There is an extensive literature focusing on the conduct of exchange rate policy in developing countries. Much of this literature, particularly in previous years, assumed that developing countries had the same unified exchange markets and unitary exchange rates that prevailed in most industrial countries. More recently, analysts have begun to acknowledge the highly segmented nature of exchange markets in many developing countries. In some of these countries, the monetary authority may peg the exchange rate for commercial transactions but allow a second exchange rate for international financial transactions to float freely. In other countries, exchange controls intended to ration scarce international reserves have given rise to black markets for foreign exchange.
Regardless of whether multiple exchange rate systems are legal or illegal, they have important implications for economic performance and policy. At the macroeconomic level, multiple exchange markets may complicate the process of stabilization, altering the effects of exchange rate and monetary policies relative to what would be expected with a unified exchange market. At the microeconomic level, multiple exchange markets may alter patterns of resource allocation and encourage rent-seeking, smuggling, and other activities for which no incentives would exist in a deregulated market environment.
The discussion that follows is divided into two parts. To begin with, I define multiple exchange rate systems more closely and discuss, at a general level, various issues regarding the operation and consequences of multiple exchange rate systems. These issues include, first, the distinction between legal and illegal multiple rate systems; second, the rationales for these systems; third, the reasons why governments follow policies that give rise to multiple exchange markets; fourth, the economic roles multiple rate systems play in different countries and the consequences of those roles for economic welfare; and finally, the unification of multiple rate systems. Many of the conclusions described in this section are based on a recent World Bank study of multiple rate systems in various countries (Argentina, Ghana, Mexico, Sudan, Tanzania, Turkey, Venezuela and Zambia).1
As it turns out, the exchange rate systems of these countries fit loosely into two categories with different characteristics. The first comprises the Latin American countries, where exchange controls were prompted by macroeconomic crises and where the primary effects took place at the macroeconomic level. The second group comprises Turkey and the African countries, where exchange controls were a response to cumulative deteriorations in competitiveness and where these controls led to profound distortions in resource allocation at the microeconomic level. Today, it is especially interesting to consider how the transition economies of Eastern Europe and the former Soviet Union might fit into this framework.
In the second part of my discussion, I summarize my case study of Argentina, which was written as part of the broader World Bank study described above.2 Argentina’s experience with exchange controls and multiple exchange markets in the 1980s is particularly interesting, for a number of reasons. First, this period was an extremely turbulent one in Argentina’s macroeconomic history, culminating in inflation rates that exceeded 100 percent per month. Second, exchange controls and exchange rate policy represented an important part of the Government’s efforts to control inflation and stabilize the economy. Finally, the widespread evasion of exchange controls that took place in Argentina in the 1980s typifies the types of problems governments face in trying to insulate their economies from market forces.
I. Multiple Exchange Rate Regimes
What exactly is meant by multiple exchange rate practices? In principle, a multiple exchange rate regime is any exchange rate regime that applies two or more exchange rates to the same currency. Many developing countries have applied separate, fixed exchange rates to different types of transactions, but this practice is, in essence, equivalent to a single exchange rate coupled with different taxes or subsidies (depending on the transaction). In the following discussion, a multiple exchange rate system will refer to a system involving one or more fixed exchange rates for current account transactions and one or more floating or market-driven rates for capital transactions. This type of system presents policymakers with numerous difficulties, because the separate exchange rates involved may react in very different ways to the same shocks or policy actions. As an example, an increase in the money supply that drives up domestic prices is likely to lead to the real appreciation of a fixed nominal exchange rate but may cause a real depreciation of any floating, market-driven rate. Models and theories devised to deal only with single, fixed exchange rates may break down in the context of multiple rates of this kind.
Legal and illegal systems
The first distinguishing characteristic of exchange rate regimes is their legality (or lack of it). Legal multiple rate systems are often referred to as dual exchange rate systems, and the Belgian example, which persisted until very recently, was perhaps a prototype of this. Most current account transactions took place at a pegged official rate, while most capital account transactions occurred in the free market, which determined their rates.
Such systems are no different in economic principle from many “illegal” systems, particularly in developing countries where official and black or “parallel” markets exist side by side. In illegal systems, transactions are often conducted much as they are in “legal” systems: trade takes place at the official fixed rate, and capital flows occur at the illegal floating rate. Legal and illegal systems are also similar in their effects on the rest of the economy. For this reason, the World Bank study addresses them jointly.
Trade barriers, quantitative restrictions, or high tariffs alone are not in themselves sufficient to generate a black market for foreign exchange, however. For a parallel exchange market to develop, there must be some restriction on transactions in foreign exchange at the official price, as well as an excess demand for or supply of official foreign exchange at that price. Only under these conditions will people go to a black market to buy or sell foreign currency.
Rationales for multiple rate systems
There is only one legitimate rationale for having a multiple rate system that routes current account transactions through a pegged rate and capital account transactions through a floating rate: to insulate domestic prices and activity from exchange rate fluctuations deriving from transitory shocks in the financial market. Such a policy, if it is to succeed, requires properly managing the official exchange rate to ensure that it remains competitive. That done, the market-driven rate for capital transactions should fluctuate around the official pegged rate, which presumably would represent the mean of those fluctuations. In fact, in developing countries, the parallel market rate is rarely more appreciated than the official rate—although there are certain exceptions. And the fact that the parallel market rate is usually more depreciated than the official rate points to one of the key dangers of adopting a multiple rate system: there is a great temptation not to adjust the official exchange rate when it needs to be adjusted, allowing it to become increasingly overvalued and the economy decreasingly competitive.
A second and related problem with multiple exchange rate systems-especially if the rate becomes overvalued—is that it is very difficult to prevent leakages of transactions from one market into the other. A typical example, which will be discussed in the context of Argentina, involves exporters who are required to surrender their foreign exchange receipts to the central bank at the official rate, but who instead hide some of those receipts and sell them on the black market at a more depreciated rate. This common practice can lead to the breakdown of a multiple rate system.
Finally (and this is a somewhat more speculative but, I think, a legitimate criticism of multiple rate systems), to the extent that they create incentives for cheating, multiple exchange rate systems can blur the line between legal and illegal activities. Illegal behavior tends to spill into other areas of economic regulation, particularly tax enforcement. When exporters hide their receipts from the government in order to sell them on the black market, they probably are also hiding their income from the tax authorities. In Argentina, there was a simultaneous breakdown of the multiple exchange rate and tax collection systems. Such occurrences do tend to be interwoven, and they underscore the principle of good governance that would avoid having laws on the books that cannot be enforced.
Aside from the one legitimate rationale for parallel systems—to allow a fluctuating capital account rate to buffer the economy from the effects of transitory shocks while maintaining a correctly valued official rate—most other purported rationales for multiple rate systems depend on the systematic overvaluation of the official rate. For this reason, it is difficult to make a case for these rationales based on considerations of efficiency. One good example is the argument that some governments—especially those with net foreign currency obligations to the rest of the world—can gain implicit tax revenues through a multiple rate system. Governments can set the official rate at an overvalued level, force exporters to turn over their receipts at the overvalued rate, and then use this cheap currency to pay off foreign currency obligations. It is true that this scheme confers a fiscal benefit on the government, but it is a highly distortionary benefit that hurts exporters while helping those importers who may have preferred access to official foreign exchange. Undoubtedly, such a benefit is inferior in terms of welfare to a more broad-based system of taxation.
Why governments really choose multiple rates
Regardless of the purported rationales behind multiple exchange rate systems, historically the choice of such regimes has tended to be motivated by more direct, practical considerations. Usually, multiple rate systems are the result of severe balance of payments difficulties. In order to protect their international reserves and avoid a substantial devaluation, governments may choose to limit the range of transactions for which they will provide foreign exchange. The World Bank study found that this process tends to differ somewhat among countries. In Latin America, the initiating problems tended to be severe balance of payments crises associated with large capital outflows.
Chart 1 indicates the path of the real parallel market exchange rate (RPER, in local currency per dollar) and the parallel market premium (the percent difference between the parallel and official exchange rates) in three Latin American countries after exchange controls were initiated. In all three countries, parallel exchange rates depreciated sharply following the imposition of exchange controls and then appreciated as the situation stabilized. Hence, at the onset of balance of payments crises, the parallel market rates tended to overshoot the level at which they would eventually stabilize, more so in Argentina and Mexico than in Venezuela.
Parallel Exchange Rate (PER)
Sources: International Monetary Fund, International Financial Statistics (various years); and World Currency Yearbook.To the extent that such overshooting is typical, it creates a rationale, at least at the onset of a severe balance of payments crisis, to impose a dual exchange market. As long as the initial steep depreciation of a market-based exchange rate is reversed soon afterwards, it makes sense to insulate domestic prices from those fluctuations by maintaining a more stable pegged exchange rate for commercial transactions. Ideally, this pegged rate would be adjusted, so that once the payments crisis was past, the rate would be close to the market rate for financial transactions. Then the exchange controls could be removed and the exchange market unified.
However, after imposing exchange controls, the authorities do not always remove them when the crisis subsides. This was the case in Argentina, Mexico, and Venezuela. Dual markets developed in each country following the establishment of exchange controls, and the authorities continued to rely on the controls to prop up somewhat overvalued exchange rates, neglecting more fundamental fiscal and monetary adjustments. Thus, in principle there may be a role for dual exchange markets, but in practice they often are not used appropriately.
In the case of Turkey and various African countries, which were also covered in the World Bank study, the process of tightening controls in the face of balance of payments deterioration was somewhat more gradual. It was also more closely linked to the deterioration in the current accounts that resulted from the overvaluation of official exchange rates than to speculative surges in capital outflows. In the absence of speculative crises and the overshooting associated with them, these countries probably had fewer justifications for multiple rate practices than Latin American countries.
The economic roles of multiple rate systems
There were also differences in the economic functions of the parallel exchange markets in Latin American and African countries. In Latin America, most current account transactions took place at the official rates, and sufficient foreign exchange usually was available to meet the needs of importers at that rate, since demands for foreign exchange for trade were not excessive. The primary role of the parallel market in Latin American countries was to finance capital inflows and outflows. Sometimes the premium of the parallel market rate relative to the official rate grew very large in Latin America, but this development usually reflected a macroeconomic crisis that gave rise to capital outflows and not a severe misalignment of the exchange rate itself.
In consequence, the multiple rate system probably did not have severe distortionary impacts on the real side of the economy in Latin America. The real problem for Argentina and other Latin American countries lay in the difficulties the multiple rate system caused for macroeconomic management. In the African countries, by contrast, exchange rates became much more severely overvalued and foreign exchange was more stringently rationed in the official market. Thus, the parallel markets played a more important role in supplying foreign exchange for importation, with very distortionary consequences. Importers with preferred access to foreign exchange enjoyed huge subsidies, whereas both importers lacking this access and exporters paid very high implicit taxes. For example, in the early 1980s much of Ghana’s economy moved underground when the official exchange rate become so severely overvalued as to be irrelevant to most prices and transactions in the private sector.
Unification of multiple rate systems
A substantial amount of the literature on exchange rate policy in developing countries is devoted to the desirability, timing, and manner of implementing exchange market unification—that is, the combining of multiple rates into a single, unified rate. To a large extent, questions of the optimality of exchange market unification and the best means of achieving unification are somewhat academic.
In Latin America, multiple exchange rate systems were often abandoned not because they were no longer needed or because they were viewed as undesirable, but because they were no longer sustainable in a macroeconomic crisis. Moreover, the most important factor in determining whether unifying exchange rates will give rise to a favorable outcome is not the method used to implement the unification, but the associated macroeconomic policies. In the late 1980s, Mexico combined unification with fiscal restraints and structural reforms, and the country’s stabilization program took hold relatively quickly. By contrast, Argentina, which unified its exchange market in December of 1989, experienced another round of hyperinflation (with inflation rates of nearly 100 percent per month) before it was able to stabilize. In essence, macroeconomic policies are far more important than exchange rate policies in ensuring stable prices, output, and financial systems.
II. The Case of Argentina
My paper on Argentina addresses the following questions: first, what motivated Argentine policymakers to impose the exchange controls that gave rise to a parallel market in 1981; second, how effective these controls were in achieving their intermediate objective, which was to allow the authorities to set the official exchange rate with some degree of independence from financial market developments; and third, how effective the exchange controls were in helping the authorities achieve their ultimate policy objective of stability combined with growth.
In talking about the evolution of Argentina’s exchange control regime, I cover approximately 12 years of Argentina’s history, from 1978 through 1990. This period can be divided into four phases (Chart 2). During the first of these, from the late 1970s through the early 1980s, no exchange controls existed. The policies of this initial phase set the stage for the balance of payments crisis and the experience with exchange controls that were to follow. Exchange controls were imposed in 1981, and the short time characterized by a spike in the parallel market rates—the second phase—represented a transition for Argentina. Once certain problems were resolved during this transitional phase, Argentina had a long period—the third phase—during which economic policy was geared toward fighting inflation. At the end of this phase, exchange rate pressures became so great that the Government had to remove the exchange controls and unify and float the exchange rate. The subsequent phase, which is shown only up to the first part of 1990, is the post-exchange control phase.
Macroeconomic Developments in Argentina
* Inflation of 115% in June 1989 and 197% in July 1989.Sources: Carlos A. Rodriguez, The Macroeconomics of the Public Sector Deficit: The Case of Argentina (World Bank PPR Working Paper Series 632, March 1991); Foundation for Latin American Research (FIEL) data base; and IMF staff calculations.The no-exchange control phase
The policies of the first no-exchange control phase ultimately gave rise to exchange controls and parallel markets. The depreciation of the official exchange rate (which was unified at that time) at a rate lower than the rate of inflation in order to bring inflation down was particularly significant in this regard. Chart 2 shows that this policy had some success, although inflation remained quite high by international standards. But because the exchange rate was being depreciated at a rate lower than the inflation rate, the real exchange rate appreciated steadily. As a result, the current account deficit grew substantially toward the end of this period. A severe recession developed in the tradable goods sector, and both private and public indebtedness ballooned, in large part because the Government was borrowing money from abroad to finance the deficit. Toward the end of this phase, investors began to realize that a devaluation was inevitable and began sending their capital overseas, prompting the Government to borrow further to support the exchange rate. These actions forced a series of devaluations that started in early 1981.
The transitional phase
By mid-1981, the Government faced four crucial and seemingly insoluble problems. First, there were severe balance of payments outflows in response to perceptions of macroeconomic breakdown; second, the private sector was deeply indebted; third, the economy had slid into recession; and, finally, inflation, prompted in part by the recent devaluations, had accelerated. The problem for the Government was that policies intended to address problems in one area tended to exacerbate problems in another area. For example, a steep devaluation would help curtail balance of payments outflows, but it would also raise inflation, deepen the contraction in the real sector, and hurt those with a high degree of external indebtedness. Conversely, monetary contraction would help the balance of payments and reduce inflation but create severe ramifications for both debtors and economic activity.
In order to reconcile these competing problems, the Government adopted the following strategy. First, it instituted a dual exchange market with a fixed commercial exchange rate and a floating rate for financial transactions—which later became, in essence, an illegal parallel market rate, also for financial transactions. Second, it depreciated the real official exchange rate substantially. Third, in order to curtail pressure on the real black market rate, which was depreciating rapidly, it imposed controls on capital outflows and effectively nationalized private external debt. To help private debtors still further, interest rate ceilings were imposed on deposits and loans; combined with inflation rates exceeding those ceilings, this tactic had the effect of inflating away most of the real value of the debt in the economy in 1982.
The results of these policies were a huge run-up in the real value of the parallel market rate, an acceleration of inflation, and an intensification of the Government’s foreign debt problems. On the other hand, what at that time was considered to be the key problem—the indebtedness of the private sector—was largely eliminated. In consequence, by the end of the transitional phase, various of the imbalances of the previous years had been corrected. The real official exchange rate was now at a relatively depreciated level, the balance of payments crisis had eased, and private debt had been largely eliminated.
Fighting inflation
The key priority for Argentina’s Government during the third phase was reducing inflation. To this end, a series of anti-inflation programs was launched that included wage and price controls, a fixed exchange rate, and some degree of fiscal adjustment. Initially, these programs succeeded in reducing inflation. During the first, the 1985 Austral Plan, inflation dropped radically. During the second, the “Plan Primavera,” inflation also fell significantly. Then, after the hyperinflation at the end of the Alfonsin period, the so-called “Plan Bunge y Born” reduced inflation again. But these successes were always temporary, because there was no sustained fiscal adjustment.
Absent a sufficient reduction of the fiscal deficit, monetary growth continued, causing not only further deterioration of the trade account but capital outflows that drove up the parallel exchange rate. In turn, foreign exchange began leaking from the official to the parallel market, forcing the Government to devalue the peso and abandon the stabilization program. Without the correct macroeconomic policies, the fixed exchange rate programs could not function, even when bolstered by exchange controls. By December 1989, evasion of exchange controls had become so rampant that the Government, faced with a ballooning public debt, rising inflation, and ongoing capital outflows, realized it could no longer control the prevailing exchange rate, and so floated the currency. That action ended Argentina’s most recent experiment with exchange controls.
The post-exchange control phase
After the exchange controls were removed and the exchange rate floated, Argentina had another bout with hyperinflation in March of 1990. This experience demonstrates that unifying an exchange market is not, by itself, a magical cure for a country’s macroeconomic problems; unification must be bolstered with the right fiscal and monetary programs. That said, Argentina did gain certain benefits from unifying the market and floating the fixed exchange rate. First, by abandoning exchange controls and the fixed exchange rate, the Government was better able to focus on its primary consideration, reducing the fiscal deficit. Second, floating the exchange rate in essence disciplined Argentine policymakers, who understood in 1990 that the smallest degree of monetary emission would result in a steep plunge in the value of the currency and another bout of hyperinflation.
The effectiveness of exchange controls
Gauging the effectiveness of exchange controls requires determining whether they allow the Government to segment the official and parallel exchange markets, thus permitting the official rate to be set independently of the financial rate. I will present data showing that Argentine exchange controls were not effective in this sense. At times, changes in the parallel market rate forced the Government to adjust its official rate, because whenever the parallel rate depreciated too much relative to the official rate, strong incentives were created for exporters to hide their export receipts and sell them on the black market instead.
Table 1 indicates the results of a simple econometric regression designed to explain movements in the level of Argentina’s exports, based on the movements of certain other variables: the real exchange rate, the parallel market premium, Argentina’s output, and output in the rest of the world. As expected, the coefficient on the real exchange rate is positive, indicating that a depreciation (increase) in the real exchange rate generally led to increases in Argentine exports. And as hypothesized, the coefficient on the parallel market premium is negative, indicating that increases in the gap between the parallel market exchange rate and the official rate tended to reduce the level of exports in the economy. The coefficient is fairly large at.27, indicating that during the periods when the gap was largest, Argentine exports may have been depressed by as much as 10-40 percent.
Merchandise Export Equation Quarterly Data: 1978/IV-1988/II
Dependent Variable: Log (Real exports)
Merchandise Export Equation Quarterly Data: 1978/IV-1988/II
Dependent Variable: Log (Real exports)
| (OLS) | (IV) | |
|---|---|---|
| Constant | -6.06 | -1.10 |
| (-1.16) | (-.17) | |
| Log (real exchange rate) | .28 | .23 |
| (2.60) | (1.74) | |
| Parallel premium | -.27 | -.39 |
| (-1.99) | (-2.06) | |
| Log (real GDP) | -.38 | -1.24 |
| (-.65) | (-1.46) | |
| Log (real foreign GDP) | .57 | .78 |
| (1.67) | (1.97) | |
| DUMQ2 | .27 | .29 |
| (5.68) | (5.66) | |
| DUMQ3 | .20 | .22 |
| (3.85) | (3.84) | |
| DUMQ4 | -.08 | -.02 |
| (-1.30) | (-.25) | |
| Standard error of regression | .10 | .10 |
|
.74 | .72 |
| Durbin-Watson statistic | 1.91 | 1.88 |
| (t-statisties in parentheses) |
Merchandise Export Equation Quarterly Data: 1978/IV-1988/II
Dependent Variable: Log (Real exports)
| (OLS) | (IV) | |
|---|---|---|
| Constant | -6.06 | -1.10 |
| (-1.16) | (-.17) | |
| Log (real exchange rate) | .28 | .23 |
| (2.60) | (1.74) | |
| Parallel premium | -.27 | -.39 |
| (-1.99) | (-2.06) | |
| Log (real GDP) | -.38 | -1.24 |
| (-.65) | (-1.46) | |
| Log (real foreign GDP) | .57 | .78 |
| (1.67) | (1.97) | |
| DUMQ2 | .27 | .29 |
| (5.68) | (5.66) | |
| DUMQ3 | .20 | .22 |
| (3.85) | (3.84) | |
| DUMQ4 | -.08 | -.02 |
| (-1.30) | (-.25) | |
| Standard error of regression | .10 | .10 |
|
.74 | .72 |
| Durbin-Watson statistic | 1.91 | 1.88 |
| (t-statisties in parentheses) |
It may also be true that when the parallel exchange rate rises relative to the official rate, officially reported imports also rise. Increases in the parallel premium provide an incentive for importers to report as many imports as possible, acquire the corresponding sum of foreign currency at the official exchange rate, and then sell the money on the parallel market. To test this hypothesis, another equation was estimated that relates the level of imports to various explanatory variables, including the real exchange rate and the parallel premium (Table 2). As expected, the coefficient on the real exchange rate is negative, indicating that real depreciations tended to depress imports. However, the parallel market premium apparently did not affect the level of imports; other factors appear to have been more important in determining imports than the value of the premium, as other researchers have found in other countries. Importers may have been more interested in evading tariffs than in boosting their reported imports in order to obtain foreign exchange.
Merchandise Import Equation Quarterly Data: 1978/IV-1988/II
Dependent Variable: Log (Real imports)
Merchandise Import Equation Quarterly Data: 1978/IV-1988/II
Dependent Variable: Log (Real imports)
| (OLS) | (IV) | |
|---|---|---|
| Constant | -4.28 | -7.73 |
| (-.87) | (-1.14) | |
| Log (Real exchange rate) | -.48 | -.40 |
| (-3.40) | (-2.35) | |
| Parallel premium | .04 | .01 |
| (.31) | (.09) | |
| Log (real GDP) | 1.07 | 1.36 |
| (2.33) | (2.16) | |
| Tarif rate | .15 | .02 |
| (.36) | (.05) | |
| Index of nontariff barriers | -.13 | -.12 |
| (-1.27) | (-1.26) | |
| Log (lagged real imports) | .42 | .46 |
| (3.96) | (3.93) | |
| DUMQ2 | .13 | .12 |
| (2.90) | (2.74) | |
| DUMQ3 | .19 | .19 |
| (4.19) | (4.02) | |
| DUMQ4 | -.10 | .08 |
| (1.90) | (1.34) | |
| Standard error of regression | .09 | .09 |
|
.92 | .92 |
| Durbin-Watson statistic | 2.12 | 2.20 |
| (t-statistics in parentheses) |
Merchandise Import Equation Quarterly Data: 1978/IV-1988/II
Dependent Variable: Log (Real imports)
| (OLS) | (IV) | |
|---|---|---|
| Constant | -4.28 | -7.73 |
| (-.87) | (-1.14) | |
| Log (Real exchange rate) | -.48 | -.40 |
| (-3.40) | (-2.35) | |
| Parallel premium | .04 | .01 |
| (.31) | (.09) | |
| Log (real GDP) | 1.07 | 1.36 |
| (2.33) | (2.16) | |
| Tarif rate | .15 | .02 |
| (.36) | (.05) | |
| Index of nontariff barriers | -.13 | -.12 |
| (-1.27) | (-1.26) | |
| Log (lagged real imports) | .42 | .46 |
| (3.96) | (3.93) | |
| DUMQ2 | .13 | .12 |
| (2.90) | (2.74) | |
| DUMQ3 | .19 | .19 |
| (4.19) | (4.02) | |
| DUMQ4 | -.10 | .08 |
| (1.90) | (1.34) | |
| Standard error of regression | .09 | .09 |
|
.92 | .92 |
| Durbin-Watson statistic | 2.12 | 2.20 |
| (t-statistics in parentheses) |
Whenever the gap between the parallel and official rates became too large in Argentina, the balance of payments deteriorated, forcing the Government to depreciate the official exchange rate. Therefore, the official rate was never fully independent of the parallel rate. A third regression (Table 3) estimates the Argentine Government’s exchange rate “reaction function,” relating the month-by-month rate of depreciation of the official exchange rate to the level of the real official exchange rate, the Argentine-U.S. inflation differential, and the parallel premium. As expected, a higher (more depreciated) real exchange rate tended to be associated with a lower rate of depreciation. Also as expected, when inflation in Argentina was higher than inflation in the United States, the Government had to depreciate its official exchange rate more rapidly. Finally, all else being equal, increases in the parallel market premium tended to induce the Government to increase the rate of depreciation of the official exchange rate. Basically, then, the Government could not pursue official exchange rate policies that differed markedly from the dictates of the market.
Commercial Exchange Rate Determination May 1982-July 1988
Dependent Variable: Official Depreciation
Commercial Exchange Rate Determination May 1982-July 1988
Dependent Variable: Official Depreciation
| Monthly | Quarterly | |
|---|---|---|
| Constant | 1.89 | 6.87 |
| (3.04) | (3.01) | |
| Log (Real exchange rate) | -.21 | -.76 |
| (3.05) | (3.02) | |
| Inflation differential | .89 | .85 |
| (7.19) | (5.29) | |
| Parallel premium | .12 | .58 |
| (2.36) | (3.27) | |
| Standard error of regression | .06 | .12 |
|
.50 | .67 |
| Durbin-Watson statistic | 2.20 | 1.79 |
| (t-statistics in parentheses) |
Commercial Exchange Rate Determination May 1982-July 1988
Dependent Variable: Official Depreciation
| Monthly | Quarterly | |
|---|---|---|
| Constant | 1.89 | 6.87 |
| (3.04) | (3.01) | |
| Log (Real exchange rate) | -.21 | -.76 |
| (3.05) | (3.02) | |
| Inflation differential | .89 | .85 |
| (7.19) | (5.29) | |
| Parallel premium | .12 | .58 |
| (2.36) | (3.27) | |
| Standard error of regression | .06 | .12 |
|
.50 | .67 |
| Durbin-Watson statistic | 2.20 | 1.79 |
| (t-statistics in parentheses) |
Of course, the pursuit of official exchange rate policies insulated from market forces was merely an intermediate goal for Argentine policymakers. Ultimately, what governments care about—and Argentina was no exception—is not the exchange rate, but economic stability and growth. And so I arrive at my final question: did the parallel exchange rate system help Argentina achieve its ultimate objectives of reduced inflation and stable economic growth?
To some degree, this is a purely academic question, since the Argentine economy exploded into extreme hyperinflation by the end of the exchange control period. But it is possible to ask whether exchange controls had a positive, marginal impact on the economic situation. The answer is essentially the same one I gave earlier in reference to the Latin American countries in general—that during the transitional period of the early 1980s, there may have been a case for having a dual exchange market system. Had the authorities floated the exchange rate at that time, it probably would have followed a path similar to the one the parallel market rate pursued over that period. Inflation would have accelerated even more than it did, and real incomes and real activity probably would have contracted more than they did, as well. Such a development would have been unfortunate, because the parallel rate appreciated a fair amount subsequent to the crisis, meaning that the economy would have experienced an unnecessary degree of turbulence and volatility.
Ideally, the authorities should have implemented a dual exchange market system—while simultaneously following proper macroeconomic policies—in order to stabilize the economy by the beginning of 1983. Then they could have abolished the dual exchange rate system and unified the exchange rate markets. Combined with appropriate fiscal adjustment and monetary restraints, these actions might have resulted in much better economic performance in the 1980s than in fact took place. Instead, the authorities retained the exchange controls, relying on them as a poor substitute for proper monetary and fiscal policies. The resulting sequence of cumulative policy mistakes erupted in the hyperinflation of 1989, when the controls had to be abandoned. In retrospect, therefore, the possibility exists that early on in Argentina’s experience, the country could have benefited from exchange controls, but these controls certainly should not have been sustained.
Comments
Mohamed A. El-Erian
My objective in commenting on Steven Kamin’s discussion is twofold: first, to explore in a selective manner his approach and findings; and second, to complement his analysis with a brief overview of the experience with multiple exchange rate (MER) regimes in Middle Eastern countries. These comments are intended to provide a more comprehensive basis for discussing related issues affecting developing countries.
Mr. Kamin provides a useful discussion of Argentina’s exchange rate policy in the 1980s, in the context of an analysis of the key factors affecting the supply of and demand for foreign exchange. After noting the main determinants of the parallel exchange rate and the channels for interaction with the “official” rate, he examines some of the policy implications for the real economy, inflation, and external capital flows.
Mr. Kamin’s analysis gives the MER policy approach a mixed report card. On the positive side, it notes that Argentina’s MER regime helped minimize the disruptive impact of financial market uncertainties in the early 1980s. On the negative side, the regime outlived its usefulness, but the associated unfavorable impact was reduced by market-related arbitrage operations.
The analysis raises several issues, especially when considered in light of the experience of other countries. Among these issues, four in particular stand out.
First, the analysis could have examined in more detail the welfare losses associated with the imposition of multiple exchange rates and controls beyond the initial period, as well as how these could have been mitigated by a more liberal exchange regime. Among the interesting points to be examined are the impact on consumption and tradeable production and the implications for segmentation of domestic financial markets.
Second, to what extent did the parallel market premium provide an adequate benchmark for determining the commercial exchange rate? Specifically, did distorting interactions between the two markets contribute to an effective exchange rate that differed substantially from the underlying equilibrium rate?
Third, to what extent were developments affected by exogenous factors beyond those analyzed in the paper (e.g., changes in U.S. dollar LIBOR rates)? Moreover, did changes in the international debt strategy—and associated prospects for improving relations with commercial creditors and limiting adverse externality effects on trade-related financing—play an important role in the interaction between the parallel and official markets?
Finally, Mr. Kamin treats the period as uniform in terms of the structure of the exchange regime. Yet the period was characterized by numerous changes, including some in the legal status of the parallel market. Could these “structural changes” have affected the estimated parameters of the equations?
The discussion’s overall framework is particularly pertinent to my second objective—to provide a brief overview of experience with MERs in Middle Eastern countries. Such an overview complements Mr. Kamin’s analysis by introducing countries that differ from Argentina in one important respect—the extent of the integration of their financial markets with those in industrial countries.
Multiple exchange rate regimes, which basically take a cost-management approach to foreign exchange, have been implemented in several Middle Eastern countries in the last 20 years, including Algeria, Egypt, Iran, Morocco, Sudan, Syria, and Yemen. In other countries in the region, with the exception of the Gulf Cooperation Council (GCC) countries, the authorities have at times resorted to direct quantitative controls without MERs. The fact that GCC countries are excepted is related in large part to the fact that the governments of those countries have direct access to the bulk of foreign exchange earnings from oil exports.
Within the countries that implemented MERs, it is possible to distinguish two broad approaches. Algeria and Morocco adopted relatively simple regimes, basically setting a premium for specific transactions. In Egypt, Iran, Sudan, and Syria, MERs tended to evolve into complex and fragmented systems of foreign exchange pricing. As a result, these systems were characterized by separate exchange rates for some capital, invisibles, and merchandise trade transactions; in some cases the same transaction was subject to different rates, depending on whether it involved the public or private sector. These MERs were supported by a relatively complex system of controls over international transactions and foreign exchange surrender requirements.
In contrast to the Argentine experience, the Middle Eastern authorities’ main motivation in choosing MERs was not to isolate the real economy from what were perceived as disruptive capital flows but to provide incentives for certain current account transactions. Specifically, the objectives were to promote the export of certain goods and services (such as workers’ remittances and tourism receipts) and to curtail specific foreign exchange expenditures, with the aim of directly reducing the balance of payments deficit and influencing its composition.
The use of MERs also reflected policymakers’ concern that a comprehensive exchange rate adjustment would put upward pressure on the prices of certain “sensitive” imports. The MER was seen as providing relatively favorable exchange rates for “essential” items and more depreciated rates for others. In some countries, fiscal considerations also came into play, as policymakers were able to contain the budgetary cost of essential expenditures by providing foreign exchange at more favorable rates. Budgetary gains from the reduced expenditures were offset, however, either by central bank losses (i.e., no net quasi-fiscal gain) or by implicit taxes on private sector foreign exchange activities.
MERs tended to remain in place for a considerable period of time, though not without changes. In fact, policy attempts to tinker with the MERs—with a view to rationalizing the system—sometimes had an opposite effect. For example, during the transition to a market-based unified rate, some countries experimented with the idea of using new exchange rates as a means of implementing a partial effective devaluation. However, continued large financial imbalances often meant that a new rate became fixed and limited in scope, requiring a yet larger number of exchange rates. Similarly, attempts to influence the rate in the parallel market in order to limit the depreciation of the “official rate” were often frustrated by the emergence of another informal market, sometimes outside the country.
In recent years, Middle Eastern countries have made significant progress toward unifying their exchange rates. Previous tentative policy steps involving shifts of products among exchange markets and a resulting depreciation of the effective exchange rate have given way to bolder efforts. These stronger steps have been taken as part of comprehensive macroeconomic adjustment and structural reform policy efforts aimed at increasing the overall responsiveness of the region’s economies and reducing domestic and external financial imbalances.
Some general conclusions may be drawn from an analysis of Middle Eastern countries’ experiences with MER regimes.
Middle Eastern countries used MERs to influence certain types of current account transactions and reduce budgetary outlays. The underlying reason for implementing these regimes was the authorities’ inability to correct the domestic financial policy imbalances that were weakening the external position, among other things.
Two related factors in particular increased the attractiveness of MERs to policymakers: (i) the severity of foreign exchange shortages in the official sector; and (ii) the relative importance of sources of foreign exchange receipts the authorities could not easily control with quantitative restrictions.
As in Argentina, MERs did not prove to be a temporary policy response. In some countries, their prolonged use was accompanied by increased complexity in foreign exchange pricing, which contributed to growing economic and financial distortions, misallocation of production and consumption resources, and, in some cases, currency substitution. In fact, rather than helping to resolve the underlying problems, the MERs tended to introduce a host of additional policy complications. The distortionary effects made medium-term external viability and sustained economic growth more difficult to achieve. Exchange rate policies were sometimes overloaded with numerous and conflicting objectives, often without adequate support from other macroeconomic policies.
MERs were associated with an inward-based economic system dominated by the public sector. The public sector, which generally had access to the appreciated exchange rate for its imports, benefited from an important subsidy at the expense of exporters and private sector importers. Private sector activities were effectively relegated to a residual role in many areas. As a result, these economies proved less adaptable than others to changes in international economic and financial conditions.
The implementation of complex MERs involved considerable administrative costs, including those associated with surrender requirements, over-invoicing of imports and under-invoicing of exports, separating transactions, and processing license requests, among others. Nevertheless, there were indications of considerable leakages between markets in some countries. Effective separation of exchange markets proved difficult, as it involved maintaining close control over the underlying transactions. While such leakages may have in some cases improved the allocation of foreign exchange, the disorderly and fragmented process also involved considerable allocative inefficiencies and equity losses.
Recognition of the economic costs of MERs led to repeated attempts to simplify the exchange system, with a view to moving toward a unified market-related rate. Until recently, many of these attempts were frustrated by the inertia and large subsidies that had developed because of the MERs. Financial imbalances also frustrated the policy effort in two important ways. First, the larger the imbalances, the greater the spreads between the prevailing average exchange rate and its underlying market value, and the greater the challenge the authorities faced in taking bold policy actions that altered the associated implicit taxes and subsidies. Second, partial policy moves (particularly moving commodities between markets) had a very limited overall impact but involved costly sector-based disruptions. The recent success that some Middle Eastern countries have had in unifying rates has much to do with the up-front adoption of a realistic market-based rate and appropriate supporting domestic economic and financial policies.
Peter J. Quirk
Steven Kamin has gone over in some detail, and with much refinement, the case of a specific country that has maintained a multiple exchange rate system. What I intend to do, in a hopefully complementary way, is to go over some of the historical and background material relating to the IMF’s involvement in the area of multiple exchange rates and to try to give a broad view of trends and remaining practices.
The IMF is involved in this area because Article VIII of its Articles of Agreement specifically calls on members not to have multiple exchange rates. That being said, there are transitional arrangements under Article XIV that do in fact permit countries to maintain multiple exchange rates for a transitional period. In some cases the practices have been sustained for a very long period, but in recent years there has been something of a watershed in the use of multiple rates. Major differentials among multiple exchange rates have virtually disappeared from the IMF member countries, and we do not usually hear arguments these days in favor of sustaining multiple exchange rates.
Among industrial countries, Belgium in particular had a longstanding dual exchange rate system that was finally unified in 1990. Chart 1 suggests some of the reasons why this practice was finally abandoned. As earlier speakers have noted, a major argument for multiple rates in the literature was that a dual rate for capital could insulate the real economy from short-run financial shocks. However, the exchange rate differential between the two Belgian markets was very limited, and thus the degree of insulation from the overall swings in the exchange rate was not a major consideration in dismantling the dual exchange market. There was another important consideration, however: the system’s administrative cost. In Belgium, that cost was considerable, because the system was changed frequently to take into account various types of transactions that had become virtually uncontrollable and therefore had been reclassified from one dual market to the other.
Exchange Rates in Multiple Markets
Source: IMF staff.1 An increase in the index denotes an appreciation.A few similar systems existed in other industrial countries, but these did not involve a specific exchange rate. Until recent years, France had a practice called the devise titre, which was a self-contained market for foreign securities. At an earlier stage, the Netherlands also had a separate market—the O-guilder market—for foreign securities. The United Kingdom had an investment currency market that was abandoned in the late 1970s. But to a degree, these markets ran into the same problem: they could not be effectively enforced, and thus a substantive differential from the official market could not be sustained. At present no industrial country has a multiple exchange rate system.
In the group of developing countries, performance with respect to the obligations under Article VIII of the Fund’s Articles of Agreement was very much affected by the debt crisis of the 1980s. Many of the official unitary exchange rates in place at the outset of the debt crisis were unsustainable, but for political or social reasons the authorities did not change them completely. Instead, the authorities split the exchange rate in various ways, maintaining the old rate for certain transactions and introducing more depreciated exchange rates for others. Long-term trends reflect these practices. Some 38 percent of world trade took place under multiple rates in 1971—which, along with the 1950s, was a peak period in the use of multiple rate regimes. However, the proportion of world trade currently subject to multiple rates fell steadily throughout the 1980s, to well below 10 percent. Moreover, the types of multiple currency practices IMF members maintain now are generally characterized by very narrow spreads between rates that apply to relatively small groups of transactions.1
In addition to multiple exchange rates, other practices can give rise to effectively different exchange rates, including taxes, subsidies, and related arrangements that are applied to exchange transactions. An important example of such a practice involves the forward foreign exchange market. In a number of developing countries, credits are available, but local borrowers are unwilling to assume the exchange rate risk connected with servicing foreign borrowing. In these countries, the central banks have sometimes taken upon themselves the role of guarantor for exchange rate risk. However, this practice has proved extremely dangerous. A 1988 IMF survey of these practices emphasized the consequent very large fiscal losses.2 For instance, by 1983 the Philippines had accumulated losses equivalent to 6 percent of GDP, adding significantly to the fiscal deficit. Indonesia at one time had guaranteed swaps that caused major budget difficulties. In all cases where an indirect tax subsidy is embedded in either a forward exchange rate, some form of interest rate subsidy connected with an exchange transaction, or direct exchange taxes on subsidies on foreign exchange transactions, the IMF investigates the practice, which in effect constitutes a separate exchange rate.
In more basic terms, it is by now generally accepted that the exchange rate is a key price for the economy. On a microeconomic level, a multiple exchange rate is similar to the separate consumer prices that would be charged for an item if one person were given a 10 percent discount and another assessed a 20 percent premium. But multiple exchange rates differentiate all goods an economy produces and therefore clearly undermine not only the core of the price system but the signals emanating from that system. However, as I stated at the outset, these arguments are quite widely accepted, and only a very small number of countries continue actively to pursue multiple rate practices.
One rationale for such systems that has been encountered recently occurs in connection with the countries of the former Soviet Union. The argument holds that the very large depreciation of the ruble raises the question of whether foreigners investing in domestic assets are paying enough. In other words, the argument maintains that with the ruble depreciating rapidly (to about 400 per dollar), the economy’s real domestic assets are greatly undervalued. This argument may be true in some instances (e.g., the cost of meals for tourists), but the difficulty is that without investment in foreign technology to update the productive base, the transition economies will never be integrated fully into world markets. The productive apparatus is there, but investment clearly has to play a strong role before the quality and acceptability of the goods can improve. In some states of the former Soviet Union where the ruble trades at around 400 per dollar, rates for direct investment have been as low as 50 and even 20 per dollar. At such rates, the exchange rate functions as a huge disincentive for inward investment. On the other hand, nonproductive investments (such as real estate that can be acquired at low prices) or mobile factors of capital (such as plants and equipment that can be removed from the country) may need protection. In this respect, a number of countries maintain direct controls on the tapes of investment or disinvestment countries find acceptable, rather than on foreign currency itself. However, such impediments on investment must also be carefully weighed, because they can slow the development process considerably.
Finally, as I have noted, there has been remarkable progress in this area recently, and 45 years after the original Bretton Woods agreement, we are finally seeing one practice that goes against the spirit of the IMF’s Articles of Agreement becoming obsolete.
Summary of Discussion
Participants discussed the advantages and drawbacks of multiple exchange rates. The point was made that countries with balance of payments deficits were often forced to choose between adopting multiple rates or exchange controls. Most participants agreed that the effects of a multiple exchange rate system on an economy were similar to the effects of restrictions on current payments, as the latter were likely to lead to the development of a black market for restricted transactions—or, effectively, to a multiple rate system. However, a multiple rate system was thought to be preferable, for several reasons. First, its costs can be explicitly identified. Second, it can protect export profitability when separate rates exist for current and capital transactions. Once a country’s financial system is liberalized and the economy becomes increasingly attractive to foreign investors, capital inflows may increase, causing the currency to appreciate and discouraging exports. A separate rate for capital transactions, however, could allow the parallel market to accommodate a wider variance in the exchange rate, while keeping the official rate for current transactions fairly stable.
On the other hand, participants expressed the view that with a dual rate system, capital inflows could enter a country disguised as export receipts (or vice versa) to take advantage of the higher rate. In Chile, the variance of the legal parallel market rate for capital transactions was much higher than the variance of the official rate for current account transactions. The official rate was considered to be properly aligned, as evidenced by the satisfactory current account outturn over the last several years.
The general consensus was that there are circumstances in which a properly managed dual rate system could have advantages. An important drawback, however, is the added difficulty a dual rate system presents, since under this system the authorities would not be aware of the market signals needed to set the appropriate official rate. Moreover, experience with multiple rate systems has shown that these systems are often subject to abuse and that political factors have influenced rate setting.
Note: This paper represents the views of the author and should not be interpreted as reflecting those of the Board of Governors of the Federal Reserve System or other members of its staff.
Miguel A. Kiguel and Stephen A. O’Connell, “Parallel Exchange Rates in Developing Countries: Lessons from Case Studies,” Policy Research Working Paper 1265 (Washington, D.C.: World Bank, 1994).
Steven B. Kamin, “Argentina’s Experience with Parallel Exchange Markets: 1981-1990,” International Finance Discussion Papers No. 407 (Washington, D.C.: Board of Governors of the Federal Reserve System, 1991).
For a listing of multiple exchange regimes as of end-1992, see Exchange Arrangements and Exchange Restrictions, Annual Report 1993 (Washington, D.C.: International Monetary Fund, 1993). Since December 1990, there have been further unifications and narrowing of the differentials between dual markets. For example, Angola, Bulgaria, China, Costa Rica, Dominican Republic, Ghana, Guyana, Haiti, Honduras, Hungary, Mexico, Sudan, and Viet Nam have unified multiple exchange rate systems.
See Peter J. Quirk, Graham Hacche, Viktor Schoofs, and Lothar Weniger, Policies for Developing Forward Foreign Exchange Markets, IMF Occasional Paper 60 (Washington, D.C.: International Monetary Fund, 1988).
