The 1980s saw an unprecedented growth in the volume and the complexity of international financial transactions (Chart 1). This has been accompanied by a significant deterioration in the coverage and quality of the data. As a result, it has become very difficult, and at times impossible, for policymakers to base judgments on reported balance of payments statistics, particularly statistics on international capital flows. Unless appropriate action is taken, there will almost certainly be a further deterioration, with inevitable consequences for policymaking.


The 1980s saw an unprecedented growth in the volume and the complexity of international financial transactions (Chart 1). This has been accompanied by a significant deterioration in the coverage and quality of the data. As a result, it has become very difficult, and at times impossible, for policymakers to base judgments on reported balance of payments statistics, particularly statistics on international capital flows. Unless appropriate action is taken, there will almost certainly be a further deterioration, with inevitable consequences for policymaking.

Chart 1.
Chart 1.

Global Capital Account Transactions Compared with Transactions in Goods and Services, 1983–90

(Index, 1983=100)

Source: Balance of Payments Statistics Yearbook, 1991.

The problem is particularly serious for two reasons:

  • (1) Net errors and omissions in the balance of payments of several of the major industrial countries have, in some years, been so large that it has been difficult to ascertain each country’s true capital (and current) account position and, therefore, how much saving the country has been providing to, or absorbing from, the rest of the world (Chart 2).

  • (2) The sharp rise in these errors and omissions after 1988 indicates that statistical problems have worsened dramatically and may well continue to worsen in the absence of a major effort to improve the data.

Chart 2.
Chart 2.

Balance of Payments Accounts of Major Countries, 1990

(In billions of U.S. dollars)

Source: Balance of Payments Statistics Yearbook, 1991.

The weakness in capital flow statistics is also reflected in the “discrepancy” in the global capital account. In the absence of errors and omissions, this discrepancy should be zero because the sum of inflows in the world should equal the sum of outflows. However, the discrepancy has fluctuated over the last decade, and it amounted to $66 billion in 1989. Imbalances in the global capital account and its principal components, for the period 1986 to 1989, are shown in Table 1.2 Although these imbalances reflect inaccuracies in the capital account statistics, they do not necessarily measure them exactly because, for example, some positive errors cancel out negative ones, and some transactions are not recorded at all. Factors causing the rapid expansion of international capital flows during the 1980s have also been a primary cause of deterioration in the statistics. For example, the removal of exchange controls by a number of countries has resulted in a loss of valuable data sources. The internationalization of markets has meant that compilers can no longer rely solely on domestic data sources; investors have increasingly used overseas intermediaries beyond the reach of the domestic compiler. The trend toward the much greater use of transactions in securities has required compilers to widen the scope of their enquiries to institutions less able or willing to provide data than banks have been. As new instruments have been introduced onto the markets, new ways of tracking and recording them have had to be found. Many countries have tried, although sometimes rather slowly, to adapt their balance of payments statistical systems or introduce new ones. However, in some countries, the changes in the international financial environment have been too large and have produced too great a burden on reporters to be met without increases in resources. Furthermore, these changes have occurred at a time when many governments have been attempting to curb growth in public spending. As a result, too little has been done too late.

Table 1.

Global Capital Account Discrepancies, 1986–89

(In billions of U.S. dollars) 1

article image
Source: Balance of Payments Statistics Yearbook, Part 2, 1990.

No sign indicates an excess of recorded inflows over outflows; a negative sign indicates an excess of outflows.

LCFAR: liabilities constituting foreign authorities’ reserves.

Against this background, the Executive Board of the Fund established the Working Party on the Measurement of International Capital Flows in late 1989. This study follows the work of the Working Party on the Statistical Discrepancy in World Current Account Balances, the report of which was approved by the board in 1987.


The Working Party has found that world capital account statistical systems are in a state of crisis. At a time when important developments in international financial markets have occurred, the systems have failed to keep pace and to provide policymakers with adequate information. Problems are widespread. Although 10 industrialized countries account for 85 percent of the total reported capital flows and the solutions therefore lie largely in their hands, other countries must also contribute to the solutions. Improvement is possible, but it will take time; it will require considerably more international coordination than hitherto seen; and it will need additional resources in many countries.

The Working Party’s studies have revealed gaps in the data, nonreporting, misclassification of many transactions, inconsistencies, and lack of coordination among countries. Some of these problems have resulted from a failure to adhere to the guidelines for balance of payments accounting specified in the IMF’s Balance of Payments Manual (Manual). The Working Party has also found many examples of poor data quality, and, in some cases, lack of resources for the compilation of balance of payments statistics.

From 1986 through 1989, recorded global capital inflows in each year have exceeded global outflows by an average of $40 billion per year (Table 1). By making a number of adjustments to improve the past data, the Working Party has been able to reduce this global discrepancy by an average of $23 billion per year. However, substantial gaps remain (Table 2).

Table 2.

Global Capital Account Discrepancies, After Adjustment, 1986–89

(In billions of U.S. dollars) 1

article image

No sign indicates an excess of recorded inflows over outflows; a negative sign indicates an excess of outflows.

LCFAR: liabilities constituting foreign authorities’ reserves.

The capital account of the balance of payments is usually broken down into four broad components—direct investment, portfolio investment, other capital, and reserves. World inflows in each of these four components—and in the capital account total—should, in principle, balance world outflows. In practice, reported inflows have not equaled reported outflows in any component. As was previously noted, consideration of only the statistical discrepancies in a component total would hide an important part of the problem because errors in inflows or outflows could be larger than the recorded discrepancy. The following paragraphs comment on the findings in each of these areas.

Direct investment is the only category in which recorded outflows have persistently exceeded inflows during the 1986–89 period. This excess has averaged almost $17 billion per year. The main reason for the excess outflow is that many countries do not report the reinvestment of earnings of multinational enterprises as direct foreign investment. These earnings should be reported by the host country as well as the investing country. There is a tendency for reinvestment of earnings to be recorded as a capital outflow by the major investing countries but not to be recorded as a capital inflow by the host countries. Indeed, during the period under study, there has been an average excess of outflows over inflows in reinvestment of earnings of $22 billion annually. The Working Party has used available counterpart data from the main investing countries to make adjustments that almost eliminate imbalances in this area. Major countries that have not reported reinvestment of earnings include Japan, France, and Canada. Besides problems with reinvestment of earnings, the Working Party has also found incomplete coverage of other direct investment, exclusion of short-term financing between affiliated enterprises, omission of investments by affiliates in their parent companies, patchy recording of real-estate investment, and imbalances arising from activities of “Special Purpose Entities” of multinational enterprises.

Portfolio investment has become one of the most difficult areas for compilers because of the liberalization of capital markets, financial innovation, and the changing behavior of investors. These changes have made it increasingly difficult for statisticians to capture all securities transactions, and statistical systems have been slow to adapt. There has resulted a deterioration in coverage. Whereas most collection systems have traditionally approached domestic intermediaries for information on international portfolio transactions, investors have been increasingly switching to new intermediaries (including those abroad) for which coverage is less complete. There have also been growing classification problems—particularly a confusion between bonds and equities and between bonds and loans. Furthermore, bilateral comparisons have revealed very large differences, particularly between Japan and the United States, in countries’ estimates of flows between each other.

Table 1 shows a relatively small measured discrepancy, averaging about $6 billion during 1986–89, for portfolio investment. As mentioned before, this discrepancy could conceal larger errors and omissions. There have also been large fluctuations from year to year, as well as large imbalances within the individual components. The Working Party has not been able to explain these discrepancies fully. Information on investment positions (that is, stocks rather than flows) has been of little help in resolving these differences—largely because, for portfolio investment, investment positions are generally derived from cumulated flows and thus come from the same sources. Independent surveys of holdings of foreign securities by large institutions are essential to provide the necessary benchmark for comparison with flows.

Other capital is a heterogeneous group of international capital flows that includes transactions of the private nonbank sector, of domestic banks, and of resident official entities. This group has shown the largest excess, averaging more than $66 billion per year of measured inflows over outflows (Table 1). With only a very partial reporting of deposits by the nonbank private sector, coverage of nonbank sectors has been generally considered poor—notably in the United States. This view has been supported by the Working Party’s examination of international banking data from the Fund and the Bank for International Settlements (BIS). These “external” statistics have provided strong evidence that cross-border flows of both assets and liabilities of domestic nonbanks have been seriously understated in the balance of payments accounts; assets have been understated more seriously than liabilities. The missing net outflows could be responsible for as much as $50 billion per year of the adjustments reflected in Table 2.

Transactions in official reserve assets generally are well measured. However, because limited information is divulged on the instrument breakdown of reserves, it has been difficult to identify the counterpart transactions in debtor countries’ capital accounts. It is not always known where these reserves are invested and whether they are held as securities, as bank deposits, or in some other form. The Working Party was able to collect details on the instrument composition of reserve transactions for the main countries, filling a major gap in the accounts.

Other findings show that balance of payments information for offshore financial centers has been found incomplete; some activities have not been included at all (for example, Cayman Islands); others have been excluded as concerning only nonresidents or have been reported on a net basis. Omission of these data has created discrepancies in a number of the components of the global capital account. The Working Party has made some broad estimates for these centers and has included them in the adjustments.

The Working Party has been unable to identify capital flows that are deliberately concealed (drug money and other illegal activities), as such flows are indistinguishable within the measured components. Therefore, it was also unable to detect which countries are involved.

Inconsistencies have been found to exist between national balance of payments data and debt statistics provided to the World Bank and the Organization for Economic Cooperation and Development (OECD). The main cause of the problem appears to have been inadequate communication among the different agencies covering external debt transactions at the national level.

A deplorable deterioration in balance of payments reporting by a number of international organizations has occurred. During the last year, the Working Party and the Fund’s Statistics Department have achieved some improvement, but the continuing cooperation of international organizations must be secured.

Some data on international asset and liability positions have been made available to the Working Party. It is clear that improvements will be required in this area for many policy purposes. Data on stocks will be essential for validating measured capital flows and, as the 1987 current account study (Report on the World Current Account Discrepancy) showed, stock data are also needed to assess investment income recorded in balance of payments statements.

Major Recommendations

The Working Party has been able to identify a number of the sources of the global discrepancies and has made adjustments to published capital flows for the recent past. However, substantial gaps remain in the net data and probably even more exist in the underlying gross data. Although the Working Party’s adjustments have been useful in identifying gaps in coverage, the prospect for improving the quality of the underlying national data in the future is more important. The Working Party has identified a number of important measures that must be taken by individual countries if the quality of the figures is to improve. The main recommendations are listed below; many others of a more technical nature are developed in later chapters.

The Working Party’s findings have indicated the urgent need to begin enhancing the world balance of payments statistical systems to an acceptably effective level for users, particularly policymakers. Improvements will take time, commitment, and resources. The following paragraphs set forth the main steps that should be taken. Some recommendations have been directed to national governments and their statistical offices and some to international organizations, particularly the Fund.

Recommendations Applicable at the National Level

These recommendations are directed to all countries, and it is essential that all countries respond positively, particularly those countries with the largest capital flows.

(1) National authorities should keep their statistical systems under review and adapt them quickly to the changing national and international financial environment. They must ensure that their systems provide comprehensive coverage of all transactions.

During the last ten years, the failure to keep pace with a rapidly changing environment has been a major cause of deterioration in the quality of the capital account statistics.

(2) The authorities should reinforce statistical agencies that compile balance of payments statistics and provide them with the necessary resources, support, and legal powers for data collection.

While compilers are chiefly responsible for overcoming data difficulties, they cannot succeed without support from policymakers. In most countries, the required improvements in statistical services can be delivered only with some increase in resources. Strong legal powers of collection are essential if compilers are to achieve high coverage and good quality statistics.

(3) National compilers should prepare their balance of payments accounts in accordance with the Manual; they must report data to the Fund in this form and on a timely basis.

This extremely important objective must be met in order to avoid inconsistencies between countries and further growth in world discrepancies, to enable the exchange of data, to facilitate the Fund’s surveillance responsibilities, and to support domestic policymakers. It is particularly important that those countries that do not collect reinvestment of earnings statistics should now do so. Certain countries should also make greater efforts to obtain information about international transactions of offshore centers within their territory.

(4) Compilers should develop regular contacts with related agencies in their own countries, with compilers in other countries, and with international organizations that hold valuable financial data bases.

National authorities should monitor for effectiveness the communication between various agencies. Collecting related statistics within their countries will help to ensure efficiency, consistency, and the quality of data, such as debt statistics. Contacts between countries should be aimed at exchanging information; compilers must be able to collect statistics on a country-by-country basis and to exchange data (subject to confidentiality constraints for individual reports). The Fund and other international organizations, especially the BIS, have rich data bases on international portfolio and banking transactions, which could be used to improve balance of payments accounts. Especially important is the information gathered by these institutions on the transactions of domestic nonbank entities with banks abroad.

(5) Countries’ statistical systems should include regular collection of position data as well as flow data.

Data on stocks of international assets and liabilities are valuable in themselves. Moreover, they are an essential check on the coverage, geographical distribution, and classification of international transactions. These data are also needed for calculating or checking the income flows generated by these assets and liabilities, a point that was stressed in the 1987 Current Account Report. There would be considerable advantage in organizing a coordinated benchmark survey of international assets and liabilities among the major countries.

(6) National authorities should make regular reports, on a confidential basis if necessary, to the Fund on the composition of their foreign exchange reserve transactions by type of financial instrument.

Access to such detail is indispensable for drawing up the global accounts by adjusting the components that are counterpart to the reserve transactions.

Recommendations Applicable to the Fund and Other International Organizations

(7) The Statistics Department and other departments of the Fund should aim to achieve a greater harmonization of available balance of payments statistics and should seek to improve estimates for countries that do not report or report late.

(8) The Fund should update the Manual on a more frequent basis to accommodate changes in the economic and financial environment.

The fifth edition of the Manual is to be published shortly, some 15 years after the previous edition. Updates of the Manual should continue to be made in consultation with national compilers. The Fund should also continue to ensure that any updates or supplements are widely disseminated and are included in regular Fund training programs.

(9) All international organizations should report their international transactions to the Fund in a timely fashion.

Substantial capital flows are associated with the operations of international organizations, a number of which have not been providing the relevant data on a standardized basis. International organizations, particularly those with large international transactions, have the same responsibility to report their balance of payments transactions as do member countries.

(10) The BIS and the Fund should consider refining their international banking statistical systems in certain areas so that the systems will be more readily usable in balance of payments accounts.

The BIS and the Fund provide an extremely valuable source of international banking data for balance of payments compilers, particularly with respect to transactions of nonbank residents with banks overseas. However, the information could be made more applicable to balance of payments accounts by harmonizing the relevant concepts and definitions; for example, banks’ cross-border claims on nonbanks should clearly distinguish between loans and securities. The next BIS meeting of central bank statisticians should consider how these requirements could be met.

(11) The OECD and the World Bank should cooperate in making every effort to ensure that the debt statistics they compile can be reconciled with balance of payments data.

External debt data, collected and disseminated by the OECD and the World Bank, could be an important source of information for balance of payments compilers to use in verifying the accuracy of debt-related flows. Research by the International Working Group on External Debt Statistics has revealed large inconsistencies between the two data bases, which are, in part, due to conceptual differences.

One Final Recommendation

Effective follow-up mechanisms are required to ensure that the recommendations of this report are followed and that the quality of capital account statistics is improved. With hindsight, it is clear that recommendations (directed toward national compilers) from the Current Account Report were not effectively acted upon. The Working Party therefore recommends that

(12) A small standing committee of senior balance of payments compilers should be set up to oversee the implementation of the recommendations in this report (and those in the Current Account Report).

The group should be constituted under IMF auspices, meet at least once a year, and report annually to the Managing Director. The group should foster greater coordination between countries and assist the Statistics Department in keeping pace with the statistical consequences of the changing international financial environment.