August 1, 2018


August 1, 2018


August 1, 2018

Approved By

Petya Koeva Brooks, Jonathan D. Ostry, and Ratna Sahay

Prepared by a staff team from the Monetary and Capital Markets Department (MCM), Research Department (RES), and Strategy, Policy, and Review Department (SPR). The MCM team comprised Chikako Baba and Svetlana Popova. The RES team comprised Tamon Asonuma, Swarnali Ahmed Hannan, Callum Jones, and Suman Basu. The SPR team comprised Masashi Saito and Silvia Sgherri. Individual authors of each chapter are indicated in the paper.



  • A. MCPs in Literature and Country Case Studies: 1925–1983

  • B. Why and How Have Countries Imposed MCPs?

  • C. Near-term Corrective Effects of MCPs

  • D. Distortive Effects of MCPs

  • E. Cross-Border Spillovers

  • F. Balancing Corrective and Distortive Effects of MCPs


  • A. Evolution of MCPs and Macro Drivers

  • B. Characteristics and Frequency of MCPs

  • C. Approval of MCPs

  • D. MCPs and Article VIII Acceptance


  • A. Introduction

  • B. Performance Before MCPs

  • C. Performance With MCPs


  • BOX

  • 1. Survey Questionnaire


  • 1. Costa Rica MCP (1980–83)

  • 2. Incidence of MCPs in Fund Membership

  • 3. Incidence of Multiple Currency Practices by Region

  • 4. Type of MCPs

  • 5. Composition of MCPs

  • 6. Approval of MCPs

  • 7. MCPs and Article VIII Acceptance


  • 1. MCP Episodes: Selected Country Case Studies

  • 2. Logit Analysis: Causes of MCPs, 1980–2016

  • 3. Regression Analysis: Macroeconomic Impacts of MCPs, 1980–2016

  • 4. Recent MCPs: Selected Experiences

  • References

Review of Literature and Country Cases: 1925–19831

This chapter describes the corrective and distortive effects of multiple currency practices (MCPs) using country case studies MCPs have been used for two main purposes: (i) stabilization of short-term BOP pressures; and (ii) revenue mobilization or resource allocation. While MCPs have had partial success in the short term in stabilizing the BOP position and capital flows, they have been associated with adverse effects in the medium to long term, especially where these systems have become a long-lasting aspect of the country’s exchange arrangement. Since MCPs can affect relative prices, they also have the potential to adversely affect other countries through trade and investment channels

A. MCPs in Literature and Country Case Studies: 1925–1983

1. This chapter describes the corrective and distortive effects of MCPs using selected country case studies over the period 1925–83.2 The literature has traditionally taken a “first-best” approach on the issue, and denounced all distortions as having negative effects. The chapter intends to take a “second-best” approach, and evaluate how MCPs may have helped to correct market imperfections as well as whether they have exacerbated or led to other distortions―both in the short term and long term. In addition, the appropriateness of MCPs in correcting market imperfections could depend on persistency of external shocks.

2. MCPs implemented in a range of middle income countries with differing degrees of trade and financial openness are studied.3 Table 1 summarizes the findings of MCP episodes in previous studies that we have examined. Specifically, the country case studies covered by our literature survey are: Argentina (1981 and 1982), Costa Rica (1980–83), Cuba (1925–50s), Honduras (late 1930s-50), Indonesia (1950s and 1960s), Jamaica (1977–80), Mexico (1982), the Philippines (1951–65), Romania (late 1970s-80), Turkey (late 1970s-80), and Venezuela (early 1960s). In our survey, we have focused on selected countries that have been found by the academic literature and past IMF papers to have used systems of multiple exchange rates for current transactions. Not all of these countries have necessarily been identified by the IMF as having had a multiple currency practice.

Table 1.

MCP Episodes: Selected Country Case Studies

article image
Sources: de la Torre and Ize (2013) for Cuba; IMF (1984b) for Argentina, Jamaica, Mexico, Romania and Turkey; Kanesa-Thasan (1966) for Indonesia; Kiguel and O’Connell (1994) for Argentina, Mexico and Turkey; Sherwood (1956) for the Philippines; Vinelli (1951) for Honduras; Woodley (1964) for Venezuela.

B. Why and How Have Countries Imposed MCPs?

3. MCPs are generally imposed to mitigate BOP pressures and/or capital flow volatility (de la Torre and Ize 2013; IMF 1984a; Kiguel and O’Connell 1994).4,5 Typically, an MCP is introduced in countries hit by a large negative external shock. These countries were also typically facing underlying vulnerabilities, including expansionary fiscal policies and/or rapid domestic credit expansion under a fixed exchange rate regime. The deterioration in the terms of trade and surge in the demand for foreign assets triggered by financial repression and/or unsustainable macroeconomic policies placed substantial pressures on exchange rates, in some cases leading to a substantial depreciation of the exchange rate. Jamaica in 1977–80 and Venezuela in early 1960s are representative episodes of MCPs having been imposed following deterioration in the terms of trade (IMF 1984b; Woodley 1964).6

4. MCPs have also been imposed to achieve revenue mobilization or special resource allocation objectives (IMF 1984a; Sherwood 1956). Countries have often used multiple exchange rates when the problem of finding sources of revenue to finance rapidly-increasing government expenditure is acute (e.g. Cuba, the Philippines and Venezuela).7 The use of multiple rates is preferred over taxation for two main reasons. First, the administration of multiple rates is easier when compared to the administrative difficulties of direct taxation (export and import taxes) particularly when state capacity is weak (Fleming 1974; Dornbusch 1986). Second, revenue collection through multiple rates can be sizable, especially in the near term. In addition, MCPs are used for special resource allocation purposes in order to promote specific categories of FX receipts or to curb certain FX expenditure.

5. Exchange taxes, export taxes/subsidies, and import surcharges have also given rise to MCPs. Honduras introduced a tax on all FX purchases, while the Philippines introduced it on FX sales (Vinelli 1951; Sherwood 1956). Similarly, Nicaragua implemented two surcharges, one on semi-essential and the other on nonessential imports (Sherwood 1956). Meanwhile, Indonesia introduced a tax on the basic rate, export taxes/subsidies, and import surcharges—together with export inducement certificates and FX entitlements (Kanesa-Thasan 1966).

6. Market structure and segmentation due to exchange controls can help explain the emergence and persistence of multiple exchange rates. IMF (1984b) finds that market segmentation due to exchange controls resulted in multiple exchange rates in Argentina, Jamaica, Mexico, Romania and Turkey in late 1970s and early 1980s. For instance, the exchange market in Jamaica was split in to a “basic” market for essential import payments, government transactions, and most transactions of the bauxite-exporting sector, and a second market where all other transactions were conducted at a depreciated rate. The same market development pattern was witnessed in Cuba and Venezuela in pre-WWII periods (de la Torre and Ize 2013; Woodley 1994).

C. Near-term Corrective Effects of MCPs

7. MCPs had some success in the short term in stabilizing the BOP position and capital flows. As Dornbusch (1986) describes, theoretically, MCPs that increase effective import prices relative to export prices, ceteris paribus, should lead to improved trade and current account balances, primarily through a reduction in imports.8 As an example, Costa Rica experienced a sharp decline in import volumes and stable export volumes (Figure 1) after it further tightened the restrictiveness of its trade and exchange rate system with a number of measures including MCPs between 1980–83 (IMF 1984b). In the case of Venezuela in the early 1960s, acute BOP difficulties were mitigated at least temporarily under MCPs and exchange controls (Woodley 1964). In a similar vein, MCPs used in Mexico in 1982 successfully insulated the economy from the impact of unsustainable capital flows (including sudden stops and reversal) and prevented further capital flight (IMF 1984a, 1984b).

Figure 1.
Figure 1.

Costa Rica MCP (1980–83)

Citation: Policy Papers 2019, 015; 10.5089/9781498318860.007.A004

8. In doing so, MCPs helped manage the trade-off between inflation and unemployment as part of the macroeconomic adjustment. MCPs can at best partially insulate domestic prices from foreign shocks in the short term (Ghei and Kiguel 1992).9 At least temporarily, MCPs in Argentina and Mexico in 1981–82 and Venezuela in the early 1960s successfully reduced the pressure to tighten monetary policy in an environment of financial distress while protecting the BOP by mitigating the transmission of external shocks to domestic prices. The countries examined, thus, appear to have had more control over domestic inflation, largely avoided a sharp drop in real wages, and were better able to stabilize output.

9. In some cases, MCPs improved the revenue-raising powers of the government. Even when the exchange tax or exchange spread is very small, multiple rates are, in general, not a small source of revenue. This, in turn, results in sizable revenue collection in comparison with revenue from other individual tax sources.

D. Distortive Effects of MCPs

10. The imposition of MCPs may hurt long-term growth. IMF (1984b) compares the macroeconomic developments of a group of non-oil developing countries with MCPs to that without such practices during the period 1979–82. It finds that real output growth was lower in three out of the four years and inflation was higher in all four years in countries with MCPs. Both export and import volumes were lower for the MCPs group, leading to a lower ratio of the current account deficit to exports of goods and services.

11. One reason for this finding is the distortive effect of MCPs on the export sector. Dornbusch (1986) outlines how an MCP can cause expenditure reducing effects through the implicit taxation associated with multiple exchange rates. Simultaneously, MCPs can also cause expenditure switching effects through the explicit changes in relative prices. Therefore, the exchange rate spread, by acting as a tax on exports and a subsidy on basic imports, can undermine a country’s exporting industries as well as import-substituting industries affected by the subsidy on imports.10 This results in resource misallocation, and a loss of economic growth and job creation. Representative episodes of these distortions include Argentina and Mexico in the early 1980s and Venezuela in the early 1960s (Kiguel and O’Connell 1994; Woodley 1964).

12. In addition, MCPs may incentivize socially-inefficient activities. The exchange rate spread can give rise to a wedge between private and social interests, resulting in socially destructive rent seeking activities (de la Torre and Ize 2013). Furthermore, the complete segmentation of exchange rates cannot be implemented in practice, with different rates giving opportunities for illegal leakage, weakening the effectiveness of dual exchange rates partly due to the over-invoicing of imports and under-invoicing of exports (Fan 2004).

13. The use of MCPs to raise tax revenue generates a negative incidence on domestic consumers. Sherwood (1956) emphasizes that the incidence of the tax would fall not only on importers but also on domestic consumers through higher prices. The tax might have been transferred to domestic consumers in the form of reduced receipt amount, since foreign investors may respond negatively to the tax on the remittance of profits abroad.

14. Countries have faced difficulties in removing MCPs, potentially delaying needed macroeconomic adjustment. Venezuela in the early 1960s was a prominent case where the MCP was maintained over prolonged periods and necessary fiscal and monetary adjustments were not implemented (Kiguel and O’Connell 1994). Similarly, MCPs in Cuba and in the Philippines became important sources of general revenue and it was consequently difficult to abolish the tax in the absence of clear alternative revenue sources (Sherwood 1956).11,12 In line with this, the 1984 IMF review of experiences documented that multiple rates had in many cases become a long-lasting device for the postponement of necessary policy adjustments, rather than a transitional arrangement:13

  • These systems have often become a long-lasting aspect of members’ exchange arrangements. In the bulk of instances, multiple rates were seen as temporary, but their elimination proved difficult and in the case of a number of members, they were maintained over an extended period of time. For example, from 1970 to 1983, 12 members had multiple exchange rates continuously, and for a total of 61 members maintaining explicit MCPs at some time over this period, the average duration for which they were in existence was five years.14

  • MCPs have been implemented using complex mechanisms subject to numerous modifications. Dual FX markets used to delay an exchange rate adjustment require either direct intervention in the second market or manipulation of the markets through administrative means, and experience pointed to associated shortcomings. Generally, the requirement to surrender FX at the official rate was rarely effective enough to prevent FX holders from profiting at the market rate.15 In these cases, as well as in others16, either the rate in the second market was pegged, or the system was modified in other ways. In general, this occurred only after a substantial amount of official FX had been used to support the rate in the second market, further stimulating capital outflows. Some countries introduced administrative exchange allocations when the rate in the free market depreciated significantly relative to the official rate (e.g., Argentina, 1981 and 1982). Partial free market mechanisms have also on occasion tended to contribute to capital outflows, as excessive exchange rate instability developed in the second market.17 In some instances, the failure of partial free markets to achieve the objectives sought by the authorities led to their rapid dismissal (Argentina, 1981 and 1982; Costa Rica, 1980).

  • Frequent changes to the system generated significant uncertainty. On occasion, rather than undertaking a depreciation of the official exchange rate, a second exchange market was established. This temporary exchange market was set up with the intention of adjusting the official exchange rate toward the exchange rate in the second market, either directly or through shifts of transactions from one market to the other, or both. This strategy has often resulted in a more appropriate average exchange rate—from that standpoint, it helped the exchange rate system perform a more effective role in the adjustment process. However, when the shift of transactions involved unduly frequent changes or when the adjustments in the official exchange rate were perceived to be insufficient, the strategy proved counterproductive for exporters and importers.18 Expectations were formed regarding exchange rate changes which tended to disrupt foreign trade, as the anticipation of a more depreciated exchange rate in the future encouraged exporters to withhold shipments and repatriation of receipts, and incentivized importers to stockpile and over-invoice imported goods. These speculative trade activities exerted added pressures on official international reserves.

E. Cross-Border Spillovers

15. MCPs also have the potential to adversely affect other countries via trade and investment channels. MCPs may be implemented to differentiate among trade partners and transfer additional payment burden, i.e. tax incidence to non-preferred partners. This may distort trade between the imposing countries and non-selected trade partners and generate retaliatory measures if used in a persistent fashion. The Romanian experience in the late 1970s-80, where a number of implicit rates were applicable to different transactions, is a representative case (IMF 1984b). In a similar vein, MCPs negatively influence investment: in the Philippines, a tax on the remittance of profits abroad was paid by foreign investors and potentially deterred foreign investment (Sherwood 1956).

16. In this context, the literature draws broad conceptual parallels between the effects of MCPs and those of direct and indirect tax/subsidy policies. Reviewing MCPs for commercial transactions across countries from the 1960s, Dornbusch (1986) notes that MCPs change the relative prices of exports and imports, with the economic consequences comparable to tax and subsidy policies that change relative prices in the same way. Furthermore, Fleming (1974) notes that taxes and subsidies on capital transactions, if enforced by exchange controls, would resemble MCPs in potential effectiveness.19

F. Balancing Corrective and Distortive Effects of MCPs

17. In summary, the review of the literature on MCPs and the experience with their operation seem to concur that these practices are likely to have partial success in the short term but negative effects in the long term. They have had partial success in the short term in stabilizing the BOP position and capital flows, as their insulation of domestic prices from foreign shocks has not been fully undone by the market However, they have also been associated with adverse effects in some instances and have had a deleterious impact on long term growth in nearly all the cases where they have been imposed in a persistent manner. They also have the potential to adversely affect other countries via the trade and investment channels. As such, the Fund should continue to discourage members from using MCPs, but there could be instances where those practices would be accepted, on a temporary basis, as a partial step in the process of undertaking the required exchange rate adjustment or to complete the necessary market development process. The desirability of avoiding reliance on multiple exchange rates or of establishing concrete plans for their elimination where they already existed should also be reiterated.

Recent Experience with MCPS: 1980–201720

This chapter discusses experience with the Fund’s MCP policy since the last comprehensive review in 1981, including the frequency and evolution of MCPs, their characteristics and main economic drivers, and their treatment under the Fund’s approval policy

A. Evolution of MCPs and Macro Drivers

18. Multiple currency practices (MCPs)21 were a frequent feature of exchange systems in the early history of the Fund, but their use dropped significantly by the early 1980s. In the latter half of the 1950s and 1960s, as international trade and payment conditions improved, considerable progress had been made toward reducing the incidence of MCPs. After the abolishment of the par value system, the incidence of MCPs increased moderately in the mid-1970s, primarily for balance of payment reasons. In the early 1980s, about 30 percent of Fund members maintained MCPs.

19. MCPs resurged in the early 1980s among emerging and developing countries (EMDCs), against the background of widespread balance of payment (BOP) difficulties. In the aftermath of the second oil shock in 1979, the global economy was hit by recession at the beginning of the 1980s. Tight monetary policy in advanced economies to rein in inflation affected the global economy and led to higher debt payment burdens for EMDCs that had borrowed in foreign currency to finance elevated oil imports in the 1970s. In addition, sharp appreciation of the U.S. dollar (about 50 percent between 1979 and 1985), reflecting high interest rates, put further pressure on the BOP of EMDCs that had used the U.S. dollar as their nominal anchor. Eventually, many EMDCs suffered from debt crises in the early 1980s, and resorted to MCPs. Correspondingly, the incidence of new MCPs during this period was concentrated among EMDCs, in particular in Latin America, while the reliance on multiple exchange rates continuously declined among advanced economies (Figures 2 and 3). The incidence of MCPs peaked in 1986, with about a third of EMDCs maintaining MCPs. As the global economy recovered toward the mid-1980s and the U.S. dollar depreciated after the Plaza Accord in 1984, MCPs started to decline.

Figure 2.
Figure 2.

Incidence of MCPs in Fund Membership

Citation: Policy Papers 2019, 015; 10.5089/9781498318860.007.A004

Figure 3.
Figure 3.

Incidence of Multiple Currency Practices by Region

Citation: Policy Papers 2019, 015; 10.5089/9781498318860.007.A004

Source: IMF staff reports and staff calculation.Notes: Countries in red maintained actual MCPs, while countries in blue maintained only potentiality-based MCPs. The findings are based on the latest Article IV staff reports issued before end-2017.

20. Developments in the number of countries with MCPs in the 1990s and early 2000s reflect a global liberalization trend. Many EMDCs gradually opened up their markets and liberalized their exchange systems, particularly in Europe owing to regional integration efforts. A sharp rise in the number of countries with MCPs in the first half of the 1990s reflected the expansion of IMF membership; 25 new members joined the IMF by 1993, following the collapse of the communist bloc in 1989. Many of these MCPs were eliminated gradually in the 1990s. Advanced economies stopped using MCPs by 2000, and many EMDCs eliminated MCPs and accepted Article VIII obligations in the late 1990s. The share of countries with MCPs in international trade declined from about 12 percent in 1982 to about 4 percent by 2017. The decline is more evident among EMDCs, with the share of countries with MCPs falling from 40 percent to 10 percent of EMDC trade during the same period. Likewise, countries with MCPs represent an increasingly smaller share of world GDP. Between 1982 and 2017, their share dropped from 14 percent to 5 percent of world GDP and from 56 percent to 13 percent of EMDC GDP.

21. Since the Global Financial Crisis (GFC), the number of countries maintaining MCPs increased marginally. At the end of 2017, there were 45 MCPs maintained by 28 countries. Recent MCPs have been found mostly in lower-middle and low-income countries, in particular in Africa and Middle-eastern regions. Some countries introduced MCPs as a result of BOP difficulties and foreign exchange shortages, but many MCPs (in particular potentiality-based MCPs) arose in efforts to move toward a more liberalized FX market. Typical examples are MCPs arising from FX auctions and the computation of official exchange rates under a floating exchange rate regime. Reflecting these, MCPs have become more prevalent among countries with more flexible exchange rate arrangements.

B. Characteristics and Frequency of MCPs

22. The evolution in the number of MCPs roughly follows the number of countries maintaining MCPs. The total number of MCPs declined from its peak of 102 in 1986 to a trough of 29 in 2006, while it has been on a rise recently (Figure 4). Countries that are found to maintain MCPs usually maintain more than one, with the average number of MCPs per country ranging between 1 and 2.

Figure 4.
Figure 4.

Type of MCPs

Citation: Policy Papers 2019, 015; 10.5089/9781498318860.007.A004

Sources: IMF staff reports and staff calculation. Notes: “Other” includes, among others, MCPs from a BPA, margin requirement, and broken-cross rates.

23. The composition of MCPs has changed since the 1980s. The use of the foreign exchange system to discriminate against certain transactions (e.g., subsidies, tax, guarantees) has declined. At the same time, the share of MCPs arising from the spread between the formal and the parallel markets has remained sizeable (9 MCPs as of end-2017). MCPs have been found increasingly on the basis of “potentiality” i.e., due to the lack of a mechanism to keep the spreads among different market segments within a two percent range, without necessarily observing de facto spreads among the effective rates (Figure 5). Recently, the share of such MCP findings represent about one third of total MCP findings.

Figure 5.

24. Potentiality-based MCPs arising from lagged calculation of official rates represent a sizable share of recent MCP findings (10 out of 13 MCPs related to official transaction rates). An MCP arises when the authorities use an official exchange rate that is calculated based on the previous day’s market transactions for official transactions, thus creating the potential for a deviation of two percent or more between the official and market exchange rates. Often, the use of lagged official exchange rates reflects the lack of capacity to calculate the market average exchange rate on a real-time basis.

25. Another type of potentiality-based MCP that has recently become common arises from multiple price auctions. Official multiple price FX auctions give rise to an MCP unless there is a mechanism in place which ensures that exchange rates of accepted bids do not deviate by more than 2 percent. Countries typically operate auctions to distribute FX among market participants when FX supply is concentrated or accrues to the authorities and the interbank market is shallow; a multiple price format is often used to promote responsive bidding behavior and facilitate price-discovery. MCPs arising from multiple price auctions were rare before 2000. Examples of MCPs from multiple price auctions in the 1980s included those found in Jamaica, Bolivia, Chile, and Ghana. Recently, multiple price auctions have been used more widely, in part reflecting the efforts to introduce a currency price mechanism in countries with less developed FX markets. Since 2000, a total of 15 countries have been found to maintain MCPs from multiple price auctions due to the potential spread among winning bids (Angola, Burundi, Egypt, Guinea, Honduras, Hungary, Jamaica, Mongolia, Mozambique, Myanmar, Nigeria, Sierra Leone, Uganda, Ukraine and Zimbabwe).

26. The use of taxes on the conversion or transfer of FX as well as subsidies and guarantees for BOP and non-BOP reasons has been declining.

  • Tax: Of 67 MCPs arising from exchange taxes in the last 35 years, about a third target invisible transactions, such as foreign exchange for travel and profit remittances. Many of these taxes were introduced for fiscal reasons and/or maintained at the time when the country joined the IMF, while some countries also introduced such taxes to address BOP difficulties. About half of tax-related MCPs target trade transactions (import or export), often introduced in the context of BOP pressures. Examples include taxes on FX proceeds from exports, and taxes on purchase of FX for import transactions to dampen demand for FX. MCPs arising from taxes targeting trade transactions were often eliminated along with trade liberalization, or replaced by a trade tariff. Some taxes introduced for fiscal reasons were similarly eliminated as the capacity for tax collection improved.

  • Subsidies: Only one subsidy-related MCP is currently in place, as a result of a legacy of schemes introduced several years ago. Exchange subsidies used to be employed as part of trade policy, with incentives given to non-traditional export sectors to promote exports, often when exchange rates were overvalued. However, the use of such subsidies largely disappeared by 2000, with only a few new MCP findings since (Rwanda, Zimbabwe, Colombia, Iran, and Mauritius22).

  • Guarantees: Exchange rate guarantees are used to reduce FX risks for certain transactions or entities, sometimes following a large devaluation. MCPs from exchange rate guarantees typically disappear naturally as guarantee schemes expire. Currently there is only one such MCP (Tunisia).

  • Preferential exchange rates: Some countries resorted to preferential exchange rates when faced with BOP pressures. In such cases, different exchange rates were applied to FX purchases for essential and non-essential imports, or for the surrender of (certain) export proceeds.

27. MCPs arising from spreads between official and parallel markets remain sizable.23 These MCPs frequently arise from the imposition of exchange restrictions in the official market (e.g. FX rationing and/or prioritization). The geographical focus shifted from Latin America, where such MCPs were prevalent in the 1980s and 1990s due to BOP difficulties, to other regions, most notably to countries in Africa with BOP difficulties typically resulting from fiscal expansion at the end of the 1990s. Recently, parallel market premia emerged in several commodity-exporting countries that experienced terms-of-trade shocks. The removal of MCPs arising from parallel market premia often took place in the process of unifying exchange rates supported by an IMF-supported program.

28. Other types of MCPs have significantly decreased in the last three decades.

  • MCPs arising from margin requirements and unremunerated advanced import deposit requirements had been frequently used for BOP reasons in the 1980s, but hardly any existed in the 1990s and 2000s.

  • MCPs arising from bilateral payment agreements (BPA) have virtually disappeared. There were 11 MCPs arising from BPA in the period since 1980, with about half adopted by Central and Eastern European countries under the Council of Mutual Economic Assistance (CMEA) agreement. As these countries established currency convertibility, the relevance of BPAs has declined. Currently, there is only one MCP arising from a BPA (Democratic Republic of Congo).

  • Broken cross rates have been rare and are usually unintentional. While broken cross rates were often used in the early years of the IMF to implement trade policy, there have been only 10 instances of broken cross rates since 1980. Six of them existed historically (before 1980 or preexisted when a country joined the IMF). Three cases of broken cross rates have been identified since the mid-1990s (Jamaica, Myanmar and Belarus); they typically arise when the authorities update the official rate only infrequently (e.g. monthly) for less commonly used currencies.

29. MCPs applying solely to capital transactions have also been rare. Some countries maintained a dual exchange system for current and capital market transactions, giving rise to MCPs. Dual systems maintained by Belgium until 1982, and South Africa and countries in its currency zone in the 1980s are typical examples. In both cases the MCPs later became subject to approval as some current transactions were also channeled to the capital market in which exchange rates were determined more freely. These episodes indicate administrative difficulties in separating FX markets for capital transactions from current transactions in practice. There were a few other types of MCPs solely related to capital transactions. For example, MCPs with respect to capital transactions arose from fees under the debt-equity conversion program in Philippines (1979–83), from separate exchange rates on the purchase and sale of proceeds by non-residents in India (1983–87), low premia on FX swap arrangements in Argentina (1983–1990), conversion and cancellation of loans covered by exchange rate guarantees in Argentina (late 1980s), tax on some capital remittances in Mauritius (1991–94), and from FX swap schemes in India (2014).24

30. The type of MCP adopted is typically determined by their motivation. Countries facing BOP difficulties and foreign exchange shortages tend to create segmented markets, including illegal parallel markets, by allowing only certain priority transactions to take place at a preferential exchange rate. In similar circumstances, other countries resort to taxes or subsidies to provide specific incentives or disincentives for particular exports or imports to improve their current account balance. Exchange taxes and subsidies were also used to help develop certain export sectors in the 1980s, while the use of MCPs for such purposes has been limited recently. In the absence of BOP difficulties, countries under a transition to more flexible exchange rate arrangement tend to introduce multiple price auctions or use lagged official rates, giving rise to relatively less distortionary MCPs.

C. Approval of MCPs

31. Members have generally quickly eliminated newly identified MCPs, but there are some exceptions. MCPs on average last for about 4.4 years, while the median and mode durations are 3 years and 1 year, respectively.25 More than 20 percent of MCPs were eliminated in a year after they had been identified. Eight percent was eliminated in less than a year; some of the MCPs were eliminated before they were discussed at the IMF board.26 The average duration is relatively long because of a few MCPs that lasted for a long time (e.g. among the current MCPs, 2 have been in place for more than 25 years).

32. Long-lasting MCPs are explained by various reasons. Taxes represent the largest group among MCPs lasting more than 10 years, suggesting that the authorities often need time to reform the tax framework before eliminating MCPs. It may also indicate the difficulty of giving up the convenience of collecting taxes through the exchange system. MCPs arising from potential spreads owing to lagged official exchange rate calculation may also be difficult to eliminate quickly if the FX markets are not yet well-developed to provide reliable market quotes instantaneously. Elimination of legacy schemes which give rise to MCPs (e.g. guarantee scheme, inoperative BPA) may also take a long time due to the length of the commitment. In some extreme cases, long-lasting MCPs merely reflect delayed assessments by the IMF, especially in countries where regular surveillance was interrupted (e.g. Eritrea, Somalia, Syria). In these cases, MCP findings remain until formal assessments revise them even though underlying measures may have been eliminated.

33. The Executive Board has sparingly approved MCPs, with only about a third of MCPs approved for at least one year during their existence. Among a total of 506 MCPs reported in staff reports since 1980, 108 measures (21 percent) were approved, 79 measures (16 percent) were approved only for a part of their duration, 298 measures (56 percent) were never approved, and the rest did not require approval by the IMF (i.e. MCPs solely related with capital transactions or Article XIV MCPs).

34. In the majority of cases, approval was not granted because the MCP was not considered temporary. When IMF staff reports elaborate on the reasons for not recommending the measure for approval, they refer to the lack of a clear time-table for the removal of the MCP in more than 70 percent of cases. MCPs are typically considered temporary when the authorities commit to a timetable for removal. For that reason, it often happens that an MCP is approved in the last few years before elimination even though it was not approved initially. The second most frequent reason for non-approval is that the authorities did not request approval.27 About 7 percent of MCPs were not approved because they were discriminatory (e.g. BPA).

35. Approved MCPs differ by type, likely reflecting the ease of eliminating the measures and how distortive they are. More than a half of parallel market MCPs and MCPs from rates for official/specific transactions were not approved. However, MCPs from multiple price auctions, which typically were based on potentiality, were more likely to be approved. Measures that usually require more time for preparing for removal (such as tax or guarantee) tend to be approved at a later stage when the authorities can commit to a plan for their removal, even though they were initially not approved.

Figure 6.
Figure 6.

Approval of MCPs

Citation: Policy Papers 2019, 015; 10.5089/9781498318860.007.A004

D. MCPs and Article VIII Acceptance

36. Countries availing themselves of the transitional arrangements of Article XIV are more likely to maintain MCPs, but only a handful of these MCPs are Article XIV MCPs. Of 171 IMF members that have accepted Article VIII status28 by 2017, 19 (11 percent) maintained MCPs, while 9 out of 18 members with Article XIV status (50 percent) maintain MCPs.29 Of the MCPs maintained by Article XIV members, only a small fraction is covered by the transitional arrangements of Article XIV. In fact, currently, South Sudan is the only country that maintains an MCP under Article XIV. Article XIV MCPs were more prevalent in the 1980s and 90s, with a sizable rise in the early 1990s when countries of the former communist bloc joined the IMF. Following the elimination of these measures, Article XIV MCPs disappeared by 2000, until South Sudan joined the IMF in 2012 with Article XIV MCPs.

Figure 7.
Figure 7.

MCPs and Article VIII Acceptance

Citation: Policy Papers 2019, 015; 10.5089/9781498318860.007.A004

Macroeconomic Performance of Countries with MCPS: 1980–201630

A. Introduction

37. MCPs are often associated with poor economic performance, but the evidence is scarce and causality is difficult to establish. Distortions and short-term benefits of MCPs are documented in the literature (see first chapter of this paper). However, it is difficult to establish causality because of strong endogeneity in economic factors related to causes and consequences of MCPs.

38. This chapter empirically investigates causes and consequences of MCPs in the past 30 years. Macroeconomic indicators in countries with MCPs behave differently (see Figure 1 in the main paper). The analysis uses a unique database on MCPs constructed from IMF staff reports from the last 30 years. The data cover 154 IMF member countries (excluding advanced economies)31 from 1980 to 2016, annually. For countries that joined the IMF at a later time, the data starts from the year of membership. One set of regressions analyzes macroeconomic factors that are associated with the introduction and maintenance of MCPs. Another set of regressions analyzes growth and inflation impacts of maintaining MCPs. Overlap in variables in the two analyses indeed suggest strong endogeneity in causes and consequences of MCPs.

B. Performance Before MCPs

39. The analysis shows that countries introduce and maintain MCPs when faced with macroeconomic concerns (Table 2). Evidence from a logit data model shows that factors contributing to a greater probability of maintaining MCPs include high inflation, a depreciating real exchange rate, lower international reserves, and worsening terms of trade. These are typical characteristics of countries with BOP difficulties, and the finding is in line with the notion that countries resort to MCPs under BOP difficulties. In addition, low trade shares32 and low income levels are also associated with the introduction of MCPs. These appear to be related to other motivations for introducing MCPs, namely trade promotion and low capacity. On the other hand, the analysis does not find a statistically significant relation between the probability of MCPs and GDP growth or the fiscal balance.33

Table 2.

Logit Analysis: Causes of MCPs, 1980–2016

article image
Source: IMF staff estimation.Note: Logit estimation. Advanced economies are excluded from the regressions. All regressors are lagged. Z-statistics are reported in parentheses. The qualitative results remain the same when country fixed effects and year fixed effects are controlled for. ** p < 0.05; * p < 0.1.

C. Performance with MCPs

40. There is a fundamental endogeneity issue in assessing the impact of MCPs, which requires caution in interpreting the results. The endogeneity originates from the observation that countries with macroeconomic concerns and BOP difficulties tend to resort to MCPs. To minimize the impact, the analysis uses only lagged variables, and includes all regressors that are found to have a relationship with the causes of MCPs in the previous section.

41. Countries with MCPs tend to experience lower growth and higher inflation (Table 3). Linear regression analysis shows that countries with MCPs experience lower growth and higher inflation. In addition, potentiality-based MCPs do not have a significant impact on growth and inflation. Lower growth after the introduction of MCPs is in line with the observation of slower and smaller post-crisis recovery in GDP for countries with MCPs, while there is no marked difference in pre-crisis performance (see Figure 1 in the main paper). Inflation, on the other hand, is both related to factors leading to MCPs and affected by MCPs after the crisis.

Table 3.

Regression Analysis: Macroeconomic Impacts of MCPs, 1980–2016

article image
Source: IM F staff estimation.Note: All regressors are lagged, t-statistics are reported in parentheses. ** p < 0.05; * p < 0.1.

Selected Recent MCPS: Insights from Survey of Country Teams34

This chapter summarizes the results of a survey of selected IMF teams covering countries that currently have—or recently had—MCPs. The survey was conducted in January 2018, and reflects information on existing or eliminated MCPs as of that date. The surveyed countries with recent MCPs were Iran, Iraq, Maldives, Nigeria, Papua New Guinea, South Sudan, and Sudan. Surveyed countries with MCPs that had been eliminated were Myanmar35, Sao Tome and Principe, and Suriname. The responses were qualitative, and based on staff assessments of the economic backdrop, policy options, and domestic and spillover consequences of the MCPs. The responses complement the first chapter of this paper, which covered experiences with systems of multiple exchange rates until 1983.

42. In most of the cases, MCPs were introduced to mitigate BOP pressures (Table 4). The conditions that led to BOP pressures included terms of trade shocks (Nigeria, Sudan, Suriname) and weak macroeconomic policies, such as lack of fiscal discipline and monetary financing (Myanmar, Suriname).

Table 4.

Recent MCPs: Selected Experiences

article image
Souce: IMF country team surveys.

43. The official actions that led to MPCs under the survey usually took the form of limiting the availability of FX. Specific measures included limiting the supply of FX at the official exchange rate to commercial banks (Maldives); allocating FX to priority imports (Papua New Guinea); and setting preferential rates for certain imports (Iran).

44. In staff’s assessment, MCPs generally did not address the fundamental reasons leading to the BOP pressures. In general, staff recognized the difficult policy trade-offs and implementation challenges typically faced by countries in these situations. In a few cases, they mentioned specific difficulties in avoiding MCPs, for instance due to international sanctions (Iran), or the need for securing foreign reserves to import socially important goods (Papua New Guinea), Nevertheless, staff often found that alternative policies could have been employed, including:

  • Fiscal consolidation and phasing out of monetary financing (Iraq, Myanmar, South Sudan).

  • Fiscal and economic reform, including mobilization of non-oil revenues (South Sudan).

  • Increased exchange rate flexibility (South Sudan).

  • Targeted subsidies to imports, instead of allocating FX or setting preferential rates for certain imports (Papua New Guinea, Suriname).

45. According to staff assessment, MCPs were associated with various forms of domestic distortions and external spillovers, including:

  • Decrease in production. MCPs contributed to: increased unemployment (Papua New Guinea); heightened volatility in exchange rates (Sao Tome and Principe); increased uncertainty of doing business (South Sudan, Sudan, Suriname); shortages in imports of intermediate goods (Papua New Guinea); a loss of export competitiveness (Iran); lower government revenues (Suriname); and reduced incentives for domestic production when imports were subsidized (Iran).

  • Increase in inflation. This occurred with depreciations in parallel market exchange rates (Nigeria, Sao Tome and Principe, South Sudan).

  • Expansion of informal economy (Sudan) and rent seeking and weakened governance (South Sudan). MPCs also reduced the authority’s capacity to monitor FX transactions (Suriname).

  • Crowding out of bank lending. Instead, banks tried to make profits from the spread between the official and parallel market rates, with adverse impacts on economic and financial sector development (Iraq). Monetary transmission was also impaired in some countries (Maldives).

  • External spillovers. MCPs contributed to: lower trade and FDI (Myanmar); higher cost of repatriating investment income from abroad (Iraq) and remittances (South Sudan); reducing capital inflows (South Sudan); and distorting trade (Sudan).

Survey Questionnaire Causes:

  • Under which circumstances has the country introduced the MCP(s)?

  • Are the MCPs related to restrictions on payments for current or capital transactions?

  • Had there been previous episodes of using MCPs in the past 10 years? When and under what circumstances?

  • Which factors have been contributing to the appearance of MCPs?

  • In the country team’s view, was the MCP necessary? If yes, why and for how long?

  • In the country team’s view, could the MCP have been avoided? If yes, how?

Impact on domestic economy and policy responses:

  • In the country team’s view, how effective have MCPs been in addressing the issue the authorities aimed to resolve? Why?

  • What other policies have been put in place to address the issue the authorities aimed to resolve?

  • What kind of distortions have MCPs per se caused to the domestic economy?

  • Which policies have been put in place to eliminate MCPs?


  • What distortions have MCPs caused to trading partners and investors?

  • How much extra costs do you estimate MCPs have effectively imposed on the country’s external trade and finance? In terms of foregone trade? In terms of foregone flows of capital? In political terms?


  • de La Torre, Augusto, and Alain Ize, 2013, “Exchange Rate Unification: The Cuban Case,” mimeo, World Bank.

  • Dornbusch, Rudiger, 1986, “Economic Adjustment and Exchange Rates in Developing Countries,” in Sebastian Edwards and Liaquat Ahamed (eds) Economic Adjustment and Exchange Rates in Developing Countries (Chicago: University of Chicago Press), pp. 14376.

    • Search Google Scholar
    • Export Citation
  • Fan, Emma Xiaoqin, 2004, “A Note on Dual/Multiple Exchange Rates,” ERD Policy Brief, Series No. 26, Asian Development Bank (Manila).

    • Search Google Scholar
    • Export Citation
  • Fleming, J. Marcus, 1974, “Dual Exchange Markets and Other Remedies for Disruptive Capital Flows,” IMF Staff Papers, Vol. 21 (1), pp. 127.

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 1981, “Review of the Fund’s Policy on Multiple Currency Practices,” The IMF Staff Report (Washington).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 1984a, “Review of Experience with Multiple Exchange Rate Regimes,” The IMF Staff Report (Washington).

  • International Monetary Fund, 1984b, “Review of Experience with Multiple Exchange Rate Regimes—Background Information,” The IMF Staff Report (Washington).

    • Search Google Scholar
    • Export Citation
  • Kanesa-Thasan, S., 1966, “Multiple Exchange Rates: The Indonesian Experience,” IMF Staff Papers, Vol. 13 (2), pp. 354370.

  • Kiguel, Miguel and Stephen A. O’Connell, 1994, “Illegal Exchange Rates in Developing Countries— Lessons from Eight Case Studies,” Policy Research Working Paper No. 1265 (Washington: World Bank).

    • Search Google Scholar
    • Export Citation
  • Lizondo, José Saúl, 1987, “Exchange Rate Differential and Balance of Payments Under Dual Exchange Markets,” Journal of Development Economics, Vol. 26 (1), pp. 3753.

    • Search Google Scholar
    • Export Citation
  • Sherwood, Joyce, 1956, “Revenue Features of Multiple Exchange Rate Systems: Some Case Studies,” IMF Staff Papers, Vol. 5 (1), pp. 74107.

    • Search Google Scholar
    • Export Citation
  • Vinelli, Paul, 1951, “The Currency and Exchange System of Honduras,” IMF Staff Papers, Vol. 1 (1), pp. 42031.

  • Woodley, W. John, R., 1994, “Exchange Measure in Venezuela,” IMF Staff Papers, Vol. 11 (3), pp. 33766.


Prepared by Tamon Asonuma, Swarnali Ahmed Hannan, Callum Jones, Suman Basu (all RES) and Silvia Sgherri (SPR).


MCPs became less prevalent after the 1980s due to both increased flexibility of exchange rates and increased capital mobility and there is limited additional studies in the literature since then.


We include Romania (late 1970s-80) as a representative country with limited trade and financial openness.


Lizondo (1987) shows theoretically that a once-and-for-all devaluation of the official exchange rate, without any accompanying policy changes, has only a transitory effect on the differential between the free and the official exchange rates and on the balance of payments. On the contrary, an increase in the rate of crawl of the official exchange rate is shown to have permanent effects on the both the differential between the exchange rates and the balance of payments.


MCPs are also adopted in response to the volatility of capital flows or one-way pressure on the capital account, particularly during periods of exchange rate volatility.


IMF (1984a) finds that episodes in Costa Rica, Dominican Republic, and El Salvador also fall in this category.


The tax revenue obtained from MCPs was 4.6, 20.5, and 5.2 percent of total revenue in Cuba, the Philippines, and Venezuela, respectively, over the years 1951–54.


Sherwood (1956) also notes that the exchange tax imposed in the Philippines in 1952 helped to restrain the effective demand for imports, which was noted as an important consideration of the authorities during a time of BOP difficulties.


Ghei and Kiguel (1992) mentions that parallel FX market can also provide insulation—though limited—in the short term.


MCPs could be designed in a such way that they may exclude intermediate goods which, in turn, affect export and FX-generating sectors.


Cuba implemented a 0.25 percent tax on most money sales in 1925, gradually increasing it to 2 percent by 1943, and levied on most sales of FX and exports of securities. In 1956, the receipts from this tax contributed about 5 percent of general government revenue.


Exchange tax in the Philippines was initially intended in 1951 to be a temporary two-year measure, but was extended at various times, and by 1954 it raised about 20 percent of the total revenue.


The review examined the historical importance of multiple exchange rates, discussed the main reasons adduced by members that resorted to multiple exchange rate systems in recent years and described the outcome of these systems (Review of Experience with Multiple Exchange Rate Regimes, March 19, 1984; and Review of Multiple Exchange Rate Regimes – Background Information, March 20, 1984 ).


See Review of Experience with Multiple Exchange Rate Regimes, March 19, 1984.


Examples are the Dominican Republic (under-invoicing of minor exports), Sierra Leone, Somalia (leakages associated with inflows of private remittances), and Sudan.


For example, Costa Rica between 1976 and 1980; Ecuador between 1981–82; and Sudan, 1982.


For example, Argentina between 1981 and 1982; Costa Rica between 1980–82; and Ecuador between 1981–82.


For example, in Costa Rica, several stabilization plans were initiated over the 1980s, but uncertainties surrounding exchange rate determination under a dual market system probably contributed to exacerbate domestic inflationary expectations and pressures. See also the examples of Romania and Sudan.


Fleming (1974) also notes that taxes and subsidies on capital transactions would be far less flexible than MCPs and would therefore be less effective in mitigating short-term BOP imbalances.


Prepared by Chikako Baba (MCM). The construction of the MCP database was assisted by Svetlana Popova (MCM) and Amanda Kosonen (external expert).


There is no established database on MCPs before 2005, when the AREAER database started to include the information on “restrictions and/or multiple currency practices” as reported in the last IMF staff report issued by December 31 of the respective year. The information on multiple currency practices for 1980–2005 was collected from IMF staff reports by reviewing all reports that contained the term “multiple currency practice(s)”. The information on MCPs is based on the last IMF staff report issued to the IMF Executive Board by December 31 of the respective year. Therefore, the first year of an MCP in the data corresponds to the first year of identification (rather than the actual introduction of the measure), and the year of elimination corresponds to the first year that the elimination was reported in an IMF staff report.


According to the guidelines of the Exchange Rate Support Scheme, no claims under the scheme will be considered after July 16, 2018.


Parallel markets in this context refer to FX markets that are not regulated or closely supervised by the authorities, and can include transactions by FX bureaus or illegal black markets for FX.


The years in parentheses refer to year of issuance of staff reports that contain information on the MCPs.


MCP findings are often made with a lag when IMF staff do not have sufficient information to finalize assessments. Hence, the duration of underlying measures that give rise to MCPs can be longer than these years.


IMF staff reports do not necessarily mention MCPs that were introduced and removed between two consecutive staff reports, so the actual number of MCPs that were eliminated shortly after their introduction may be larger.


There are overlaps, because the authorities may not seek approval unless they consider they satisfy conditions for approval (e.g. time-table for removal).


Members that chose to avail themselves of the transitional arrangements under Article XIV may maintain and adapt to changing circumstances exchange restrictions and MCPs that they had in place when they joined the Fund. The provision applies only to the measures maintained at the time of membership, and all members are subject to Article VIII obligations with regard to new exchange restrictions and MCPs as soon as they join the Fund; no formal acceptance of these obligations is necessary. Once the member formally notifies the Fund of acceptance of Article VIII obligations, it can no longer rely on Article XIV, Section 2 to maintain or adapt the exchange measures it had in place when it joined the Fund.


As of end-2017, the member countries that make use of the transitional arrangements under Article XIV are Afghanistan, Angola, Bhutan, Bosnia and Herzegovina, Burundi, Eritrea, Ethiopia, Iraq, Kosovo, Liberia, Maldives, Myanmar, Nigeria, São Tomé and Príncipe, Somalia, South Sudan, Syria, and Turkmenistan. Kosovo accepted Article VIII obligations in January 2018.


Prepared by Chikako Baba (MCM).


Advanced economies are excluded because their use of MCPs almost disappeared before the analyzed period. The results are qualitatively similar if advanced economies are included.


Trade openness can affect the likelihood of MCPs in different ways. For MCPs introduced to deal with BOP difficulties, countries with higher trade shares would expect greater policy impact of the measure and hence are more likely to introduce MCPs. On the other hand, MCPs introduced to promote trade are more relevant for countries with lower trade shares. The latter impact appears more relevant for MCPs in the early periods (in the 1970s and 80s), and the regression results showing the relationship between lower trade shares and the probability of MCPs only in the early period are consistent with the view that the motivation for MCPs has shifted over time.


The regressions with the fiscal balance are reported separately because inclusion of this variable significantly reduces the number of observations.


Prepared by Masashi Saito (SPR).


Myanmar maintains one MPC from multiple price auctions, but the survey focuses on those eliminated in 2013.

Review Of The Fund’s Policy On Multiple Currency Practices: Initial Considerations
Author: International Monetary Fund. Legal Dept., International Monetary Fund. Monetary and Capital Markets Department, International Monetary Fund. Strategy, Policy, &, Review Department, and International Monetary Fund. Research Dept.