis deux siècles et ont donné lieu, au cours des 15 dernières années, à une multitude d’analyses théoriques consacrées à ces problèmes. Les ouvrages publiés récemment à ce sujet sont souvent regroupés sous le chapitre de l’”Imposition optimale” de la théorie économique. Le document se propose de donner un aperçu de caractère non technique de cette question afin de faciliter la compréhension générale des méthodes d’approche, du type d’arguments utilisés et des conclusions qui s’en dégagent.

L’étude de l’imposition optimale des biens remonte à Ramsey (1927) et a été développée par Samuelson et Boiteux et, plus récemment, par Diamond et Mirrlees (1971), ainsi que par d’autres économistes. La loi de Ramsey, dans le cas d’une économie à un consommateur, est que, dans le cas d’impôts de faible importance, la réduction proportionnelle de la demande compensée (opposée à la demande en l’absence d’impôts) doit être la même pour tous les biens. Dans le cas d’une économie à plusieurs consommateurs, il faut tenir compte également de l’importance que revêt la consommation de chaque bien dans le budget des groupes de population concernés. La théorie de l’imposition optimale du revenu a été créée par Mirrlees (1971) et les cas non linéaires et linéaires sont présentés, ainsi que les calculs en ce qui concerne le deuxième cas. Il apparaît que le dosage optimal entre impôts directs et impôts indirects est sensible à la complexité des instruments fiscaux utilisés (par exemple, au caractère linéaire ou non linéaire de l’impôt sur le revenu), à la manière dont les revenus sont produits et à la forme des fonctions d’offre et de demande.

La théorie de la réforme (c’est-à-dire l’évaluation de mouvements de faible importance à partir d’une condition initiale) est présentée et comparée à celle de l’optimalité. Les données nécessaires et les hypothèses à retenir dans le cadre d’une telle analyse sont moins nombreuses que dans le cas de la théorie de l’optimalité.

Certaines conclusions importantes et générales sont présentées : par exemple : 1) des considérations ayant trait à l’efficacité militent en faveur de l’imposition des biens faisant l’objet d’une demande inélastique (les élasticités croisées étant cependant traitées de manière appropriée); et 2) la relation mutuelle entre impôts est un déterminant important de leur niveau optimal.

Politique des pouvoirs publics et investissement privé dans les pays en développementmario i. blejer et mohsin s. khan (pages 379–403)

Le comportement de l’investissement privé dans les pays en développement a, de toute évidence, des répercussions tant sur le développement à long terme que sur l’élaboration des programmes de stabilisation à plus court terme. Quels sont les facteurs déterminant le taux de l’investissement privé et comment ce taux réagit-il aux modifications de la politique appliquée par les pouvoirs publics? Ce sont là des questions qui revêtent une importance considérable tant pour les responsables de la politique économique que pour les théoriciens. Par exemple, un resserrement de la politique monétaire se traduira-t-il par une baisse de la formation de capital dans le secteur privé, en valeur réelle, ou n’aura-t-il aucun effet sur cette variable? De même, une augmentation des dépenses d’investissement de l’Etat aura-t-elle un effet négatif ou positif sur l’investissement privé? Il est clair que, pour être valable, toute analyse de la croissance dans les pays en développement doit tenir compte des questions de cette nature.

S’il est vrai que le comportement de l’investissement privé dans les pays industrialisés a fait l’objet d’études approfondies, rares sont les analyses qui ont été, jusqu’à présent, systématiquement consacrées à ce sujet dans le cas des pays en développement. Dans ce document, un modèle de l’investissement privé dans les pays en développement a été construit en vue d’analyser la relation empirique existant entre l’investissement privé et certains de ses principaux déterminants. L’analyse est principalement consacrée à l’influence que les variations du crédit bancaire et de la formation de capital dans le secteur public exercent sur les décisions du secteur privé en matière d’investissement. Etant donné qu’elle porte principalement sur ces deux facteurs, l’analyse permet d’examiner de manière explicite le phénomène d’éviction réelle et financière, question qui est largement controversée. En outre, l’analyse cherche à établir une distinction empirique entre l’investissement public qui est lié au développement de l’infrastructure, et qui, selon toute vraisemblance, est un complément de l’investissement privé, et d’autres types d’investissement public qui peuvent se substituer à la formation de capital dans le secteur privé.

Le modèle est estimé pour 24 pays en développement à partir d’une série de données relatives à l’investissement privé et public qui a été établie spécialement aux fins de cette analyse. Les données relatives au secteur public, défini de manière à inclure l’administration générale, les institutions autonomes et les entreprises d’Etat non financières, ne sont pas directement disponibles et doivent être établies au moyen de renseignements provenant de différentes sources. Selon les résultats empiriques, il est possible d’identifier, pour les pays en développement, une fonction de l’investissement privé possédant d’assez bonnes propriétés. Les conclusions principales de l’étude, sur le plan de la politique économique, sont les suivantes : en premier lieu, les variations du crédit bancaire au secteur privé ont sur l’investissement privé une incidence significative, telle que, si le montant du crédit accordé au secteur privé diminue, en valeur réelle, pour n’importe quelle raison, l’investissement privé aura, lui-aussi, tendance à diminuer en valeur réelle. En deuxième lieu, une augmentation de la composante infrastructure de la formation de capital dans le secteur public (représentée par différentes variables empiriques de remplacement) aura pour effet d’accroître l’investissement privé; en revanche, une augmentation analogue d’autres types d’investissement public semble avoir, dans une certaine mesure, un effet d’éviction sur les investissements du secteur privé.

Contrôles quantitatifs et marchés non officiels de change:un cadre théoriquemichael nowak (pages 404–31)

Le document examine les effets des politiques monétaire, budgétaire et de taux de change dans le cadre d’un régime de double marché des changes qui est apparu en réaction à l’existence de contrôles de change sur le marché officiel. Une des caractéristiques essentielles distinguant un tel régime des autres régimes de double marché des changes est l’absence de mécanismes automatiques, tels que les variations des réserves ou les mouvements du taux de change, permettant d’assurer la réalisation de l’équilibre sur le marché officiel des changes. En conséquence, il existe des facteurs incitant à effectuer des transactions illégales de change à un taux comportant une prime, sur le marché secondaire ou parallèle. L’analyse, d’abord effectuée dans le cadre d’un modèle d’équilibre partiel et ensuite étendue au cadre d’un modèle d’équilibre général, aboutit à des résultats qui diffèrent sensiblement de ceux que font apparaître les modèles types d’économie ouverte dans lesquels l’équilibre est réalisé sur le marché officiel des changes.

Les conclusions principales de l’étude sont les suivantes. En premier lieu, compte tenu d’une série d’hypothèses plausibles, une dévaluation du taux de change officiel se traduit par une nette appréciation du taux de marché parallèle. En deuxième lieu, lorsque l’équilibre ne peut pas être réalisé sur le marché de change officiel, les variations du crédit intérieur et du taux de change officiel ont des effets réels permanents. En outre, étant donné que l’existence de contrôles empêche toute détérioration de la balance des paiements lorsque des politiques financières inflationnistes sont appliquées, la monnaie nationale subira une dépréciation effective sur le marché parallèle. Dans de telles conditions, une dévaluation du taux de change officiel n’aura pas d’effet inflationniste étant donné que la hausse des prix et la baisse des salaires réels, qui accompagnent normalement une dévaluation, auront déjà eu lieu. De fait, l’analyse montre que des facteurs déflationnistes commencent à jouer si l’on s’attend généralement à ce qu’une dévaluation ait lieu ou si celle-ci est annoncée préalablement. En troisième lieu, l’analyse montre que, dans le cadre d’un régime de change restrictif, le fait de ne pas tenir suffisamment compte des anticipations concernant les variations du taux de change officiel peut conduire à la conclusion erronée qu’une dévaluation du taux officiel est inflationniste et que son influence sur le taux du marché parallèle est négligeable.

RESUMENES

La hipótesis de desequilibrio del salario real: Una evaluación empíricajacques r. artus (páginas 249–302)

En el curso de los años setenta y principios de los ochenta, en la mayoría de los países industriales aumentaron tanto la participación del costo de la mano de obra en el valor añadido como la tasa de desempleo. Dichos aumentos han inducido a muchos observadores a sostener que actualmente los salarios reales son demasiado altos y que ocasionan desempleo “clásico”. Tales conclusiones no son necesariamente acertadas. El aumento en la participación del costo de la mano de obra puede obedecer a cambios de largo plazo en ciertos factores tales como: las técnicas de producción, el precio de la energía, o la disponibilidad relativa de la mano de obra y del capital. Además, el desempleo registrado pudo haber sido estructural o keynesiano, y no clásico. En el presente estudio se emplea un enfoque de función de producción para examinar tales posibilidades y para someter la hipótesis de desequilibrio del salario real a una prueba empírica más rigurosa que las que se han aplicado hasta ahora. Este estudio se limita al sector manufacturero, considerándose como dados el tipo de cambio y la masa de capital.

De los resultados se infiere que, en cuanto a Francia, el Reino Unido y la República Federal de Alemania existen ciertamente razones poderosas para creer que los salarios reales del sector fabril en la actualidad son excesivamente elevados, en el sentido de que son incompatibles con un nivel “alto” de empleo. En especial, en este estudio se llega a la conclusión de que no parece que una parte considerable del aumento registrado en la participación del costo de la mano de obra en el valor añadido obedezca a variaciones de largo plazo en las técnicas de producción, el precio de la energía o la disponibilidad relativa de la mano de obra y el capital. Ahora bien, en lo que se refiere a Canadá y Estados Unidos, de los resultados se deduce que el conjunto del sector fabril no padece un problema de salarios reales, aunque sí puede haberlo, naturalmente, en determinadas industrias. En cuanto a Japón e Italia, los resultados son más dudosos. En lo que atañe a Japón, de los resultados se deduce que el gran aumento de la participación del costo de la mano de obra en el valor añadido no está justificado enteramente en virtud de la variación concomitante de los factores que se tienen en cuenta en el estudio. Al propio tiempo, la participación inicial de la mano de obra era tan reducida que puede ser que este aumento constituya en Japón un problema menor que en Francia, el Reino Unido o la República Federal de Alemania. En lo que hace a Italia, los resultados sugieren que no existe un problema de salarios reales, por más que la imperfección de los datos impida llegar a conclusiones firmes en este sentido.

Precios relativos, salarios reales y políticas macroeconómicas: Algunas conclusiones obtenidas de la industria japonesa y británicaleslie lipschitz y susan m. schadler (páginas 303–38)

El ajuste a las crisis de precios de productos básicos durante el decenio de los setenta ha sido diferente en cada país. En algunos apenas hubo ajuste salarial, incidiendo el deterioro de la relación de intercambio especialmente en las utilidades; en otros, un mayor ajuste salarial permitió una distribución más pareja del deterioro de la relación de intercambio entre el trabajo y el capital. La medida del ajuste salarial es importante desde el punto de vista del problema actual de desempleo y para dilucidar la eficacia de la regulación de la demanda. En los países en que el ajuste salarial ha sido limitado, la deficiente evolución del producto y el empleo probablemente se ha debido al excesivo costo salarial (factor “clásico”) y no sencillamente a deficiencias de la demanda (factor “coyuntural”).

En este estudio se utiliza un modelo econométrico para distinguir la influencia de los factores coyunturales y de los factores clásicos sobre el producto y el empleo industriales. Se comparan los resultados de Japón y el Reino Unido, países escogidos por las importantes diferencias que presentan sus respectivos mercados de trabajo. De la comparación se deduce que la evolución del costo de la mano de obra deprimió el crecimiento del producto y el empleo en el Reino Unido durante el último decenio, mientras que en Japón ejerció una pequeña influencia positiva sobre ambos durante el mismo período.

A partir del modelo estimado se elaboran dos medidas de variación “justificada” de los salarios reales. La primera indica variaciones salariales congruentes con participaciones fijas de los factores en el valor añadido, y la segunda, variaciones salariales que equilibrarían la oferta y demanda de trabajo. La diferencia entre el nivel de salarios justificados y efectivos indica la medida adecuada del ajuste salarial, y también, si es sostenible una reacción determinada del producto y del empleo ante una expansión de la demanda. Durante el último decenio, y especialmente a partir del segundo aumento de precios del petróleo, esa diferencia salarial aumentó mucho más rápidamente en el Reino Unido que en Japón. Las estimaciones de esta diferencia entre nivel efectivo y nivel justificado de salarios ponen de relieve los peligros de una expansión excesivamente rápida de la demanda, particularmente en el Reino Unido.

Tributación óptima y política fiscalnicholas h. stern (páginas 339–78)

¿Qué clase de productos deben gravarse? ¿Qué grado de progresividad debe tener el impuesto a la renta? ¿Cuál debe ser la combinación óptima de la imposición de los bienes y servicios y la de la renta? En los últimos dos siglos muchos destacados economistas se han planteado estas preguntas, y en los últimos 15 años ha aumentado notablemente el estudio sistemático de estos problemas. A este tipo de análisis más reciente sobre el tema se le conoce frecuentemente como “tributación óptima”. El trabajo presenta una introducción no técnica al tema a fin de ofrecer una presentación general de los métodos de análisis, la clase de argumentos utilizados y las conclusiones a las que se llega.

El estudio de la tributación óptima de los bienes se remonta a Ramsey (1927) y fue desarrollado por Samuelson y Boiteux y, más recientemente, por Diamond y Mirrlees (1971), entre otros. La regla de Ramsey referente a una economía con un solo consumidor indica que, para niveles bajos de tributación, la reducción proporcional de la demanda compensada (comparada con lo que sería si no existieran impuestos) debe ser igual para todos los productos. En el caso de una economía con muchos consumidores también se debe tener en cuenta la importancia del consumo de cada artículo en el presupuesto de los grupos de bajos ingresos. El estudio de la imposición óptima de la renta fue iniciado por Mirrlees (1971); en este trabajo se presentan los casos lineal y no lineal, así como cálculos relativos a este último. Se muestra que la combinación óptima de impuestos directos e indirectos es sensible a la sofisticación de los instrumentos tributarios disponibles (por ejemplo, si el impuesto a la renta es o no lineal), a la forma en que se generan los ingresos y a la forma de las funciones de oferta y demanda.

Se presenta también la teoría de reforma (es decir, la evaluación de pequeños cambios a partir de una condición inicial) y se la compara con la teoría de optimalidad tributaria. Dicho análisis presenta menores requerimientos de datos y de supuestos que el análisis de la optimalidad.

Se presentan ciertas conclusiones de validez general. Por ejemplo: 1) desde el punto de vista de la eficiencia, deben gravarse los productos de demanda inelástica (teniendo debidamente en cuenta las elasticidades cruzadas de estos bienes) y 2) la interrelación entre los impuestos constituye un factor central determinante del nivel óptimo de tributación.

Política estatal e inversión privada en los países en desarrollomario i. blejer y mohsin s. khan (páginas 379–403)

El comportamiento de la inversión privada en los países en desarrollo tiene repercusiones obvias tanto sobre el proceso de desarrollo a largo plazo como para la formulación de programas de estabilización a más corto plazo. Cuáles son los factores que determinan el coeficiente de inversión privada y de qué manera reacciona éste ante las variaciones de las políticas nacionales son interrogantes que revisten una importancia considerable tanto para las autoridades como para los teóricos. Por ejemplo, ¿ocasionaría una política monetaria más restrictiva un descenso en la formación real de capital privado? Igualmente, ¿un aumento del gasto de capital del sector público tendría una repercusión negativa o positiva en la inversión privada? Evidentemente, en todo análisis del crecimiento económico de los países en desarrollo se deben tener en cuenta interrogantes de esta naturaleza.

Si bien el comportamiento de la inversión privada en los países industriales ha sido estudiado detalladamente, son muy pocos los análisis sistemáticos consagrados a este tema en el caso de los países en desarrollo. En este trabajo se formula un modelo de inversión privada en los países en desarrollo con el fin de analizar las relaciones empíricas entre la inversión privada y algunos de sus principales determinantes. El análisis se concentra en la influencia sobre las decisiones del sector privado en materia de inversión de las variaciones del crédito bancario y de la formación de capital por parte del sector público. Al concentrarse en estos dos factores es posible efectuar un análisis explícito de la “exclusión” (crowding out) real y financiera, tema que ha suscitado grandes controversias en estos últimos tiempos. Además, mediante el análisis se trata de establecer una diferencia empírica entre la inversión pública destinada al desarrollo de la infraestructura, la cual en general es un complemento de la inversión privada, y otros tipos de inversión pública, los cuales pueden sustituir a la formación privada de capital.

El modelo se ha estimado para 24 países en desarrollo usando una serie de datos sobre inversión privada y pública preparada especialmente para los fines del presente análisis. Los datos sobre el sector público—que según esta definición incluye las administraciones públicas, las entidades autónomas y las empresas estatales no financieras—no son fáciles de obtener y se los ha extraído de diferentes fuentes. Los resultados empíricos indican que es posible identificar una función de inversión privada bastante específica para los países en desarrollo. Son dos las principales conclusiones del estudio en materia de política: en primer lugar, las variaciones del crédito bancario al sector privado tienen una repercusión considerable en la inversión privada, motivo por el cual toda reducción de la corriente de crédito real al sector privado—cualquiera sea el motivo—ocasiona una disminución de la inversión privada real. En segundo lugar, un aumento del componente infraestructural de la formación de capital público (componente representado por diferentes variables empíricas) da lugar a un aumento de la inversión privada, mientras que aumentos similares en las otras categorías de inversión pública parecen traer aparejada una cierta “exclusión”.

Controles cuantitativos y mercados no oficiales de divisas: Un marco téoricomichael nowak (páginas 404–31)

En este trabajo se examinan los efectos de la política monetaria, fiscal y cambiaria en un régimen de tipo doble de cambio, surgido como consecuencia de la imposición de controles cambiarios en el mercado oficial. Una característica fundamental que diferencia a este régimen de otros sistemas de doble tipo de cambio es la ausencia de un mecanismo automático, como las fluctuaciones de las reservas o del tipo de cambio, que asegure que el mercado oficial de divisas pueda alcanzar un equilibrio. Como consecuencia de la ausencia de ese mecanismo, surgen incentivos para efectuar transacciones cambiarias ilegales con prima en un mercado secundario o paralelo. A partir de un modelo de equilibrio parcial y ampliando más adelante el análisis a un contexto de equilibrio general, en el trabajo se presentan resultados que difieren en forma significativa de los resultados obtenidos con los modelos más usados de economía abierta que permite el equilibrio del mercado oficial de divisas.

Las principales conclusiones a las que se llega en el trabajo son las siguientes. Primero, con un conjunto aceptable de supuestos, una devaluación del tipo de cambio oficial da por resultado una apreciación inequívoca del tipo de cambio del mercado paralelo. Segundo, cuando las condiciones del mercado oficial de divisas impiden que se equilibre, las modificaciones del crédito interno y del tipo de cambio oficial tienen efectos reales permanentes. Además, debido a que la imposición de controles impide el deterioro de la balanza de pagos cuando se aplica una política financiera inflacionaria, la moneda nacional experimenta una depreciación real en el mercado paralelo. En estas circunstancias la devaluación del tipo de cambio oficial no tendrá efectos inflacionarios, ya que el aumento de los precios y la reducción de los salarios reales, normalmente relacionados con una devaluación, habrían ocurrido antes. En efecto, el análisis demuestra que, si la devaluación se anuncia de antemano o se prevé en forma generalizada se ponen en movimiento fuerzas deflacionarias. Tercero, el trabajo demuestra que, en un régimen cambiario restrictivo, si no se tienen debidamente en cuenta las expectativas sobre las modificaciones del tipo de cambio oficial, se puede llegar a la conclusión equivocada de que una devaluación del tipo oficial tiene efectos inflacionarios y de que no ejerce mayor influencia en el tipo de cambio del mercado paralelo.

REFERENCES

  • Argy, Victor, and Michael G. Porter, The Forward Exchange Market and the Effects of Domestic and Foreign Disturbances Under Alternative Exchange Rate Systems,” Staff Papers, International Monetary Fund (Washington), Vol. 19 (November 1972), pp. 50328.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bhagwati, Jagdish N., Fiscal Policies, the Faking of Foreign Trade Declarations, and the Balance of Payments,” Bulletin of the Oxford University Institute of Economics and Statistics, Vol. 29 (February 1967), pp. 6177.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bhagwati, Jagdish N., Foreign Trade Regimes and Economic Development: Anatomy and Consequences of Exchange Control Regimes (Cambridge, Massachusetts: Ballinger, 1978).

    • Search Google Scholar
    • Export Citation
  • Blejer, Mario I., Exchange Restrictions and the Monetary Approach to the Exchange Rate,” in The Economics of Exchange Rates, ed. by Jacob A. Frenkel and Harry G. Johnson (London: Addison-Wesley, 1978), pp. 11728.

    • Search Google Scholar
    • Export Citation
  • Braga de Macedo, Jorge, Exchange Rate Behavior with Currency Inconvertibility,” Journal of International Economics (Amsterdam), Vol. 12 (February 1982), pp. 6581.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cumby, Robert E., Monetary Policy Under Dual Exchange Rates” (unpublished, International Monetary Fund, May 17, 1983).

  • Dickie, Paul M., and David B. Noursi, Dual Markets: The Case of the Syrian Arab Republic,” Staff Papers, International Monetary Fund (Washington), Vol. 22 (July 1975), pp. 45668.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dornbusch, Rudiger, and others, The Black Market for Dollars in Brazil,” Quarterly Journal of Economics (Cambridge, Massachusetts), Vol. 98 (February 1983), pp. 2540.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fleming, J. Marcus, Dual Exchange Markets and Other Remedies for Disruptive Capital Flows,” Staff Papers, International Monetary Fund (Washington), Vol. 21 (March 1974), pp. 127.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flood, Robert P., Exchange Rate Expectations in Dual Exchange Markets,” Journal of International Economics (Amsterdam), Vol. 8 (February 1978), pp. 6577.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Flood, Robert P., and Nancy P. Marion, The Transmission of Disturbances Under Alternative Exchange-Rate Regimes with Optimal Indexing,” Quarterly Journal of Economics (Cambridge, Massachusetts), Vol. 97 (February 1982), pp. 4366.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kami, Edi, On the Specification of Asset Equilibrium in Macroeconomic Models: A Note,” Journal of Political Economy (Chicago), Vol. 87 (February 1979), pp. 17177.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Krueger, Anne O., Foreign Trade Regimes and Economic Development: Liberalization Attempts and Consequences (Cambridge, Massachusetts: Ballinger, 1978).

    • Search Google Scholar
    • Export Citation
  • Lanyi, Anthony, Separate Exchange Markets for Capital and Current Transactions,” Staff Papers, International Monetary Fund (Washington), Vol. 22 (November 1975), pp. 71449.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marion, Nancy P., Insulation Properties of a Two-Tier Exchange Market in a Portfolio Balance Model,” Economica (London), Vol. 48 (February 1981), pp. 6170.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Michaely, Michael, A Geometrical Analysis of Black-Market Behavior,” American Economic Review (Nashville, Tennessee), Vol. 44 (September 1954), pp. 62737.

    • Search Google Scholar
    • Export Citation
  • Sargent, Thomas J., and Neil Wallace, The Stability of Models of Money and Growth with Perfect Foresight,” Econometrica (Evanston, Illinois), Vol. 41 (November 1973), pp. 104348.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sheikh, Munir A., Black Market for Foreign Exchange, Capital Flows, and Smuggling,” Journal of Development Economics (Amsterdam), Vol. 3 (March 1976), pp. 926.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Swoboda, Alexander K., The Dual Exchange-Rate System and Monetary Independence,” in National Monetary Policies and the International Financial System, ed. by Robert Z. Aliber (Chicago: University of Chicago Press, 1974) , pp. 25870.

    • Search Google Scholar
    • Export Citation
  • Turnovsky, S.J., Monetary Policy, Fiscal Policy and the Government Budget Constraint,” Australian Economic Papers (Adelaide), Vol. 14 (December 1975) , pp. 197215.

    • Crossref
    • Search Google Scholar
    • Export Citation
*

Mr. Nowak, economist in the Exchange and Trade Relations Department, is a graduate of the University of Manchester and also studied at the University of Western Ontario.

1

Examples are the United Kingdom in 1947–79, Belgium since 1951, France in 1971–74, Italy in 1973–74, and the Rand Monetary Area in 1974–83.

2

For example, Krueger (1978, p. 55) has stated that “…exchange rate changes under QRs [quantitative restrictions] cannot be analyzed in exactly the same manner as would a devaluation undertaken from an initial position of currency convertibility.” See also Bhagwati (1978).

3

Dickie and Noursi (1975) estimated a model of the parallel foreign exchange market in Syria, but, while acknowledging the existence of formal controls, did not explicitly take them into account. A simple flow model of the parallel market has been developed by Bhagwati (1967), but attention in that study is restricted to an examination of the effects that illegal transactions have on the officially recorded and actual balance of payments. Dornbusch and others (1983) have constructed a portfolio-balance–currency-substitution model of the black market for foreign exchange in Brazil, but the partial equilibrium approach they adopted precludes an examination of the economywide impact of policy shifts in the presence of quantitative controls.

4

This conclusion is applicable not only to unified exchange rate models but also to dual market models in which the official foreign exchange market clears through movements in central bank reserves. For example, Argy and Porter (1972) and Swoboda (1974) use a pure-flow model, and Cumby (1983) employs a portfolio-balance framework to examine the implications of segregating current and capital transactions into fixed and floating rate markets. The short-run insulation properties of this type of dual system are dealt with in taxonomic fashion by Marion (1981). See also Fleming (1974), Flood (1977), Blejer (1978), and Flood and Marion (1981). With regard to the effectiveness of monetary policy under conditions of high capital mobility, this class of model generates results that in essence correspond to those of the basic model for a flexible-unified exchange rate in the short run and to those of the fixed rate model in the long run. The long-run effect of a devaluation in all these models is a once-and-for-all increase in reserves and an equiproportionate rise in all prices. Leakage between the two markets as a result of illegal transactions weakens, but does not qualitatively alter, the basic properties of the dual market model; see Lanyi (1975) and Braga de Macedo (1982).

5

This description corresponds to the fixed-unified exchange rate case and to the official market segment of a dual rate regime without import controls.

6

This may be recognized as a case of the Marshallian equivalence proposition. Bhagwati (1978) has argued that, in practice, this equivalence will break down because the two prices (in this example, e1 and e¯1) cannot be adjusted parametrically in response to changes in the supply and demand schedules. The problem can, however, be overcome simply by auctioning import licenses. The price paid for these licenses would correspond to the shadow parallel exchange rate and would fluctuate in response to variations in the supply and demand for foreign exchange.

7

This subsection is in the spirit of a model by Michaely (1954), which presents a geometric exposition of the determination of black market prices. For an application of this model to black markets in foreign exchange, see Sheikh (1976). The presentation here is somewhat simpler than in these two papers because it avoids construction of an exceedingly complicated demand curve for the black market. It also allows the path of the black market exchange rate to be traced as it responds to changes in the official rate.

8

It is assumed throughout this subsection that the economy is on a given production possibility frontier, so that aggregate production, appropriately measured, is fixed. The assumptions underlying the curves drawn in Figure 2 are set out fully in the Appendix.

9

In this deterministic model, the (known) marginal cost of smuggling varies with distance from the frontier. For an elaboration of this relationship, see the Appendix.

10

The conclusions of this subsection, however, obviously are sensitive to the assumptions relating to the penalty costs of transacting in the parallel market. In this regard, the existence of smuggling costs is a crucial assumption underlying the negative slope of the unification path. In the absence of such costs, all producers would sell on the parallel market whenever the premium in this market was positive. Thus, the supply curve OA in Figure 2 would be flatter, passing through the point R. Without official reserve intervention, the unification path would, therefore, be a horizontal line passing through the points e¯2 and R, with the parallel rate equal at all times to the restriction-free equilibrium rate. Penalty costs on the demand side (that is, for buyers of foreign exchange in the parallel market) may introduce an element of indeterminacy into the analysis if such costs are sufficiently large. A rise in the parallel rate relative to the official rate may cause an increase or decrease in the quantity demanded on the parallel market. Hence, the demand curve for imports may shift upward or downward if demand-side penalty costs are present. The assumptions adopted above, which give the unambiguous result that an official devaluation results in an appreciation of the domestic currency in the parallel market, should, however, be regarded as empirically plausible.

11

The following is a national income accounting treatment of aggregate output. Measurement errors arise when relative prices in the current period differ from those in the base period. For small changes in relative prices over time, however, this error may be safely ignored.

12

All base-period prices (including the parallel exchange rate e) have been set equal to 1 (with the exception of the official exchange rate e, which must be less than e); e therefore represents not only the price obtained by selling the export good on the parallel market but also the price of the imported good in local currency.

13

Points on IS0 to the northeast of Z satisfy equation (1), with the condition that the partial derivative X1 and the function X( ) are set equal to zero.

14

The greater-than-unitary elasticity of IS below point Z reflects the fact that equal proportionate changes in pn and in e, but not in e, will result in a reallocation of resources from production of the export good to the production of the nontraded good.

15

Static expectations mean that the value expected in the next period is equal to the current (known) value; that is, e+1* = e. The expected nominal return on holding foreign currency is then always zero.

16

Because the analysis throughout assumes that there are no interest-bearing assets, increases in domestic credit that are not associated with the financing of government purchases of commodities arise out of “helicopter” operations or, equivalently, cash grants from the government to the private sector.

17

An inspection of equations (1) and (2) reveals that point B is both a short-run and a long-run equilibrium position. This result is independent of the assumptions made about wealth effects.

18

The indeterminate effect on the parallel rate in the impact period reflects the adoption of an end-of-period, asset-equilibrium specification. In a beginning-of-period formulation, asset supplies (including the stock of money) are held stationary in the impact period. Hence, an increase in government spending initially involves only a leftward shift of the IS curve and leads unambiguously to an appreciation of the parallel rate. As in the end-of-period formulation, however, money financing in subsequent periods results in a rightward shift of the LM curve and produces a progressive depreciation of the parallel rate. A discussion of the issues involved in the alternative specifications of asset market equilibrium is contained in Kami (1979). For an example of an end-of-period model that incorporates a government budget constraint, see Turnovsky (1975).

19

The analytical techniques used in this subsection are discussed in detail in Sargent and Wallace (1973).

20

The decline in e+1*/e before t + θ is illustrated in Figure 5 by the concave (in relation to the origin) shape of the time path for e. This concavity is necessary for stability because the demand for money must rise, and therefore e+1*/e must fall over time, if e is to decline to its long-run equilibrium level.

21

Net sales of foreign exchange by the central bank produce a leftward shift in the LM curve and are thus identical in their effect to a contraction in domestic credit.

22

The real wage level will tend to rise with a devaluation because, as discussed above, the price level will decline with the increase in output.

IMF Staff papers: Volume 31 No. 2
Author: International Monetary Fund. Research Dept.