In the process of deriving adjustments to the reported balance of payments data to correct the statistics and reduce some of the global discrepancies, the Working Party has also attempted to allocate these adjustments by region and, in a few cases, by country. It must be emphasized, however, that such allocations are, at best, partially judgmental, and the basis for making them ranges from fairly good for the investment income data to rather poor, at this stage, for the shipping, transportation, and transfers accounts. Moreover, while introducing such adjustments may reduce the global discrepancies considerably, this does not necessarily lead to a clear-cut improvement in the analysis of the external situation of particular countries. Since the allocation problem is different for each of the main categories studied by the Working Party, it seems best to describe each one in some detail before attempting an overall summary.

a. Investment Income

The Working Party analyzed the international investment income accounts in considerable detail and was able to suggest adjustments that eliminated most of the discrepancy. However, major problems remain for the geographic analysis of the direct investment portion of those adjustments. For reinvested earnings, the indicated adjustments can be based on any one of three premises: (1) the credits are likely to be reported more completely and accurately, so the appropriate adjustment would be to add debits where they seemed to be inadequately reported; (2) since an excess of credits seems to be characteristic of this sector, the discrepancy could be reduced by eliminating the credits for which there are no corresponding debits; and (3) since there is a built-in bias in this sector, it would be appropriate simply to eliminate all the data on reinvested earnings when aggregating the world current account. Clearly, the balances of regions, as adjusted, would be considerably affected by the choice of the adjustment procedure—at least until the reporting of reinvested earnings was greatly improved.

There is a further complication for reinvested earnings, because whichever adjustment is made for any country or region there should be, by definition, an opposite and equal adjustment in the capital account, so that the net arithmetical effect on the total balance of payments statement would be zero. Nevertheless, there would be some gain in understanding the economic significance of the operations of these enterprises, for both creditors and debtors, if an effort were made to include all their earnings, whether distributed or not, in the international accounts. On that basis, the first column of Table 78 shows the effect of retaining the amounts reported as credits, and compensating, on balance, by inserting debits in the locations where the analysis completed up to this point suggests that they should be. The result is to reduce somewhat the net credits of industrial countries but to add a sizable amount to the debits of developing countries.

For “other,” or distributed, direct investment income, the geographic allocation of adjustments is shown in the second column of Table 78. For this category, the adjustments tend to raise credits or reduce debits, thus adding a considerable amount of net credits to industrial countries and Middle Eastern oil exporters. However, a considerable global net debit, centered on the developing countries, remains under this heading. As noted in Chapter IV, these remaining discrepancies seem to be traceable to a few countries, and further analysis may show whether the debits were reported erroneously or whether they should be shifted to another payments category without affecting the overall current account discrepancy.

In the case of portfolio investment income, which is comprised mainly of returns on positions with banks and on holdings of securities (including assets held as official reserves), results are shown in the third column of Table 78. The suggested adjustments almost eliminate the discrepancy under this heading, but this result should not be taken too literally—it is the result of a rather complicated process (discussed in Chapter V), and no doubt there are many offsetting and accidental errors remaining. Nevertheless, the body of statistics and the methodology underlying this set of adjustments seems to the Working Party to be relatively well founded. The net effect of these adjustments is to add to the net income credits reported by all areas except Eastern Europe, but to add rather more net credits to industrial countries and offshore banking centers than to developing countries. It should be noted that $5.9 billion of the suggested increase in net credits cannot be allocated geographically, reflecting mainly the lack of information on residences of asset holders in the banking position data. For convenience, the adjustments for investment income data, apart from the more problematic allocation of reinvested earnings, are shown in Table 77.

Table 77


(In billions of U.S. dollars)

article image

Excluding reinvested earnings.

Includes those countries that were not Fund members.

b. “Shipment” and “Other Transportation”

In the case of the large net debit shown for the “shipment” and “other transportation” categories, taken together, the Working Party at this stage can offer only interim answers on the likely geographic distribution. However, some general comments may help in the further pursuit of this subject. As was shown in Chapter VII, the largest discrepancy is an excess of debits (averaging over $30 billion per year recently) in the “shipment” account. This discrepancy results primarily from the lack of reporting of freight revenues by several countries or economies whose residents operate very large merchant fleets. On the basis of this information, it could be inferred that most of the missing credits could be attributed to ships owned or operated by persons or enterprises associated with Greece, Hong Kong, the U.S.S.R., and other countries of Eastern Europe that were not members of the Fund. In addition, there is probably some understatement of credits by industrial countries that may have some difficulty in capturing information on all the ship operators resident in their countries. Further, sizable revenues are probably accruing to ships operated by multinational companies via subsidiaries organized in countries that do not compile data of this kind. Such ships are probably not counted among the fleets operated by residents of the home countries of such multinationals.

The Working Party has not been able to carry through a detailed evaluation of the shortfall in reported revenues, but the first of the possibilities mentioned above is undoubtedly the most important. Further research will be required to assess the importance of the other factors mentioned. The sizable portion of net revenues generated by chartering ($6.7 billion) cannot be allocated by residence and remains unallocated at this time.

Corresponding to the missing revenues traceable to nonreporting by countries where operators reside, there is also a set of unreported debits by these countries representing the foreign operating costs of these fleets. Such debits should appear in the “other transportation” category. However, such an adjustment raises the existing discrepancy in that category, confirming that a major problem also exists because of underreporting by many countries of “other transportation” credits. Some estimates of where these missing credits are located are given in Chapter VII, but these are still rather impressionistic. A summary of the indicated adjustments is given in Table 78.

Table 78


(In billions of U.S. dollars)

article image
Note: Details may not add to totals because of rounding.

International Monetary Fund, Balance of Payments Statistics Yearbook, Vol. 36 (Washington, 1985), Part 2.

Includes countries that were not Fund members.

Even if all the suggested estimates for the shipping and transportation accounts could be made, however, the improved arithmetic results would still be inadequate for the analysis of the economic impact of these transactions on individual countries. For instance, the fleet operated by residents of Greece may generate large revenues and expenditures, but the net result may be relatively small and could be either a credit or debit balance in any given year. Moreover, the cash flow representing the net operating result, as well as the flows associated with the purchase and financing of ships, may in large part be retained abroad by the operators and may not affect the Greek economy, apart from actual requirements for expenditures in Greece. To some extent, this result parallels the situation in the investment income accounts, where identifying the countries whose credits or debits are not correctly reported is a necessary step in understanding their underlying economic situation, but is not sufficient to enable one to understand the full significance of the missing data.

c. Unrequited Transfers and “Other” Services

Attempts to reduce the discrepancies in the reported data on unrequited transfers are also in an early stage, so that the analysis of geographic allocations is obviously sketchy. The Working Party believes, however, that adjusting for a few of the major anomalies may go some way toward correcting the record. There is also a considerable element of asymmetry as between unrequited transfers and “other services,” depending on the point of view of the reporting country. Since a few of these asymmetries are discussed in the Report (although the Working Party has not investigated “other services” as such), they are included in this brief discussion of the geographic allocation problem.

As shown in Table 78, the main adjustment derived from Chapter VIII involves an addition of $7.0 billion to cover the excess of net reported transfers to international organizations for which no data are included in the present Fund tabulations. It will also be noted that $3.1 billion of credits was added to net portfolio income for these institutions. Of this combined total of about $10 billion, it is very roughly estimated that about $7 billion should be inserted as a debit adjustment under “other services” to represent the unreported expenses of these organizations; these expenses are, in principle, among the credits reported by countries where the expenditures are made. In fact, little is known about the accounting for these operations, and additional research is needed to establish an adequate basis for a final adjustment.

A further net credit, amounting to $3.4 billion, is added to transfer receipts of other developing countries, representing a minimum amount by which their reporting appears to understate receipts of development assistance. On the other hand, $2 billion is added to the transfer debits of Middle Eastern oil exporters to make up for data not reported separately but perhaps now recorded under “other services.” It is not yet possible to estimate what the geographic distribution should be for the recipients of reported pension payments or for the payers of workers’ remittances as reported by the receiving countries.

When these adjustments to transfers and other services are inserted in Table 78, the effect is to enlarge the discrepancy under “other services” to $17.0 billion and to shift the total under transfers from minus $6.2 billion to $2.2 billion. Since these categories are intermixed to some extent, it might be best to combine them, noting that the combined net debit was $16.2 billion as originally reported and remains quite large, at $14.8 billion, after adjustment.

The net result of this preliminary set of adjustments is to reduce the overall discrepancy on services and transfers from minus $85.0 billion to about minus $27 billion. Including the merchandise trade balance the result, shown in Table 79, is to reduce the overall current account discrepancy from minus $75 billion to minus $17 billion. In the process, all areas except Eastern Europe emerge with larger net credits or reduced net debits. The allocation of these adjustments should be viewed as experimental at this stage, but is probably accurate enough to enhance the analysis of the data. Even on an experimental basis, however, a considerable part of the adjustment remains unallocated.

Table 79


(In billions of U.S. dollars)

article image

International Monetary Fund, Balance of Payments Statistics (BOPS) Yearbook, Vol. 36 (Washington, 1985). Part 2.

See Table 78. Adjustments cover only selected current account sectors.

Includes only those countries that were not Fund members.

Not included in BOPS Yearbook data, adjusted balances reflect only indicated adjustments.


a. Analytical Aspects of Allocations

The principal finding of this study, from the point of view of the analysis of world economic developments, is that the negative discrepancy in the world current account is not concentrated in any individual country or grouping of countries to such an extent that the interpretation of basic economic problems has been distorted. Nevertheless, adoption of the adjustments suggested in this report, together with continued emphasis on improving the quality of the available data and estimates of international transactions, will help to sharpen the focus of advice and guard against any further deterioration of the basic data. In a few cases the adjustments suggested may cause some reappraisal of economic policies, and they may be useful in the context of Article IV consultations. However, it should be remembered that the Working Party cannot reconstruct national accounts as they would be if there were no statistical problems, but can only suggest adjustments that would alleviate certain evident discrepancies. Moreover, the range of uncertainty in the adjustments widens as the focus shifts to individual countries, though the direction of the indicated change may be fairly clear.

Another finding of the study that carries analytical implications is that there would be considerable benefit from enhanced integration of the research and data collection functions in individual countries and at the level of international organizations. When the aggregated data are submitted to analysis not only on a regional basis but also in terms of the substantive content of the various types of transactions, the perspectives of those responsible for the compilation and analysis of data would be brought closer together, and there would be earlier awareness of emerging discrepancies and of changes in the economic environment that might call for corresponding adjustments in compilation techniques.

A specific contribution of the procedures adopted by the Working Party is that they should enhance the dialogue between the Fund and member countries in the context of the Fund’s surveillance responsibilities. For such purposes it is essential not only that a common body of statistics be available but also that the parties be aware of any gaps or discrepancies in basic data that might lead to analytical errors or disagreements. In particular, the results of this study should be taken into account in the context of analyses of the problems of international indebtedness.

Finally, it is recommended in this report that the Fund and the OECD apply adjustments regularly to the current account balances of geographic groupings as reported, so as to effect corrections of the discrepancies that have been identified. Such adjustments may of necessity have a large judgmental element, but they will also provide an opportunity for observing the kinds of abnormalities in national reporting that have been the cause of some of the large year-to-year swings in the global discrepancy. As noted in Chapter I, such swings introduce a considerable additional uncertainty into the projecting of economic activity and savings/investment balances.

b. Allowance for Hidden Assets

The geographical allocation of the discrepancies in the investment income accounts in the Report is based primarily on readily available data on stocks of assets and liabilities. There is a general perception that there is also a large stock of hidden cross-border assets and liabilities that would generate unrecorded income flows. This subject is discussed briefly in Appendix VI, but the results are necessarily inconclusive. On the one hand, it would appear that the stock of external assets, at least for developing countries, that underlies the income estimates in this report is lower than various estimates of unrecorded plus recorded capital outflows, but is nevertheless very large. On the other hand, it is clear that estimates of “capital flight” do not generate an asset stock usable as a basis for income estimates for countries or even regions.

c. Effects of Offshore Financial Centers and Financial Innovation

One of the major concerns leading to the initiation of this study was the possibility that the growing use of international financial centers, especially those relatively newly established in offshore locations, might be causing a deterioration in the statistics on international transactions. As noted in Chapter VI, the tendency to pass transactions through these centers does create asymmetries in the current account (with the countries where the ultimate transactors reside possibly accounting for the same amount of funds, but under different transaction categories) and may also create offsetting discrepancies as between the capital and current accounts as recorded in third countries. Of course, the use of these centers makes it difficult, if not impossible, to trace the flow of funds between particular final parties to transactions. Such tracing is not essential for accuracy in the overall balance of payments accounts of a country, but as these centers encourage the breakdown of geographic boundaries, they undermine the essential resident/nonresident demarcation of the international accounting framework.

It appears that the major immediate problem created by these centers is not their functioning as intermediaries (which is, after all, a long-standing feature of the international scene) but their failure to supply information to the Fund on their balance of payments accounts and, in some cases, their failure to supply geographic data on the assets and liabilities of the banks and other financial intermediaries operating in their jurisdictions. The lack of data on their own balance of payments accounts leads to sizable omissions respecting the activities of their own residents and of the enterprises or intermediaries established in these locations by foreigners. The extent of the omissions is indicated in Chapter VI. The lack of geographic detail on banking positions hampers the attempt to allocate income by country or area, as was done in Chapter V, and, as these positions grow, they will loom larger as a barrier to geographic analysis.

In summary, the operations of international financial centers cause difficult statistical problems but do not seem to be responsible for a large part of the growth in the current account discrepancy over time.

A parallel development has been the introduction of new methods for conducting international financial transactions, which have been associated with the diminution of barriers to capital flows and the ease of conducting business via computers and improved communications, together with the remarkable spread of innovative financial instruments. These developments challenge the statistical apparatus designed to capture international capital flows, partly because of the sheer size and speed of the transactions and partly because it becomes more difficult to fit the transactions into the traditional categories of the balance of payments accounts, and especially into geographic boundary lines.

Probably the most disturbing elements of the innovation process from the point of view of balance of payments compilation are the following: (1) the nature of the transactions becomes obscure as a chain of transformations occurs via swaps of all kinds, so that at any given time the instrument involved may be short term or long term, denominated in local or foreign currency, bank loan or marketable security, etc.; (2) the identity of the ultimate creditors or debtors may be more difficult to establish, threatening a breakdown of the residence principle essential for international accounting; and (3) the transactions are increasingly carried out between the final parties, with banks arranging the transactions in many cases, but not carrying the resulting international assets or liabilities on their own books. This last factor is especially important, because it is much more difficult to obtain the relevant information from the universe of potential transactors than from the more limited, and more statistically experienced, banking sector. A related facet of this tendency to diversify financial transactions is that the largest multinational companies, which are often the major parties in the transactions, also have less incentive to pay attention to national boundaries when accounting for their financing operations.

It may be that these difficulties will be accommodated over time by modifications in the statistical process, but there is some evidence that the ability to measure overall capital flows has degenerated, reinforcing the tendency to capture inflows more effectively than outflows, which has become obvious since 1979. The ability of compilers to collect data on such flows in meaningful geographic breakdowns has never been robust and must be increasingly doubtful. The effect of innovation on the geographic allocation of current account transactions would be less severe, but for countries making such allocations based on information regarding the currency of payment or the location of the account being credited, the record could become increasingly misleading.


The regional allocations of discrepancies attempted in the Report can certainly be refined and extended to more categories of transactions. We believe the Fund could adopt the procedures we have developed and could prepare estimates of the appropriate adjustments. The size of the adjustments needed will vary over time depending on (1) whether the underlying factors causing the discrepancies are becoming stronger, and (2) whether the countries whose accounts are deficient are beginning to make the called-for corrections. For broad analytical purposes these adjustments need not be precise, but some careful analysis will be required.

One major element in refining the allocation process for the income accounts would be continuing efforts by the Fund to obtain more geographic information from some countries reporting substantial banking positions.

There is a connection between the process of allocating the current account discrepancy to countries and regions and the efforts by individual countries to produce geographic breakdowns of their international accounts. This is seen most clearly in the work on direct investment income, where the existence of discrepancies, and their location, can be effectively traced only through comparisons of individual country data generated by a few major creditor countries with the data available in host countries. In principle, the matching of bilateral data should help to identify discrepancies in other accounts as well, but experience with such comparisons suggests that they are difficult to achieve and time-consuming. Regional allocations of such transactions as income on positions with foreign banks, for instance, will be almost notional if they largely reflect the flows vis-à-vis international financial centers.

From the point of view of reducing, or at least understanding, the discrepancies in the current account, the Working Party can recommend intensified work on sorting out direct investment transactions geographically. In addition, more complete country detail on banking positions and on the ownership of securities would be very useful. Progress could surely be made in sorting out the shipping and transportation accounts if better data on the industry could be assembled. It also appears that discrepancies in the unilateral transfer accounts could be explored more effectively with greater country detail. However, whether each country should be urged to prepare geographic distributions of a wide range of transactions is a broader question that does not seem to be critical for dealing with the discrepancy problem.