Abstract

1. INTRODUCTION AND SUMMARY

1. INTRODUCTION AND SUMMARY

Reported “portfolio investment income” is the fastest growing, and is now the largest, of all individual current account discrepancies. The term as used in this chapter is defined to include all types of investment income, other than income related to direct investments, which is covered in Chapter IV. The two main categories of portfolio investment are bank positions, and investments in bonds and equities. While the official-private breakdown might be useful for some purposes, it will be seen that it is necessary to combine these categories for this study since, for instance, credits on official accounts of one country may appear as debits on private account in the debtor country. However, a discussion of resident official income is given in a later section of this chapter.

a. Discrepancies in Reported Income Data

As a starting point for our analysis we use the numbers published in the Fund’s Balance of Payments Statistics Yearbook, 1985, Part 2, particularly Table C-8, “Other Investment Income.” (For many countries, a more detailed breakdown is published in Part 1 of the Yearbook.) According to that table (summarized in Table 17), reported portfolio investment income debits exceeded the corresponding credits by $32 billion in 1983 and $42 billion in 1984, compared with excess debits of only $7 billion in 1979. Not only have the missing amounts increased, but the quality of reporting has also worsened as measured by the relative size of the gaps. In 1979, reported credits were only 5.5 percent below reported debits. This gap widened to 9 percent in 1981, 13 percent in 1983, and more than 15 percent in 1984. In addition, the quality of information also became worse relative to the reporting of other services. While the portfolio investment income gap accounted for 25 percent of the world services gap in 1979, this share increased to 40 percent in 1983, and to 43 percent in 1984.

Table 17

WORLD SUMMARY OF PORTFOLIO INVESTMENT INCOME

(In billions of U.S. dollars)

article image

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

Calculated from Yearbook data.

b. Methods of Adjusting Reported Data

As noted in Chapter I, the bias in the reporting of portfolio income that results in a rising excess of debits over credits is consistent with the bias in the reporting of capital flows. For 1977–83, reported net capital flows showed a cumulative net inflow of nearly $300 billion, indicating that the countries receiving capital were in a better position to measure the flows than the countries where the creditors resided and were thus better able to record the related investment income flows.

The cumulative discrepancy in recorded capital flows described above indicates the likely sign and magnitude of the world investment income discrepancy. But it is still necessary to identify specific gaps according to the different types of investment and to allocate them geographically to the extent possible. For instance, even if the reporting of credits is less complete than the reporting of debits, it does not necessarily follow that the debits are not overstated. Moreover, the stocks of assets and the income received on them (and the outstanding liabilities and the income paid) may fit together reasonably well, as reported by a given country, but both statistics may be wrong. Also, while comparing debits and credits as reported by pairs of countries might seem a promising way of identifying discrepancies, few countries have such bilateral information, and its validity is limited by the growing tendency to invest via intermediary financial centers. Even if bilateral data exist, it is impossible to decide which country’s amount is correct unless there is some exogenous indication that one country has better information than the other.

Thus the Working Party felt that the only credible basis for checking reported income credits and debits would be to rely on independent information on outstanding stocks of cross-border assets and liabilities, and to use available information on appropriate yields to judge the reasonableness of reported credits and debits. Fortunately, there now exists a comprehensive and detailed body of data on the cross-border assets and liabilities of banks that is published in the Fund’s monthly International Financial Statistics. These data give extensive figures, by country, for interbank positions and for assets and liabilities of nonbanks vis-à-vis banks, based on the records of the banks.

Unfortunately, data for outstanding stocks of other types of assets held by foreign investors are not well organized and estimates must be based on heterogeneous sources. The results of these estimates, and the sources used are shown in Table 18. In addition to using published sources, the Technical Staff prepared its own estimates of outstanding cross-border holdings of bonds and equities, and information on other types of assets (such as trade credits) was derived largely from the replies to the income questionnaire circulated by the Working Party. A description of the estimating procedure for bonds and equities is given in Subsection 2.c of this chapter; a copy of the questionnaire appears in Appendix II.

Table 18

ESTIMATE OF WORLD CROSS-BORDER FINANCIAL POSITIONS,1 1983

(In billions of U.S. dollars)

article image
Note: For sources and more details, see Appendix V.

Excluding direct investment.

Based on outstanding amounts at beginning of year (totaling roughly $4,250 billion) and cross-border interest rates.

Includes positions between deposit banks ($1,582 billion), claims of deposit banks on monetary authorities ($98 billion), and claims of monetary authorities on deposit banks ($198 billion).

Excluding $110 billion of bonds that are held or issued by banks and included in bank-related accounts, and excluding securities that represent official reserves.

Not included elsewhere but excluding claims on banks and including securities. SDR holdings, and reserve positions in the Fund.

Mostly World Bank and government loans but also including Fund credits.

Roughly 6 percent of world trade, presumably at no or very low interest.

Reported flows plus adjustments derived in this chapter, (See also Table 19).

Another important source of information was the data on the debts of developing countries assembled by the Development Assistance Committee of the OECD. This is described in an appendix in which data on investment positions and debt are assembled and evaluated (Appendix IV).

It is important to note that the estimates of outstanding amounts of cross-border positions cover only those positions recorded by the debtor or creditor country, and often, but not always, both. In particular, the estimates of outstanding amounts are not analogous to the numerous estimates of the amounts of flight capital that may have escaped statistical recording. That may, for instance, be the case if cross-border funds are directly invested in real estate, or if the foreign investor or his intermediary is regarded as a resident. Also funds that are spent, rather than invested, do not add to external assets. The Working Party has taken note of the various estimates of “flight capital” but has not found them useful in calculating income adjustments. (See the discussion in Appendix VI.)

In addition to the revisions based on the general procedures described above, a number of revisions were based on specialized analyses or other sources. These adjustments are described in Section 2 of this chapter.

c. Summary of Results

On the basis of the various types of adjustments made to the reported data, the discrepancy in world portfolio investment income can be reduced to a minor amount. (See Table 19.)

Table 19

PORTFOLIO INVESTMENT INCOME: SUMMARY OF 1983 ADJUSTMENTS

(In billions of U.S. dollars)

article image

Mainly a U.K. adjustment owing to gross reporting which was not yet incorporated into the Fund’s 1985 Balance of Payments Statistics Yearbook.

Excluding major offshore banking centers.

In the adjustment procedure gross income credits and debits were both raised substantially above the reported amounts, mainly because data for interest received and paid by banks in certain international financial centers were obtained or estimated on a gross, rather than net, basis. However, that change in itself had little effect on the net result. Most of the total net adjustment of $32.8 billion was accounted for by additional net credits attributable to claims on banks by nonbanks in a wide range of countries. Another sizable addition to net credits reflects the return on cross-border investments in bonds and equities. A considerable net credit was also determined to be appropriate for the international institutions themselves, since their accounts had not yet been integrated into the Fund’s Balance of Payments Statistics Yearbook.

The Working Party has also attempted to establish the geographic distribution of the adjustments. (See Table 20.) However, the allocation by country grouping is inherently more conjectural than the overall adjustments, as discussed in Subsection 2.g of this chapter. The main impression that results from this exercise is that the addition of net credits is spread among many country groupings, with the exception of Eastern Europe. Moreover, a substantial amount of the 1983 adjustments cannot be allocated by area because the basic data (in this case the country distribution of banks’ liabilities to foreign nonbanks) contain large unallocated elements. Finally, the adjustment procedures adopted for 1983 were applied to the entire 1979-84 period. As shown in Table 54, the discrepancies for those years could be reduced to minor amounts, except for 1980, where a sizable net credit residual remained. Further analysis also indicates that these results would not be greatly affected by the adoption of a somewhat different range of yields to generate income flows.

Table 20

ALLOCATION OF PORTFOLIO INVESTMENT INCOME DISCREPANCY BY AREAS1

(In billions of U.S. dollars; net amounts)

article image

For gross amounts (credits and debits separately), see Table 56

2. DERIVATION OF ADJUSTMENTS

The analytical process used in determining the adjustments derived in this chapter is a mixture of the application of the general principles outlined above and the working out of corrections for a variety of special situations. In practice this required approaching the problem from a number of directions, sometimes overlapping. For instance, one focus was on different groups of countries with distinctive reporting problems: industrial countries, oil exporters, other developing countries (especially major debtors), and international financial intermediaries. Another focus was on the separation between direct-investment income and other income, and a third focus was on the two broad types of portfolio assets: banking positions and holdings of securities. The results of following these various avenues of study are presented below in a somewhat simplified way, in order to describe the main lines of research rather than the details of each stage. Nevertheless, the underlying complexities should be borne in mind.

The first stage of research was to establish a credible base for either the debit or credit side of the income accounts—and for reasons already mentioned the debit side was chosen. Thus the Technical Staff began by analyzing the income debits on a country-by-country basis, using all the available information on the liabilities of each country and on the yields appropriate to each kind of debt. The results of that study are not given separately in the following material, but they are implicit in the adjustments shown for the income debits of various country groups and special situations. On balance, the re-estimation of bank-related and other portfolio income debits (excluding some grossing-up) added about 15 percent to the debit amounts included in the Fund’s world income tables. Having established an acceptable total for debits, the main task was to locate the gaps in the income credits as reported, by type and by geographic location, to the extent possible.

a. Revisions and Reclassifications

(1) Revisions Based on Questionnaire

The data that are regularly reported to the Fund and published in its Balance of Payments Statistics Yearbook, which are used as a starting point, are not always sufficiently detailed to permit one to make adequate comparisons with other related data. Consequently, at the beginning of the study, a questionnaire on investment income was sent to 61 countries that either already reported substantial amounts of investment income or were assumed to have substantial cross-border assets or liabilities. (See Appendix II.) Responses to the questionnaire in many cases supplied new or revised data on investment flows and outstanding amounts of cross-border assets and liabilities. On balance, they increased income credits more than debits, as shown in Table 21. Among the larger revisions, the United Kingdom reported interest income related to the Eurocurrency business gross on the questionnaire, rather than net as in their balance of payments reports. While the grossing-up is now recognized in all U.K. balance of payments publications, it had not yet been incorporated into the Fund’s 1985 Balance of Payments Statistics Yearbook. The United Kingdom also revised the data on other private investment income, thereby contributing $1.2 billion of net credits. In addition, adjustments to Canadian data generate a net credit of $0.9 billion, reflecting the elimination of withholding tax from the debits and, to a minor extent, a more precise breakdown between direct and non-direct investment income. The gross transactions were also affected by the inclusion of foreign currency interest transactions of Canadian banks on a gross, rather than a net basis. Other questionnaire responses indicated some reclassifications of data—for example, for Singapore the Fund staff estimates a shift of $1.5 billion from non-direct investment income to direct investment income debits. (For country details, see Table 21.) Unfortunately, it cannot always be determined to what extent revisions reported in the questionnaire represent reclassifications rather than new data.

Table 21

REVISIONS INDICATED BY SPECIAL INVESTMENT INCOME QUESTIONNAIRE1

(In millions of U.S. dollars)

article image

Differences between responses to investment income questionnaire and the Fund’s Balance at Payments Statistics Yearbook data as published in 1985. Small differences also arise from the conversion of Yearbook data stated in SDRs into US dollars.

Representing a shift from non-direct to direct investment income debits, based on Fund staff estimates.

(2) Reclassification of Direct Investment Income

Certain adjustments to the non-direct investment account reflect the reclassifications set forth in Chapter IV. Correction for misclassifications generates $0.7 billion of net non-direct investment income debits in 1983, and $2.3 billion of net credits in 1979 (Table 23). These shifts do not affect the overall investment income or the global current account. While it is likely that the “other investment income” still includes some “direct investment income” that cannot be identified, such flows are thought to be negligible on a net basis.

Table 22

SUMMARY OF QUESTIONNAIRE REVISIONS, 1983

(In billions of U.S. dollars)

article image
Table 23

RECLASSIFICATION OF DIRECT/NON-DIRECT INVESTMENT INCOME DATA1

(In billions of U.S. dollars)

article image

For further details see Chapter IV.

Excluding the revisions in response to the questionnaire, which are already recognized in Table 22.

b. Adjustments Based on Banking Data

(1) Data Sources

(a) International Banking Statistics Data on Outstanding Amounts.

The largest net adjustments to reported portfolio income are based on the data for cross-border bank deposits of nonbanks as reported in the International Banking Statistics (IBS) published by the Fund in its monthly publication, International Financial Statistics. The IBS data cover more than 70 percent of the known world cross-border financial positions (excluding direct investment). The statistics on international (cross-border) banking positions have been developed by the BIS and the Fund. The BIS statistics, published quarterly in International Banking Developments, are based mainly on reports from banks in industrial countries and in offshore financial centers. In addition to that information, the Fund includes information on the cross-border bank accounts of almost all Fund member countries (as a part of the regular money and banking data supplied to the Fund). Moreover, the Fund supplements the data by using information on trustee accounts that are established by nonbanks and channeled through Swiss banks. The comprehensive results are published monthly in the Fund’s International Financial Statistics (IFS); see Table 24.

Table 24

SUMMARY OF INTERNATIONAL BANKING STATISTICS FOR THE END OF 1983, AS PUBLISHED BY THE FUND

(In billions of U.S. dollars)

article image
Source International Monetary Fund. International Financial Statistics (Washington), Vol. 39 (June 1986).Coverage:

Total cross-border assets and liabilities as reported by deposit banks of individual countries (including deposit banks’ positions vis-à-vis monetary authorities) These tables exclude, however, cross-border positions as reported by monetary authorities This series is used here primarily to check the interest income flows of developing countries.

The “interbank accounts” refer to all cross-border positions between deposit banks and also include all positions between deposit banks and monetary authorities The tables exclude the positions between monetary authorities, however Assets should equal liabilities, and to the extent they do not, data are not yet quite consistent.

Cross-border bank credits to nonbanks as reported by the lending banks. Data are shown separately by residence of lending banks (bank assets) and by residence of the borrowing nonbanks (nonbanks’ liabilities derived from the reports of lending banks).

Cross-border bank deposits of nonbanks as reported by the banks receiving the deposits. Data are shown separately by residence of borrowing banks (bank liabilities) and by residence of depositing nonbanks (nonbanks’ deposits derived from the reports of borrowing banks).

The two main sectors of the banking data are (1) interbank positions and (2) nonbanks’ deposits with, and borrowings from, banks. It appears that reporting by the banks themselves is usually comprehensive, and detailed analysis indicates that interest credits and debits of the banks themselves are fully incorporated into national balances of payments, either gross or net. The only major exceptions are the major offshore banking (MOB) centers, for which adjustments are required (Subsection 2.b.(6) of this chapter). That reporting by banks other than those of MOB centers is essentially accurate can be verified on the basis of the responses to the investment income questionnaire. Cross-border assets of banks in industrial countries reported in the questionnaire totaled $1,583 billion at the end of 1983, and bank liabilities amounted to $1,554 billion. These amounts are consistent with data from other sources (mainly deposit banks’ “total assets” and “total liabilities” as reported in the IBS). Banks’ reported interest income ($146 billion of credits and $136 billion of debits) implies overall rates of return of 9.6 percent on the asset side, and 9.1 percent on the liability side, leaving a spread of 0.5 percent (Table 25). Since these are plausible rates that are based on verifiable amounts outstanding, banks’ interest income and expenditure, as reported in the questionnaires and presumably as reported in national balance of payments accounts, appear to be acceptable.

Table 25

CROSS-BORDER POSITIONS AND INTEREST INCOME OF BANKS OF INDUSTRIAL COUNTRIES,1 1983

(In billions of U.S. dollars)

article image

Limited to questionnaire respondents, which account for 90 percent of the total cross-border bank positions of industrial countries.

Derived from questionnaire responses.

Interest incomes related to amounts outstanding at the beginning of the year.

While little correction is required for interbank interest flows, a major adjustment is required for interest paid to, and received from, banks by non-banks. In the IBS tabulations, liabilities of banks visa-vis nonresident nonbanks totaled $673 billion at the end of 1983, including trustee accounts channeled through Switzerland and deposits held in MOB centers. Corresponding interest payments amount to approximately $52 billion (Table 26). Using country detail available in IBS tables and rates of return as indicated by market rates (see below), it is possible to check the income of nonbanks reported in balance of payments statements country by country. Since deposits in the banking data are allocated according to the residence principle, their definition is consistent with balance of payments procedures. However, since the banking statistics leave substantial amounts of nonbank accounts unallocated, it is likely that deposits allocated to individual countries and the derived interest attributed to the deposits are minimum amounts.

Table 26

CROSS-BORDER BANK DEPOSITS OF NONBANKS, 1983

(In billions of U.S. dollars)

article image
Source of Amounts Outstanding: International Monetary Fund, International Financial Statistics, world tables.

Including $83 billion in trustee accounts.

Major offshore banking centers.

See Table 27.

It is convenient, using the same data source, to check simultaneously the interest paid by nonbanks on their borrowing from banks. Nonbanks’ interest payments to nonresident banks seem to be relatively well recorded in national balances of payments, but some adjustments are necessary. As a basis for estimating interest paid to banks by nonbanks, the IFS tables showing $760 billion of bank lending to non-banks at the end of 1983 were used (Table 24).

(b) Derivation of Interest Rates.

To check the reported interest income against estimated actual flows requires not only a solid record of outstanding amounts but also some assumptions about the level and structure of interest rates. Obviously most of the interest flows are determined by market rates, allowing for a certain lag of rate adjustments and of actual payments, and taking into account the currency mix and the special characteristics of some areas (MOB centers). Interest rates related to bank positions are derived from the interbank rates (U.S. dollar, London). A six-month lag is applied, since the actual rates on current positions with borrowers and depositors are only changed at agreed intervals and actual interest payments are usually due at the end of these periods. A further adjustment is made to allow for the currencies. Most of the cross-border positions are denominated in U.S. dollars, but 10 to 12 percent is in currencies that usually bear lower interest rates (deutsche mark, Swiss franc, Japanese yen). Therefore, on average, a downward adjustment of roughly 0.5 percent is necessary. (See Table 27.)

Table 27

STRUCTURE OF CROSS-BORDER INTEREST RATES, 1979–85

(In percent per annum)

article image

London interbank offered rates on U.S. dollar deposits (six months).

Taking into account that rates vis-à-vis nonbanks are adjusted with a lag and that interest payments are due at the end of each basis period.

Roughly 10 percent of cross-border bank accounts is denominated in deutsche mark, Swiss francs, yen, and pounds sterling.

Positions are almost exclusively in U.S. dollars, and spreads are assumed to be smaller.

For more details, see Tables 42 and 46.

The resulting rate is taken as a basis for calculating the actual bank rates. vis-à-vis nonbank borrowers, a positive spread of 1 percent is assumed, and vis-à-vis depositors, a negative spread of 1.5 percent. Relative to the positions of major offshore banking center banks, it is assumed that the spreads are smaller and that all positions are in U.S. dollars.

Derived rates used for estimates are slightly above the rates of return implied in the questionnaire responses. (See Table 25.) Questionnaire rates may be distorted by some net reporting, and banks of the responding countries (including the Federal Republic of Germany, Japan, and Switzerland) presumably have above-average shares of low-interest currencies in their cross-border positions.

(2) Industrial Countries

While analyzing the interest related to banking positions, it is useful to deal separately with industrial countries and developing countries, which are here divided into Middle Eastern oil exporters, major offshore banking centers, and other developing countries. The overview for the estimated income of industrial countries is given in Table 28; country details are given in Table 29; and similar information on the debit side is shown in Table 30.

Table 28

INDUSTRIAL COUNTRIES: SUMMARY OF NONBANKS’ CROSS-BORDER BANK DEPOSITS AND RELATED INCOME CREDITS, 1983

(In millions of U.S. dollars)

article image

International Financial Statistics.

Calculated on the basis of details given in questionnaire responses.

Table 29

INDUSTRIAL COUNTRIES: NONBANKS’ CROSS-BORDER BANK DEPOSITS AND RELATED INCOME CREDITS, 1983

(In billions of U.S. dollars)

article image

Reported in International Financial Statistics. Nonbank deposit data are restricted to geographically allocated deposits, which represent minimums because a considerable part of total deposits cannot be allocated.

Based on outstanding amounts (beginning of year) and cross-border interest rates.

The response to the questionnaire does not give this breakdown, but a maximum allocation to interest received from banks has been made.

Assuming a relatively high share of low-interest-bearing domestic currency.

Includes deposits of Dutch special financial institutions.

Amounts as given in the questionnaire response represent solely the cross-border trustee accounts established by Swiss residents through Swiss banks. These accounts are excluded from the table to make data comparable to the numbers in International Financial Statistics. (See also footnote 7.)

Excluding $11 billion trustee accounts established by Swiss residents. For interest flows related to trustee accounts, see Table 14.

Assuming that almost all deposits are denominated in U.S. dollars.

Table 30

INDUSTRIAL COUNTRIES: NONBANKS’ CROSS-BORDER LIABILITIES VIS-À-VIS BANKS AND RELATED INCOME DEBITS, 1983

(In billions of U.S. dollars)

article image

Line 57 (if not available, see figures in parentheses on line 56), and line 66. See sample questionnaire in Appendix II.

International Financial Statistics.

Line 26 (if not available there, see figures in parentheses on line 23), and line 14 (if not available there, see figures in parentheses on line 13). Responses to the questionnaire are not sufficiently detailed, and data may include interest payments other than those to nonresident banks.

For Switzerland and the United States, based on amounts outstanding at beginning of period (IFS) and cross-border interest rates. For all other countries, missing debits in total are roughly estimated at 10 percent of reported debits.

Amount presumably included in Balance of Payments Statistics Yearbook data.

Fund staff estimate.

Mainly representing New Zealand.

Cross-border bank deposits of industrial countries’ nonbanks totaled $301 billion at the end of 1983, generating interest income credits of at least $23 billion, calculated on the basis of average outstanding amounts and cross-border interest rates (Table 28). Questionnaire responses indicate that at most $10 billion is reported in national balances of payments. Wherever detailed figures for interest from banks were not shown separately in the questionnaire, an estimate was made of the amount included, assuming for this purpose a questionnaire coverage of bank-related income that was as broad as possible. Consequently, the partly estimated $10 billion of credits shown for the questionnaire responses tends to overstate the actual reported amount of interest receipts on bank deposits, because it may, in fact, include some other portfolio income. Therefore, the indicated adjustment of $13.1 billion is likely to represent a minimum.

The greatest absolute difference occurs for the United States. According to national data reported in the questionnaire, U.S. nonbanks’ cross-border bank deposits totaled $52.8 billion at the end of 1983, generating $5.3 billion of interest credits. However, international banking statistics show $168 billion of U.S. nonbank deposits, which would yield about $13 billion of interest income (Table 29). The indicated credit adjustment of $7.7 billion is partly offset by about $4 billion of underreporting on the debit side (Table 30). Most other countries also understate interest receipts of nonbanks, although to a lesser extent and in most cases by not more than $1 billion. For instance, missing income for Belgium/Luxembourg, France, the Federal Republic of Germany, Italy, and Spain may be estimated at between $0.4 and $0.8 billion each (Table 29), and the adjustment for Switzerland is $1.6 billion. Swiss nonbanks’ deposits with foreign banks (excluding trustee accounts which are channeled through Swiss banks) totaled $23.8 billion at the end of 1983, but corresponding interest receipts are not yet included in the Swiss balance of payments.

In many cases, nonbanks’ interest incomes are omitted by national authorities because the receipts are not recorded by the exchange control system. In other cases underreporting occurs because income estimates are based on domestically reported assets only; the United States, for instance, uses a reporting system that mainly covers large multinational companies that can easily be identified. Still other countries (the Federal Republic of Germany, the Netherlands, and Japan) rely on individual income reports from nonbanks, but even these extensive systems leave gaps.

The underreporting of credits is partly offset by underreporting of interest payments on nonbanks’ borrowing from nonresident banks. Unfortunately, verifying these debits is much more difficult than checking the credits. Questionnaire responses were less detailed on the debit side, and data reported by national authorities systematically differ from information derived from the International Banking Statistics. Banks’ reported cross-border claims on non-banks include claims on official agencies and also include some bonds, so that there may be some overlapping in the estimates. Because of these limitations, it appears that interest payments of industrial countries’ nonbanks to foreign banks are understated by $6-7 billion (Table 30). Again, the largest missing amount is allocated to the United States ($3.7 billion). U.S. nonbanks’ liabilities vis-à-vis foreign banks are reported at $20 billion at the end of 1983 in the national statistics, but the International Banking Statistics show $58 billion. Moreover, Switzerland seems to omit the interest payments of nonbanks to nonresident banks. For all other countries the upward adjustment is estimated at $2.2 billion (i.e., roughly 10 percent of the actual payments), but questionnaire responses do not give a sufficient basis for allocations by country. Part of the debit adjustments may compensate for overadjusted receipts if countries partly report on a net basis—for example, as in the Netherlands. Underreporting on the debit side is clearly on a smaller scale, but here, too, there is probably some evasion of exchange control records, and methods that rely on reports by individuals probably have gaps, especially where there are exemption limits.

(3) Swiss Trustee Accounts

Trustee accounts channeled through Switzerland are mainly established by non-Swiss nonbanks, and the funds are held by Swiss banks as deposits in nonresident banks. While interest rate and exchange rate risks fall entirely on the depositing nonbanks, interest income is free of withholding tax, which is normally 35 percent if deposits are held directly in Swiss banks. The volume of such accounts totaled $94 billion at the end of 1983, of which 12 percent was held by Swiss residents and 88 percent ($83 billion) by nonresident nonbanks (Table 31). The Swiss trustee accounts are fully integrated into the International Banking Statistics that are published by the Fund and that serve as a basis for estimating missing income credits. However, the net adjustment of the world investment income is zero, insofar as Switzerland is concerned.

Table 31

SWISS TRUSTEE ACCOUNTS, STOCKS, AND RELATED INTEREST FLOWS, 1979 AND 1983

(In billions of U.S. dollars)

article image

Net receipts entered into the Swiss balance of payments. Liechtenstein is considered resident.

Switzerland publishes information on trustee accounts (including a regional breakdown), but since these accounts are not in their balance sheets, Swiss banks do not include them in their regular reports on cross-border positions. On the other hand, banks of other countries fully include Swiss trustee accounts in their reports under “liabilities vis-à-vis Swiss banks,” because they are unable to separately identify those liabilities which are related to the Swiss trustee business. As a result, non-adjusted interbank accounts imply a certain asymmetry, with interbank liabilities vis-à-vis Switzerland approximately $100 billion higher than interbank assets of Swiss banks. However, in the Fund’s International Banking Statistics, the trustee accounts are treated as if they were balance-sheet items of Swiss banks. They are added to the interbank assets reported by Swiss banks and thereby equate these positions with other reporters’ interbank liabilities vis-à-vis Switzerland. Moreover, Swiss banks’ liabilities vis-à-vis nonresident nonbanks are raised by $83 billion to $108 billion by this addition. Correspondingly, these amounts are also reflected in the derived IFS world table on “cross-border bank deposits of nonbanks by residence of depositor.”

Swiss banks’ interest receipts on trustee accounts held with nonresident banks totaled $9 billion in 1983, of which $8.5 billion was passed on to nonresidents and $0.5 billion to Swiss residents. (Residents of Liechtenstein are considered Swiss residents.) The amount of $0.5 billion is entered into the Swiss balance of payments under “investment income” (and commission fees are included under “other services”). As far as Switzerland is concerned, no further (net) adjustment is necessary. The gross amount of interest income which is channeled through Switzerland ($8.5 billion in 1983) may be regarded as a “memorandum item” that is useful when using the International Banking Statistics, since those statistics as published in International Financial Statistics fully integrate the trustee accounts.

The real problem is how to adjust for the nonreported interest income of the beneficial owners of the trustee accounts. This, however, does not affect the Swiss balance of payments, but only those of the creditor countries to which these accounts have to be allocated. This is taken into account in the corresponding discussion of nonbanks’ receipts on deposits with banks.

(4) Middle Eastern Oil Exporters

For purposes of this discussion, this grouping includes Iran, Iraq, Kuwait, Libya, Oman, Qatar, Saudi Arabia, and the United Arab Emirates. It is an analytical grouping used in the Fund’s World Economic Outlook, and it consists of those Middle Eastern countries that, over a relatively long period, had sustained current account surpluses from oil and thereby accumulated a considerable amount of foreign assets. At the end of 1983, total gross assets of this group, emerging from current account surpluses and borrowing from nonresidents, are estimated at $350 billion (Table 32). Reported investment income of Middle Eastern oil exporters (roughly $26 billion) is rather difficult to verify, since these countries do not give much information on outstanding amounts or on the methods of estimating income. However, some data are published in the IFS world tables on foreign exchange and international banking assets of residents of these countries, and valuable estimates have been made by the Bank of England (published, for instance, in its March 1985 Quarterly Bulletin). Moreover, discussions with authorities indicate that most of the Middle Eastern oil exporters do not include in their balances of payments the cross-border investment incomes of private nonbanks.

Table 32

MIDDLE EASTERN OIL EXPORTERS: FOREIGN ASSETS AND RELATED INCOME ESTIMATES, 1983

(In billions of U.S. dollars)

article image

Estimate based on cumulated current account balances (1972–83) and borrowing from nonresidents.

Saudi Arabian Monetary Agency.

Unidentified securities, property, real estate, and lending to developing countries.

One third of the estimated $350 billion in gross assets can be precisely identified with some confidence: according to IFS data, there is $34 billion in official foreign exchange holdings, $45 billion in assets of banks, and $48 billion in bank deposits of nonbanks. An additional $90 billion is held in securities, according to data for Saudi Arabia published by the Saudi Arabian Monetary Agency (SAMA). Roughly $120 billion is, by implication, held in other assets—that is, securities other than official Saudi holdings, real property, and direct investment, or credits to developing countries. (See Table 32.)

Middle Eastern oil exporters report $26.5 billion of non-direct investment income. Taking into account the fact that Libya, Qatar, and Saudi Arabia report their investment income on a net basis, an appropriate amount has to be added to compensate for the subtracted debits (roughly $1 billion). Resulting receipts ($27.5 billion) related to outstanding assets imply a rate of return of some 8 percent. This is not too low, compared with prevailing market rates and taking into account the fact that lending to developing countries bears interest at concessional rates. Moreover, a country-by-country analysis does not reveal larger inconsistencies (Table 33). However, it is known that interest receipts of private nonbanks are not recorded and that particularly the interest income of the United Arab Emirates seems to be clearly understated ($2.2 billion in receipts, while assets derived from current account surpluses and outstanding liabilities may be estimated at $65 billion). The U.A.E. balance of payments appears to cover neither private capital outflows nor private interest receipts. For the latter, roughly $2 billion has to be added. The overall adjustment for oil exporters’ unrecorded interest receipts (on bank accounts, securities, and other investment) is tentatively estimated at $3.5 billion. However, these adjustments do not attempt to account for the external assets and incomes that may not be shown in these countries’ reported international accounts.

Table 33

MIDDLE EASTERN OIL EXPORTERS:1 CROSS-BORDER ASSETS AND RELATED NON-DIRECT INVESTMENT INCOME CREDITS, 1983

(In billions of U.S. dollars)

article image

Country grouping according to the Fund’s World Economic Outlook. Country figures do not always add to the totals shown.

International Financial Statistics world tables on reserves and international banking.

Includes purchased SDRs.

Based on cumulated current accounts and borrowing from nonresidents. Includes all kinds of cross-border assets, such as reserves, bank-related assets, lending to developing countries, portfolio investment, direct investment, and real estate.

According to International Monetary Fund, Balance of Payments Statistics Yearbook, Vol. 36 (Washington, 1985), Part 2, and questionnaire revisions.

Based on average outstanding amounts and cross-border interest rates.

Credits are reported on a net basis.

Does not include foreign securities held by the Saudi Arabian Monetary Agency that amount to roughly $90 billion.

Total adjustment is $3.5 billion, of which an estimated $1.5 billion may be related to nonbanks’ bank deposits, $1.6 billion to holdings of securities, and $0.4 billion to other investments.

According to OECD data, Middle Eastern oil exporters’ liabilities total $54 billion, of which $40 billion is bank-related (Table 34). The adjustment of $1.5 billion is the result of estimating reasonable payments in relation to outstanding liabilities. Presumably most of the adjustment represents payments of private nonbanks that are omitted from balance of payments recording.

Table 34

MIDDLE EASTERN OIL EXPORTERS:1 CROSS-BORDER ASSETS AND RELATED NON-DIRECT INVESTMENT INCOME CREDITS, 1983

(In billions of U.S. dollars)

article image

Country grouping according to the Fund’s World Economic Outlook.

International Financial Statistics world tables on international banking.

OECD estimates, including trade credits.

Vis-à-vis nonresident banks.

According to International Monetary Fund, Balance of Payments Statistics Yearbook, Vol. 36 (Washington, 1985), Part 2, and questionnaire revisions.

Technical Staff estimate based on average outstanding amounts and cross-border interest rates.

Netted against credits (Table 33).

(5) Other Developing Countries18

Balance of payments reporting of nonbanks’ interest receipts is also incomplete in the developing countries, and there is a smaller shortfall in reported debits. While estimates for individual countries, as implicit in Tables 35 and 36, are based on geographically allocated positions, considerably more assets may be included in unallocated bank positions or be implicit in receipts on securities. Taking this into account, the asset base implicit in these calculations may not be very far below some of the estimates of accumulated capital outflows. (See Appendix VI.)

Table 35

OTHER DEVELOPING COUNTRIES:1 OUTSTANDING ASSETS AND ADJUSTMENT OF INCOME CREDITS, 1983

(In billions of U.S. dollars)

article image

Excluding major offshore banking centers and Middle Eastern oil exporters.

Table 36

OTHER DEVELOPING COUNTRIES:1 OUTSTANDING ASSETS AND RELATED OTHER INVESTMENT INCOME CREDITS, 1983

(In billions of U.S. dollars)

article image

Excluding Middle Eastern oil exporters and major offshore banking centers.

International Financial Statistics world tables on international banking.

Includes purchased SDRs.

According to International Monetary Fund, Balance of Payments Statistics Yearbook, Vol. 36 (Washington, 1985), Part 2, and questionnaire revisions.

Technical Staff estimate based on outstanding amounts (at the beginning of the period) and cross-border interest rates.

To estimate the missing interest income of developing countries (excluding MOB centers and Middle Eastern oil exporters) the procedure is somewhat different from that used for industrial countries. In the balances of payments, as well as in the questionnaire responses, the income is usually not reported in detail. Therefore, total other investment income of individual countries is compared with the income estimated on the basis of total foreign assets—that is, official foreign exchange, plus bank assets, plus non-banks’ bank deposits (Table 36). It is not likely that developing countries will have large assets in forms other than bank claims for which the host countries might record an income debit. Some additional assets, however, are implied in the calculations for missing income receipts on cross-border security assets. (See Subsections 2.c and 2.g of this chapter.)

Few countries report amounts close to the estimated figures. Mexico employs the IFS world table as a basis for its income estimates, and no sizable adjustment is necessary. Nevertheless, the largest differences occur in Latin America ($3.4 billion). Some unreported receipts can only be allocated by area, not by individual country—for example, “other Asia” (more than $1 billion). Some countries seem to ignore nonbanks’ assets entirely (e.g., Liberia and Greece), while Lebanon does not report them at all. Estimates for Lebanon are incorporated in the Fund’s Balance of Payments Statistics Yearbook, Part 2, but these do not cover income on nonbanks’ deposits in foreign banks. No income data at all are reported for Bermuda, but Bermuda contributes at least $0.9 billion of income credits to the adjustments; banks’ and nonbanks’ foreign assets total $11 billion and seem to be related to offshore businesses. (See Chapter VI.) (In IFS, Bermuda is not classified as a “major offshore banking center.”)

As a result, total interest income of developing countries (excluding MOB centers and Middle Eastern oil exporters) may be estimated at $23.8 billion, compared with a reported income of $15.2 billion. The adjustment of $8.6 billion is almost entirely related to nonbanks’ deposits in foreign banks; a series of case studies established that interest income on official exchange and bank assets is reported rather comprehensively.

Developing countries’ missing income credits are only partly offset by missing income debits. Reported debits generally appear to be accurate; reporting has improved, as attention in recent years has been focused on debt problems, rescheduling procedures, and the interest burden. Outstanding debt of developing countries totaled more than $800 billion at the end of 1983 (Table 37). Interest payments reported to the Fund amounted to $66.9 billion. However, in some individual cases, reporting seems inadequate. For instance, Liberia omits interest paid by nonbanks on their $8 billion liabilities vis-à-vis foreign banks that is probably used to finance shipping investments.

Table 37

OTHER DEVELOPING COUNTRIES:1 CROSS-BORDER LIABILITIES AND RELATED OTHER INVESTMENT INCOME DEBITS, 1983

(In billions of U.S. dollars)

article image

Excluding Middle Eastern oil exporters and major offshore banking centers.

International Financial Statistics world tables on international banking.

Including trade credits (roughly 11 percent), data for which was supplied by the OECD.

Also includes the amounts by which SDR holdings are below SDR allocations.

Vis-à-vis nonresident banks.

According to International Monetary Fund, Balance of Payments Statistics Yearbook, Vol. 36 (Washington, 1985), Part 2, and questionnaire revisions.

Based on outstanding amounts (beginning of period) and cross-border interest rates.

OECD data cover only part of the liabilities.

Thus, in the case of Liberia, $0.7 billion is added to compensate for nonreported interest debits. In addition, and partly for similar reasons, debits have to be supplemented for Bermuda, Greece, Portugal, Nigeria, and Lebanon (Table 37). As a result, total interest debits of developing countries are increased by $2.6 billion, to $69.5 billion. The implied rate (roughly 9 percent) is below the lending rate of banks, but consistent with other information according to which a substantial part of loans is given at concessional terms.

(6) Major Offshore Banking Centers

Major offshore banking centers are financial centers characterized by extensive external activities of banks that are not really integrated into the domestic economies, and whose external business is usually disregarded for purposes of balance of payments reporting. (See Chapter VI.) According to the IFS definition, there are seven MOB centers: Hong Kong, Singapore, Bahrain, The Bahamas, the Cayman Islands, the Netherlands Antilles, and Panama. The balance of payments reporting of the MOB centers is usually either incomplete or, as in the cases of Hong Kong and the Cayman Islands, nonexistent. Even if nonreported amounts are small in net terms, adjustments of individual categories of income are useful, since the offshore centers transform the character of many transactions and global nonreporting creates discrepancies in various income categories. In addition to the banks, there are nonbank offshore companies that receive income on investments and for insurance and shipping services, and, on the debit side, pay interest on borrowings and transfer direct investment income to their head offices.

Some estimates of investment income can be made on the basis of available banking statistics. Cross-border positions of banks in MOB centers—that is, assets as well as liabilities—amounted to well over $500 billion at the end of 1983 (Table 38). On the basis of average amounts outstanding and prevailing cross-border market rates (Table 27), interest credits of banks in these centers may be estimated at $55 billion in 1983, and interest debits at $51 billion, generating $4 billion of net credits. Probably there is not much net return from the pure interbank accounts, or there may be a net return only to the extent that money is borrowed from, and lent to, smaller banks. On the other hand, business with non-resident nonbanks, roughly one third of the total, is certainly more profitable. Much of the business may be the channeling of funds from the nonresident head offices of banks to nonresident nonbanks, and vice versa.

Table 38

OTHER DEVELOPING COUNTRIES:1 CROSS-BORDER BANK POSITIONS AND RELATED INTEREST INCOMES, 1979–85

(In billions of U.S. dollars)

article image

Hong Kong, Singapore, Bahrain, The Bahamas, the Cayman Islands, the Netherlands Antilles, and Panama.

Data were take, from International Monetary Fund, International Financial Statistics (Washington), Vol. 39 (June 1986).

Liabilities of Hong Kong banks vis-à-vis nonresident nonbanks are not reported and are therefore not included.

Vis-à-vis nonresident banks.

Based on outstanding amounts (beginning of year) and cross-border interest rates. (See Table 27.)

Balance of payments transactions taken from International Monetary Fund, Balance of Payments Statistics Yearbook, Vol. 36 (Washington, 1985), Part 2, plus, for 1979 and 1983, questionnaire revisions and reclassifications.