II Capital Flight: Concepts, Measurement, and Issues

International Monetary Fund
Published Date:
January 1987
  • ShareShare
Show Summary Details

Until recently, the issues raised by capital outflows from developing countries have been overshadowed by the concerns over the growth of their gross external debt. In the past few years, however, there has been an increasing recognition of the significance of these capital outflows, reflecting in part greater awareness of the sheer volume of assets that have been transferred abroad. For the group of capital importing developing countries, for instance, external assets are estimated to have amounted to some $500 billion at the end of 1985, of which only $150 billion was accounted for by official reserves. These countries’ total external assets, therefore, amounted to well over half of their external liabilities.

In the light of these magnitudes, capital flight has received increased attention as a possible cause for the protracted slow growth experienced by many developing countries in recent years and for the debt crisis. Moreover, capital flight has been identified by foreign creditors as a justification for not lending to developing countries. A recent Morgan Guaranty newsletter noted, for instance, that “Creditors, both private and official, are understandably reluctant to provide fresh funds unless the debtors put a stop to the capital flight.”1

Capital flight is, however, a somewhat elusive concept. From a conceptual standpoint, it is hard to distinguish from those outflows that can be considered “normal” transactions, while it is difficult to measure both empirically because statistics are weak and because it requires classifying data according to characteristics that are only loosely and indirectly related to the constructs being measured. Finally, the significance of capital flight is very much a function of circumstances that reach well beyond the asset transactions in question. In particular, whether capital flight engenders a transfer of real resources hinges largely on accompanying developments in external liabilities. In much of the 1970s, for instance, when foreign creditors were in effect willing to finance capital flight, its harmful effects stemmed more from the distortions that caused it than from the outflows per se. As foreign creditors became increasingly unwilling to lend to developing countries, however, the private sector’s demand for foreign assets was effectively “financed” through reductions in net imports, a different and much more serious consequence of capital flight.

In an attempt to shed some light on these matters, this chapter begins by focusing on capital flight in the way it is most often perceived—as a particular type of capital outflow. The first section defines the concept, and the chapter goes on to describe and assess the various techniques used to measure capital flight, to sketch the main trends since the mid-1970s, and to identify some of its proximate causes. In the two closing sections a broader view is taken, and the economic implications of capital flight are examined within the context of the overall external situation of developing countries.

Definition of Capital Flight

Although bouts of capital Right have been identified for periods dating back to the 17th century, and possibly before, there is not yet a generally accepted definition.2 One difficulty in formulating an acceptable definition is that “capital flight,” in its broadest sense, covers capital outflows that are motivated by widely different considerations that cannot easily be distinguished empirically. Consequently, a particular definition of capital flight as comprising the acquisition and holding of certain claims on nonresidents, or as arising out of investors’ motives for acquiring such claims, inevitably involves a large element of judgment. The variety of definitions that has been proposed reflects, in the main, the variety of observers’ judgments on the line dividing “normal” capital outflows and capital flight.

Although this distinction cannot be drawn finely, it is clear that while all capital flows are motivated by endeavors to maximize returns on capital for any given level of risk, the particular motivation for capital flight is rather more specific. Essentially, it is motivated by a resident’s concern that, if his wealth were held domestically, it would be subject to a substantial loss or impairment. Fears of capital loss tend to arise from risks of expropriation, debt repudiation, or exchange rate depreciation. Fears of capital impairment arise from the risks of new market distortions—for example, capital controls, taxation, and financial repression—that would reduce the value of an asset compared with its value if invested abroad.3

However, these general concepts are rather difficult to apply in practice. The concept of a market distortion that is sufficiently serious to cause capital flight, for example, is somewhat elastic, like the concept of capital flight itself. Many markets are subject to some distortion and this is accepted by the participants. However, if the distortions become sufficiently large, then participants are likely to seek to transfer their activities elsewhere. This transfer might involve simply a geographical relocation of transactions; for example, if domestic interest rates on bank deposits are controlled and uncompetitive, investors might shift their bank deposits offshore. Alternatively, the transfer might involve a shift from one class of transactions to another. Investors might seek, for example, new domestic or foreign financial instruments in preference to domestic financial instruments that are subject to interest rate controls. Similarly, the imposition of a tax on one class of assets might cause holders to shift to less heavily taxed domestic assets (tax-avoidance) or to transfer their wealth abroad (tax evasion or tax avoidance, depending on the legal circumstances).

To further complicate the matter, a shift of assets abroad may or may not involve capital flight, as the following example demonstrates. The emergence of favorable tax and regulatory regimes in offshore banking centers in the 1970s provided incentives for many banks to relocate some of their activities from the traditional financial centers in industrial countries to the offshore banking centers. This geographical relocation was accompanied by massive outflows of funds from the traditional financial centers into the offshore banking centers, which were matched by deposits of funds by entities located in offshore centers in traditional centers. Although banks were responding to a chance to avoid capital impairment—-in the form of taxation of financial transactions and regulatory requirements on the use of their funds—the large capital outflows were not generally viewed as capital flight. One reason may be the perception that the banks were responding to opportunities to increase their profits rather than to the risks of incurring losses.4 Another may be that the banks involved in these transactions were essentially exchanging assets of similar risk characteristics, whereas capital flight typically involves a discrete shift—an exchange of high-risk for low-risk assets in the portfolio of the investor. Lastly, shifting transactions offshore did not result in any significant change in the opportunities and constraints facing the ultimate holders of liabilities and claims, nor did it result in a net change in the availability of foreign exchange, or net savings, to the country concerned.

This example provides a useful insight into the nature of capital flight. A capital outflow must be motivated by the owner’s concern that, if the capital were retained as a direct claim on other residents, it would be subject to a reduction in value that could be avoided by acquiring claims on nonresidents. But although a necessary condition for flight capital, this is not a sufficient condition. To be classed as capital flight, the outflow must also be a response to losses and risks that are perceived to be “large” in relation to the capital deployed. The relatively small differentials between rates of return on banks’ capital obtainable in traditional and offshore banking centers are not sufficient to justify the resulting capital flows being treated as capital flight.

Capital flight is not, of course, necessarily embodied only in a current outflow of capital. It may, as discussed by Dooley, occur when there is a shift in residents’ motives for holding their stocks of foreign assets.5 Changes in circumstances—prompting fears of the introduction of a tax on domestic financial assets, for example—may cause investors to reappraise the value of their foreign assets relative to domestic assets; if the expected losses incurred as a result of repatriating foreign assets were sufficiently large, assets that had previously been accumulated abroad for nonflight purposes might be retained abroad for flight reasons. Consequently, capital flight may occur without any current outflow of capital across the exchanges, and, by the same token, it may be reversed without any capital reflows to the country in which the capital owner resides.

The above analysis suggests that capital flight may be defined as the acquisition or retention of a claim on nonresidents that is motivated by the owner’s concern that the value of his asset would be subject to discrete losses if his claims continued to be held domestically.6 Capital flight is not necessarily a problem from the individual’s point of view; indeed, it offers the capital owner the opportunity to protect his wealth from expected losses, and is therefore welfare-improving from his perspective. However, the avoidance of losses by some residents may merely shift the burden to others, either individually or, in aggregate, to society. When these costs become unacceptable—usually to the authorities—capital flight is viewed as a problem. The problem with capital flight is. therefore, not so much that there is a capital outflow, nor that there is a shift in sentiment for retaining outstanding stocks of foreign assets abroad; the problem is. rather, that resources escape those who seek to exercise some degree of control over how the funds may be used.

In general, governments have wanted to have some control over the disposition of private capital because they take a broader, and sometimes more optimistic, view of the returns obtainable on domestic investments than does the private sector, calculating social rates of return on domestic investments that are often higher than the equivalent private rates of return. They perceive, therefore, social gains from inhibiting private capital outflows. Consequently, many governments have imposed controls on capital outflows, with the intention of increasing the domestic savings available for domestic investment. Moreover, many governments have been willing to offer guarantees to foreign creditors and to pay market-related rates of return in order to attract inflows of foreign savings that would further increase the resources available for development.

The experience of the 1970s and 1980s indicates that, despite governments’ use of capital controls and moral suasion, private individuals in some developing countries have for a long period exported capital in order to avoid the foreign currency losses associated with holding wealth in domestically issued financial assets.7 Governments tolerated these private outflows for several reasons, the main one being that the countries retained ready access to foreign loans. This enabled them to remain net capital importers and to pursue their development plans without serious disruption. However, as foreign creditors became unwilling to undertake further lending to many of these countries in the wake of the debt crisis, governments more forcefully confronted the opportunity costs of continued private capital outflows. At this point, capital flight, and the stocks of private foreign assets that had been acquired earlier, began to be systematically described as a problem.

Recognizing the existence of a problem and appreciating its extent, however, are rather different matters. In this context, it is important to distinguish between “normal” capital outflows and capital flight. As argued above, “normal” capital outflows comprise all outflows that are not motivated by the attempt to avoid large losses. Normal outflows, therefore, include those resulting from households’ attempts to maximize returns through international portfolio diversification; enterprises’ efforts to promote trade through providing export credits, accumulating working balances abroad, and investing directly in the acquisition of productive capacity abroad; and commercial banks’ efforts to expand their activities through accumulating deposits with foreign correspondent banks and acquiring claims on nonresidents through portfolio and direct investment.8 Clearly, one would not wish to overstate the magnitude of the problem of capital flight by incorporating those normal outflows in a measure of capital flight.

Measuring Capital Flight

This discussion of the definition of capital flight suggests that a motivation-based definition is likely to have limitations, both conceptually and empirically. Indeed, capital flight is inordinately difficult to measure. Its relationship to capital outflows, as noted above, hinges on motives and perceptions that are both difficult to measure and subject to abrupt shifts in response to changing circumstances. Hence, very narrow measures that might be appropriate for certain countries or in certain circumstances might be wholly unrealistic for other countries or in other circumstances. The approach taken in this study is to consider a variety of measures that are representative of current empirical work on capital flight, with a view to finding a statistical measure that most closely corresponds to the motivation-based definition discussed above. The estimates of capital flight produced by others on the basis of these various measures are not reproduced here since, as Cumby and Levich found, significant differences in results may be attributed to differences in data used by the various analysts.9

Four measures of capital flight are examined: (1) a broad measure, defined as the identified acquisitions of external assets except official reserves, plus recorded errors and omissions in the balance of payments accounts. Errors and omissions are thus implicitly attributed in their entirety to capital transactions that can be regarded as capital flight;10 (2) a private claims measure, defined as the acquisition of external claims by the private sector (which includes deposit banks and the nonbank private sector), including recorded errors and omissions;11 (3) a narrow measure, defined as net short-term capital outflows of the nonbank private sector plus recorded errors and omissions, a measure which views capital flight as essentially similar to “hot money” flows;12 and (4) a derived measure. defined as that part of the whole economy’s stock of foreign assets that does not yield a recorded investment income, a measure which therefore encompasses both capital outflows and shifts in the motives for holding stocks of foreign assets.13

These measures obviously differ significantly in their analytical content. In the first place, the first three are, strictly speaking, measures of capital outflows, with no attempt to distinguish between flight and nonflight components. The underlying assumption of this approach is that it is reasonable to impute a motive to certain types of asset transactions at the time they occur. Judgments differ as to which set of transactions arc the most likely vehicle for capital flight, hence the important differences in the sectoral coverage of The three measures (Chart 1).

Chart 1.Measures of Capital Flight: Sectoral Coverage of Foreign Assets

The narrow measure is restricted to the identified acquisitions of short-term foreign assets by the non-bank private sector plus the errors and omissions recorded in balance of payments statistics. The presumption here is that capital flight is essentially a short-run, “hot money” phenomenon, and that a large number of capital flight transactions are hidden from the authorities and will only appear residually as errors and omissions in balance of payments statistics.14 The private claims measure has a broader coverage and is designed to capture any outflow of capital flight that originates from the private sector;15 it includes, in addition to the assets covered by the narrow measure, the foreign assets of deposit banks, outward equity investments (direct and portfolio investment) and the long-term assets of the nonbank sector.16 In principle, the broad measure incorporates a still wider range of foreign asset transactions, because it includes the non-reserve asset transactions of the monetary authorities and the asset transactions of the nonbank official sector. At first glance, the inclusion of these public sector asset transactions appears odd, since for most purposes it is not usual to consider the public sector as participating in capital flight. In practice, however, the non-reserve asset transactions of the monetary authorities and of the nonbank official sector, which does not include public enterprises, are usually relatively small for most countries; thus the sectoral coverage of the broad measure is often equivalent to that of the private claims measure,17 Moreover, there are significant statistical advantages that stem from the superior availability of data on this measure, and these more than offset the problems caused by the inclusion of a small number of public asset transactions.

MemorandumEstimated stocks of foreign assets at end 1985(billions of US$)


Of which:Countries








Monetary authorities16666101
Official reserves1505595
Nonbank official20911
Deposit banks752451
Equity investment963
Nonbank other1308743
Errors and omissions1118625

The main conceptual weakness of these three measures of capital flight is their inability to distinguish between outflows motivated by fear of loss and those caused by other reasons, which leads to inaccurate measurement. The broad and private claims measures, for example, include as capital flight private capital outflows that promote external commercial relations, both trade and financial. Conversely, the narrow measure is probably too restrictive; it includes nonflight short-term transactions and. more important, omits capita] flight effected through acquisitions of long-term assets, which may be relatively close substitutes for the short-term assets included in this measure. Is the acquisition of a house or an equity portfolio in the United States, for example, any less capital Right than a bank deposit? This omission of the acquisition of real and long-term financial assets is a serious shortcoming of the narrow measure.

Another conceptual weakness in these three measures is that they are only able to measure capital flight embodied in a capital outflow, and fail to identify the flight occurring when asset holders change their motive for retaining abroad their outstanding stocks of foreign assets. Indeed, the broad and private claims measures are formulated in such a way as to eliminate the possibility of considering how shifts in investors’ sentiments might cause capital flight, since by definition all stocks of private foreign assets recorded by these measures are viewed as having been accumulated by flight capital. However, this is inherently implausible because, as noted above, while all private foreign assets may at the limit be considered potential flight capital, it is evident that some assets are not originally accumulated for flight reasons. A measure covering this aspect of capital flight would clearly be advantageous.

In an attempt to avoid these conceptual problems. Dooley proposes the “derived” measure of capital flight as that part of a country’s stock of foreign assets which does not yield a recorded inflow of investment income credits.18 The presumption here is that only the retention of investment income abroad—oral least its retention outside the reach of the authorities of the country in which the asset owner resides—is indicative of flight concerns. An asset owner may wish to protect his property and investment income from domestic taxation, for example, or he may wish to avoid surrendering his foreign exchange revenues from investment income to the monetary authorities, perhaps because of fears of an imminent depreciation; he can obviously avoid both types of losses (and other losses) by accumulating foreign assets and retaining the consequent investment income beyond the jurisdiction of his authorities.

This measure is intuitively appealing on conceptual grounds for a number of reasons. First, capital flight measured in this way becomes a clearly distinct element of the private sector’s stock of foreign assets and allows for the fact that some private capital outflows and some private foreign asset-holdings are not motivated by flight considerations. Second, the measure captures capital flight that does not involve current exchange transactions and therefore allows for the shifts in investors’ sentiments that do not necessarily result in capital flows. And third, the measure incorporates an estimate of the government’s loss of control over resources—the investment income that is not subject to domestic taxation and the foreign currency accruals that are not made available to the authorities for reserve-management purposes, for example. The measure therefore is useful in indicating the extent to which capital flight might cause budgetary or reserve-management problems for the authorities and thereby lead them to view flight as a problem.

However, this measure, too, is conceptually unsatisfactory in some respects. The acquisition of foreign assets does not cease to be a problem or cease to be capital flight merely because the reinvestment abroad of the investment income earned on those assets is reported to the relevant authorities in the capital exporting country. If, for example, a resident were to report the investment income earned on foreign assets to the authorities, but reinvest the income abroad, he would eliminate his scope for tax avoidance, but would not eliminate the opportunities to avoid other forms of capital loss or impairment, such as losses caused by an exchange rate depreciation or by severe domestic financial repression. However, the measure is empirically useful for constructing estimates of the minimum level of capital flight, since it seems unlikely that the assets it includes would have been acquired as a result of “normal” capital flows. Thus, the derived measure can be used in conjunction with one of the other measures of capital outflows to delineate statistically an upper and lower bound within which actual capital flight probably falls,19 The upper bound (or maximum potential flight-motivated asset transactions) can be defined by one of the broader measures of capital outflow, and the lower bound by the derived measure.

These conceptual differences between the four measures, and especially the differences in coverage, naturally lead to different estimates of capital flight. In general, one might expect the broad measure to produce the highest estimate, followed by the private claims and the narrow measures, in that order, with the ranking of the derived estimate being dependent on the measure of outflows to which it is applied. However, data problems can be large enough to offset differences in coverage and produce an unexpected ordering of the estimates.

The most serious data problems affecting the four measures discussed here stem from limitations in the sectoral disaggregation of money and banking statistics, and shortcomings in national balance of payments statistics. The broad measure, for example, is particularly sensitive to the accuracy of data on external debt, since these data, suitably adjusted for valuation effects, are used to measure external borrowing. However, the quality of the debt data is such that one cannot be confident that annual changes in stocks of debt necessarily reflect accurately the underlying capital movements. The private claims measure, which is based primarily on directly measured stocks of foreign assets, as recorded in the Fund’s money and banking statistics, is prone to error because these statistics inadequately record the stocks of claims held by the resident nonbank sector on the nonresident nonbank sector. Moreover, unlike the debt data used in the broad measure, the data on stocks of private claims cannot be adjusted for valuation changes; this must be considered a serious deficiency when using the data to analyze developments in capital flight during a period characterized by pronounced exchange rate fluctuations among the currencies of the main industrial countries.

The narrow measure is subject to particularly pronounced measurement problems because of the patchy coverage of the short-term asset transactions of the “other” sector (which is the nonofficial, nondeposit bank sector) in national balance of payments statistics. Furthermore, the assumption that the recorded errors and omissions item in national balance of payments statistics pertains mainly to unrecorded asset transactions may be incorrect if, as Gulati argues, the systematic outflow under this item is primarily attributable to tariff- and quota-evasions by importers.20 The derived measure is particularly sensitive to the accuracy of the balance of payments statistics on investment income credits, the choice of the interest rate used to capitalize the investment income credits, and the assumption that all assets yield a market rate of return. Since the interest rate cannot be measured directly, a proxy interest rate series must be constructed carefully in the light of the limited available information on the composition of the relevant stocks of interest-bearing foreign assets. Any deviation of the proxy series from the underlying interest rate series is likely to result in large errors in the stock estimates, because the interest rate series is used to capitalize the stream of recorded investment income.

Finally, capital flight is a surreptitious activity that is likely to be particularly prone to mismeasurement. Accordingly, the estimates presented below should be viewed as orders of magnitude only and interpreted with caution. They are based on staff estimates compiled from the money and banking data shown in international Financial Statistics, balance of payments data from the Balance of Payments Yearbook, and debt data from the World Economic Outlook.21

Table 1.Capital Importing Developing Countries: Summary Estimates of Capital Outflows and Capital Flight, 1975–85(In billions of U.S. dollars)

(Annual averages)
1983–85Increase from End-1974 to

End-1985 in Stocks Outstanding
Capital importing countries
Capital outflows
Broad measure15.330.017.9235
Private claims15.533.225.3271
Narrow measure6.722.39.2153
Capital flight, “derived measure” and:
Broad measure5.626.412.5165
Private claims5.829.619.9201
Countries with recent debt-servicing problems
Capital outflows
Broad measure10.526.77.4171
Private claims10.323.213.7175
Narrow measure4.819.86.5124
Capital flight, “derived measure” and:
Broad measure5.020.79.3131
Private claims4.817.215.6135
Countries without recent debt-servicing problems
Capital outflows
Broad measure4.83.410.564
Private claims5.210.111.796
Narrow measure1.92.52.728
Capital flight, “derived measure” and:
Broad measure0.65.73.134
Private claims1.012.44.366
Note: Data may not sum to totals owing to rounding errors. Data and country classifications are from International Monetary Fund, World Economic Outlook, International Financial Statistics, and Balance of Payments Yearbook. Capital outflows from the offshore sectors of the offshore banking centers are excluded.

Developments in Capital Flight

This section presents a range of estimates for broad groups of developing countries for the period 1975–85. In line with the discussion in the preceding section, emphasis is given to the so-called broad and private claims measures of capital outflows, but eslimates on the narrow basis are also presented. Further, extensive use is also made of the derived measure to attempt to distinguish between the flight and nonflight component of any outflow. Indeed, in the balance of this study, the broad, private claims, and narrow measures are viewed as measures of capital outflows rather than capital flight, with the latter term being reserved for estimates based on the derived measure. Needless to say, in interpreting the following estimates, the reader should make due allowance for the reservations expressed above concerning the measurability of capital flight, and reservations to be discussed regarding its interpretation.

Capital outflows have been a prominent feature of the external situation of capital importing developing countries for a considerable number of years. For the group as a whole, outflows over the II years to 1985 appear to have amounted to roughly $250 billion (Table 1). The broad measure points to a somewhat lower figure ($235 billion) and the private claims measure to a somewhat higher one ($271 billion).22 Thus, while the shortcomings of the basic data are considerable, these two estimates are sufficiently close and their statistical underpinnings sufficiently independent to warrant the conclusion that capital outflows from these countries have almost certainly been in the range of $200-$300 billion. These are large sums by any standard. A figure of $250 billion over the years 1975–85 would have been about a third of the increase in these countries’ external debt over the period and about three fifths of the cumulated financing obtained from private creditors. Viewed in flow terms, an average annual outflow of $23 billion would have been equivalent to 6½ percent of merchandise imports.

The pace of these capital outflows varied considerably over the period under review. They averaged some $15 billion a year in the relatively tranquil years immediately following the first oil price increase and the 1974–75 recession (1975–78). With the onset of the second round of oil price increases, the steep rise in international interest rates, and the extended global recession, the rate of capital outflows doubled to some $30 billion per annum, an amount equivalent to about 8 percent of export earnings and about half of new borrowings from private creditors. With the onset of the debt crisis in 1982 and the reduced availability of foreign exchange, however, capital outflows receded significantly, perhaps to two thirds their former rate. Nevertheless, capital outflows appear to have remained large, both in absolute terms (some $20 billion per annum) and relative to flows of new lending from private creditors.

How much of these outflows can properly be attributed to capital flight? In the view of some observers, all outflows other than those that increase official reserves are viewed as capital flight. As was suggested earlier, however, capital outflows reflect a wide variety of motives. Some, such as trade credits and working balances, are essential lubricants of international commerce that are sanctioned by governments and ought not to be considered as capital flight. Other flows, on the other hand, are flight, prompted by a desire to avoid the effects of domestic financial repression or expected exchange rate depreciations, or simply taxation. It is, however, very difficult in practice either to separate these various motives or, assuming that were possible, to translate the resulting categories into meaningful statistical criteria. The approach espoused above, therefore, is to distinguish flight from nonflight outflows on the basis of whether or not they yield a recorded stream of investment income credits. If they do not, the flows are viewed as capital flight, whereas, if they do, they are viewed as being a normal feature of international commerce.

On this basis, it appears that perhaps three fifths of the capital outflows from capital importing countries over the period 1975–85 can be viewed as capital flight. The broad measure points to an estimate of $165 billion and the private claims measure to one of some $200 billion.23 Both estimates are highly sensitive to assumptions regarding the investment income that the outflows ought to have recorded. To the extent that the rate of return on nonflight foreign assets has been overestimated, the capital flight estimates will be on the high side and conversely if the rate of return has been underestimated. Nevertheless, even after allowance for these uncertainties, it seems clear that capital flight has been large, perhaps $I50-$200 billion for the 1975–85 period as a whole. By the same token, however, it should also be noted that, at least as estimated here, the nonflight component of capital outflows has also been large. Although amounting probably to under half the outflows, nonflight capital outflows nevertheless appear to have been in the order of $50—S100 billion over the period, a magnitude which underscores once again the need to discriminate among various types of outflows.

The pace of capital flight from capital importing developing countries varied by considerably more than that of the overall capital outflows during the period under review. Capital flight, according to the derived definition, amounted to a relatively modest $5 billion a year over the 1975–78 period, or only about a third of the aggregate outflows. With the upheavals of the late 1970s and early 1980s, however, capital flight accelerated sharply, to some $25-$30 billion a year, or about five sixths of the total outflows. The pace slackened in 1983–85. but to an uncertain extent. The estimates based on the private claims measure suggest a fairly modest deceleration to $20 billion per annum while the broad measure points to a somewhat sharper deceleration to some $12-$13 billion a year. Both estimates, however, represent significantly larger fractions of the corresponding aggregate capital outflow measures than those for the mid-1970s, suggesting that capital flight remained proportionately high. Moreover, flight became a much larger proportion of the accompanying capital inflows. Indeed, both estimates of capital flight account for all of the contemporaneous flow of financing from private creditors, which averaged $13 billion per annum over the period.

Just as the scale of capital outflows from capital importing developing countries has varied considerably over time, so has it varied across countries. This is demonstrated by comparing the estimates for the countries classified for World Economic Outlook purposes as having experienced debt-servicing difficulties with those for countries not so classified. Not surprisingly, the estimates indicate that the debt-servicing problem countries accounted for about two thirds of the cumulative outflow of capital of the entire group of capital importing countries over the period 1975–85, even though they account for only about 40 percent of the group’s exports of goods and services. Further, the problem countries accounted for an even more disproportionate share of capital flight according to the derived measure over the period (closer to three fourths). This proportion rose somewhat further for the most recent period covered in Table 1. with capital flight among the problem countries having receded relatively less over the past few years than among the non-problem countries.

It is important to note, however, that while capital outflows and capital flight appear to have been particularly prominent among countries that did in the end experience debt-servicing difficulties, such outflows were not limited to these countries. Thus, capital outflows among countries that avoided debt-servicing difficulties are estimated to have cumulated to something of the order of $75 billion—equivalent to one half of the amount received from private creditors over the period. Moreover, although the estimates diverge significantly, it is apparent that capital flight as measured here was a significant problem for countries that avoided rescheduling as well as for those that did not.

The regional pattern of capital outflows and capital flight also reflected the underlying financial characteristics of the regions during 1975–85. Using for simplicity an average of the broad and private claims measures, it is evident that, with allowance for differences in the scale of the various regions’ external transactions, the countries in the region most troubled by debt problems, the Western Hemisphere, were also the main source of capital outflows over the period. Indeed, in absolute terms the outflows from the Western Hemisphere were larger than for all other regions combined. By contrast, as a percentage of exports, outflows were smallest in Asia, probably the region least troubled by debt problems (Table 2).

Table 2.Capital Importing Developing Countries—by Region: Summary Estimates of Capital Outflows and Capital Flight, 1975–85(Annual averages, except as noted)
1975–781979–821983–85Increase from End-1974

to End-1985 in Stocks

Capital outflows(In billions of U.S. dollars)
Non-oil Middle East1.03.41.923.1
Western Hemisphere7.119.97.5131.0
(In percentage of exports of goods and services)
Non-oil Middle East6.711.05.88.3
Western Hemisphere12.
Capital flight2(In billions of U.S. dollars)
Non-oil Middle East0.2–
Western Hemisphere3.714.711.0106.6
(In percentage of exports of goods and services)
Non-oil Middle East1.5–
Western Hemisphere6.
Note: Both outflows and flight data are expressed as an average of the levels indicated by the broad and private claims measures.

The driving force behind the capital outflows from Africa and the Western Hemisphere was capital flight. Capital flight accounted for over four fifths of the outflows from countries in the Western Hemisphere and Africa during 1975–85 as a whole, and actually exceeded recorded capital outflows during some sub-periods, indicating a shift in private investors’ reasons for holding assets abroad.24 On the other hand, capital flight from countries in Asia and the Middle East accounted for about one third and one quarter, respectively, of capital outflows from these regions during 1975–85, indicating that nonflight considerations were paramount in explaining outflows. Nevertheless, even these regions experienced significant levels of flight at some points in this period: capital flight more than fully explained capital outflows from Asia in 1979–82 and from the Middle East in 1983–85. The estimates for Europe are similar to those for Africa, but otherwise Europe appears to be something of an exception to the overall pattern since the general level of outflows from this region remained relatively subdued. The explanation may be structural, for it was generally the centrally planned countries of Eastern Europe that experienced debt problems. The private sector of these economies is neither as prominent as in many other countries nor is it generally as fully integrated with the world’s financial markets as the private sector in, say, the Western Hemisphere.

Causes of Capital Flight

The discussion in the first section suggested that capital flight is a particular form of international portfolio diversification, and that not all foreign asset transactions are flight motivated. Clearly, all portfolio diversification is motivated by investors’ efforts to obtain the maximum risk-adjusted return on their capital, but at some point along the spectrum of asset transactions, flight occurs as investors attempt to avoid discrete losses. When seeking the causes of capital flight, the difficulties of determining the borderline between nonflight and flight-motivated transactions must therefore be borne in mind and the existence of a large indeterminate zone must be recognized. In attempting to distinguish between the factors that drive nonflight and flight transactions, it is perhaps useful to approach the problem from both ends of the spectrum of asset transactions, outlining first the causes that definitely do or do not cause flight before moving to a discussion of the causes of the transactions that lie in the indeterminate zone.25

Factors causing increases in private capital outflows, but not flight, include the growth of private wealth, external trade, and external financial integration. Increases in private wealth are generally associated with increases in private holdings of financial claims on nonresidents, because risk-minimization considerations argue in favor of investors holding an internationally diversified asset portfolio, and because international diversification of assets, by broadening the range of assets available to the investors, enables them to tailor the structure of their portfolios efficiently to reflect individual preferences for risks and returns,26 If the growth of real output could be taken as indicative of the growth in real wealth, and if the elasticity of demand for foreign claims with respect to total wealth were unity, then the 4¼ percent annual growth averaged by the capital importing countries during 1975–85 would be expected to cause roughly similar growth of real private foreign assets.

The growth of trade also causes capital outflows. Enterprises engaged in trade will wish to increase their stocks of working balances held abroad; export growth may involve the provision of suppliers’ credits to importers; and direct and portfolio investment abroad may be stimulated by resident enterprises’ willingness to promote their external trade through integration with nonresident private enterprises. The growth of the capital importing countries’ total external trade, which averaged 9 percent annually in dollar terms during 1975–85, would consequently be expected to lead to significant increases in nonflight capital outflows. One factor underpinning these commerce-related outflows was that domestic financial markets were often less efficient than financial markets abroad, and domestic residents and enterprises therefore shifted abroad some of their financial transactions in order to reduce costs. Profitable opportunities in international financial intermediation also cause nonflight capital outflows from financial enterprises. Such opportunities are exploited by resident commercial banks, necessitating increases in deposits at correspondent banks abroad, and perhaps an expansion into overseas offices and branches via outward direct and portfolio investment. The capital flows that accompany these various efforts to exploit international differences in comparative advantage are clearly at the nonflight end of the spectrum of international asset transactions.

Equally clearly, some asset transactions are entirely flight motivated. Domestic social pressures, for instance, may stimulate capital flight because of the increases in risks and the losses that they impose on certain classes of assets or asset holders. Social persecution of minority religious or racial groups, for example, is a persistent theme in the history of capital flight; and capital flight inevitably ensues when the domestic social structure breaks down to the extent that there is an outbreak of civil war. Similarly, changes in circumstances that are external to a country may cause it to experience capital flight. Increases in international political tensions and outbreaks of war are likely to stimulate capital flight from the affected countries, and perhaps from their regional neighbors. And shifts in exogenous economic circumstances—large changes in the terms of trade, for instance—cause residents to reappraise the attractiveness of the domestic economic environment unless domestic economic policies are promptly adjusted to reflect the change in the external environment. Similarly, changes in domestic economic policies, such as attempts to increase taxes on private capital, clearly stimulate capital flight. These policy changes are discussed in more detail below.

Turning now to the large gray area of factors that cause capital outflows and may also cause flight, it is evident that judgments must be formulated on a case-by-case basis as to whether any particular change in external or domestic circumstances causes a capital outflow or capital flight.

On the external side, shifts in economic policy abroad can engender capital outflows and, at one remove, capital flight. The changes in U.S. policies that led to the sharp rise in U.S. interest rates in the early 1980s, for example, widened interest differentials between domestic financial assets issued by capital importing countries and foreign financial assets in favor of the latter (after allowing for exchange rate movements). Moreover, the introduction of international banking facilities in the U.S. during the early 1980s had the effect of reducing U.S. taxes on certain nonresident investors, and thereby further shifted relative post-tax rates of return in favor of foreign assets.27 In part because of this change in U.S. financial policies, the average post-tax differential in favor of foreign assets for a sample of 66 capital importing developing countries widened from about 2 percentage points in 1977–78 to. on average, around 20 percentage points during 1979–82 (Table 3).28 The combination of highly attractive interest differentials in favor of foreign assets and the strongly negative real interest rates typifying domestic financial assets issued by many capital importing countries was an important contributory factor to the surge in both capital outflows and capital flight during 1979–82.29 However, not all countries were equally affected by the change in U.S. interest rates; in particular, countries that took steps to maintain the international competitiveness of their domestic financial assets and thereby eliminated this type of loss to holders of domestic assets forestalled this source of capital flight.

Table 3.Capital Importing Developing Countries: Selected Economic Indicators, 1975–85(Changes, in percent, except where otherwise indicated)
Real GDP5.
Total merchandise trade (in dollars)3.310.817.215.728.828.94.8–8.4–2.47.1–0.9
Central government fiscal deficit (in percent of GDP)
Public and publicly guaranteed long-term debt (in percent of GDP)
Broad money27.633.731.629.635.637.137.638.244.056.855.9
Consumer prices21.123.222.519.923.629.628.627.337.944.346.9
Interest rates (in percent)1
Domestic real rate2–6.2–2.8–3.8–2.4–4.1–8.3–6.1–1.6–4.4–2.4
Nominal differential in favor of foreign assets372.618.
Exchange rates
Real effective rate–4.3–0.1–2.6––4.1–6.2–1.5–5.8
Black market premium on official rate4370.0348.9106.589.873.443.788.778.580.4
Sources: International Monetary Fund, World Economic Outlook, and International Financial Statistics.

Another external development that may have caused capital flight, depending on countries’ circumstances, was the growth of international financial intermediation during the 1970s and 1980s. As noted above, this growth facilitated investors’ efforts to diversify their asset portfolios internationally for nonflight reasons. But in doing so, the costs of domestic financial repression and the inefficiencies of the domestic capital market were highlighted for residents of many countries. To the extent that financial integration was accompanied by financial innovations that effectively weakened national barriers to capital mobility, the increase in international financial integration probably encouraged and facilitated capital flight, In this context it is interesting to note that commercial banks, who tend to decry the aggregate level of capital flight, are themselves competing for deposits from flight-motivated investors resident in developing countries.

Changes in domestic economic circumstances, however, are also frequently sources of capital flight. Of particular note is the impact of governments’ economic policies on investors’ willingness to hold domestically issued assets, because fiscal, monetary, exchange rate, and structural policies can quickly affect the expected returns and risks associated with holding claims on residents.

Fiscal policy impinges directly on post-tax rates of return on assets. Income taxes on interest, profits, and dividends and property taxes on private wealth introduce a distortion as compared to market rates of return. Because such distortions are absent or are negligible in countries that implement “tax haven” policies, fiscal policies in countries that include taxes on wealth and on income from wealth therefore introduce a wedge between the post-tax rates of return on domestic assets and the rates of return on assets issued in tax havens. Moreover, if the tax wedge is sufficiently large, differentials may also appear vis-a-vis the post-tax rates of return on assets issued by non-tax haven countries. Clearly, the larger the tax wedge and the greater the confidentiality attaching to the transactions, the greater the incentive for the private sector to shift its assets to tax havens and to countries that impose low taxes on assets and on incomes from assets.

The structure of the tax system may also induce capita] outflows. Taxation policies that bear more heavily on residents than nonresidents may induce large and offsetting capital flows. Some countries, for example, tax residents’ investment income from domestic assets, but provide tax relief on interest paid on liabilities to nonresidents. This type of tax structure provides residents with an incentive to simultaneously acquire claims on and liabilities to nonresidents in order to maximize post-tax rates of return on wealth.

Fiscal policy also influences investors’ expectations as to future taxes. Fiscal deficits that are financed by borrowing, for example, imply increases in future tax liabilities because of the government’s need to service its debt. Large current fiscal deficits may cause residents to fear that tax rates will be increased, perhaps to confiscatory levels, in order to meet future debt service obligations. In this context ii is noteworthy that for the capital importing countries the ratio of public and publicly guaranteed debt to GDP more than doubled between 1975 and 1985, implying a corresponding increase in residents’ contingent tax liabilities (Table 3), It is reasonable to suppose that during the period in which debt increased rapidly some residents of these countries acted to protect the foreign currency value of their wealth from the expected increases in the tax wedge by shifting part of their wealth abroad. And the historically high debt-to-GDP ratios that have been recorded since 1982 may in part explain why large stocks of flight capital, some of which were accumulated for other reasons prior to 1982. continued to be retained abroad even after some of the initial causal factors had been corrected.

In extreme cases, residents may have feared that fiscal deficits would persist at levels that could threaten the public sector’s ability to service its liabilities. Fears would then have been aroused that the public sector would at some point take measures to reduce its real debt service burden through repudiation, default, or the monetization of liabilities denominated in domestic currency. Such developments would raise the perceived riskiness of holding claims on all residents of such countries, not just the public sector, because of the externalities that would be associated with the public sector’s efforts to cope with its debt problem. Consequently, domestic and foreign creditors would attempt to reduce their claims on residents, both public and private sector, and residents would attempt to protect the foreign currency value of their wealth by acquiring flight-motivated claims on nonresidents.30

Monetary policy, like fiscal policy, may also be a source of capital flight, most often as a result of an inflexible interest rate policy or lax control of the monetary aggregates. Interest rate policies prompt capital flight when domestic interest rates are held at rates that are unattractive to domestic savers, because domestic interest rates are usually then at levels which create an interest differential in favor of foreign assets (after adjusting for expected exchange rate changes). The resulting outflows are likely to be viewed as capital flight by governments unable to finance their fiscal plans. The consequent increases in monetary financing of deficits will lead to inflation and a further incentive to capital flight as real interest yields fall further and the domestic currency depreciates.

The stance of monetary policy in the capita) importing countries was clearly an important explanatory factor underlying the capital flight of 1979–82. During this period the growth of broad money increased substantially, inflation accelerated, and domestic interest rates became increasingly negative in real terms despite an increase in nominal international interest rates.31 Residents were therefore confronted with the choice of accepting large losses in real terms if they held claims on residents—amounting to an average 5 percent annually during 1979–82 (Table 3)—or substantially positive real rates of return on foreign assets—amounting to some 17½ percent annually, after taking into account exchange rate movements.

Expectations of future exchange rate movements play a particularly powerful role in generating cross-border capital flows. In particular, exchange rate policies that imply short-run real appreciations of a currency often prompt private capital outflows as residents seek to avoid capital losses when the overvaluation is corrected. The real appreciation of the capital importing countries’ exchange rates during 1979–81, for example, was obviously inappropriate and unsustainable in view of the inflationary conditions in those countries, as was shown by the subsequent depreciations; asset holders were therefore able to avoid significant capital losses—amounting to a cumulative 19 percent of their capital during 1982–85—by shifting resources abroad in 1979–81. and particularly in 1981.

More generally, exchange rate policies that are perceived as inconsistent with fiscal and monetary policies give rise to strong incentives for residents to acquire foreign assets. The exchange rate policies introduced in a number of Western Hemisphere developing countries in the late 1970s, for example, are viewed by some observers as having provided private speculators with the possibility of achieving large capital gains almost risk free because of the inconsistencies between the exchange rate policy and the implementation of fiscal and monetary policies.32

The authorities’ policies vis-à-vis financial markets are also important when analyzing the causes of capital flight because asset-holders are sensitive to the changes in rates of return and risks on domestic assets that can be caused by structural adjustments. Policy changes that are likely to lead to capital flight arc typically those that are associated with an intensification of financial repression, including raising taxes on domestic financial intermediation, imposing high reserve requirements on banks, interest rate ceilings, quotas on credit allocation that provide insufficient credit to the private sector, and resorting to greater use of the inflation tax. The intensification of any of these distortions prompts residents to reappraise their willingness to seek financial intermediation services in the domestic economy, and is likely to drive some residents to seek financial intermediation offshore.

While data on some forms of financial repression are difficult to obtain, one can use the data on inflation and on real interest rates on domestic financial assets as illustrative of the extent of financial repression. By these measures, financial repression increased in the capital importing countries between 1975–78 and 1979–82, and this probably stimulated flight. Financial repression almost certainly continued to motivate strongly investors’ portfolio allocation decisions during 1983–85, as inflation accelerated further, to 43 percent annually, and real interest rates remained negative, albeit less negative than in 1979–82 (Table 3).

Implications of Capital Flight

As noted earlier, concerns about capital flight are grounded in a number of considerations, including the stated unwillingness of certain creditors to provide funds that might simply end up financing capital placements abroad by the wealthier strata of the country in question. The fundamental economic concern about capital flight, however, is that it reduces welfare in the sense that it leads to a net loss in the total real resources available to an economy for investment and growth. That is, capital flight is viewed as a diversion of domestic saving away from financing domestic real investment and in favor of foreign financial investment. As a result, the pace of growth and development of the economy is retarded from what it otherwise would have been.

These perspectives on capital flight are well founded, and certainly have been the operative ones over the past few years. These are not, however, necessary consequences of capital flight per se. For these consequences to follow, another condition must also be met, namely, that nonresidents be unwilling to indirectly finance the capital flight through the acquisition of offsetting claims on the country in question.

Over much of the period under review, there was an underlying willingness of nonresidents to invest in developing countries. Net lending by private creditors increased rapidly throughout the second half of the 1970s. Indeed, private foreign creditors were sufficiently willing to invest in these economies to finance substantial increases in both real resources transfers and the demand for foreign assets (Chart 2). Presumably, this reflected a positive assessment by those creditors of the prospects of the economies in question and a view that, adjusted for risk, these claims would yield returns at least as favorable as those available on alternative placements. Fundamentally, these views were consistent with the classical notion—borne out by the statistics for a number of countries—that rates of return on capital ought to be higher in developing than industrial countries because of the relative scarcity of capital in developing countries.

Chart 2.Capital Importing Developing Countries: Net Borrowing and Capital Flight

1 Data plotted refer to “other net borrowing” estimates shown in World Economic Outlook, April 1987. Table 40, p. 168.

2 Data plotted are centered three-year moving averages of the average of the broad and private claims measures of capital flight shown in Table 1.

It might be argued that foreign creditors could usefully have cross-checked their favorable perspectives on the economies in question with those of residents, who were seemingly intent instead on acquiring external rather than domestic claims. To do so, however, would be to misinterpret the motivations behind the capital outflows. As explained earlier, the capital flight of the 1970s can be substantially explained on the basis of considerations that have little to do with the underlying rate of return on capital in the country concerned. Some of the outflows stemmed from the requirements of international commerce, outflows that presumably enhanced the return on capital invested in the tradable goods sector. Some of the outflows were motivated by ordinary portfolio diversification considerations. Similarly, as the advent of offshore banking centers made offshore financial intermediation more efficient than onshore intermediation, some of the outflows stemmed from residents’ desires to reduce transactions costs. Finally, some of the outflows reflected capital flight as residents sought to avoid the wedge introduced by domestic policies between the underlying rate of return on capital in these economies and the real return paid to resident holders of domestic financial assets. Indeed, these policies, which often resulted in financial repression, excessive fiscal deficits, inflation taxes, and unrealistic exchange rates, were often such as to make foreign assets significantly more attractive than domestic assets. It is these considerations—considerations pertaining to the return to private residents rather than to that of nonresidents or to the underlying prospects of the economy—that motivated the capital outflows.

Under these circumstances, where both resident and nonresident investors retained an essentially positive perspective on the prospects for an economy, capital flight can perhaps best be viewed as reflecting the internationalization of the intermediation between domestic savings and investment—an internationalization that was given impetus by national policies that discriminated against resident holders of domestic assets and by the parallel explosion in the provision of international intermediation services. As a result, domestic savings went offshore, but returned, often in The form of increased bank claims on national governments. Consequently, the implications for the real economy of this “intermediated” form of capital flight tended to be smaller than they would otherwise have been, at least in the short run. While there was a grossing up of both claims and liabilities, the net effect on the balance of payments and the exchange rate was limited. For the same reason, the pool of resources available for domestic investment was also probably little changed so that the underlying growth path of the economy was, for a time at least, largely unaffected.

The policies that underpinned the flight component of the outflows were of course harmful to these economies. They led to distortions and misallocations of resources that necessarily weakened growth relative to what it otherwise would have been. More fundamentally, however, capital flight would in time increase external indebtedness and contribute to changing the perspectives of foreign creditors, a change which would have highly adverse implications for these economies.

Even before the onset of financing constraints, capital flight is likely to have adverse implications for the public finances and the allocation of resources. Thus, capital flight is likely to raise the carrying costs of public debt for two reasons. First, the exchange rate risk borne by private residents when they hold domestic assets is effectively shifted to the public sector as the intermediation of domestic savings becomes internationalized. Second, foreign creditors are unlikely to accept the imposition of the domestic taxes or uncompetitive interest rates that were typically imposed on domestic financial assets in many capital importing countries so that the internationalization of intermediation is again likely to raise public sector borrowing costs. Capital flight is also likely to lead to an erosion of the tax base because it removes resources from the government’s tax jurisdiction, and thereby weakens public revenues. Further, capital flight may have led the public sector to undertake a larger share of investment than it might otherwise have done. Because the internationalization of intermediation tended to result in increases in private foreign claims offset by increases in public liabilities, the share of domestic investment controlled by the public sector may have increased. This tendency may have been further exacerbated by official actions intended to ensure that profits from new, externally financed investments not be lost through further capital flight and be available for debt service payments. Finally, capital flight may have had regressive effects on income distribution. The shifting of private wealth beyond the government’s tax jurisdiction probably led to a shifting of the lax burden away from the internationally mobile factor of production, capital, and onto the less mobile factors of production, labor, and land. On balance, although assessment is difficult, such a shift in the tax burden is likely to be regressive.33

However, given the postulated willingness of offshore intermediaries to act as intermediaries among residents, the adverse implications of capital flight just described may not have been very evident. They take a long time to build up to a level where they become noticeable and a priority concern of national authorities. Accordingly, concerns regarding capital flight were somewhat muted during the 1970s. Those that were expressed tended to be with reference to particular countries at times when they experienced balance of payments difficulties, that is at times when foreign creditors were less accommodating.

This generally low-key view of capital flight was reversed in the 1980s, especially in the aftermath of the debt crisis. As foreign creditors became increasingly unwilling to lend to developing countries, they in effect also became unwilling to serve as the intermediaries between domestic savers and investors. That is, while foreign banks continued to accept deposits from developing country residents, they no longer were willing to acquire the offsetting claims, and capital flight became immediately translated into a loss of resources. The real resources available to the economy for investment and growth—already shrunken by weak purchasing power of exports, high debt service payments, and the virtual cessation of private lending—shrank further as a result of outflows through the capital account.

This is a much more serious situation than the “churning” of claims that prevailed in the 1970s. At root, it reflects a fundamental shift in the perception of creditors of the expected returns on investments in the country in question. Under these circumstances, it can no longer be presumed that capital flight is motivated primarily by differentials in favor of foreign assets stemming from ill-advised financial policies. Rather, the presumption must be that, just as foreign savers drastically altered their perceptions of expected returns on investments in the country concerned, so did domestic savers. As a result, neither residents nor nonresidents wished, at the margin, to acquire claims on the affected economies. Consequently, residents became obliged to exchange real resources in order to effect their desired acquisitions of foreign assets.

This shift from the “churning” type of capital flight in the period to 1981 to the net real resource transfer form of capital flight that is evident since then has had a number of implications, all of them unpleasant. In the first place, capital flight under these circumstances invariably leads to additional unwanted pressures on the foreign exchange markets. As a result, the authorities are forced off their intended path for official reserves or the exchange rate and thereby forced to reconsider monetary and fiscal policies in a context of diminishing growth prospects and accelerating inflation. Typically, these dilemmas prompt the authorities to adopt more restrictive policies than they would have pursued in the absence of capital flight. Separately, the private sector’s appetite for foreign assets is likely in and of itself to lead to a curb in domestic spending since its demand for foreign assets can only be satisfied through increases in net exports. Overall, therefore, real growth in the capital flight country is likely to fatter. Moreover, the toss might well carry over into the medium term. In the circumstances postulated, it is likely that the funds in flight are at the expense of private sector investment, so that capital flight might impair medium-term growth prospects as well.

Finally, it is to be noted that the “intermediated” outflows of the early period contributed to the “real” capital flight of the 1980s and exacerbated its consequences. The intermediated outflows led to increases in gross indebtedness which contributed significantly to the adverse shift in the perceptions of investors as regards countries’ creditworthiness. Moreover, the nonflight outflows of the 1970s led to large accumulations of assets abroad which, when sentiments changed, were readily transformable from nonflight to flight assets. Thus, although the consequences of the intermediated outflows were secondary throughout the 1970s, they helped both to trigger the changing perspectives of the 1980s and to effect a negative real transfer once sentiment changed.

Capital flight has, of course, implications for the capital receiving country as well as for the capital exporting country. The consequences of an inflow of flight capital are largely the mirror image of those for the flight country. To the extent that Right inflows are associated with the internationalization of financial intermediation, the commensurate capital outflows will ensure that resource-transfer implications are minimal, although welfare gains would be achieved through residents’ acquisitions of the range of domestic and foreign assets that they wish to hold. When “real” capital flight occurs, however, recipient countries obtain a net inflow of foreign savings from the flight country, whose real resource counterpart is achieved through a trade deficit, or at least through a reduction in the recipient’s surplus compared to what otherwise would have been implied by the balance of economic forces. By the same token, the capital inflows would be likely to depress interest rates, and increase domestic absorption in the recipient country, thereby setting in motion economic adjustments which, if sufficiently large, could prove cosily for the recipient country’s tradable and nontradable goods sectors. The balance of advantage probably favors the recipient country, otherwise residents would not have been willing to incur the net increase in liabilities to nonresidents. But the issue is not clear cut, and governments of recipient countries (Switzerland, for example) have sometimes resorted to controls on capital inflows in order to reduce the domestic social costs of being a “safe haven.”

Turning to the implications of capital flight for national policies, the starting point must be the realization that such flight emanates from distortions that must be eliminated or at least eased if flight is to be curbed. Accordingly, a first implication for policies is that national authorities need to adjust policies so as to avoid the pronounced discrimination against resident holders of domestic assets that has been so prevalent among developing countries. In order to mobilize and retain domestic savings, savers must be remunerated at rates that are more in line with those available internationally. Avoidance of financial repression together with the development of an efficient system of domestic financial intermediation would seem to be especially useful in achieving this objective and in reducing countries’ vulnerability to flight in the future. Shifting to wholly market-determined interest rates may of course be unrealistic for some countries in the near term, for instance because of the undeveloped character of their financial markets. Nonetheless, even in these cases, policies need to correct the large negative real rates of return that have sometimes characterized domestic financial instruments. It is only once those negative rates are substantially eliminated that one can reasonably expect a sustained reduction in the capital flight.

A second implication is the need to adopt policies to keep the exchange rate in line with the expectations of the private sector. Capital flight has repeatedly been prompted over the years by the expectations of residents that they could avoid large losses on their domestic currency-denominated assets if these were converted into foreign currency at some point prior to the expected depreciation. In principle, of course, this cause of capital outflows could be eliminated by simply letting the exchange rate float. Moreover, a floating rate would automatically neutralize the effects on capital flows of distortions elsewhere in the economy, such as the regulation of domestic interest rates, and thereby reduce incentives for flight. In practice, however, most countries exercise some control over the path of the exchange rate, a control that opens avenues for private gain and capital Right. In these circumstances, financial policies need to be carefully controlled to keep the exchange rate in line with expectations, a requirement which goes hand-in-hand with the need to pursue sound financial policies.

The financial policies required to keep capital flight in check are very familiar. In most cases, there is a need to keep fiscal deficits at prudent levels in order to avoid triggering expectations of increased taxation, tightened capital controls, and acceleration of inflation, each of which would lead to a prompt reassessment of the exchange rate. It is especially important to keep a rein on inflationary expectations since, besides instigating capital flight by building up expectations of exchange rate depreciation, they also tend to magnify the distortive effect of financial market regulations. It is therefore important that monetary policy and the financing of the budget play a stabilizing role by keeping the growth of the monetary aggregates on a stable and in most cases decreasing path.

Needless to say, the reasons for adopting more market-oriented structural policies and orthodox financial policies extend far beyond considerations relating to capital flight. The case for market-oriented structural policies rests rather on the generalized improvement in resource allocation they would permit. While one consequence of market interest rates might well be increases in the public sector’s borrowing costs and increased fiscal pressure, any adverse implications from this source would be overshadowed by the benefits in the form of increased investor confidence, improved efficiency of investment, and faster growth. This in turn would lead to increasingly buoyant tax revenues. Similarly, while cautious financial policies are important to efforts to reduce capital flight, their main benefit would of course be in terms of the creation of an economic environment that would facilitate economic growth on the basis of increased investments of voluntarily-mobilized domestic and foreign savings. This requires, in the first instance, the adoption of an appropriate mix of fiscal, monetary, and exchange rate policies, and a willingness to quickly adjust policies to changes in the domestic and foreign economic environment.

While market-determined interest rates and exchange rates would go far toward eliminating capital flight, they would not eliminate flight stemming from outright tax avoidance. As noted earlier, capital flight originates in part in the desire of the private sector to avoid taxes, including the more implicit forms of taxation such as those associated with inflation and the discrimination against resident holders of domestic assets. As regards the avoidance of direct forms of taxation, the proper corrective action is less clear cut. One view is that it is a mistake to attempt to tax capital to begin with. Given the fungibility of loanable funds, attempts to tax domestic capital at rates significantly higher than those applicable elsewhere are bound to fail. In this view, this consideration is especially pertinent today given the efficiency and confidentiality attaching to transactions with offshore banking centers. Another view is that tax-evading capital flight, like other forms of capital flight, reduces public revenues and creates a distortion of its own. Making up the revenue from other taxes or increasing existing taxes on capital will further distort the allocation of resources. In this view, therefore, the solution is to control tax-evading capital flight using the same regulatory means that are used to control other forms of tax evasion.

Which of these two views is to be preferred in any particular case is likely to depend heavily on the perceived mobility of domestic capital. In countries where financial markets are at a rudimentary stage of development, capital mobility is likely to be perceived as quite low. In these countries, the authorities’ desire to maximize public revenues, minimize the borrowing cost of the public debt and, at one remove, increase the resources available for investment and growth is likely to lead them to a reliance on regulations and capital controls. In countries where financial markets are better developed, however, the case for gearing policies to the low tax solution is stronger, especially over the longer term. In these countries, the explosion of offshore international intermediation services has rendered increasingly ineffective the regulations and capital controls that have supported taxation differentials in the past. The private sector has proved itself willing and able to circumvent such controls when necessary in order to protect the real purchasing power of its wealth from significant loss.

The case for using capital controls to curb capital flight is, of course, more general than is suggested by the considerations of tax evasion alone. The starting point is the view that, because of externalities, the social rate of return on domestic investment is higher than the private rate of return. Consequently, investment spending needs to be pushed to the point where the marginal cost of funds is equated to the social rather than the private rate of return—an expansion that can, in this view, be achieved by bottling up investable funds at home through use of controls. This view is buttressed by the conviction that the productivity of investment in developing countries is greater than that in industrial countries so that there is an underlying economic case for attempting to maximize investment in developing countries. In some countries, capital controls are also justified on the basis of infant industry arguments—that controls are necessary during the transition period while the domestic financial sector is being built up to be competitive in the global market for intermediation services.

The use of capital controls is also justified on the basis of more short-run considerations. The main one is that controls are useful in preventing transitory strains and stresses in financial markets and, at one remove, private sector expectations from filtering forward into the balance of payments, exchange rates, inflation, and the real economy. The premise of course is that the disturbances are random and non-systematic. If so, controls may be effective in the short run in countries with underdeveloped financial markets. In point of fact, however, the disturbances are more likely to stem from inconsistencies in financial policies as governments seek to achieve a multiplicity of short-run objectives that are not always easily reconcilable. Under these conditions, controls substitute for the required correction to policies and are unlikely to be effective. Indeed, the increasing tension brought about by the controls, on the one hand, and the inconsistency of policies, on the other, is, in the absence of policy adjustments, likely to generate expectations of a further tightening of controls. It is expectations of this type which in the past have led to some of the most pronounced bouts of capital flight. Thus, even in the short run, controls are unlikely to be effective unless they are buttressed by reasonably sound financial policies to begin with.

The usefulness of capital controls over the longer term in the sense of whether they increase the availability of real resources for development is also questionable. In the first place, the proponents of controls tend to make insufficient allowance for the adverse effects controls have on resource allocation and the level of saving. In addition, these arguments tend to lose sight of the implications of controls for the availability of foreign saving. In the long term, the allocation of global savings among countries is presumably governed by the perceived marginal efficiency of investment in the various countries, with the allocation being such as to equate returns at the margin. Hence, given some dependence on foreign creditors, attempts to increase investment in one country through controls on outflows are likely to drive down the return on investment and prompt foreign investors to seek more profitable opportunities elsewhere. On balance, therefore, controls on capital flight—while they may, depending on circumstance, retain domestic savings that would otherwise have gone elsewhere—are unlikely to increase the overall availability of real resources over the long term. Indeed, since controls are likely to lead to a perceived leftward shift in the marginal efficiency of investment schedule for that country, they would probably reduce the availability of real resources.

Information on capital Right can be useful to the authorities as a signal of the degree of distortion that exists in the economy and of the need to adjust policies. As was seen earlier, the ebb and flow of capital flight over the years has, in part, been proportional to those tensions. Accordingly, governments might consider collecting and tracking the relevant data more systematically. In retrospect, it is clear that it would have been advantageous to pay this signal greater heed than it actually received. However, care needs to be exercised in the interpretation of such data. One reason is, of course, that the available data are typically very imperfect measures of capital flight. Another reason, however, is that the interpretation of the data hinges on the surrounding circumstances. For instance, at any point in time, certain distortions may be large but not reflected in capital outflows because of a recent real exchange rate change. By the same token, “rear” capital flight may be associated with much smaller observed outflows than “intermediated” capital flight. However, this is more likely to reflect the difficulty of effecting the real transfer rather than any differential in the tensions underlying the two flows. Hence, while capital flight data can be useful as a warning signal for policy, such data will often be misleading unless it is assessed within the context of a comprehensive analysis of the factors underlying the outflows.

In sum, the policies that are likely to be most effective in curbing capital flight are, in the main, the same market-oriented and cautious policies that can be justified on much wider grounds. Such policies minimize the distortions and disturbances that prompt capital flight and are over the long run likely to be the most effective. A final question, however, is whether such policies are likely to be equally effective against the “intermediated” and “real” forms of capital flight observed in the 1970s and 1980s, respectively. Clearly, the preferred policies are likely to be quite effective in curbing “intermediated” flight. The willingness of foreign creditors to invest in the country in that case makes it clear that the flight is primarily due to the wedge brought about by government policies between the return to savers and the return on investments. Elimination of that wedge will eliminate the need for the external intermediation at the levels seen in the 1970s and hence that source of capital flight. Indeed, it might well be expected that, under these circumstances, past capital flight would be repatriated.

“Real” capital flight, however, is not necessarily rooted only in domestic distortions. As noted earlier, this form of capital flight stems rather from an underlying unwillingness of both foreign and, at the margin, domestic savers to invest in the economy in question. The adoption of sound policies would certainly contribute to easing the country’s difficulties. It would curb that part of the outflow stemming, for example, from any misalignment of the exchange rate. It would also significantly improve prospective rates of return on investments in that economy, a shift that might be sufficient to offset the bearish views of investors and lead to a resumption of voluntary lending by both foreign and domestic creditors.

However, the intractability of the debt problem and the evident unwillingness of private foreign creditors to lend to a range of developing countries suggest that the belated adoption of sound national policies may not in itself always be sufficient to eliminate real capital flight. Existing levels of indebtedness may be such as to continue to discourage investors, be they domestic or foreign, and thereby sap the potential for improved creditworthiness through higher growth. As such, real capital flight is a part of the much wider set of issues associated with the debt situation and the need to restore growth in the developing world—issues that are beyond the scope of the present paper. Like those issues, however, real capital flight in the sense just described needs to be addressed within a collaborative framework, tailored to each country’s circumstances, that deals simultaneously with the reluctance of both residents and nonresidents to invest in the country in question.


    Conesa, E.R.,“The Causes of Capital Flight from Latin America, 1970–85” (unpublished, Inter-American Development Bank, August1986).

    Cuddington, John T.,Capital Flight: Estimates, Issues, and Explanations, Princeton Studies in International Finance 58 (Princeton, New Jersey: Princeton University, 1986).

    Cuddington, John T.,“Economic Determinants of Capital Flight: An Econometric Investigation,” in Capital Flight: The Problem and Policy Responses, ed. by DonaldR. Lessard and JohnWilliamson (Washington: Institute for International Economics, June1987).

    Cumby, R. and R.Levich,“On the Definition and Magnitude of Recent Capital Flight,” in Capital Flight: The Problem and Policy Responses, ed. by DonaldR. Lessard and JohnWilliamson (Washington: Institute for International Economics, June1987).

    Dooley, Michael P.,“Country-Specific Risk Premiums, Capital Flight and Net Investment Income Payments in Selected Developing Countries” (unpublished, International Monetary Fund, March1986).

    Dornbusch, Rudiger,“External Debt, Budget Deficits and Disequilibrium Exchange Rates,” in International Debt and the Developing Countries, ed. by GordonW. Smith and JohnT. Cuddington (Washington: World Bank, 1985), pp. 21335.

    Duwendag, Dietor,“Kapitalflucht aus Entwicklungsländern: Schatzprobleme und Bestimmungsfaktoren,” in Die Internationale Schuldenkrise: Ursachen, Konsequenzen, Historische Erfahrungen, ed. by ArminGutowski (Berlin: Duncker & Humblot, 1986).

    Erbe, Susanne,“The Flight of Capital from Developing Countries,”Inter economics (Hamburg), Vol. 20 (November1985), pp. 26875.

    Grubel, Herbert G.,“Internationally Diversified Portfolios: Welfare Gains and Capital Flows,”American Economic Review (Nashville, Tennessee), Vol. 58 (December1968), pp. 1299314.

    Gulati, S.K.,“Capital Flight Through Faked Trade Invoices: A Statistical Note” (unpublished, Columbia University, May1985).

    Gulati, S.K.,“Adjusting Capital Flight Measurement for Over and Underinvoicing: Some Surprises” (unpublished, Columbia University, October1986).

    Harberger, Arnold C.,“Lessons for Debtor-Country Managers and Policymakers,”in International Debt and the Developing Countries, ed. by GordonW. Smith and JohnT. Cuddington (Washington: World Bank, 1985), pp. 23657.

    International Monetary Fund, “Foreign Private Investment in Developing Countries,” Occasional Paper 33 (Washington, January1985).

    International Monetary Fund, Government Finance Statistics, Vol. 9 (Washington: IMF, 1985).

    International Monetary Fund, Balance of Payments Statistics Yearbook, Vol. 37 (Washington: IMF, 1986).

    International Monetary Fund, International Financial Statistics Yearbook (Washington: IMF, 1986).

    International Monetary Fund, World Economic Outlook (Washington: IMF, October1986).

    Ize, Alain and GuillermoOrtiz,“Fiscal Rigidities, Public Debt and Capital Flight,”Staff Papers, International Monetary Fund (Washington), Vol. 34 (June1987), pp. 31132.

    Katz, Menachem,“Impact of Taxation on International Capital Flows: Some Empirical Results,”in Taxation, Inflation and Interest Rates, ed. by VitoTanzi (Washington: International Monetary Fund, 1984).

    Khan, Mohsin S. and NadeemUl Haque,“Foreign Borrowing and Capital Flight: A Formal Analysis,”Staff Papers, International Monetary Fund (Washington), Vol. 32 (December1985), pp. 60628.

    Kindleberger, Charles P.,“Capital Flight—A Historical Perspective,” in Capital Flight: The Problem and Policy Responses, ed. by DonaldR. Lessard and JohnWilliamson (Washington: Institute for International Economics, June1987).

    Lecraw, D.,“Direct Investment by Firms from Less Developed Countries,”Oxford Economic Papers (Oxford), Vol. 29 (November1977), pp. 44257.

    Mattione, Richard P.,OPEC’s Investments and the International Financial System (Washington: The Brookings Institution, 1985).

    Morgan Guaranty Trust Company, World Financial Markets (New York: March and September1986).

    Nolling, Wilhelm,“Combating Capital Flight from Developing Countries,”Intereconomics (Hamburg), Vol. 21 (May1986), pp. 11723.

    Rodriguez, M.A.,“Consequences of Capital Flight for Latin America” in Capital Flight: The Problem and Policy Responses, ed. by DonaldR. Lessard and JohnWilliamson (Washington: Institute for International Economics, June1987).

    Williamson, Martin C.,“Acquisition of Foreign Assets by Developing Countries, 1970–84: Empirical Evidence,”International Monetary Fund Working Paper (Washington, forthcoming).

    World Bank, World Development Report (New York: Oxford University Press, 1985).

Morgan Guaranty Trust Company (September 1986).

See Kindleberger (1986) for a historical review.

The term “financial repression” is used in a broad sense in this paper to refer to the losses occurring to asset holders occasioned by all types of distortions imposed on the domestic financial system.

This distinction possibly owes more to an accountant’s view of profit and loss than to the economic concept of loss, for the opportunity cost of failing to transfer banking activities offshore is economically equivalent to a loss, although such “losses” are not directly entered on balance sheets.

Alternatively, one could rephrase this definition in terms of a class of risks which pertain to a resident’s claims on other residents and which are not compensated with an adequate risk premium. Khan and Haque (1985), for example, define capital flight in terms of residents’ attempts to avoid some broadly defined “expropriation risk.”

Outward direct investment from developing countries is discussed in, for example, Lecraw (1977), International Monetary Fund (1985), and Mattione (1985).

Cumby and Levich (1986) and Nolling (1986) compare various estimates of capital flight. Examples of important studies based on particular measures are noted below, as appropriate.

This definition corresponds to that used by the World Bank (1985), Erbe (1985), and Rodriguez (1986), for example. It is measured statistically as a residual—gross capital inflows less the current account deficit less the acquisition of official reserves.

This definition corresponds to Conesa (1986).

This corresponds to the approach of Cuddington.

The nonflight stock of foreign capital is estimated as the capitalized value of recorded investment income credits, using an appropriate interest rate; flight capital is therefore computed as the difference between total foreign assets and the stock of nonflight capital. This method follows the approach proposed by Dooley (1986).

The inclusion of cumulated recorded errors and omissions data from balance of payments statistics is common in the literature on capital flight because these data are generally believed to pertain to the unrecorded asset transactions. However, this presumption may be incorrect, as will be discussed below. (Dornbusch (1985) measures capital flight as a “hot money” flow, but excludes the errors and omissions item.)

Including nonfinancial public enterprises which, because they are often motivated by the same economic considerations as their private sector counterparts, are hereafter included in the private sector for analytical and statistical purposes.

Some measures based on the private claims approach introduce certain restrictions on the types of private claims that are included. Morgan Guaranty Trust (February/March 1986), for example, excludes the foreign assets of the resident banking system, presumably on the grounds that these interbank deposits are a necessary element in international financial intermediation and are not motivated by flight considerations.

For the (net) capital importing developing countries, the growth of the foreign assets held as non-reserve assets by the monetary authorities and by the nonbank official sector probably accounted for some 9½ percent of the growth of total assets covered by the broad measure between the end of 1974 and the end of 1985; at the end of 1985 these public sector assets were equivalent to 10 percent of the total assets held by the sectors covered by the “broad” measure. See the memorandum item to Chart 1.

Note that because capital flight is unlikely to originate from either the monetary authorities or the nonbank official sector, the derived measure probably has roughly the same sectoral coverage as the private claims measure.

Gulati (1985, 1986). This qualification applies also, either explicitly or implicitly, to the broad and private claims measures, although the problem is less significant for these measures because the bulk of the capital outflows that are included are not attributable to the errors and omissions item, whereas for many countries, the errors and omissions constitute the greater part of the outflows covered by the narrow measure.

See Williamson (1987). The interest rate used to capitalize investment income for use in the derived measure is a weighted average of the Eurodollar rate in London on three-month deposits (LIBOR) and the U.S. Government bond yield on three-year maturities, in order to allow for differences in the maturities of the foreign claims held by residents of developing countries. These weights were more or less arbitrarily set at five sixths and one sixth, respectively. The composite rate is lagged six months to allow for lags in the responsiveness of interest payments to interest rates.

This is the inverse of the relationship one would expect between the two estimates since the broad measure includes the “nonmonetary” official sector. The result is due to inconsistencies between, on the one hand, the debt and balance of payments statistics and, on the other, money and banking statistics.

As noted earlier, the stock of flight capital is estimated residually as the difference between the specified stock of foreign assets and the estimated stock of nonflight assets arrived at by capitalizing investment income receipts. For this purpose, one requires investment income estimates corresponding to the assets encompassed by the various outflow measures. In fact, however, only aggregate investment income series are available. Moreover, those data include the income earned on reserve assets, assets that are not included in any of the outflow measures. Accordingly, capital flight has been estimated by broadening the outflow-based asset estimates to include the stock of reserve assets and deducting from that total the capitalized value of investment income. One implication of this procedure is that the difference between the broad and private claims measures of capital flight is just equal to the difference between the broad and private claims measures of capital outflows.

For example, capital flight was 30 percent greater than capital outflows for Africa in 1979–82 and for the Western Hemisphere in 1983–85.

The few empirical studies of the determinants of capital flight are not particularly helpful in clarifying what prompts flight since the data problems cause the respective authors to heavily qualify their empirical conclusions. See, for example, Conesa (1986), Cuddington (1985, 1986), Dooley (1986), and Williamson (1987).

Assuming that rates of return on financial assets are imperfectly correlated internationally. See Grubel (1968) for a theoretical statement of the issue, and Williamson (1987) for empirical evidence on imperfectly correlated interest rates.

Although it should be noted that the U.S. was largely responding to earlier changes in other countries’ willingness to ease the tax burden on nonresident investors.

Due care should be exercised when interpreting these data, however, because the samples are weighted by GDP and shifts in policies or circumstances of the largest countries may result in changes to the aggregate data that do not accurately reflect the changes affecting all countries in the sample. Moreover, the actual change in the exchange rate is counted as part of the financial return, whereas it is the ex ante expected return that is relevant for portfolio allocation decisions.

Real interest rates on one-year time deposits became increasingly negative between 1977–78, when real rates were – 3 percent, and 1979–82, when real rates averaged – 5 percent (Table 3).

See Ize and Ortiz (1987) for a formal model of the effects of fiscal deficits on capital flight, and for a discussion of the relevance of these issues to Mexico.

The growth of broad money accelerated from 30½ percent annually during 1975–78 to 37 percent annually during 1979–82; similarly, inflation increased from 20½ percent annually to 27 percent over the same period (Table 3).

See Dornbusch (1985) and Harberger (1985) for a detailed discussion of the experience of individual countries.

The regressive effect of capital flight depends mainly on the ratio of land to capital in the asset portfolios of the rich and on the reaction of the government to the loss of taxes on capital. If land is a large component of the asset portfolios of the rich and if the government fully offsets the lost capital taxes by increased land taxes, any regressive impact of capital flight would be minimized.

    Other Resources Citing This Publication