6 Recent Experience with Floating Exchange Rates in Developing Countries
- Richard Bart, and Chorng-Huey Wong
- Published Date:
- June 1994
A major change in the exchange regimes of developing countries in the 1980s and early 1990s has been the adoption of floating exchange rates. Prior to the 1980s, it was widely believed that operating a competitive floating exchange rate regime required a level of institutional development these countries did not possess, although there had been a few isolated instances of floating among developing countries in the postwar period. Thirty have since adopted independently floating arrangements.
In 1987, the IMF examined its experience with these regimes through the mid-1980s and concluded that the early experience, although not definitive, indicated that floating exchange rate systems can be operated satisfactorily by developing countries with diverse economic structures, including those with a small number of commercial banks.1 Floating was not found to lead to free fall or greater exchange rate instability, which instead depend on the quality of domestic economic policies.
The IMF has continued to play a major role in the adaptation and application of market-determined exchange rate systems in developing countries. The purpose of the present paper is to provide a brief update of this experience since 1985, a period in which a number of developing countries have adopted floating exchange rates.2 Experience with floating exchange rates in Eastern Europe and the former Soviet Union (FSU) is covered in this paper only to a limited extent because these exchange rate regimes have been operating for a very short time.
In large part, the main issues with floating exchange rate regimes continue to be those relating to the long-standing debate between the merits of a floating exchange rate regime and the alternative, fixed or “anchor” regimes. To some extent this debate has been conducted along parallel tracks: supporters of floating regimes argue that such arrangements are inevitably implemented when, for example, international reserves are low or nonexistent, and supporters of fixed rates underline the optimality of fixed exchange rates because of their stability. Another issue that has emerged in some of the more successful “floaters,” with perhaps surprising force, has been how to handle intervention and sterilization policies so as to counter the expansionary effects of large capital inflows.
The paper is organized as follows: Section I examines the experience with adopting and operating market-determined exchange rates, including the reasons for floating, institutional arrangements in the exchange market, the role of the IMF, and accompanying measures. Section II reviews economic developments under floating exchange rate regimes, both in the exchange markets themselves and in terms of macroeconomic performance. The Annex sets out some technical considerations in setting up market arrangements for floating.
I. Introducing and Operating Floating Regimes
All developing countries that adopted floating exchange rates in the period 1985–92 did so in response to severe balance of payments difficulties. Most made the change as a prior action or performance criterion in the context of discussions for an IMF-supported economic program. The exceptions were Brazil, Haiti, and Peru (Table 1).
|Floating Arrangements Linked to Program|
|Country||Date of Program||Date of Adoption|
|Other||Prior Use of Official|
Multiple Exchange Rate
(Introduced as part of
|Afghanistan||—||December 1991||—||—||Yes (—)|
|Albanie||August 1992||June 1992||Yes||No||Yes (No)|
|Bolivia||June 1986||August 1985||Yes2||Yes||Yes (No)|
|Brazil||August 1988||March 1990||No||No||Yes (No)|
|Bulgaria||April 1992||February 1991||No||Yes||Yes (No)|
|Costa Rica||—||February 1992||—||—||No(—)|
|Dominican Republic||August 1991||January 1991||Yes||No||Yes (No)|
|EI Salvador||August 1990||June 1990||No||Yes||No|
|Gambia||September 1986||January 1986||Yes2||Yes||No|
|Ghana||August 1934||September 1986||Yes||No||Yes (Yes)|
|Guatemala||October 1988||November 1989||Yes||Yes||Yes (No)|
|Guyana||July 1990||February 1991||Yes||No||Yes (No)|
|Haiti||September 1989||June 1991||No||Yes||Yes (No)|
|Honduras||June 1992||February 1992||Yes||No||No|
|Lithuania||—||November 1992||—||—||Yes (—)|
|Mozambique||June 1990||April 1992||No||No||Yes (No)|
|Nigeria||December 1986||September 1986||Yes2||Yes||Yes (No)|
|Paraguay||January 1991||February 1989||No||No||Yes (No)|
|Peru||March 1954||August 1990||No||No||Yes (No)|
|Romania||May 1992||June 1992||No||Yes||Yes (Yes)|
|Russia||August 1992||June 1992||Yes||Yes||Yes (No)|
|Venezuela||June 1989||March 1939||No||No||Yes (No)|
The reference here is to a dual exchange market other than the illegal parallel market that was present in most cases at the time floating rates were instituted.
Program was tentative at time of introduction.
The reference here is to a dual exchange market other than the illegal parallel market that was present in most cases at the time floating rates were instituted.
Program was tentative at time of introduction.
Reasons for floating
These governments’ reasons for adopting independent floating exchange rates rather than other exchange arrangements—such as pegs to single currencies or currency baskets, or managed floating, whereby the exchange rate is moved administratively according to various economic indicators—have become clearer in recent years as debate on the issue has grown. As noted above, the arguments for floating regimes have tended to concentrate on the inevitability of such arrangements in a range of circumstances, particularly those that marked the debt crisis of the 1980s and its aftermath.
Lack of reserves has been the most conspicuous reason for allowing the exchange rate to float. Without sufficient reserves, a commitment to defend a fixed or crawling peg exchange rate is not credible and is quickly tested by the exchange markets. Because foreign exchange transactions now top US$1 trillion a day worldwide and information on even the remotest speculative or arbitrage possibilities flows almost instantaneously by satellite or wire, governments need large resources to hold a rate against market sentiment. Before they floated their exchange rates, Bolivia, El Salvador, The Gambia, Guatemala, Guyana, Haiti, and Venezuela had the equivalent of three months’ or less imports in gross official international reserves. Brazil, Ghana, Nigeria, and Paraguay had more, but they were limited in their ability to sell the reserves because of large external payments arrears and other liabilities.
Lack of information
Another major reason for adopting floating systems is a shortage of information that will aid in determining a sustainable equilibrium exchange rate under fixed or crawling regimes. Particularly in the context of extensive structural reforms and liberalization—such as those underway in Eastern Europe and the FSU—computation of a fixed or crawling rate is subject to considerable uncertainty. Errors that require resetting a fixed or crawling rate undermine the often tenuous credibility of government policies in the early stages of reform.
Continuing high rates of inflation in the early stages of reforms or in the absence of sufficiently strong programs necessitated independently floating arrangements in several countries in the 1985–92 period (Brazil, Peru, Romania, Russia, and Zaïre). In these circumstances, fixed or crawling exchange rates could not be adjusted quickly enough to keep up with prices and squeeze out significant arbitrage possibilities against the black market exchange rate. In order to avoid the flight of economic activity to the black market—with adverse consequences in the form of tax evasion, criminalization, and loss of economic control—the authorities have had little option but to allow the market to determine the exchange rate directly. Full currency board arrangements (such as those recently adopted in Estonia) can provide strong support for the credibility of macroeconomic policies, provided they do not prove overdeflationary—and therefore politically unsustainable—in their effects on output.
The choice of an exchange regime cannot be divorced from political considerations. In most instances, the authorities saw considerable merit in relinquishing political responsibility for adjusting the exchange rate and allowing it to be determined by market forces.
The main issue facing the authorities in setting up a floating exchange market has been the extent to which the system will be centralized. In essence, deciding this issue has involved two parameters of institutional arrangements: (1) auction versus private sector (“interbank”) markets, and (2) within the latter, whether or not to permit the functioning of nonbank foreign exchange dealers.
The 1987 IMF study concluded that the initial experience with auction foreign exchange markets had not been particularly satisfactory.3 Auctions have been subject to destabilizing intervention, ad hoc controls, and discontinuities in the supply of foreign exchange to the market. Since the mid-1980s, auction markets have been introduced only as supplements to extensive commercial bank markets or as limited secondary dual markets. In Nigeria, an auction market was used to distribute the foreign exchange proceeds of oil to banks and from there to the foreign exchange market, and in Russia the auction market has played a similar role for export proceeds of enterprises (although nonbank residents also participate in the auctions). The Nigerian auctions were marked by large spreads between the auction and market exchange rates, leading to sustained excess profits for participating banks. The Russian auctions, which were introduced only in April 1991, have not been marked by such problems. So far, the main lesson appears to be that auctions require a significant commitment on the part of the authorities to ensure transparency and competitiveness. However, the very choice of this centralized arrangement has in some instances been a signal that the commitment is not sufficiently strong.
Most of the larger developing countries have enough banks to make cornering of the foreign exchange market by a bank or group of banks unlikely. Nevertheless, most also have licensed nonbank dealers that work alongside the banks. On the other hand, the continued use of capital controls has led some authorities to question whether or not nonbank dealers can be monitored as well as banks, given the extensive reporting requirements and prudential oversight of banks.4 Providing that a comparable system is extended to the licensed dealers, the gains of a better customer service network and insurance against market rigging greatly outweigh any costs in terms of monitoring requirements.5
The role of the IMF
Exchange rate flexibility has played an important role in many members’ IMF-supported financial programs. Conversely, these programs have played an important role in the adoption of independently floating exchange arrangements. The 1987 review noted that one quarter of all programs in the period 1983–86 called for the adoption or maintenance of such arrangements. This close association has continued: 13 of the 44 stand-by, extended, or Structural Adjustment Facility/Extended Structural Adjustment Facility (SAF/ ESAF) arrangements outstanding as of September 30,1992 were with countries maintaining independently floating arrangements. Further, of the 22 independently floating arrangements adopted in 1985–92, 15 were implemented either as a performance criterion or an associated action in the context of an IMF program.
Multiple exchange rates run contrary to members’ obligations under Article VIII of the IMF’s Articles of Agreement, and in most instances the adoption of a floating exchange rate gave members an opportunity to unify a relatively appreciated official or commercial exchange rate and a freely (or almost freely) determined secondary exchange rate used primarily for nontrade transactions. In the previous review, it was noted that some countries took a gradualist approach to unification: six temporarily adopted dual exchange markets in the context of IMF programs. In contrast, in the period 1985–92, only two such programs (Ghana’s and Romania’s) included the introduction of temporary dual markets, reflecting increased recognition of the serious distorting effects, lack of transparency, and difficulties in enforcing the practices such an arrangement involves.
In those cases where the new rates were adopted in the context of an IMF-supported macroeconomic program, the IMF staff provided assistance in adapting the technical foreign exchange market arrangements to the particular circumstances of the member. Such assistance, previously provided by the IMF’s Exchange and Trade Relations Department and, since 1992, by the newly constituted Monetary and Exchange Affairs Department, has included providing advice on the broad policy and technical aspects of market design, drafting foreign exchange laws and regulations in association with the IMF’s Legal Department; integrating the new practices with quantitative trade, capital restrictions, and monetary policies; establishing central bank foreign exchange operations; developing regional payments systems; introducing new national currencies; and developing forward foreign exchange markets.
II. Developments Under Floating Exchange Rates
It must be emphasized, for it is often neglected in discussions of exchange rate regimes, that it is the rate itself and not the regime that is important. Large differences between official and market-determined (black or parallel) exchange rates signal that the market does not view the official exchange rate as realistic. Before proceeding with a discussion of floating rate cases, it is useful to consider some evidence that parallel market-determined rates are more than mere indicators of sentiment.
Monetary determinants and black market rates
Limited cross-country investigation has been undertaken of the behavior and determinants of black market exchange rates, although regional and single-country studies tend to confirm the importance of monetary determinants.6Table 2 sets out the results of econometric tests of the relationship between black market exchange rates and broad money (lagged one period) for a broad sample of 13 major developing countries over the period 1977–89. These results show a remarkably consistent and strong bivariate relationship in all the countries except Korea and suggest that, while governments have maintained official exchange rates at differing and often inappropriate levels, at least one part of the private sector has consistently weighed and acted on the realities of monetary policy strengths or weaknesses.7
Broad money (annual average) in period preceding exchange rate.
Serial correlation of errors is indicated.
Results are generally insignificant.
Broad money (annual average) in period preceding exchange rate.
Serial correlation of errors is indicated.
Results are generally insignificant.
Such results point to the value of investigating the role of differences between the black and official market rates as a disequilibrium variable in prompting other macroeconomic developments. A number of studies have incorporated the variable into tests of covered interest parity. One finds the variable to be a significant determinant of long-run growth performance in a group of East Asian and Latin American economies.8
Exchange market developments
With the caveat that underlying developments in market-determined exchange rates—be they official or black market rates—are determined not by the system but by the stance of monetary policies, it is nevertheless interesting to look at the evolution of rates before and after the adoption of floating arrangements.
A striking feature of exchange rate movements after floating is that they are broadly similar to those that occurred before floating. The fixed or managed regimes did not eliminate or even greatly smooth official exchange rate movements, which broadly followed those of the black market rates, but with a considerable lag.9 The exceptions are Brazil—and Zaïre some five years after it adopted the float—where the depreciation of official rates has accelerated sharply in recent years. In several countries, the official exchange rate remained less depreciated than the black market rate after floating, reflecting either market imperfections (Nigeria, as discussed above) or an illegality premium associated with continuing exchange controls (Guatemala, Guyana, Philippines, Peru, South Africa, and Uruguay). In some others, the black market rate was less depreciated than the official exchange rate after floating (Brazil, Venezuela, and Zaïre), indicating either the influence of strong capital inflows into the entire system (Venezuela) or growing avoidance of the official arrangements.
In many countries, the independently floating arrangements continued a process of real effective depreciation and improving competitiveness that was already underway (Guyana, Nigeria, Philippines, South Africa, Uruguay, Venezuela, and Zaïre). The exceptions were Brazil and Peru, where inflation was particularly rapid; El Salvador and Guatemala, where a small appreciation in the one or two years following floating reflected a strengthening of economic policies; and Paraguay, owing to large depreciations in neighboring countries. In Bolivia, the shift to floating reversed a deterioration in competitiveness beforehand.
The consequences of systemic reforms for specific macroeconomic variables are difficult to trace. First, economic developments depend on many factors, of which the exchange rate is but one. Second, there are well-known problems with “before and after” tests, but better methodology requires more data. Noting these caveats, the 1987 review found somewhat better balance of payments performances under floating arrangements and broadly similar inflation and output.
Updated data are shown in Charts 1 and 2. Of the 12 countries surveyed, inflation declined in 6 following floating (Bolivia, Brazil, The Gambia, Peru, Philippines, and Venezuela) and accelerated in one (Nigeria). Inflation was broadly unchanged in Paraguay following an initial upturn. The adoption of independently floating exchange rates (most often in the context of IMF-supported programs) between 1985 and 1992 has been associated with the surprisingly positive output performance of the 11 countries shown in Chart 2. Six experienced faster GDP growth after floating (Bolivia, with a one-year lag; Nigeria; Peru; Philippines; Uruguay; and Venezuela, after a one-year lag). In two countries (Brazil and Paraguay), growth performance deteriorated.
Chart 1.Consumer Price Developments in Selected Countries, 1980–91
Sources: IMF, International Financial Statistics (various years).
Note: Percentage change; 12-month change in consumer prices. Vertical slashes indicate the beginning of floating exchange regime.
Chart 2.Developments in GDP of Selected Countries, 1980–91
Sources: IMF, International Financial Statistics (various years).
Note: Percentage change in GDP (1985 prices). Vertical slashes indicate the beginning of a floating exchange regime. Guyana moved to a floating regime in February 1991.
Intervention policies and sterilization
Questions of appropriate intervention policies arise immediately following the adoption of floating. In industrial countries, intervention (in the context of floating exchange rates) usually means purchases and sales of foreign exchange by the central banks. In many developing countries, however, central banks will probably undertake official intermediary transactions in the exchange market that must be distinguished from direct intervention aimed at achieving a specific exchange rate. Basically, it is convenient to define as intermediation rather than intervention (1) sales of foreign exchange by the central banks in accordance with the normal business of applying public sector foreign exchange receipts to public sector current payments,10 and (2) purchases toward an established quantitative target for the recumulation of foreign exchange reserves over a period of time, within the context of the economic program.11 In contrast, intervention involves the use of international reserves—either by adding foreign exchange to reserves, selling from reserves to the market, or taking a position in the forward market—with the principal objective of changing or supporting the exchange rate.
Under the market exchange system, a central bank trading desk should generally refrain from direct intervention in the early stages of exchange reform and limit foreign exchange transactions to those consistent with its intermediary function for the government and with the overall target for the stock of international reserves. Intervention to forcefully appreciate the currency cannot be sustained during the initial reform period, because reserves are generally low.
Later in a successful stabilization and reform process, there may be a need for direct intervention to prevent the exchange rate from appreciating and undermining competitiveness in case of reverses in terms of trade improvement or an unsustainable reflow of private capital. Such intervention should be coordinated with monetary policy actions to sterilize any undue expansion of liquidity. A central bank absorbs all or part of the reversible inflow of foreign exchange by (1) purchasing foreign exchange from the spot market into reserves and (2) offsetting money aggregates by reducing net domestic assets.
While reflows of capital in several of the countries that floated and liberalized were anticipated in principle, they were nevertheless surprising in terms of the speed of the turnaround and the size of the inflows. Reversals to capital inflows (indicated by short-term private capital and errors and omissions in the balance of payments) occurred in El Salvador, Nigeria (in the first year of the economic program), and Venezuela. Such a reversal has also taken place recently in Jamaica, which floated its currency in 1983 and eliminated the remaining exchange controls and interest rates in 1991. The main lesson of these episodes, as noted earlier, has been that it is necessary to adjust fiscal policies by reducing credit to government in order to sterilize the threat that foreign exchange purchases for international reserves pose for domestic liquidity. Sterilization by contracting credit to the private sector not only raises interest rates but accelerates and perpetuates the capital inflows; it therefore tends to be ineffective beyond the very short run.
Technical Considerations in Establishing Floating Regimes
Under a market-determined exchange system, the central bank undertakes regulatory and prudential functions that include monitoring for “fair trading” (spreads, rates for small transactions, and the like), monopolistic practices, collusion, capital adequacy, and dealer licensing arrangements.
Fair marketing requires that dealers make open and honest quotations to all customers. The central bank requires that dealers post openly, in a public place, quotes for smaller transactors, including “live” (not indicative) buying and selling rates for major currencies, without excessive spreads and/or commissions. (Existing exchange rate spreads narrow considerably with the elimination of exchange taxes.) The market is monitored for abuses on a sample basis by the central bank, and continued abuses can lead to the suspension or withdrawal of dealers’ licenses. Customers are free to choose their dealers and dealers to make transactions with one another. However, nonperformance is penalized—first, by the imposition of penalty fees and, with continuing nonperformance, by exclusion from the market. The nonperforming customer or dealer bears all exchange losses, if any. Dealers make suitable arrangements for monitoring and exchanging information on creditworthiness and performance, and the central bank similarly monitors the dealers.
For purposes of macroeconomic policymaking, the banking system and other dealers need to provide the central bank with up-to-date information on foreign exchange flows and main financial aggregates. Timely and accurate information is critical to prevent the destabilization of output, prices, and the balance of payments. For example, the central bank should receive, on a daily or weekly basis, aggregate data on each bank’s or dealer’s long and short positions in foreign exchange (spot and forward separately). These data provide early information on inward capital movements that are larger than expected, possibly requiring sterilization.
Controls on the spot and forward positions of market participants, which need to be handled delicately, pose several problems. Experience has shown that it is difficult to impose limits that prevent the undue accumulation of foreign exchange but permit dealers to amass working balances sufficient for conducting business efficiently. The central bank can monitor dealers’ exposure relative to capital adequacy through normal prudential supervisory functions, consulting with the affected dealers and the association of dealers when exposure exceeds levels the central bank considers healthy for the bank in question. In any event, dealers will not accumulate excessive balances or attempt to corner the market, because they will be subject to two-way exchange rate risk under a floating system and, once they have been punished by losses, will keep their exposure to a minimum. The central bank can also act against any monopolistic or collusive tendencies by licensing further dealers, including nonbank dealers, to ensure competitiveness in the market.
The most important feature of the foreign exchange market in terms of ensuring competition is allowing creditworthy applicants free access to the market. To deepen the market, participation should be extended to all banks not experiencing financial difficulties that could call into question their ability to honor exchange contracts. In addition, in order to improve retail service, head off any collusion between the banks, and tap into the parallel market, bureaux de change could be licensed.
Consideration also needs to be given to delegating to the licensed dealers responsibility for ensuring compliance with foreign exchange monitoring requirements. The central bank supervises the bank’s performance of this responsibility by spot checks, as well as by aggregative data.
Although most basic operations necessary for the new market will be extensions of the existing exchange market, there are several areas in which strengthening and new development of the market will generally be needed. Communications links in the interbank market may need to be strengthened, including telephone connections and a Reuters or similar screen for most recent representative transactions (banks’ and dealers’ bid-ask prices, amounts, and so forth).
Opening and closing hours for the exchange market need to be decided primarily with the interests of customers in mind. In general, this will mean longer rather than shorter market hours. Arrangements also need to be made for foreign exchange services at major points of entry and exit from the country.
All interdealer purchases and sales orders are drawn up on a final basis. However, if both parties agree, orders may be drawn up on a provisional basis specified clearly at the outset. Orders in the interbank market may be made by telex or by telephone; in the latter case, the selling dealer confirms the sale on the same day by telex (with a test number) or letter. The confirmation provides the type of currency, the amount bought or sold, the name of the correspondent bank abroad, the value date, the applicable exchange rate, and the local currency equivalent.
Foreign exchange for a valid customer purchase order must be delivered within two working days from the day of payment, in local currency in respect of spot transactions or within two working days of the fixed future date in respect of forward transactions.
Dealers should move in time to offer services in forward exchange (primarily three and six months). This activity is essential to achieving the maximum benefit from the increased confidence and availability of the credit lines expected to result from the new private sector-based arrangements. Following the introduction of the market, the central bank should generally consider coordinating seminars for market participants, particularly the smaller ones, in more sophisticated aspects of the market, including forward operations. (Correspondent banks often offer assistance from their home offices in this respect.)
Reinold H. van Til
Ghana’s experience with the reform of its exchange and trade systems is an interesting one.1 The country worked through a number of steps in its move from a fixed exchange rate to a floating regime. It was some seven years before the system reached the stage of maturity, and the road was at times bumpy. To understand why it took so long, we must take into account the dismal initial economic and financial conditions. In 1983, the economy had a severely weakened external sector, large external payments arrears, a minimal official market for foreign exchange, a flourishing parallel market, and virtually no official international reserves. Since then, as a result of the steadfast implementation of a coherent set of financial and structural policies, Ghana’s economic and financial situation has improved steadily, as evidenced by its sustained economic growth and greatly strengthened external position.
The first wave of exchange rate reforms took place between 1983 and 1986. Ghana’s currency, the cedi, had been at a fixed rate of
On September 19, 1986, as a precursor to a full float, the Ghanaian authorities introduced a dual exchange rate system consisting of a fixed rate and a floating rate. As in 1983, when the move was made from a rigid fixed rate to a more flexibly managed rate, the authorities exercised caution, testing the market before relinquishing all controls. For the fixed rate, which applied mainly to government transactions and export receipts from sales of cocoa and gold, the cedi was traded at
In conjunction with this reform of the exchange system, the import licensing system was liberalized considerably, increasing access to the official foreign exchange market and reducing the spread between the auction and parallel market rates. Under the new licensing system, import licenses and access to the auctions could be obtained for virtually all nonconsumer goods. Traders in consumer goods were allowed to import with their own foreign exchange resources, which meant in practice through the parallel market.
Having tested the auction system, the authorities felt confident enough to unify the exchange system in mid-February 1987 by introducing an auction market for all official foreign exchange transactions. This market operated until the beginning of 1990, but the progressive liberalization of the trade regime—which ultimately led to the elimination of the import licensing system at the beginning of 1989—broadened the official market demand for foreign exchange over the years. At the same time, the strengthened external position allowed the central bank to increase the supply of foreign exchange to the auctions.
Despite the significant broadening of the official foreign exchange market, the parallel market for foreign exchange remained an important phenomenon during these years, and integrating and legalizing this market became an urgent objective. Accordingly, on February 1, 1988, the authorities allowed the establishment of foreign exchange bureaus. This decision was viewed with some skepticism in certain quarters, as it abandoned the “official” unified market and reintroduced a dual exchange system. This skepticism was reinforced by the strict separation of the foreign exchange bureaus and the auction market, which led to a persistent spread between the auction and bureau rates. The foreign exchange bureaus did not have access to the auction market, and the “no-questions-asked policy” and minimum reporting requirements governing the bureaus contrasted sharply with the rules in the official market.
Nonetheless, the bureaus brought what had been hidden in the parallel market into the open and were ultimately instrumental in establishing an integrated and market-determined foreign exchange system. Most importantly, however, they changed the public’s perception that access to foreign exchange was reserved for the privileged. Once the bureaus were established, foreign currency, like any other commodity, became available to everyone. Moreover, the bureaus underscored the point that market forces, not a central authority, were responsible for exchange rate developments.
The process of integrating the foreign exchange bureaus and the auction market started at the end of 1989 and was completed by the end of April 1990, when licensed bureaus were granted access to the market. Authorized dealer banks and bureaus were then able to buy foreign currency for themselves and their customers, transforming the retail auction into a wholesale market. In addition, foreign exchange dealers could trade among themselves and with their customers, and thus an interbank market for foreign exchange developed. Not unexpectedly, the unification of the foreign exchange market caused the spread between the auction and foreign exchange bureau rates virtually to disappear. The last step in this process came with the development of the interbank market for foreign exchange: the auction market lost its importance in the pricing and distribution of foreign exchange and was replaced by a fully operational interbank market in March 1992.
Ghana’s transition from a fixed exchange rate system to a floating regime was a carefully planned and controlled experiment, in which the policymakers closely monitored every step with a view to consolidating the gains achieved in previous stages of the reform process. I do not wish to suggest that this approach is the model for all developing countries, especially since I have often wondered whether a more rapid transformation would not have contributed to a speedier and more successful adjustment process in Ghana. In this respect, several factors need to be considered.
First, the degree of gradualism a country can usefully follow depends on the external resources available to finance the adjustment process. The considerable external support that Ghana obtained as a result of its successful adjustment program gave the authorities more time than they would otherwise have had.
Second, the speed of exchange rate adjustment is a function of the speed of trade liberalization, and coordination between the two is essential in controlling the process.
Third, political and social consensus on the need for and speed of macroeconomic and structural adjustment dictates the phasing of reforms. In many cases, economic policymakers may be ahead of the rest of the political establishment, which must be fully in agreement for the program to be successfully implemented.
As Peter Quirk noted, the exchange rate itself, not the regime, is the primary factor in achieving stability. However, it is also clear that certain institutional settings may inhibit needed reforms, and in that sense a reform of the exchange system may be a necessary condition for achieving an equilibrium exchange rate.
A question was raised about the extent to which a floating regime contributes to exchange rate instability and whether it is preferable to maintain a fixed exchange rate regime to assure a stable exchange rate. It is important to emphasize that exchange rate stability also depends heavily on the stance of monetary and fiscal policies. Exchange rate stability can be meaningfully achieved only if the exchange rate is consistent with external and internal equilibria.
Lorenzo L. Perez
Assessing Venezuela’s experience with a floating exchange rate system requires discussing the exchange regime that existed prior to the floating of the currency in March 1989. In the face of a balance of payments crisis in 1983, Venezuela imposed a multiple currency practice system and introduced payment restrictions. By the end of 1988, the year before the floating of the bolivar, there were three exchange rates pegged to the U.S. dollar, plus a market-determined exchange rate. The bolivar was seriously overvalued in the three-pegged exchange rate markets, which were used primarily for servicing private and public debt and for authorized imports of goods and services. Tourism, nontraditional exports, and nonregistered private capital flows were channeled through a free exchange market.
This system of multiple exchange rates entailed complex administrative controls and, together with associated trade restrictions, led to severe distortions that curtailed growth and international trade. The cost of the system was borne by the public sector through the quasi-fiscal losses of the Central Bank, as well as through the decreased oil revenues of the government and the state oil company (the state petroleum company’s oil export receipts were converted to bolivars at one of the appreciated exchange rates). The system, of course, created large economic rents for its beneficiaries, and by 1988 it had become increasingly clear to the private sector that the system was not sustainable. Along with capital flight, there was an anticipatory buildup of inventories of imported goods brought in at preferential exchange rates and financed by guaranteed letters of credit. These developments led to a sharp drop in international reserves and a balance of payments crisis.
I. The 1989 Reform
By the beginning of 1989, with official reserves at a very low level, transactions outside the free market were reduced to minimal amounts, and de facto devaluation and unification began. In these circumstances, and in support of reforms to liberalize the trade and financial systems, the new administration of President Carlos Andres Perez abolished the multiple exchange rate system on March 13, 1989, allowing the bolivar to float against the U.S. dollar in a free interbank market. The authorities’ action was prompted by the difficulty of establishing an equilibrium exchange rate, as well as by the impending significant liberalization of the trade and financial systems, which would require an exchange rate settled at the outset of the float at a level close to that prevailing in the free market. With the adoption of the unified rate system, all exchange controls were eliminated, and all external transactions of the public and private sectors were channeled through the interbank market.
As was the practice prior to 1989, under the new exchange rate system the state oil company (PDVSA) continues to surrender its foreign exchange earnings—about 80 percent of total exports in recent years—to the Central Bank, which, in turn, intermediates by selling foreign exchange to the interbank market. As a result, the Central Bank has been and continues to be both a net seller of foreign exchange to the interbank market and a major participant. Foreign exchange proceeds of loan disbursements to the public sector are also surrendered to the Central Bank to service external debt. Although some 30 commercial banks and several exchange houses now participate in the interbank market, forward exchange market transactions have developed only on a limited basis.
In summary, the new exchange rate system has not been operating as a clean float, with the exchange rate clearing the market and no changes in net international reserves. As an operating principle, the Central Bank has set a target for its average daily sales of foreign exchange in the market, taking into consideration monetary objectives, including growth of base money consistent with growth in output and prices, and a net international reserve target for the year in question. Given the level of intermediation of foreign exchange established by the Central Bank, market forces play a role in determining the value of the bolivar in the foreign exchange market. However, in practice, the Central Bank occasionally intervenes in the interbank market, varying the amount of its daily foreign exchange sales with a view to achieving exchange rate objectives.
II. Experience with the Flexible System
When the bolivar was floated in March 1989, the authorities intended not just to implement a flexible exchange rate system; they also believed that the level of reserves needed to be raised significantly, given the experience of the late 1980s. With this idea in mind, the Central Bank limited its sales of foreign exchange receipts from the oil sector. Yet the reserves increased only marginally due to the large repayments the private sector was making against import letters of credit and the delays in completing a commercial bank financing package for the public sector. However, in that year the Central Bank was able to clean up the exchange system by honoring the exchange rate guarantees and thus to reverse its own quasi-fiscal position. The total cost was close to 5 percent of GDP, most of it falling due in 1989, the rest in 1990 and 1991. The bolivar depreciated by about 30 percent in real effective terms during 1989, facilitating the initiation of the trade liberalization program.
Overall, the new exchange rate system and the adjustment program adopted by the authorities in 1989 were successful in strengthening the balance of payments and reducing inflation, although the contraction in output was larger than expected. Measured average inflation for 1989 was much higher than in 1988, but in June 1989 the inflation rate began to decline significantly. The implementation of the flexible exchange rate system was successful despite the fact that coordination of exchange and credit policies was hampered during that year by the continuation of restrictions on interest rates, which were not liberalized until April 1990.
Gradual implementation of a flexible exchange rate policy continued until late 1990. But in mid-1990, the bolivar began appreciating in real terms due to an increase in inflationary pressures related to slippages in the authorities’ fiscal program and to the oil price boom associated with the crisis in the Middle East. Only about half the oil revenue windfall was saved, and in general fiscal policy was relaxed in 1991. The Central Bank tightened credit and tried to avoid the real appreciation of the bolivar by limiting the effect of the oil revenue windfall on the foreign exchange market. In the event, the currency did appreciate in 1990 and 1991, as no progress was made in reducing inflation and there were large increases in net international reserves, facilitated in part by the government’s privatization program.
During the first nine months of 1992, the authorities changed their exchange rate strategy and tried to reduce price pressures by stabilizing the exchange rate. However, a further weakening in public sector finances (associated with the decline in oil revenues, which credit policy could not offset fully) fueled pressure on prices. The Central Bank increased its average daily sales of foreign exchange from the level of the previous year—the level that would have been necessary at least to maintain international reserves unchanged during 1992. Under these circumstances, the exchange rate settled at about 65 bolivars to the dollar between March and September 1992. Large amounts of foreign exchange were sold in the interbank market during several episodes when political instability and increased uncertainty about economic policy put pressure on the bolivar. The Central Bank lost about US$1 billion in net international reserves, and the inflation rate edged up.
In early October 1992, the Central Bank reduced the amount of foreign exchange sold in the market in an effort to slow the loss of reserves and regain competitiveness. Unfortunately, this action coincided with renewed political and fiscal uncertainties, and there was a new attack on the bolivar. The Central Bank raised interest rates sharply but also increased its sales of foreign exchange; a further large decline in net international reserves occurred, and the currency depreciated significantly. By mid-November 1992, however, the Central Bank had succeeded in halting the loss of reserves and in stabilizing the currency.
III. Lessons from the Venezuelan Experience
The managed floating exchange rate system has contributed to Venezuela’s fiscal adjustment in the face of external and domestic shocks and facilitated the implementation of an ambitious trade liberalization program. A flexible exchange rate system has allowed the country to cope with the effects of unstable oil export prices, changes in the direction of capital flows (by minimizing the effects of foreign interest rate disturbances on domestic demand), and shifts in the relative prices of tradable goods caused by the trade reform. It also can facilitate the use of a monetary anchor to help reduce inflation. As a side comment, it should be noted that the use of a fixed exchange rate as a nominal anchor does not appear to be a viable course as long as a fiscal problem persists. However, the implementation of the flexible exchange rate system in Venezuela has been uneven, with the exchange rate lagging inflation at times when fiscal and credit policies have been weak and inconsistent with the goal of a stable exchange rate. Short periods of relative exchange rate stability have been followed by rounds of sometimes large depreciations, with unsettling effects on markets.
The Venezuelan experience illustrates the crucial role that fiscal policy plays in determining how successful exchange rate policy is in achieving its aims. Intervention in the foreign exchange market to achieve a real depreciation of the currency cannot be effective when macroeconomic fundamentals are not right. Attempts to reduce exchange rate flexibility to lower inflation are likely to be short-lived, especially in a country with the degree of capital mobility that exists in Venezuela, where the monetary authority’s reduced control over base money makes monetary policy less effective in reducing inflation.
It would appear that a successful program to reduce inflation, protect the balance of payments, and create sustainable conditions for economic growth must be based on a strong fiscal position as well as on suitable credit and wage policies. In the case of Venezuela, these policies might usefully be complemented by an oil stabilization fund that would help reduce the economy’s vulnerability to oil shocks and smooth out government expenditures over the medium term. Hedging techniques for oil prices in future and options markets could also help to reduce the impact of oil price shocks.
Mexico is an interesting case, inasmuch as it provides examples of a number of different exchange regimes, some of which failed and some of which worked rather well. It is important to keep in mind, of course, that the exchange rate is but one of many elements in an economic program and that it is the whole program that matters. Also, each country has its own peculiarities, and some things worked in Mexico simply because of the nature of the country; they can hardly be replicated elsewhere.
The period of this case study dates from 1982 onward—that is, from the onset of the debt crisis up to 1992. Because some background is required to put this period into perspective, I will first briefly analyze developments in the years leading to the crisis.
I. The Period Leading to the Crisis
Mexico has long had a fixed exchange rate. In the years after the Great Depression and into the 1950s, the country had what was virtually a fixed rate. In the 22 years from 1954 until 1976—almost a whole generation—the value of the peso remained constant in U.S. dollar terms. We have to bear in mind, then, that we are dealing with a society that is very used to a fixed exchange rate.
As they were for many other economies, these years were also a period of inward-looking growth based on import substitution and high protective barriers. While it lasted, this strategy was quite successful in producing low inflation and high growth, but the inefficiencies of the model eventually caught up with it. The high cost of protectionist barriers and other distortions, together with the inefficiencies of growing with a small market, began to limit the possibilities for growth. By 1970, exports were almost stagnant, and the economy was in decline. When the Government stepped into the breach, growth resumed for a short period, but this time at the cost of an increasing fiscal deficit and rising inflation. By 1976, the financial situation had become unsustainable. The fixed exchange rate system had to be abandoned.
There was a large devaluation in 1976, with the peso dropping in value from Mex$12.50 to Mex$20 per U.S. dollar. The authorities also put in place a major adjustment program. However, the following year large new petroleum reserves were discovered, and Mexico again became very creditworthy. As the country found itself swamped with offers of foreign loans, the urgency to adjust disappeared. Over the next five years, the Government managed not only to spend the new petroleum revenues but also to quadruple the foreign debt—from US$22 billion in 1976 to US$88 billion in 1981.
After the 1976 devaluation, a “fixed but adjustable” exchange rate system was initially put in place. The new system was intended to maintain a stable exchange rate, which would be moved only if necessary. Given the petroleum reserves and newly found riches, however, the system leaned more toward the fixed than the adjustable. While the fiscal deficit ballooned and inflation rose from 12 percent to the 20–30 percent per annum range, the exchange rate moved very little. By 1981, the rate had risen only to Mex$26 to the U.S. dollar, resulting in substantial appreciation in real effective terms—some 30–40 percent—under circumstances that included a large fiscal deficit of almost 17 percent of GDP and an external debt of US$88 billion. It required only the increase in world market interest rates in 1981 for expectations to turn around. Creditors recognized that the situation was unsustainable, capital started to flow outward, and Mexico entered the crisis that soon extended to all other debtor countries.
II. The Crisis and Early Adjustment
Thus begins the period I want to look at. In February 1982, when capital outflows began, the authorities tried to respond with another major devaluation of 40 percent. But it was not sufficient; capital continued to flow out, forcing the Government to suspend debt payments in August 1982. The authorities then started organizing a drastic adjustment program. The exchange rate was adjusted further (the cumulative depreciation was about 73 percent for 1982 as a whole), and, at the end of 1982, a dual exchange system was introduced. The dual system that was put in place comprised a rate for trade and debt transactions—the official rate—and a free rate, with which the official rate was supposed to converge. The official rate was arranged in terms of a crawling peg—basically a variant of the tablita—that would preannounce the path of future inflation.
But the new exchange arrangements were only part of the overall adjustment program, which was centered mainly around fiscal policy and was to achieve impressive results. From 1982 to 1983, the primary balance of the public sector—that is, the overall balance less interest payments—was adjusted by 10 percentage points of GDP, from a deficit of 5 percent to a surplus of 5 percent. Few countries have managed such an improvement. Much of the adjustment was accomplished by cutting expenditures and raising taxes, but a good part of it was the result of setting public prices at more realistic levels. But these adjustments, together with the massive devaluation—73 percent—made in 1982, pushed both the inflation and interest rates higher than had been expected. In the end, despite the 10-point improvement at the primary level, the overall deficit was cut by a little over 8 percentage points, declining from almost 17 percent of GDP in 1982 to 8.6 percent in 1983. Growth, however, was negative, with GDP dropping by almost 4 percent in 1983.
The new exchange rate regime worked relatively well. The current account responded immediately, moving into surplus. But, as it often does, the tablita had some unfortunate effects. A lot of wishful thinking was involved; in preannouncing the rate, the authorities always seemed to be signaling that inflation would be lower than anticipated—not only because there were expectations to massage, but because policymakers tend to be optimistic in such situations. The result was that the exchange rate lagged behind inflation, appreciating gradually during 1983–84 and into 1985. At the same time, after the significant fiscal adjustments of 1982 and early 1983, the fiscal position started to weaken slightly. The primary surplus declined from almost 5 percent in 1983 to 3 percent in 1985. The current account, which remained in surplus, also began to weaken. The situation was then compounded by two exogenous shocks: the earthquake of 1985, which involved substantial fiscal costs; and the drop in petroleum prices in 1986. From one year to the next, exports dropped by about US$8 billion—the equivalent of 40 percent of export proceeds—and 26 percent of public sector revenues vanished.
Mexico had to start the whole reform process over again. After three years of adjustment, most of what had been gained was lost. A new program was put in place, based—as the previous one had been—on fiscal and exchange rate adjustment. On the fiscal side, retrenchment measures worth about 3 percentage points of GDP were implemented. The primary balance was brought up to about 5 percent of GDP, while the depreciation of the peso was accelerated during 1985–86, for a cumulative 35 percent depreciation. Again, the balance of payments responded. By 1987, reserves were up, and the current account was showing a strong surplus; not surprisingly, however, inflation had climbed to 160 percent a year. Thus, when the world’s stock markets crashed on “Black Thursday” in October 1987, Mexico was on the verge of hyperinflation. Within the country itself, the crash renewed mistrust in the currency. Capital started leaving the country, and by November 1987, the Banco de Mexico—the central bank—was forced to step out of the market and allow the rate to move freely. This situation caused some controversy in Mexico, given that the exchange rate crisis had occurred at a time when the country was gaining reserves and showing a current account surplus. The lesson the authorities drew from this situation was that as long as inflation existed, particularly triple-digit inflation, no exchange system would be stable. Stopping inflation had to be the priority.
III. Inflation and Devaluation
Twice—in 1982 and 1986—adjustments in the exchange rate brought about sharp increases in inflation in Mexico. In 1982, inflation climbed from the 20 percent range before the devaluation to the 50 percent range afterward. As we have seen, in 1986–87, it rose again, from 50–100 percent to 160 percent. In the context of Mexico in the 1980s, these increases were also manifestations of a more general problem: the resource transfer associated with the country’s external debt. Because of the debt’s size and its high real interest rates, the servicing costs were basically unsustainable and contributed heavily to inflation. We get an idea of the magnitude of the problem by looking at its history. Between 1970 and 1981, Mexico experienced an inward resource transfer; on average, during these years, the country could afford a trade deficit of about 1.3 percent of GDP per annum. Between 1982 and 1988—the six years after the crisis—this trend reversed itself. Mexico was transferring resources out of the country and running a trade surplus that averaged 6.3 percent of GDP annually. In the course of the debt crisis, a shift took place in the transfer of resources of close to 8 percentage points of GDP; in other words, the annual supply of goods in the Mexican economy declined an average of 8 percentage points of GDP. This phenomenon also had other dimensions. In the fiscal area, the Government had to resort to the “inflation tax” in order to service the debt, since the usual revenue sources were not sufficient to cover the added expense. In terms of competitiveness, restoring the external equilibrium required a large drop in real wages and, given the lag with which nominal wages adjust, inflation was the way to bring this about.
Despite the effects of the devaluations on prices, Mexico’s overall inflationary inertia was less than in many other countries, and the exchange rate adjustments did result in improvements in the current account. Real wages dropped. In part, the reduced degree of inflationary inertia is explained by the fact that because Mexico had for generations lived with a fixed exchange rate, mechanisms of indexation were not that well established. Another factor, the cooperation of the labor unions, was the result of Mexico’s particular political structure.
IV. Successful Stabilization: The 1987–88 Program
As we have seen, the exchange rate crisis in October 1987 convinced the authorities that as long as inflation existed, no exchange regime would be stable enough. A more comprehensive approach was needed to cope with the problems of inflation and growth in the Mexican economy.
A new program was devised in 1987, based on four planks1. The first and foremost of these was the fiscal effort. The second was the social pact—an agreement among business, labor, and Government aimed at breaking the inertia in the system. The third element involved finding a more permanent solution to the debt problem, while the fourth included a number of structural reforms.
The fiscal effort
A major new effort was made to improve public finances, and the primary surplus was increased by an additional 3 percentage points of GDP to a surplus of 8 percent, mostly by cutting expenditures and adjusting public prices.
The social pact
This agreement was something new in the overall strategy. It involved an initial drastic realignment of wages, public sector prices, and the exchange rate, which were then frozen at the new levels. This approach was not new; the same kind of package had been tried elsewhere. But what was singular in Mexico’s case was that the freeze was for one quarter only, from November 1987 through February 1988. Mexico once again had a fixed exchange rate system, but this time one with a very short time horizon. The solution was a compromise between the desire to provide stability (a fixed rate) and credibility (a short horizon). With the country just coming out of 160 percent inflation, an indefinite freeze would not have been credible; on the contrary, it would have backfired, as it had elsewhere. But a three-month freeze had some credibility; most economic agents could believe it would hold for that long.
In February 1988, policymakers agreed that the freeze was working and decided to continue it for another three months. The freeze was renewed by mutual agreement every three months, so that the exchange rate and all other key prices remained frozen for the whole of 1988. By the end of 1988, fiscal adjustment and the freeze had helped to turn expectations around. Inflation decelerated sharply, dropping from 160 percent in 1987 to 52 percent in 1988. The social pact was renewed for 1989 but was made more flexible: after a minor initial adjustment of the exchange rate, Mexico would move to an exchange regime with a preannounced peso slide. According to this system, the price of the U.S. dollar would be raised by one peso a day, the equivalent of about a 16 percent depreciation for 1989 as a whole. Mexico returned to a type of tablita. This system, with its preannounced slide, is the one that, by and large, continues to this day. It has been renewed every year, although with different rates of slide—first one peso a day, then 40 centavos a day, then 20 centavos, and once again 40 centavos.
The last change in the exchange regime was made in 1991, when the authorities decided to provide some more flexibility by creating a band within which the exchange rate could move. This decision was implemented by maintaining the November 1991 buying rate and applying the slide only to the selling rate, with the idea that by the end of 1992, the spread between the buying and selling points—the lower and upper intervention points—would be about 4 percent.
The reasoning behind this decision reflects the solution to a policy dilemma. On the one hand, to attract direct investment and long-term capital, it was imperative to have a more or less fixed rate system that would build confidence. On the other hand, the preannounced scheme, which made known how much the exchange rate would move each month, invited short-term speculative capital inflows. If speculators learn that a rate will move by half a percentage point during a certain month, and the interest rate is about 1 percent a month, they know that the best thing to do is to enter the market for a month and gain a free half point in interest. In such circumstances, it is necessary to introduce some risk into the system by creating the possibility of exchange losses, at the same time keeping the overall system more or less flexible. Thus Mexico established its band.
In late 1992, the agreement was renewed for 1993, with one slight difference: the band was widened further, from the initial 4 percent to 9 percent. To do this, the rate at which the upper intervention limit moves was increased from 20 centavos to 40 centavos a day. At present, with a band about 4 percentage points wide, and with the exchange rate moving close to the middle, there is the risk of a depreciation of up to 2 percent at any time—exactly what is needed in order to deter speculation.
Solving the debt problem
A comprehensive debt deal struck with the commercial banks in 1990 reduced the debt significantly. The bulk of what remained is now basically a fixed-interest loan with a 30-year maturity and guaranteed repayment of principal through zero-coupon bonds. This solution had two important repercussions. First—and most obvious—the annual transfer of debt service was cut by more than half, from 5.5 percent to 2.4 percent of GDP. But more important, uncertainty about servicing the debt was eliminated, greatly increasing confidence in Mexico’s economic performance. This fact introduces an interesting consideration in terms of exchange rate policy. All through the 1980s, while Mexico had a huge debt overhang and an associated large debt service, expectations about the sustainability of the exchange rate could never really take hold. It was always feared that the next debt crisis would precipitate an exchange rate adjustment. Once the problem of the debt was put to rest (for 30 years at least), confidence in the sustainability of the rate rose substantially. This renewed confidence has also affected domestic interest rates, which came tumbling down.
These reforms had actually been started during the very first program in 1983, particularly in the areas of privatization and trade. By 1985, Mexico had substantially reduced quantitative restrictions and lowered the maximum tariff from 100 percent to about 45 percent. In 1986, Mexico joined the General Agreement on Tariffs and Trade (GATT). Further progress was made with the 1988 program, which eliminated quantitative restrictions and further reduced the maximum tariff to 20 percent. Similarly, the 1988 program included a major liberalization of exchange restrictions, the unification of rates, and the formal reopening of the capital account, which had been closed at the time of the debt crisis. The privatization process accelerated after 1990: in 1991, enterprises worth about US$10 billion were privatized, representing close to 3 percent of GDP for that year.2 Financial reform, including the liberalization of interest rates and the elimination of credit controls and reserve requirements, spurred financial reintermediation and lowered costs, particularly for the export sector. There have also been revisions to the foreign investment law and, in general, a number of initiatives to deregulate and reduce red tape.
In summing up, I would make two observations about the more recent features of Mexico’s external position. Thus far, this fourpronged approach has been remarkably successful. Inflation is down,3 economic growth has been quite acceptable in the last few years, and per capita GDP has been growing. However, after five years with a predetermined exchange rate, and given how slowly inflation has come down, the exchange rate has appreciated substantially. Mexico now faces a rather large current account deficit, raising the question of what caused this deficit to increase in an otherwise very successful inflation-fighting program. One possibility is a decline in private savings that has exerted pressure on the current account. From 1987 to 1991, private savings dropped by more than 4 percentage points of GDP, from 14 percent to 10 percent. At the same time, private investment moved up by some 3 percentage points of GDP, from 13.6 percent to 16.9 percent. Thus, the private sector’s current account deteriorated by 7.6 percent of GDP. The relatively small deterioration in the overall current account during those years is attributable to the improvement in public sector performance.
Exactly why developments in the private sector can offset public sector adjustment is not clear. One explanation may be that at the same time an economy stabilizes and public finances improve, the inflation tax declines, permitting private sector consumption to expand. A related manifestation could be that as inflation declines and interest rates drop, the private sector receives fewer interest payments from the government. If the private sector treats these interest receipts as transitory income, the result is a corresponding decline in savings. Another possibility is a “wealth effect”: stabilization could bring about large increases in stock market values or real estate, increasing private consumption and decreasing private savings. Whatever the explanation, the fact remains that in Mexico, as in a number of other countries, as the economy stabilizes and public finances improve, the private sector has an offsetting effect that tends to cause the current account to deteriorate and puts additional pressure on the exchange rate. On the other hand, as inflation settles down at a low rate, most of the factors just mentioned tend to disappear, private savings recover, and the current account deficit narrows again.
How have the structural reforms affected the equilibrium level of the exchange rate and competitiveness? Liberalization of the exchange system, reduced financial costs to exporters due to financial reform, reforms of the foreign investment code, and deregulation have all had definite, positive effects on competitiveness. The effect of trade liberalization on the exchange rate is more ambiguous, of course. On the one hand, there is a need to offset the impact on imports, but on the other hand, exporters’ costs have been lowered, particularly for imported raw materials, and exportables have been diverted from the previously protected domestic market (where they were bringing in high profits) to the open foreign market (where they attract lower profits). Considered in this light, liberalization has increased export possibilities. Over the last ten years, real wages have come down dramatically and have remained low, so that in terms of unit labor costs, the real effective exchange rate is still much lower than the rate based on relative prices. Eliminating the debt overhang has increased confidence in the peso and reduced the risk that the exchange regime may prove untenable. Considering all these factors, at the moment we cannot say for sure what the equilibrium level of Mexico’s exchange rate should be. Only time and the evolution of the balance of payments will provide an answer.
Mexico has been used as a case study partly because its relative success during the last few years has attracted interest. This interest is by no means entirely flattering—first, because success brings with it new problems, such as capital inflows, that are difficult to cope with; and second, because after ten years of struggling, we Mexican policymakers are by no means sure that our current policy has finally succeeded. Indeed, many internal elements indicate that something negative may still happen, and the external environment is, as always, uncertain. But it is useful in terms of exchange arrangements to study Mexico, since it has had both fixed and floating rates and single and multiple regimes—all with a certain degree of success for some time, and then with clear failures.
Because Thomas Reichmann and Claudio Loser have already given a clear chronological description of Mexico’s exchange rate situation, I will talk only about six lessons that can be drawn from Mexico’s experience. The first lesson is that it is often difficult for policymakers to see events clearly while the events are taking place. It is very easy today to look back over the last ten years and see where the mistakes were made. But during those years, we were coping with a potential crisis every day. We were never sure where we were heading or what the internal and external environments for our exchange rate policy would be in the future. So our policy was developed crisis by crisis, and short-term considerations very often overrode decisions that had been made from a long-term perspective. It is therefore not easy to talk about our “strategy,” because we had to adjust our actions to the events taking place around us. We also faced very serious short-term demands. We had to service a very large foreign debt, for example, and did not have the resources to do so. If our exchange rate policy had a specific orientation, then, it was primarily to use very large devaluations to close the external gap generated by inflation.
The debt overhang is a good example of the kinds of problems we faced. In 1982, we were not even aware of the seriousness of this problem. The then Minister of Finance told the Congress that we were experiencing what he termed “a liquidity crisis,” but then we were forced to suspend our debt service payments. It took a long time for Mexico to realize that the size of our debt was incompatible with growth and stability. Once we realized that, we had to convince the rest of the world—starting with the IMF—that we could not survive with the existing level of debt. The next step was to reach an agreement with creditor governments and banks. Settling our debt problem took almost seven years, from 1982 to 1989, when an agreement in principle was reached; the final agreement was not signed until February 1990, almost eight years after the debt crisis began. The uncertainty added to our problems. We were never really sure, for example, what level of external savings we could count on.
The second lesson involves the use of privatization to channel resources to the Government. In 1982, we understood that the Mexican Government was too large and had to be reduced. But it took time to set up the privatization of so many different public enterprises—including 18 banks, the enormous telephone monopoly, and many other large firms—in a way that would maximize Government revenues. Aside from the revenues, however, the process of privatization was important in itself because of the marketoriented signal it gave the economy.
We then had to decide what to do with the privatization revenues in a poor and underdeveloped country like Mexico. There were many demands on the money, and it took great courage for the President and ministers to decide that all of it would be used to amortize the domestic debt. It was a very difficult situation: peasants and workers were asking for a chunk of the country’s resources, the Government was saying the country had no money, and at the same time the Government was paying back billions of dollars worth of domestic debt. In retrospect, however, I believe there was no other way the Government could have channeled resources from the private sector to our foreign creditors. We needed money and would not have been able to change our tax system quickly enough to generate more resources. Privatization thus played an important role in our efforts to reduce our debt and create a climate that would help stabilize our exchange rate.
The third lesson that can be drawn from the Mexican experience is that no matter how undervalued the real exchange rate is, it is still susceptible to a run against the currency. At the end of 1982, we had devalued by more than 40 percent—and in the short term it was mostly real devaluation—and yet there was a run against the Mexican peso. In 1987, we had even had a real exchange target for some time and had been devaluing at almost the same rate as inflation. Internal and external investors alike recognized that the rate was undervalued. But when the stock crisis developed in the United States in October 1987, the Mexican market was immediately contaminated, and investors fled. In the United States, the flight was out of the stock market into safer bonds, but in Mexico, despite the undervalued exchange rate and the fact that foreign reserves were at their highest level ever, there was an exchange rate crisis.
This crisis was particularly threatening because we had been told earlier that based on the experiences of other countries, we should reduce our domestic debt and increase real interest rates in our financial market. In fact, after several years of high inflation, our domestic public debt had been to a certain extent liquidated. For many years, we had had negative real interest rates, and the proportion of currency over GDP was only about one third of our historical level. By all measures, the financial market had contracted. And yet when faced with this lack of certainty from investors, the internal financial market was subject to a crisis that was reflected in our exchange rate market. We learned the hard way that an undervalued exchange rate does guarantee a firm base for planning everything else, because of the danger of a currency run.
The fourth lesson is that the external sector is extremely important. This lesson was also a painful one for Mexico, because we as a country have absolutely no authority or control over the external sector. One important determinant of the exchange rate is the terms of trade, and Mexico suffered deteriorations in its terms of trade in 1981, 1985, and 1986. As expected, our exchange market reflected these deteriorations.
There are other external determinants more subtle and difficult to identify and measure that are nevertheless important. Among them are the business cycles of larger economies, crises in foreign markets—like the recent problems with the European exchange rate mechanism—and factors such as the very weak recovery of the U.S. economy, which has led to a significant amount of dumping on the Mexican market. Because trade liberalization in Mexico was very orthodox, we did not pay enough attention to mechanisms that could legitimately have prevented such dumping. The result was a large inflow of capital and goods that caused the current account to deteriorate. It took us some time to realize that there was a problem, and more time to design a mechanism that would prevent dumping but not be overly protective.
Another important determinant of the exchange rate is tax arbitrage, especially in a country like Mexico, which is the base for many foreign companies. In order to maximize their profits worldwide, these companies register them in the country with the lowest marginal tax rate. As part of our structural adjustment, Mexico has decreased the average tax rate on corporations to a level that is very competitive vis-à-vis the U.S. market, so that taxes on corporations are lower in Mexico than in the United States. However, the automotive industry, which is in crisis in the United States, has not been making profits there recently and thus has an incentive to shift profits made elsewhere to their U.S. operations. Accordingly, the companies tend to overprice exports to their subsidiaries in Mexico, so that the profits are registered in the U.S., where the companies are not paying taxes. This behavior leads to a surge in the value of imports into Mexico and creates a signaling problem, because many observers interpret the increased import growth as a deterioration of the current account—when it may in fact be nothing more than a disguise for capital transfers.
The fifth lesson is that expectations cannot be ignored, and once again, Mexico offers a variety of examples of this fact. At the peak of speculation in 1962, Mexico had a floating regime, and the exchange rate went up from 22 pesos per dollar at the beginning of the year to 150 pesos per dollar—a depreciation of more than 80 percent. That exchange rate was not validated by any of the real factors, and yet the only way the central bank could improve the currency’s credibility was to validate the rate. In fact, the Government chose to move from a multiple exchange rate regime to a dual exchange rate regime, leaving the higher exchange rate at 150 pesos per dollar. We understood that such an exchange rate was irrational, but once it had been reached, we found convincing the market that it was unreasonable very difficult. The same thing happened in 1987. After the U.S. stock market crashed, the Mexican floating rate rose to 2,200 pesos per dollar, and it was very difficult to convince the market that this rate was not sustainable. We actually had to prove that such a rate would lead to higher inflation.
The lesson here is that somehow the Government has to explain policy to the public, pointing out strengths and weaknesses, in order to avoid this type of irrational behavior, which is difficult to deal with after the fact. In Mexico, and probably in other countries, the exchange rate is a symbol of economic policy and must therefore be treated with care. It is not just another price that can be liberalized and left to move up and down as the market dictates. I am by no means saying that authorities should manipulate the market, but on the other hand, the exchange rate cannot just be ignored and the public left uninformed. It is difficult to judge whether our Government has had any success in this respect, for the fact that the exchange rate has not fluctuated strongly does not mean that our communication program has been successful. However, we have made a point of going to the public and describing what we are trying to do.
And, finally, the sixth lesson that can be drawn from our experience is that price stability is important per se because it affects the exchange rate. If the fundamentals are correct and price stability is sustained, then the exchange rate, whether fixed or floating, will also be stable. What is important is that economic policy be geared toward stability. We tried unsuccessfully in 1984–85 to stabilize through the exchange rate, and the result was an exchange rate crisis in 1985. Then, we reversed our policy almost completely, so that in 1986, when the price of oil was approaching its lowest level in many years, we tried first to devalue the exchange rate and then to establish a real exchange rate target. Although that target was successful in producing a current account surplus—despite the fact that our terms of trade had deteriorated dramatically—inflation did not come down. In fact, that year marked the beginning of a period of hyperinflation in Mexico; we had no anchor to help us in the important task of reducing inflation.
As a result, we gave the exchange rate a central role in our stabilization program of December 1987. From 1982 onward, we had been following an ongoing process of fiscal discipline that included cutting expenditures, privatizing, and completely revising the tax and public price structures. At the beginning of our stabilization program, most of what we call controllable prices were frozen in both the public and private sectors, and gradually private prices were liberalized. But it was difficult to liberalize the exchange rate immediately, so we took almost the opposite approach, setting a real exchange rate target supported by strong fiscal and monetary policy. This approach was responsible for our success in bringing down inflation.
We were able to stabilize the exchange rate only after achieving a certain degree of price stability. Each supported the other. Only very recently have we been able to allow more flexibility in the exchange rate, moving slowly from a fixed exchange rate or preannounced crawling peg to a system that allows the exchange rate to float within a band. The band has been expanded over time; the ceiling depreciates while the floor remains fixed.
We do not know what the future holds. The falling price of oil on the world market is an important concern for both our current account and public finances, because Mexico depends on oil revenues. Thus far in 1992, rains and floods have affected agricultural production. Because of developments like these, we know we will never be able to start with a clean slate when designing our policy. We must be prudent and leave ourselves enough margin to respond to potential internal and external crises. Despite our desire for a long-term strategy, we will have to go on “muddling through.” Mexico’s experience shows that instability is costly in terms of creating and sustaining long-run exchange rate policy. As I said earlier, Mexico’s lack of a planning horizon was so painful and costly in terms of welfare, growth, and savings that I feel stability should be a goal in itself. Mexico is a vivid example of how important stability is in extending the planning horizon.
Summary of Discussion
In the discussion of issues relating to exchange rate systems, most participants agreed that the distinction between a freely floating regime and a managed float is not clear, especially if central banks do not make their targets or indicators clear. There was also agreement that capital account convertibility could be tolerated, depending on the exchange rate and interest rate regimes. For example, freely flexible interest rates serve as shock absorbers under fixed exchange rate regimes. Indonesia’s fixed exchange rate regime had attributes of a crawling peg, with occasional devaluations that transferred what could have been exchange rate volatility to the interest rate.
In this connection, the relative merits of interbank foreign exchange markets and central bank auctions were discussed. The point was made that many countries—particularly those in which a large share of the foreign exchange earnings accrues to state enterprises—have foreign exchange surrender requirements, and in these countries the auction system not only works well but allows the government to pursue a managed float. In a number of developing countries, a managed float with supporting monetary and fiscal policies has contributed to external viability, diversification, low inflation, and economic growth.
Discussing Mexico, participants agreed that the country’s open capital account had been a major factor in the authorities’ decision to adopt a flexible rate system. Since the authorities felt that financial shocks would magnify exchange rate movements and send confusing signals to the real sector, the system Mexico adopted established a mechanism to stabilize the exchange rate but still provide flexibility. This combination of an exchange rate band and a flexible interest rate policy was designed to create a degree of uncertainty among speculators, thereby removing some of the pressure on the exchange rate.
Participants emphasized that foreign savings were attracted to Mexico and that the policy response to this inflow was either to allow foreign reserves to build up or to permit the current account to go into deficit. The capital inflows were judged to be a response to structural reforms and fiscal consolidation, as well as to potentially bright economic prospects (in part as a result of the North American Free Trade Agreement). A long-term current account deficit was therefore considered sustainable, and the exchange rate was allowed to appreciate.
Finally, it was agreed that the exchange rate had been a central instrument of Mexico’s stabilization efforts but that the Government had also undertaken strong fiscal adjustment and followed a tight monetary policy. For this reason, exchange rate policy had not been misused, as the fact that the exchange rate had remained well within its band showed.
Peter J. Quirk, Benedicte Vibe Christensen, Kyung-Mo Huh, and Toshihiko Sasaki, Floating Exchange Rates in Developing Countries; Experience with Auction and Interbank Markets, IMF Occasional Paper No. 53 (Washington: International Monetary Fund, 1991).
Afghanistan, Bolivia, Brazil, Bulgaria, Costa Rica, Dominican Republic, El Salvador, The Gambia, Ghana, Guatemala, Guyana, Haiti, Honduras, Mozambique, Nigeria, Paraguay, Peru, Romania, Russia, and Venezuela.
Quirk and others, pp. 32–33.
However, the widespread capital flight of the 1980s—despite extensive capital controls in developing countries—raises questions regarding the effectiveness of any monitoring of banks or nonbanks.
Technical considerations in setting up a competitive foreign exchange market are summarized in the annex to this paper.
See Sanjeev Gupta, “An Application of the Monetary Approach to Black Market Exchange Rates,” Weltwirtschaftliches Archiv (Germany), Vol. 116 (No. 2, 1980), pp. 235–52; and Peter E. Koveos and Bruce Seifert, “Market Efficiency Purchasing Power Parity, and Black Markets: Evidence from Latin American Countries,” Weltwirtschaftliches Archiv, Vol. 122 (No. 2, 1986), pp. 313–26.
Exchange rate data are provided by correspondents to Pick’s Currency Yearbook and, more recently. International Currency Reports.
Peter J. Quirk, “Exchange Rate Policies and Management: A Model for Successful Structural Adjustment: The Experience of Southeast Asia,” in Strategies for Structual Adjustment: The Experience of South Asia, edited by Ungku A. Aziz (Washington: International Monetary Fund, 1990).
Chart 1 shows official and parallel bilateral exchange rates against the U.S. dollar, as well as real and nominal exchange rates, in the period 1980–91 for the following group of countries: Bolivia, Brazil, EI Salvador, Guatemala, Guyana, Lebanon, Nigeria, Paraguay, Peru, Philippines, South Africa, Uruguay, Venezuela, and Zaïre. Effective exchange rate series are from the IMF’s information notice system.
Examples are sales of oil receipts by the Bank of Venezuela and the auctioning of enterprise receipts by the Moscow exchange.
The aim of increasing reserves to a certain minimum level should not be considered intervention because the focus is on a quantity of reserves considered necessary (with the exchange rate as a “residual”).
A detailed account of developments is contained in Ishan Kapur, Michael T. Hadjimichael, Paul Hilbers, Jerald Schiff, and Philippe Szymczak, Ghana: Adjustment and Growth, 1983–91, IMF Occasional Paper 86 (Washington, D.C.: International Monetary Fund, 1991).
See Claudio Loser and Eliot Kalter, Mexico: The Strategy to Achieve Sustained Economic Growth, IMF Occasional Paper 99 (Washington, D.C.: International Monetary Fund, 1992).
In 1992, about the same amount was privatized.
At the end of 1993, inflation had dropped below 10 percent.