Some capital flows are deliberately concealed by the transactors. “Capital flight” from certain countries and movements associated with illegal transactions, such as the drug trade, are examples of such activities. This chapter examines how such capital flows may be reflected in balance of payments statistics. The Working Party finds that “concealed” capital flows are so intermixed with “normal” flows that balance of payments entries can provide little insight into the scale of such activity. International banking statistics indicate that both industrial and developing countries have difficulties recording the transactions in their nonbank sectors.


The international debt crisis in the early 1980s drew increased analytic attention to capital flowing out of some countries in search of “safe haven” in others. Although the phenomenon of “capital flight” is not new, a considerable literature has developed on this subject in recent years.136 The literature has concentrated on important developing countries, especially in Latin America—Argentina, Brazil, Chile, Peru, and Mexico—and certain others, such as the Philippines and Korea. Various studies have presented empirical estimates of the magnitude of capital flight (and sometimes reflows) for the affected countries.

Much of the existing literature deals with the welfare and policy implications of capital flight against a background of economic development. Capital flight usually is examined in the context of developing countries with “capital shortages.” Typically these are countries that encourage net capital inflows to finance domestic investment and therefore have, or should have, current account deficits.137 Although capital flight usually is assumed to escape measurement in a country’s external accounts, the literature typically has little to say about balance of payments measurement.138

It is difficult to define “capital flight” and distinguish it from “normal” capital outflows. A definition that seems to express a reasonable consensus focuses on “capital that flees from the perception of abnormal risks at home.139 Flight capital is that which is influenced primarily by the owner’s desire to escape actual or expected domestic risks, rather than capital that moves in response to simple differences in investment incentives among countries. Thus, the emphasis is on the abnormality of existing or anticipated conditions in the country of the capital owner.

Motivations for flight range from macroeconomic mismanagement—such as exchange rate overvaluation or reckless fiscal policies—to fear of confiscation, desires to evade exchange controls, or simple evasion of domestic taxes on wealth or income. Putting private assets safely beyond the knowledge and control of the authorities is the common theme in this analysis. The incentive to hold assets in foreign havens is abetted by tax regimes in various countries and offshore centers that effectively exempt the investment income of nonresidents from taxation.

The Measurement Problem

Flight capital moves covertly if possible, and the movers of concealed capital can be presumed not to have an interest in accurate economic statistics. Can such flows nonetheless be detected in the balance of payments accounts of the country of origin or in the accounts of destination countries? Various studies have made the effort and attempted to quantify such flows.140 Although methodologies differ—sometimes sharply—the results typically support the view that private nonbanks (companies and individuals) sometimes have exported large amounts of capital from certain countries to safer havens. Estimates of the total flow between the mid-1970s and the mid-1980s range from $50 billion to $200 billion or more.

Table 58 illustrates the wide range of these estimates for the principal “flight” countries. The balance of payments accounts of these countries, in fact, suggest capital flight, insofar as they show negative errors and omissions during the periods studied and, in most cases, the recorded outflows of their private nonbank sectors are small. Further, their current accounts show suspiciously low amounts of income receipts from foreign investments. Against a background of a positive global capital account discrepancy, the movements associated with capital flight seem a promising mine in which to prospect for unmeasured outflows.

Table 58.

Range of Cumulative Capital Flight Estimates for Selected Countries, 1976–841

(In billions of U.S. dollars; outflows ( – ))

article image

Tabulated by Sunil Gulati, “Capital Flight: Causes, Consequences, and Cures,” Journal of International Affairs, June 1988, p. 173.

Balance of Payments Accounting Framework

Balance of payments compilation systems have special difficulties capturing transactions of private non-banks, and private nonbank sectors usually are thought to carry out most capital flight. The statistical problem is to identify how flight capital can elude compilation and what the repercussions might be.

When, for whatever reason, nonbank residents of a country determine to acquire foreign assets, several avenues are available. For those involved in exporting or importing, a well-known device is to underinvoice exports or overinvoice imports and accumulate the difference in foreign assets. In consequence, the country’s current account will register a smaller surplus, or larger deficit, than the true balance, which statistically is equivalent to the omission of a capital outflow. There will be no direct evidence in the reported capital account that an outflow has occurred.141 Capital flight can also occur if domestic residents abandon domestic assets in favor of foreign assets. Residents may, for instance, purchase foreign exchange from other domestic sectors and retain it in deposits (or securities) abroad. In that case, there may be a reduction in the recorded foreign assets of other resident entities—for example, of banks or of the government. When such a reduction occurs, the statistical record may show—paradoxically—a capital inflow as recorded foreign assets decline, but the counterentry for the increase in the capital exporter’s foreign assets is missing. Such a sequence of transactions results in negative errors and omissions in the balance of payments accounts. A similar result is obtained if the capital flight is “financed” by an inflow of foreign capital that is captured in the balance of payments statement while the flight outflow is not.

Many other avenues for exporting capital can be cited, but the essential point is that a concealed capital outflow may not be evident at all in a country’s balance of payments accounts, or it may be reflected only in a residual category. Ex post, the overall net capital flow of the country will be equal to its true current account balance (the recorded balance may be altered because of capital flight). Consequently, the capital flight will have to be absorbed within the overall capital flow, sometimes as a reduction in other private or official external assets of residents, or sometimes by the receipt of larger capital inflows from foreign lenders or investors. Strong incentives such as exchange rate depreciation may be required to clear the markets whenever there is great political or economic uncertainty and a tendency for capital to flee. Strong downward pressure on an exchange rate, in fact, often is a symptom of incipient or actual capital flight from affected countries.

These observations suggest that capital flight is preeminently a sectoral phenomenon; in some cases, the capital export from one sector is balanced by a capital inflow to another. The sectoral aspects of capital flight have drawn some attention in the literature. It is widely held that in some countries the public sector’s foreign borrowing has financed capital flight via “round-tripping” mechanisms.142 For example, foreign exchange arising from government borrowing can wind up in the hands of the private sector if the authorities attempt to defend an exchange rate that the market regards as unrealistic. Round-tripping potentially leaves the authorities with more external liabilities (or reduced assets, such as reserves) and private entities with higher external assets than before the capital flight was accomplished. Mohsin Khan and Nadeem Ul Haque (1985) and others describe cases of government borrowing that facilitate capital flight by private entities. One consequence for the flight country is that income that would have been received from foreign assets will be lost (or more interest will have to be paid on a larger debt), while no income is received—or, at least, recorded—from the flight capital.143

National and global balance of payments discrepancies can easily arise from the kinds of transactions that take place under these conditions. Even in the best of circumstances, compilation systems are far likelier to capture government capital imports than private exports, especially if the latter are concealed. Depending on the specific sequence of transactions, the traces of capital flight that might be inferred from balance of payments statistics are not always the same, and even the sizes and signs of discrepancies can sometimes be misleading.

Alternative Data Sources

The general conclusion of this short review is that balance of payments data are of very limited use in assessing capital flight, even though the fact of flight is not in doubt. Private sectors in some countries undoubtedly export capital in a manner that deliberately eludes reporting systems, and national balance of payments figures give only slight insight into the magnitudes of such movements. However, this is a problem that “capital flight” countries share in common with many others. Reference to “partner country” or “haven country” statistics is the only feasible check on national data to roughly quantify the flows that may not be captured in balance of payments figures.

Flight capital can be directed into various kinds of assets, including portfolio securities, bank deposits, and direct investment (such as real estate). Limitations in partner country balance of payments statistics usually preclude direct comparisons with statistics of specific “flight” countries, but for the important category of banking flows some information can be derived from the international banking statistics published by the Fund. A broad view of the “derived” IBS statistics already has been presented in Chapter 6 in connection with statistical adjustments to capital flows of nonbank sectors for major country groups. The discussion that follows focuses on statistical contrasts for a few countries that often are mentioned in capital flight discussions.

Table 59 compares, for several countries and country groups, IBS data on stocks and flows of “nonbank deposits with foreign banks, by location of depositor” with balance of payments versions of net changes in assets of private nonbank sectors. For reasons mentioned in Chapter 6, balance of payments outflows should be at least as large as flows derived from the IBS series.144 Usually the reverse is true; IBS statistics commonly suggest larger increases in banking liabilities to nonresident nonbanks than balance of payments figures show for nonbank assets.

Table 59.

Comparison of IBS and Balance of Payments Statistics on External Assets of Nonbanks for Selected Countries and Groups

(In billions of U.S. dollars; outflows ( – ))

article image

IBS flows derived by the Fund.

Derived from Balance of Payments Statistics Yearbook data on “other assets of other sectors.”

As noted in Chapter 6, Mexico uses IBS data as the basis for its balance of payments estimates.

A comparison of the last two columns of Table 59 suggests a rough scale for unrecorded capital outflows. For the five developing countries, the IBS-derived outflows total about $23 billion during 1984–89 and exceed the comparable balance of payments figure by about $9 billion. For several reasons, this $9 billion difference is probably a minimum bound on the size of unrecorded outflows. One reason is the different coverage of the banking and balance of payments statistics mentioned earlier. Another is that IBS show large increases in liabilities to “nonbanks” ($54 billion in flow terms) in important offshore financial centers. Flows of this size are far too large to have originated in such centers. Most of these outflows must have been financed by funds from elsewhere in the world that were simply channeled through the OFCs. Both industrial and developing countries—including those listed in the table—likely provided portions of these capital flows.

The purpose of Table 59 is not to suggest specific figures for “capital flight” from certain countries but to underscore the limitations of balance of payments data as a basis for the analysis. As explained in previous chapters, the banking figures have their own limitations but, even so, the comparisons provide some perspective on the size of unmeasured outflows.145 U.S. figures are shown separately in the table mainly to highlight the danger of using statistics on nonbank assets held in banks to draw conclusions about capital flight, as opposed to “unmeasured outflows” generally. Although the United States is not considered a “capital flight” country, the contrast between IBS and balance of payments figures causes the appearance of capital flight to be greater for the United States than for the developing countries. These comparisons merely emphasize the fact that important problems afflict balance of payments compilation systems in industrial and developing countries alike; countries of very different kinds face similar statistical problems with respect to measuring transactions of private nonbank sectors.

Although this discussion has dealt only with traces of outward capital flows, inflows through nonbank sectors likewise are poorly recorded, especially in some industrial countries (see Chapters 5, 6, and 9). Indeed, in some cases, the degree of underreported inflows through nonbanks appears to exceed the scale of under-reported outflows.146

Drug Trafficking and Other Illegal Activities

Estimates of Market Size and Capital Flows

Capital flows associated with the investment of proceeds from the so-called underground economy, drug money transactions, money laundering, and other illegal activities are similar to capital flight in the sense that they are deliberately hidden. Empirical work, however, generally has focused on the amounts of drug transactions within individual countries. Thus, various estimates have been made of the scale of such transactions in the United States and elsewhere. For example, the recent Report of the Financial Action Task Force on Money Laundering (FATF)147 examined and put into a rough quantitative framework the typical sequence of actions—beginning with crop production and extending through subsequent stages of processing, pricing, and retail sales—involved in the drug trade and laundering dirty money. Part I of the FATF Report estimated the annual size of the U.S. cocaine market at almost $29 billion. A number of other recent books and articles also have treated aspects of the dirty money business.148

As the FATF Report makes clear, however, it is easier to describe how such markets work than to quantify the results. Large margins for error lurk at almost every stage of the estimating process. The uncertainties in quantification, alone, would make efforts to trace flows of dirty money through balance of payments statistics most difficult. The remainder of this chapter will illustrate some of the analytic problems associated with using available balance of payments figures to isolate drug money flows or, indeed, those associated with other illegal activities. Such statistical problems are quite similar to those posed by capital flight.

An important initial point is that estimates of the scale of drug activities, money laundering, and the like are not the same as the probable amounts of associated international capital flows. Many of the drug transactions considered by the FATF are domestic, although some do have an international dimension. Much of the gross turnover (and profit) is “domestic” to both producing and consuming countries. At some point in the chain of transactions, of course, the residence of transactors changes; one country is an exporter and another is an importer of the drugs. At that point, there occurs a potentially recordable international transaction, although actual recording is not likely.

Some of the literature on the drug trade tends to confuse the estimated gross value of drug transactions (at the street level) with profits generally and, more specifically, with profits made by foreign operators or cartels. A more accurate picture is drawn by considering the entire chain of dealings from initial producer to final consumer in value-added terms. In the drug trade, for example, each participant in the chain faces costs (inputs, processing, bribes) that are included in the markup on the product. Only actual profits at each stage constitute investible proceeds, however. Profits are only a portion of sales, and, at most, only a fraction of all estimated transactions are sales across international borders.

This chain of events can be expressed in simple arithmetic form. The “street value” of drugs has a “value added” of V, consisting of the sum of imported (V1) and domestic (V2) components. The variable V1 can be thought of as the value added within the country of origin and includes markups when drugs are sold across international boundaries. The variable V2 is the further increase in value in the consuming country. The FATF and other analysts of these markets think that the highest value added occurs in the consuming country: that V2 is the larger of the two terms. Thus, the estimated sales value of drugs across international borders is thought to be less than half the dollar size of the domestic final market, probably much less than half. According to the FATF study, only about $5 billion of the estimated $29 billion in U.S. cocaine sales consists of international transactions; domestic markups account for the remaining $24 billion, or more than 80 percent of the retail value of this drug.

Two aspects of this example are relevant in connection with the problem of capital flow imbalances in balance of payments statistics. One is that estimates of market size should not be confused with potential capital flows. Second, even the gross export values of drugs (and other illegal goods) are not a reliable measure of associated capital flows. In terms of balance of payments accounting, the export value of drugs is like the value of any other export. This export must have as a counterpart either offsetting current account imports or a capital flow, or some combination of the two. It is recording asymmetries, rather than legality, that give rise to discrepancies. Misrecording of conventional, legal transactions—or, indeed, of flight capital—have the same statistical effect.

The total export value of drugs undoubtedly is shared among a number of domestic transactors in the exporting country. In each instance these transactors face costs of doing business that account for some portion of their proceeds and further diffuse the receipts before profits can be spent or invested. Among the costs are likely to be expenditures on imports (processing materials, weapons, and so forth) that in an accounting sense would diminish the potential capital flow associated with the trade. In that respect, the actual capital flow related to drug transactions may well be much less than the export value of the drugs themselves.

In addition, whatever net capital flows result may take highly variable forms that may or may not be recorded in some set of statistics. Even if recorded, these net capital flows may not appear in the accounts of either of the countries between which the initial transaction took place. To illustrate, a farmer-producer in the exporting country may be paid in, or acquire and hoard, currency—a form of capital flow that will not be registered in any country. Middlemen in the chain, perhaps in another country, might deposit their profits in an OFC, or they may import goods from a third country, with the exporter of those goods putting his or her profit into U.S. Treasury securities. Top members of the cartel may acquire real estate somewhere else in the world, and so forth.

The list of possible permutations of such cases is extensive and merely illustrates two things. First, even if the export value of drugs and the net value of associated total capital flows could be reliably established, the distribution of such flows into various assets is likely to be highly dispersed. Second, this dispersion may also have complicated geographical aspects, sufficient to obscure in any available statistics the ultimate origin of the funds in question.149 To summarize, even specific knowledge that one country exports, say, $100 million of drugs to the United States or elsewhere is insufficient to establish any of the following: whether an associated capital flow takes place; how large the flow might be; what forms such a flow might take; and even whether the associated flow is between the original transactor countries.

Drug Flows and Balance of Payments Statistics: Two Examples

The foregoing section suggests in general terms the conclusion that tracing drug money flows through balance of payments accounts is an intractable problem. This section will briefly illustrate aspects of measurement and the related issue of discrepancies in published balance of payments accounts. The measurement of illegal transactions obviously is imperfect. When all elements of such transactions are missed by both domestic and foreign compilation systems, no discrepancies arise in either national or global balance of payments accounts. More likely, however, is the case that only some portions of the related current and capital account flows are missed. As with capital flight, when this occurs, national and global balance of payments discrepancies are a result.150 Even in the best of circumstances, isolating capital flow totals or specific countries associated with such transactions will be most difficult.

Tables 60 and 61 illustrate numerically just two of the many possible balance of payments consequences of exports and imports of illegal goods. In each example, it is assumed that country A imports illicit goods with a value of 10 from country C. The third country E has related current account transactions or is an offshore center or haven providing financial services. Both examples, however, are applicable to both legal and illegal transactions in the sense that the focus is on the imperfection of measurement rather than on the legality of the transactions. The example shown in Table 60 is one in which no statistical imbalance arises in the global current account. However, there is a positive imbalance in the global capital account, and negative errors and omissions result from the described transactions. This example corresponds to intuitive expectations about the mismeasurement of illegal transactions and associated concealed capital flows. In the case shown in Table 61, the pattern is somewhat more complicated and the results are different. The global current account is unbalanced, but there is no trace of excess capital inflow. Again the global errors and omissions are negative, but whether this arises because of undermeasurement of imports or unrecorded capital outflows, or both, cannot be determined from the recorded data.

Table 60.

Illegal Trade as Reflected in Balance of Payments Accounts: Example 1

article image
Table 61.

Illegal Trade as Reflected in Balance of Payments Accounts: Example 2

article image

These are only two of many possible scenarios. Entirely different results can be produced by merely changing the assumptions about which portions of related international transactions are recorded and which are not. Some combinations will leave counterintuitive discrepancies in the principal account balances, and some will leave none.

The “typical” result of unmeasured drug transactions (and unmeasured capital flight, generally) will probably be consistent with the generally positive capital account discrepancy that has emerged in global accounts over the past decade. However, the size of that discrepancy should not be taken as a proxy for the amount of capital flow related to illegal drug trafficking. Apart from analytic factors mentioned in this chapter, this is because many other kinds of capital outflows are also more poorly measured than the corresponding inflows to receiving countries, as is illustrated here and in other portions of this report.


In some instances capital flows, especially outflows, may be deliberately concealed from national authorities and compilers. Both outright “capital flight” and flows resulting from illegal activities may lead to imbalances in global capital accounts. There is clear evidence that balance of payments compilers fail to record large amounts of both legal and illegal capital outflows from nonbank sectors. Discrepancies in the global data arise when haven countries record increases in foreign liabilities that are not compiled as assets in the exporting country’s accounts. A comparison of international banking data with balance of payments figures suggests a scale for capital outflows from “flight” countries but also reveals similar gaps in recording for transactions of nonbank sectors in important industrial countries. The statistical record of “capital flight” in both industrial and developing countries is quite similar to that of “normal” unrecorded capital flows—both inflows and outflows—of nonbank sectors. Consequently, extreme caution is warranted in drawing conclusions, on the basis of available balance of payments data, about the magnitude of capital flight from any country.

Capital flows resulting from drug money and other illegal activities are also deeply intermixed with other kinds of poorly measured flows. The Working Party found that the size of these flows probably is overstated in popular conjecture and in the literature. To a limited extent, the size of domestic “drug markets,” and even the gross value of purely international transactions that feed these markets, can be evaluated, but capital account statistics provide little insight into capital movements specifically related to this trade. The intrinsic flaws in existing compilation systems are such that discrepancies and imbalances would arise even in the total absence of drug trafficking and similar illicit activities.

Although the scope for improving estimates of such flows is limited, the Working Party nonetheless notes that better balance of payments coverage of certain haven countries—especially OFCs—and better geographic detail in international banking statistics might be helpful in improving estimates of “unrecorded” capital flows of nonbank sectors. Recommendations about such improvements are detailed in Chapters 6 and 9 of this report.