The Role of the IMF
Financing and Its Interactions with Adjustment and Surveillance

The Changing International Environment

Compared to the world conceived at Bretton Woods, four major developments have changed the international environment in which the IMF operates. First, a system linked to gold of fixed parities with infrequent adjustments has given way to much greater exchange rate flexibility. Second, the size and agility of private capital flows have increased greatly. Third, countries have become more integrated, through increased openness to both trade and capital flows, as barriers to trade in goods and assets have been removed and technological changes have reduced transaction costs. Fourth, the membership of the IMF has expanded enormously, and the institution has become universal, increasing the potential demand for IMF resources on the part of many of the newer members, which sometimes need substantial balance of payments financing.

Exchange Rate Flexibility

The move to floating exchange rates among the currencies of the largest industrial countries in the early 1970s fundamentally altered the nature of the international monetary system by removing any link between the value of money and the price of gold, and by freeing monetary policies to be directed toward domestic objectives. For those countries that adopted floating exchange rates--including the three largest industrial countries--maintaining a fixed parity was no longer a critical concern of economic policy; these countries typically used the increased flexibility to give greater weight to key domestic policy objectives. Current account imbalances did not disappear, nor were they more rapidly eliminated, as a consequence of floating exchange rates. Official reserves continued to be held by countries with floating exchange rates (Table 1), and governments continued to intervene to influence the course of market-determined exchange rates. However, without the obligation to intervene to defend pegged rates, the objective of most intervention changed, and official settlements imbalances lost most of their earlier significance.

Table 1. Ratios of Nongold Reserves to Imports of Goods and Services1
(In percent)

1970 1975 1980 1985 1990 19942

Industrial countries 13.4 13.0 13.2 13.4 15.3 17.6
Developing countries 19.4 32.0 29.2 27.0 28.6 31.0
 Countries with recent debt-
  servicing difficulties
17.0 25.5 31.2 19.3 20.8 39.1
 Small low-income economies 17.1 12.1 10.0 12.0 10.6 27.4
Countries in transition . . . . . . . . . . . . . . . 16.3
 Former Soviet Union3 . . . . . . . . . . . . . . . 13.2

Source: IMF, World Economic Outlook and International Financial Statistics databases.
1End-of-year reserves as ratios of imports of goods and services during the year. Imports include interest payments on debt where data are available. Country groups are consistent with those used in the May 1995 World Economic Outlook; some of these groups have experienced changes in membership during recent years.
2Data for some countries are staff estimates.
3Excludes Tajikistan. Imports exclude trade within the former Soviet Union and, in most cases, do not include services.

The change in the fundamental character of the international monetary system brought by floating exchange rates necessarily created important changes in the IMF's role. IMF surveillance over the international monetary system and over members' economic performance and policies (especially their exchange rate policies) replaced the earlier responsibility to supervise exchange rates under the par value system. As the monetary policies of the major countries now provided the nominal anchor for the international monetary system, an essential part of the IMF's role was to ensure that those policies were neither inflationary nor deflationary. IMF financing of payments imbalances of the industrial countries, in support of established or adjusted exchange rate pegs, generally ceased, and, over time, the focus of IMF financial support shifted to developing countries.7

It is difficult to reach a definitive judgment about whether, on balance, the move to a system of floating exchange rates has improved or worsened the overall performance of the world economy.8 The experience with the defense of pegged exchange rates since the collapse of the Bretton Woods system has made clear that the relevant counterfactual for assessing the consequences of maintaining a global system of pegged exchange rates would have required significant modifications in the policies actually pursued by the largest industrial countries. Simulation exercises to assess such counterfactuals using multicountry econometric models generally suggest that exchange rate pegging is suboptimal relative to policies directed at domestic economic stabilization.9 However, the argument remains that discipline exerted on national economic policies would have been stronger under a global system of pegged exchange rates and that such discipline would have helped to avoid some important policy mistakes, notably the excessive monetary accommodation of the upsurges of inflation in the mid- and late 1970s. Nevertheless, as a practical matter, it is clear that policymakers in the three largest industrial countries generally have no interest in, and see no benefit from, moving to an international monetary system that would impose significant constraints on their ability to direct their own national monetary policies in accordance with their domestic objectives.

While floating exchange rates have the virtue of allowing greater independence of national monetary policies, at least for the largest countries, it is far less certain that floating exchange rates necessarily produce optimal patterns of adjustment in countries' balance of payments. In particular, there appear to be instances in which market forces drive exchange rates in a manner contrary to the needs of orderly balance of payments adjustment, as seems to have occurred in the United States in the first half of the 1980s, and especially in late 1984 and early 1985 (Chart 1). Alternatively, when financial markets suddenly sense the need for a correction in a country's balance of payments, they may drive exchange rates in the right direction but to an extent that appears to force an unduly large or rapid correction. For example, the large depreciations of the Italian lira and the Swedish krona since the summer and autumn of 1992 have helped to bring large and rapid improvements in the current accounts of both countries (see Chart 2). However, it is arguable that somewhat less exchange rate depreciation would have allowed better-balanced economic recoveries with less danger of an acceleration of inflation.

Chart 1. United States: Real Effective Exchange Rate, Current Account Balance,
and General Government Fiscal Position

chart 1
Source: IMF, Research Department and World Economic Outlook database.
1Based on normalized unit labor costs in manufacturing.

Chart 2. Italy and Sweden: Real Effective Exchange Rate
and Current Account Balance

chart 2 (two graphs)
Source: IMF, Research Department and World Economic Outlook database.
1Based on normalized unit labor costs in manufacturing.
2Based on consumer price index.

Another example of the problems caused by floating exchange rates is the sharp depreciation of the Mexican peso during the early part of 1995, which has contributed to a rapid reduction of Mexico's current account deficit, but not without large output losses and an upsurge of inflation. On the other side of the ledger, the strong appreciation of the Japanese yen since the summer of 1992 appears to be self-defeating in the short term because Japan's current account surplus has risen, as a result of the J-curve, and because the yen appreciation has tended to undermine economic recovery.

More generally, it is often the case that exchange rate adjustment alone is not the most efficient or desirable method for correcting a payments imbalance. Currency depreciation generally needs to be accompanied by other measures to bring about the reduction in aggregate demand required to close the saving-investment imbalance and to resist the excess demand resulting from the increase in net exports. Market forces that induce changes in exchange rates generally cannot by themselves assure durable improvement in a country's external payments position, but they may in some cases reinforce the needed policy discipline.

Beyond these theoretical considerations, the empirical evidence clearly shows that for most countries reducing payments imbalances under the regime of generalized floating exchange rates has not been a wholly painless process. Typically, correction of an unsustainable current account deficit has been accompanied by large output losses, as well as by real exchange rate depreciation. Table 2 presents data for selected countries that substantially reduced large current account deficits. For the industrial and developing countries presented, the current account change was associated with both a slowdown in output growth and a real exchange rate depreciation. Output losses were particularly large for some developing countries in recent years, notably Turkey, Mexico, and the Philippines.

The continued holding and use of international reserves, especially by developing countries, is further evidence that a purely market-driven adjustment of exchange rates is not viewed by those countries as optimal (see Table 1). For industrial countries, which generally enjoy favorable access to world financial markets, the cost of holding reserves (the difference between borrowing costs and the return on reserves) is usually relatively low. For many developing countries, in contrast, the cost of holding reserves is quite high. Accordingly, the decision by many developing countries to hold reserves presumably reflects a preference for using reserves to meet payments imbalances, rather than relying exclusively on the alternative of exchange rate adjustment. In this regard, it is noteworthy that the level of reserve holdings relative to imports has risen since 1970 (Table 1), and that the use of reserves (Table 3) by both industrial and developing countries is relatively large in comparison with IMF quotas.

Table 2. Selected Corrections of Large Current Account Deficits
and Associated Output, Real Exchange Rate, and Reserve Changes
(In percent)

Years Change in Ratio of Current Account to GDP Change in Rate of Growth of GDP Change in Real Effective Exchange Rate1 Change in Reserves2 Memorandum:
Initial Current Account Ratio

Industrial countries
Iceland 1982­83 6.3 ­4.3 ­17.0 ­30.9 ­8.3
Greece 1985­86 4.2 ­1.5 ­9.6 ­8.2 ­10.3
Italy 1992­93 3.5 ­1.4 ­17.1 ­46.4 ­2.3
Finland 1992­93 3.6 ­2.63 ­26.6 ­21.9 ­4.6
Developing countries
Mauritania 1988­89 21.4 ­0.9 ­8.9 ­28.0 ­6.4
Chad 1986­87 14.9 ­2.83 ­24.5 ­54.0 ­7.9
Venezuela 1988­89 14.7 ­1.63 ­3.3 ­58.7 ­9.2
Ecuador 1982­83 7.5 ­4.1 ­7.0 ­51.5 ­8.5
Mexico 1982­83 7.3 ­7.4 ­40.3 ­70.0 ­3.7
Turkey 1993­94 5.6 ­13.2 ­17.0 ­7.0 ­3.7
Mali 1986­87 5.3 ­7.3 ­19.6 ­57.7 ­10.3
Philippines 1983­84 4.8 ­9.2 ­16.8 ­5.9 ­5.3
Kenya 1980­81 4.0 ­1.8 ­4.5 ­33.8 ­12.3

Source: IMF, World Economic Outlook and International Financial Statistics databases.
1Change of relative consumer price indices over the two years.
2In first year.
3Average output growth in the two years.

Expanded Private Capital Flows

As has been well documented elsewhere (for instance, by Mussa and others (1994), and Goldstein and others (1993)), the size and agility of private capital flows have vastly expanded in recent decades. This development has had two major implications for the international monetary system. First, it has made it more difficult to maintain pegged exchange rates when the market comes to believe that the exchange rate peg may not be viable--a difficulty that may have been compounded by the fact that the stock of reserve assets has not grown commensurately with private capital flows.10 Second, it has allowed countries to finance current account deficits (and surpluses), presumably benefiting both creditors and debtors, but also apparently delaying desirable adjustment in some circumstances.

Table 3. Selected Industrial and Middle-Income Developing Countries:
Monthly Changes in Gross Reserves (1985­93)
Relative to IMF Quota

Standard Deviation Maximum Loss (In percent of currentquota) Month/Year of Maximum Reserve Loss

Industrial countries
Australia 19 56 1/92
Austria 30 108 3/91
Belgium 16 49 9/92
Canada 16 49 9/92
Denmark 61 215 1/93
Finland 60 222 10/91
France 20 64 11/93
Germany 65 281 10/92
Greece 61 116 10/93
Iceland 32 99 9/92
Ireland 54 240 9/92
Italy 36 127 7/92
Japan 22 61 3/90
Netherlands 22 33 3/93
New Zealand 31 97 6/88
Norway 55 240 11/92
Portugal 106 579 9/92
Spain 83 530 9/92
Sweden 78 317 11/92
Switzerland 46 120 1/93
United Kingdom 14 31 9/92
United States 5 13 3/91
Middle-income developing countries
Argentina 24 57 3/91
Bolivia 18 55 1/90
Brazil 30 60 1/86
Chile 23 35 1/86
Colombia 23 71 10/93
Costa Rica 23 72 5/90
Egypt 34 153 11/90
El Salvador 21 61 5/90
Hungary 27 45 12/92
Indonesia 18 31 5/90
Israel 46 98 10/91
Jordan 57 334 8/91
Korea 58 196 12/89
Malaysia 68 193 12/92
Mexico 52 167 11/93
Morocco 20 38 3/89
Paraguay 39 160 10/92
Peru 18 47 9/91
Philippines 30 96 5/93
Poland 15 54 12/90
Singapore 93 161 3/91
South Africa 8 23 9/92
Thailand 37 59 7/92
Tunisia 27 93 4/91
Turkey 36 112 12/93
Uruguay 19 103 11/86
Venezuela 15 28 7/90

Source: IMF staff calculations.

Maintaining exchange rate pegs

The crises of 1992­93 involving the Exchange Rate Mechanism (ERM) of the European Monetary System (EMS) are the clearest recent example of the difficulties of maintaining exchange rates in narrow bands around central parities. The experience of this period suggests the importance of early adjustment of unsustainable exchange rates, the need for consistent and credible macroeconomic policies to validate a fixed peg, and the massive force of speculative flows when parities are not considered to be sustainable by the private capital markets. Indeed, the experience during the ERM crises indicates that, in the modern situation of international capital mobility, massive pressures can arise because of doubts about the sustainability of exchange rates, even if those exchange rates appear to be in line with fundamentals.

However, access to private capital flows can also assist in the defense of pegged exchange rates by enabling governments to secure additional (borrowed) reserves with which to intervene in the money markets. For example, during the ERM crisis in 1992, the United Kingdom arranged a foreign currency borrowing program valued at ECU 10 billion, and Sweden arranged for credits totaling ECU 31 billion, to help defend their exchange rates. In both instances, efforts to defend the official parities proved unsuccessful. Thus, while access to private capital flows may assist in defending pegged exchange rates, it is clearly no guarantee of success.

Financing current account deficits

Current account deficits can arise for various legitimate reasons (that is, not related to inappropriate policies), and increased availability of private capital flows can help to finance those imbalances. First, these deficits can result from long-term differences between saving and investment across countries. Countries with more profitable investment opportunities but inadequate domestic saving can benefit from foreign capital; likewise, countries with excess saving can obtain a higher return by investing abroad. Second, different cyclical positions of countries will lead to current account surpluses and deficits, as saving and investment typically move differently over the cycle. As a result, paths for consumption can be smoother than they would be if the current account were always balanced. Third, current account surpluses and deficits can result from unsynchronized portfolio diversification across countries (for instance, if faster liberalization in one country leads to net capital outflows). Portfolio diversification generally improves the trade-off between risk and return.

However, greater private capital flows may not always or necessarily lead to greater economic efficiency and welfare. Inflows may delay needed policy adjustment. For example, the persistent inflows of capital to the United States since the early 1980s have had the beneficial effect of helping to finance U.S. investment in circumstances of low national saving (and a substantial fiscal deficit). It might have been preferable, however, if sterner external discipline had induced more strenuous efforts to correct the underlying causes of the U.S. current account deficit. Alternatively, the euphoria concerning the prospects for monetary union in Europe led to massive inflows into high-inflation, high-interest rate EMS countries during the 1987­92 period. These inflows were often funded in currencies with lower interest rates (mainly the deutsche mark).11 Similarly, there was a sharp increase in capital flowing into emerging markets in 1992­93, complicating the task of economic policy in these countries because of their limited ability to sterilize the inflows and contributing to overvalued exchange rates. The counterpart of excessive inflows is the rapid withdrawal of capital when sentiment changes, which occurred in both cases cited above. Changes in sentiment led to a massive speculative attack against a number of European currencies, including some whose central parities were not obviously overvalued on competitiveness grounds, and to indiscriminate selling of Latin American bonds and stocks in the immediate aftermath of Mexico's devaluation of December 1994.

The point here is not that international capital flows are generally misdirected or that relatively free international capital mobility is generally bad--quite the contrary. The point is that the modern regime of international capital mobility is not perfect and that the IMF's financing and surveillance activities may therefore be able to help improve its performance.

Increased Integration of National Economies

Since the Second World War, trade in goods and in financial assets has increased enormously. To some extent, growth in the early postwar period involved merely a return of exports and imports to prewar levels as reconstruction proceeded and wartime controls were abandoned. However, trade continued to grow in later decades at a sustained pace that has exceeded GDP growth in both industrial and developing countries (Chart 3). This sustained growth has reflected a fall in tariff and nontariff barriers as well as a reduction in transport costs. The IMF has contributed to this process through its success in promoting widespread current account convertibility.

Chart 3. Trade Flows1 Relative to GDP
(In percent)

chart 3
Source: IMF, World Economic Outlook and International Financial Statistics databases, and staff estimates.
1 In goods and nonfactor services.

Integration of financial markets has been even more dramatic. One feature of this integration has been expanded capital flows, as discussed above. The immediate postwar period began with most countries (with the notable exceptions of the United States and Canada) having numerous restrictions on capital account transactions. Capital account convertibility has now become virtually complete among industrial countries and has been adopted by a number of developing countries. This expansion of capital account convertibility, coupled with technological and financial innovations, has increased cross-border claims enormously. For instance, the stock of international loans rose from 5 percent of industrial country GDP in 1973 to 19 percent in 1993.12 Developing countries have also gained access to private capital, in the form of bank lending in the 1970s and 1980s, and more recently in the form of portfolio flows and direct investment (Table 4).

Table 4. Net External Financing Flows to Developing Countries
and Economies in Transition
(In billions of U.S. dollars)

1987­88 1989­90 1991­92 1993­94

Developing countries
Non-debt-creating flows, net 31 53 66 118
Net credit and loans from IMF ­9 ­3 1 ­1
Net external borrowing 72 97 199 214
From commercial banks
21 29 26 ­28
From official creditors
38 43 35 29
13 25 138 213
Total, net external financing 94 147 265 332
Countries in transition
Non-debt-creating flows, net 1 1 11 24
Net credit and loans from IMF ­2 ­1 5 5
Net external borrowing ­25 2 4 9
From commercial banks
7 8 ­8 9
From official creditors
­4 10 41 10
­28 ­16 ­29 ­10
Total, net external financing ­26 2 20 37

Source: International Monetary Fund (1995b, Table A33).

Increased economic integration has resulted in gains from trade in both goods and asset markets that, in turn, have surely contributed to postwar prosperity. This increased integration has also meant that countries are increasingly exposed to shocks from abroad, through both trade and capital market linkages. To be sure, the game has been worth the candle; the benefits of integration have generally far outweighed the costs. However, the appropriate speed and necessary conditions for full capital market liberalization remain important policy issues. To the extent that the external shocks are small or are negatively correlated with domestic shocks, greater external exposure may not be harmful; nevertheless, the concern about external shocks or about shocks that may be amplified because of openness is legitimate. For example, neither the debt crisis of the 1980s nor the recent economic crisis in Mexico could have occurred to economies that were effectively closed to trade and capital movements.

For several reasons, developing countries tend to be more exposed to external shocks than the larger industrial countries. First, a number of developing countries are heavily dependent on export earnings of a few primary commodities, making their current account positions especially vulnerable to fluctuations in world prices and foreign demand. Second, these countries often do not have well-developed domestic financial markets, so they cannot fall back on efficient domestic sources of financing for shortfalls in government revenues. Third, developing countries cannot borrow abroad in their own currencies and generally have to rely on foreign currency financing (or reserve use) to deal with balance of payments shortfalls. When an adverse external shock or a domestic disturbance (including policy errors) generates a balance of payments financing problem, foreign currency financing tends to dry up, and the magnitude of the adjustment required to correct the payments imbalance is magnified. Fourth, many developing countries have accumulated large external debts in foreign currency; servicing these foreign currency debts requires either trade surpluses or other capital inflows. The existence of a large volume of international foreign currency claims (or, more generally, large foreign portfolio investments) exposes a country to changes in foreign investor sentiment, which could force the authorities to restrict domestic demand drastically to adjust the balance of payments. Fifth, analysis of data for internationally traded assets of developing countries suggests that these assets are subject to problems of financial market inefficiency. In particular, there is evidence of bandwagon effects, showing up in serially correlated asset returns, perhaps because of the portfolio shifts into and out of institutions specializing in emerging markets (International Monetary Fund (1995a)). Moreover, contagion effects were present in the weeks immediately following the December 1994 Mexican devaluation, as internationally traded asset prices of different Latin American countries facing different economic fundamentals moved together much more closely than during earlier periods (Table 5).

Table 5. Correlation of Returns on Selected Latin American Brady Bonds
with Returns on Mexican Brady Bonds,
January 1993­March 1995
(In percent)

Jan. 1993-
Jan. 1994
May 1994
Mid-May 1994-
Mid-Dec. 1994-
Early March 1995

Argentina 0.56 0.84 0.77 0.89
Brazil 0.22 0.51 0.45 0.73
Venezuela 0.55 0.65 0.47 0.78

Source: Folkerts-Landau and others (1995,Table I.4)

Expanded Membership

The context of the IMF's operations has changed greatly as a consequence of the substantial expansion of its membership. This expansion resulted from two principal events in the postwar period: the process of decolonization, which had begun in the late 1940s and largely run its course by the 1970s, and the recently initiated transformation of former centrally planned economies into market economies. These two events have made the IMF into a universal organization. At the same time, because most of the IMF's new members have not established assured access to private capital flows, the number of members that are likely potential users of IMF resources has continued to expand.

In particular, most transforming economies are still much less integrated into the world economy than many developing countries. Correspondingly, these transition countries have generally not been able to draw on private capital flows to anywhere near the same degree as middle-income and more advanced developing countries (Table 4), although some individual countries have attained market access or are soon to do so. Accordingly, many transition countries--like the poorer developing countries--have needed to rely on official financing for much of their balance of payments support. As most transforming economies do not benefit from large amounts of official bilateral financing flows, the IMF has been heavily involved in making balance of payments financing available to them, subject to comprehensive conditionality to help ensure that necessary structural policies are implemented in a framework of macroeconomic stability.

In sum, when projecting the potential need for IMF resources, the Bretton Woods conference did not anticipate the move to greater exchange rate flexibility and the pervasive access of industrial countries to private financing. However, it also did not anticipate either the growth in world trade and capital movements and the greater integration of national economies, which has left them more exposed to external shocks, or the expansion of the IMF's membership, which has created new demands for IMF resources by countries where the correction of payments imbalances may typically require somewhat more time than was envisioned at Bretton Woods. These evolving needs explain why the membership has approved periodic increases in IMF quotas.

7The last IMF-supported programs for industrial countries were implemented in Italy and the United Kingdom in 1977, after the general move to floating exchange rates. Both countries were experiencing balance of payments difficulties in the aftermath of the first oil shock, and the Italian lira and the U.K. pound were under severe downward pressure. The IMF, at that time, was still seen as playing an important role in helping to deal with these problems. Since the summer of 1992, the U.K. pound and especially the Italian lira have depreciated substantially, particularly against the currencies of other large European countries. These depreciations have helped to bring large improvements to the current accounts of both countries, and they have not involved the use of IMF resources.
8It is sometimes suggested that the general slowdown of industrial country growth since the early 1970s may be attributed in some significant measure to the collapse of the global system of pegged exchange rates. There is, however, no serious empirical evidence to support this hypothesis. For a review of the evidence on the effects of exchange rate volatility under floating rates, see Goldstein (1995).
9See, for instance, Taylor (1989), and Frenkel, Goldstein, and Masson (1989).
10Turnover on the three largest foreign exchange markets increased threefold between 1986 and 1992, while foreign exchange reserves of industrial countries increased by only 77 percent in U.S. dollar terms over that period. In April 1992, a survey of foreign exchange markets estimated that global daily turnover was $880 billion, compared with a stock of nongold foreign exchange reserves of industrial countries equal to roughly $414 billion (Mussa and others (1994)). Preliminary analysis of a more recent survey by the Bank for International Settlements of foreign exchange markets suggests that daily turnover now exceeds $1 trillion.
11This was termed a "convergence play" by Goldstein and others (1993).
12Figures quoted in Mussa and others (1994).