Information on stocks of international assets and liabilities is a valuable adjunct to data on capital flows. Position statistics, based on surveys or other sources independent of the flow data, can be used to verify the accuracy of flow data and to provide geographic allocations not usually available from flow measures. Stocks are also analytically useful in themselves, since they cumulate the results of cross-border investment over the years and they give a basis for generating and evaluating estimates of international investment income.

Global Asset and Liability Positions

The Role of Position Statistics

There are several ways in which the collection of data on cross-border assets and liability positions can contribute to the evaluation of a country’s international economic situation.151 First, such data present a comprehensive picture of the cumulative results of a country’s international transactions (and other factors affecting values) over time. They can also give some perspective on the types of assets and liabilities in a country’s investment position. In particular, data on stocks of liabilities, by type, can alert market participants and monetary authorities to potential problems of liquidity or excessive debt burdens. Also, data on international stocks of assets and liabilities can be used for across-the-board country and regional comparisons of important economic structures. Another use that is important to the Working Party’s assignment is that statistics on international assets and liabilities can serve as a check on the capital flow data that enter the balance of payments accounts. Position statistics can also serve as a basis for estimates of investment income receipts and payments.

At an early stage in its studies, the Working Party concluded that as much data as possible on international stocks of assets and liabilities should be brought together and that an effort should be made to reconcile the data on capital flows and stocks. The Special Questionnaire circulated by the Working Party therefore contained a detailed breakdown of countries’ stock data, which was designed to match flow categories, and a reconciliation section. This emphasis on stock data and reconciliations with flow data is in accord with the draft of the Fund’s revised Balance of Payments Manual. The Fund has collected similar data, but they vary in detail and completeness and have not been published in aggregated form.152 The following discussion draws primarily on data supplied in the Special Questionnaire and shown in Table 62.

Table 62.

International Investment Positions, Year-end 1988

(In billions of U.S. dollars)

article image
Source: Special Questionnaire on International Capital Flows.

Includes some short-term securities.

Includes some securities.

Includes public corporations.

Includes reserve assets.

Includes securities.

Several caveats apply to the use of these statistics:

  • a. The coverage is relatively incomplete for developing countries. Consequently, the total liabilities reported for nonindustrial countries omit large debtor positions, on balance. If this gap were filled, the world totals and the developing country sector would likely show a larger excess of liabilities over assets than is reflected in Table 62. Also, in some cases, resident holdings of domestic bonds acquired in foreign markets continue to be counted as holdings by nonresidents. Considering the large indebtedness of developing countries, it is surprising that the net asset position of industrial countries is nearly in balance.

  • b. Valuation principles—especially for marketable securities and direct investment—employed by different countries vary widely, so the results of the tabulations are somewhat impressionistic and are not to be taken too literally. The results do not have the quality of standard balance sheet compilations. Also, reduction of all these data to a common numeraire (the U.S. dollar) introduces a further element of difficulty if different conversion rates and principles have been used by reporting countries.

  • c. For many individual countries, the tabulations are incomplete, so the totals include only those items for which values are given. In the case of Hong Kong, for instance, only banks’ positions are given.

  • d. By type of asset, nearly all reporting countries could provide data on bank positions and on their official assets and liabilities. The data on nonbank positions is less complete, and many countries did not provide data on their direct investment positions.

  • e. In addition to the general weakness of coverage of developing countries, offshore centers are largely omitted (apart from certain data for The Bahamas, Hong Kong, Panama, and Singapore), and there is no coverage of Middle East oil producers.

Although the results should be treated with caution because of these limitations, certain aspects of the worldwide debtor-creditor situation are reasonably well delineated. The global discrepancy between reported assets and liabilities is a small portion of the outstanding amounts—for instance, the difference in 1988 between reported assets and liabilities totaling $15,300 billion was only $215 billion. However, the reported assets include $143 billion of gold, which has no counterpart on the liability side, and liabilities are understated because of the limited coverage of developing countries. A review of World Bank data on outstanding debt indicates that, for major debtors, the questionnaire coverage omits about $250 billion in 1988. If adjustment is made for these two factors, the reported net global liability figure in 1988 would be about $550 billion.

Major Categories of Investments

Direct Investment

Among the types of investment shown, direct investment is probably the most completely covered on a global scale. (For individual countries, the cross-border assets and liabilities of banks probably are measured more completely than other private investments but, on a global scale, there are difficulties because of the inadequacies of data for offshore financial centers.) Most of the industrial countries obtain data on direct investment positions from enterprise surveys, which most often yield valuations on a book value or historical cost basis rather than on a replacement or market value basis. A few industrial countries, and many developing countries, base these statistics on cumulative capital flows or official authorizations. One result of using book values for direct investment is that comparisons with assets and liabilities measured at current market values (notably marketable securities) are severely hampered. For instance, in Table 62, the 1988 value of direct investment ($1,053 billion) is less than the value of securities ($1,382 billion), but this lesser value greatly understates the relative economic significance of direct investment.

Table 62 shows that, as of 1988, a few countries—notably the United States, the United Kingdom, and Japan—dominated the creditor side of direct investment, while the United States stood out on the liability side. For Japan, inward direct investment totaled only about 10 percent of the value of outward direct investment, with the result that Japan was the largest net direct investor. Developing countries had very partial data on direct foreign investment, and it is clear that they have not yet produced adequate statistics on investment positions.


Data on holdings of foreign securities, including government issues, are shown for industrial countries in Table 63. Cross-border positions in foreign securities are an important part of the international investments of industrial countries, but they were reported in relatively insignificant amounts for developing countries. Because countries were not required to identify separately in the Special Questionnaire the amount of foreign securities held as part of their foreign exchange reserves, there was an overstatement of net liabilities in the form of securities (see Chapter 8). Also, very large amounts of securities held through Swiss custody accounts (and possibly similar accounts elsewhere) were more likely to be reported on the liability side than on the asset side. Consequently, the large net liability position shown in the table ($450 billion at the end of 1988) exaggerates the imbalance in the actual positions.

Table 63.

Holdings of Securities by Industrial Countries, 19881

(In billions of U.S. dollars)

article image

Includes only amounts reported on the Special Questionnaire; includes some short-term issues. See Chapter 4 for additional data on outstanding securities.

Excludes securities held in official reserves.

A number of countries were able to report on equity securities separately. Reported resident holdings of foreign equities were $360 billion, while domestic equities held by nonresidents were reported at $594 billion. The shortfall of reported assets may reflect holdings through custody accounts or other intermediaries, or valuation differences, as well as relatively incomplete reporting in creditor countries of resident purchases made in international markets.

Banks’ International Positions

By far the largest gross amounts of outstanding cross-border assets and liabilities were reported by banks (see Table 62), even though the gross positions of banking offices in offshore and some other financial centers were not included. Of course, there are other, more comprehensive sources for data on cross-border positions of banks (see Chapter 6). Banking data presented in balance of payments accounts usually do not indicate the nature of the counterparties (whether claims or liabilities are vis-à-vis other banks, other private parties, or official accounts). The enormous size of banks’ cross-border assets and liabilities relative to other types of investments may result in part from the redepositing of funds among banks in different financial centers.

Other Private Capital

Data reported in the questionnaire for “other” foreign assets and liabilities of resident private nonbanks displayed a mixture of elements. A large part was funds placed with, or borrowed from, foreign banks; another part was trade or other credits granted or received between nonbanks; and another was short-term financial instruments of all kinds. Collecting data from private resident nonbanks on this diverse collection of assets and liabilities is no doubt one of the most difficult compilation problems of balance of payments accounts, and coverage differs so much among countries that the data have limited analytical value.

Official Assets and Liabilities

Amounts reported in the Special Questionnaire for resident official assets and liabilities are also shown in Table 62. Amounts reported for industrial countries are relatively complete, while the coverage of other countries is less complete and details are often missing. Reported assets are dominated by reserve holdings and, for industrial countries, loans and credits outstanding. With respect to liabilities of the official sector, the dominant feature is the amount of securities issued by official agencies and held by nonresidents (also see Table 63). Over three-fourths of cross-border resident official liabilities of industrial countries consist of securities, primarily United States government obligations.153

Reconciliation of Flows and Stocks

A major conclusion of the Working Party is that the quality of data on capital flows can be validated when comparable independent data on outstanding amounts of cross-border assets and liabilities are also assembled. In some countries, data for capital flows—at least for some types of capital—are derived from data on outstanding stocks; in others, estimates of outstanding stocks are constructed by accumulating flows. Under these procedures, the relationship between stocks and flows is consistent but, since the data series have a common basis, there is no built-in mechanism that would detect emerging discrepancies.

Countries that do construct stock and flow data independently can readily expose discrepancies that need to be corrected. Of course, there are several reasons (which need to be allowed for in making comparisons) why stock and flow data can diverge. A major difference is that changes in position figures include valuation effects, including foreign exchange translations, variations in market values, capital gains or losses, write-offs, and changes in coverage. To overcome these difficulties, it is useful to develop (as recommended in the draft of the revised Manual) reconciliation accounts that show the effects of each of these factors separately and therefore tie in closely with the flow accounts.

Part IV of the Working Party’s Special Questionnaire invited countries to show the stock-flow reconciliations for major elements of their international assets and liabilities for 1987 and 1988. Changes in year-end stocks were to be explained on the basis of capital flows, exchange rate changes, price changes, write-offs, and other adjustments. This section proved useful in some cases by drawing attention to some of the sources of discrepancies in the flow accounts but, on the whole, the response was disappointing. Of the industrial countries surveyed for 1988, 13 completed the reconciliation section, but only a few did so convincingly. Of the developing countries surveyed, only five carried through plausible reconciliations.

On the basis of this experience, the Working Party is convinced that countries, using whatever techniques are best suited to their situations, should take steps to measure their stocks of international assets and liabilities. Moreover, stock and flow data should be linked by a reconciliation account, as recommended in the revised Manual and foreseen in the System of National Accounts.

Relation to Income Accounts

The 1987 report submitted by the Working Party on the World Current Account Discrepancy demonstrated that a main avenue for reducing the global current account discrepancy was to improve the measurement of investment income receipts and payments, if necessary by adjusting recorded figures. Part of the adjustment involved comparisons of countries’ bilateral accounting for direct investment income, but the greater part involved adjustments to the income received on cross-border assets held with banks and dividends and interest on holdings of securities. For these latter categories, the procedure was to use as a basis the data (as reported to the Bank for International Settlements) on banks’ cross-border liabilities to non-banks and totals constructed by the Working Party on cross-border holdings of securities, to which relevant yields had been applied. This same process has continued to be employed in recent years by the Statistics Department to carry forward adjustments to the world and regional income accounts.

The adjustments, as constructed by the Statistics Department, to reported international investment income are shown in Table 64. The growing deficiency in the reporting of “other investment income” reflects the growth in the stock of unrecorded assets, together with changes in estimated yields. In its investigation of direct investment (Chapter 3 of this report), the present Working Party developed significant additional adjustments to reinvested earnings beyond those shown in the table.

Table 64.

Adjustments to Reported Current Account Net Investment Income, 1986–89

(In billions of U.S. dollars; additional credits ( + ))

article image
Source: Balance of Payments Statistics Yearbook, Part 2, Table 3.

The Working Party reviewed the procedures used to carry forward the income adjustments as now published in the Yearbook. It appears that the approach of the earlier Working Party has been carried forward very carefully; that is, in the case of nondirect investment income, the estimates of outstanding cross-border assets and liabilities have been adjusted annually, and the rates of return have followed current developments in the relevant financial markets. However, the assumption by the Fund that the ratio of reported to unreported income in country submissions remains the same as it was in 1983 is now outdated. The Working Party recommends that the Fund carry out a specialized survey to establish a current basis for the adjustment exercise. Similarly, Fund adjustments to reinvested earnings should take advantage of the new data developed by this Working Party.

In the view of the Working Party, national compilers should compare their data on investment income receipts and payments to the underlying stocks of assets and liabilities in order to establish the validity of the investment income component of their current accounts. Consistent categorization of income flows and stocks is essential for this purpose.

Geographic Allocations

Some principles for allocating international transactions geographically are discussed in Chapter 4 and Appendix VI of this report, and guidance is given in the draft fifth edition of the Manual. There is general agreement that the most useful information for analysis is the location of assets (direct investment, securities, banking positions, and so forth) according to the country where the obligor resides. Likewise, for a country’s liabilities (including direct investment and equities), it is most useful to know the residence of the beneficial owner. To some extent, such information can be extracted—in principle—from flow data, but there are many well-known difficulties in obtaining the necessary geographic information from that source. To obtain geographic data on external assets, surveys of the resident holders of such assets are indispensable. Surveys of resident issuers of liabilities held by nonresidents are also useful, but the results are less revealing because the identity of beneficial owners is often concealed.

For outward direct investment, flow data generally can be traced only as far as the first enterprise outside the country. When an enterprise survey is conducted, however, it is possible to obtain accounts for each of the affiliates, down to the country level. Similarly, for inward direct investment, surveys can indicate the location of the ultimate beneficial owners, rather than the intermediate country through which ownership may be held.

The flow data on portfolio transactions now available in most countries are not well suited to identifying ultimate owners or issuers of securities. Flow data are also deficient in determining the stock of resident holdings of foreign securities by country of issue, unless issuers of the securities are specifically identified. As discussed in greater detail in Chapter 4, benchmark surveys of resident issuers, and custodians and other holders, of cross-border securities are better able to reveal geographic details of assets and liabilities. While admittedly not perfect, surveys of outstanding stocks are the most likely method for allocating securities investments to ultimate debtor or creditor countries.

The Working Party believes that information on the geographic distribution of international investments is a valuable tool for policymakers and for economic analysis in general. Therefore, it urges national authorities and international organizations to support strongly efforts to develop such data.

Conclusions and Recommendations

This review of a somewhat limited and fragmented data set indicates that a greater effort to collect information on stocks of international assets and liabilities would be desirable. Such information is valuable for many kinds of economic inquiries, including some questions that go beyond the mandate of the Working Party. Since the compilation of position data goes hand in hand with the compilation of flow and income data, the Working Party is pleased to find this connection emphasized in the Fund’s draft Manual and carried forward in its draft Compilation Guide.

The following is a brief recapitulation of the recommendations mentioned in this chapter: