Some countries omit entirely from their balance of payments accounts certain entrepot transactions that pass through the domestic financial system but that are viewed to be unrelated to the domestic economy. These missing capital flows affect world balances of recorded flows in several ways. For six major offshore financial centers, the totals of unrecorded capital transactions show sizable net capital outflows that are not included in world balances. These outflows appear to be financed by direct investment inflows that are well above the amounts reported by partner countries.

Some industrial countries also omit capital transactions between foreigners that are carried out in domestic capital markets because they have no net effect on their capital accounts. However, these flows affect the measurement of capital flows in other countries, as well as global balances, and they should be brought into the global picture.


Offshore financial centers raise a specific problem for world capital flow statistics in that they omit from their balance of payments statements the large volume of financial flows that pass through their entrepôt facilities. The offshore centers are small economies, and these flows are considered to be unrelated to activity in their countries and inappropriate for balance of payments statements. Some OFCs do not report to the Fund at all. As a result, capital flows in the Balance of Payments Statistics Yearbook omit most financial transactions of offshore centers.

Omission of these flows would not affect world balances of capital flows if OFCs acted only as clearinghouses for transactions. Transactions change form as they pass through OFCs, however, perhaps entering as direct investment or interbank flows and leaving as portfolio investment transactions. Moreover, offshore activities produce sizable net earnings that accrue to direct investors abroad but that are largely retained in the accounts of the OFC affiliates. While OFCs omit such activities from their statements, partner countries report them in their own balance of payments as specific transactions with OFCs. In order to balance capital flow statistics at the world level by form and type of claim, the OFC entrepot flows must be brought into totals as specifically as possible.

Data from OFCs are valuable well beyond their function in balancing types of capital flows at the world level. As choke points in the world’s financial system, the OFCs can shed light on large volumes of capital flows that are not reported in other countries’ statistics. Several important OFCs participate in the BIS/IBS programs discussed in Chapter 6, and they provide geographic detail on about one-fourth of the bank positions in the BIS totals. To fit the geographic information from these reports into the world balance of payments structure is a substantial addition of capital account information.

This chapter brings together the major forms of data that are available for a balance of payments statement of OFC activities, and it presents estimates of amounts omitted from the Yearbook. Chapters 3 through 6 of this report included a number of adjustments for OFCs, and this chapter combines the adjustments for major OFCs with data on transactions of their banks to present as complete a picture as possible for these centers.

The problem of omitted entrepot flows is not limited to OFCs. Many other international financial centers, such as the United States and Switzerland, also omit from their balance of payments statements many transactions among foreigners that pass through domestic markets. This chapter also discusses aspects of the problems raised by such omissions.

Table 48 summarizes all of the capital account adjustments proposed in this report for OFC activity. All together they increase net capital outflows by an annual average of $8.8 billion, for 1986–89, with a major effect in “other capital” through the use of banking statistics.126 At least half of this adjustment is for the six OFCs discussed later in this chapter. Adjustments for Hong Kong result partly from use of BIS/IBS banking data for 1987–89, but they also include other adjustments such as the large portfolio adjustments that are discussed in Chapter 4.

Table 48.

Summary of Capital Account Adjustments for Activity in Offshore Centers

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

article image

Discussed in later section of this chapter.

Mainly portfolio investment transactions that are not allocated to specific OFCs. Includes adjustments for Bermuda and Gibraltar.

Special financial institutions are not included in totals. The SFIs are discussed in a later section.

Adjustments for “other” OFCs in Table 48 include portfolio transactions that cannot be allocated to individual OFCs, but most of these transactions are probably in the six OFCs. The adjustments in the memorandum line of Table 48 are a separate issue that arises from the Netherlands treatment of certain cross-border transactions of its special financial institutions. International transactions of SFIs are included on a net basis in the Netherlands balance of payments statement. They are combined into a single figure in “other capital,” but this figure includes SFI current account results. The memorandum adjustments in the table remove this component and align the Yearbook figures to match reporting by other countries. The SFI adjustments are attributed to industrial countries in Table 6.127

Forms of Offshore Activity

OFCs typically have legal, tax, regulatory, and exchange control systems that are designed to encourage or promote financial transactions among foreigners through domestic intermediaries. OFCs have special classes of business licenses that exempt holders from most forms of domestic regulation and taxes, provided such firms carry on only international business activities and do not participate in the local economy. A few licensees have actual offices in the OFCs, but most are no more than “brass plate” companies that have no staff or facilities in the OFC other than a local representative. The transactions of these companies are conducted from other countries, and records are often kept elsewhere. It is the transactions of these special-license firms that are excluded from OFC balance of payments statements and that constitute offshore financial activity in the most common meaning of the term.

International banks dominate the financial activity of OFCs. The banks establish themselves in OFCs through branches or subsidiaries and run important parts of their business through those offices for a variety of purposes. Many of the OFC banking positions are with affiliates in other countries. Some banks pass funds from other offices through OFCs to lend to non-banks in both home and foreign countries, while others use their OFC offices to raise funds from nonbanks for worldwide activity.

There are also many nonbank special-license firms in OFCs. On average, the nonbanks are much smaller than OFC banks but, as a group, they develop significant cross-border financial positions and flows. The nonbanks, like the banks, usually are branches or subsidiaries of companies in other countries. Insurance company branches and subsidiaries are important, including captive insurance companies established by industrial or other corporations elsewhere. Many non-bank licensees are not affiliated with companies in other countries; they are also established to carry out a variety of personal financial activities for which the conveniences of OFCs are valuable.

Six Offshore Financial Centers

This section presents some speculative and incomplete estimates for the offshore activities of six OFCs: The Bahamas, Bahrain, Cayman Islands, Netherlands Antilles, Panama, and Singapore. The six are all major banking centers, and all are small economies. These OFCs omit most or all of their activities as offshore centers from their balance of payments statements as submitted to the Fund. All except Panama also report international banking statistics to the BIS and IBS; for Panama there are data for branches of U.S. banks. This section presents estimates, for the six OFCs together, of the balance of payments flows that are omitted from the Yearbook and that should be added back into world capital flows.

To put these OFCs into world perspective, Table 49 compares their international banking positions in 1985 and 1989 with banks in BIS-reporting industrial countries and Hong Kong. Hong Kong is a seventh major OFC in the BIS/IBS statistics, but its statistical relation to the Yearbook is different from that of the other six. Hong Kong is discussed in a later section of this chapter.

Table 49.

Cross-Border Banking Positions for Offshore Financial Centers, Years-end 1985 and 1989

(In billions of U.S. dollars)

article image
Source: International Financial Statistics, 1991.

International Banking and Financial Developments, Bank for International Settlements, Table 2.

Table 49 presents a number of similarities and differences between banking in OFCs and in industrial countries. In 1989, cross-border assets of banks in the six OFCs were about 25 percent of those in the industrial countries, and OFC nonbanks held about 10 percent of the claims on nonbanks reported by all international banks. OFC positions were thus a sizable part of total world banking positions and a source of possibly large balance of payments adjustments.128 Assets of OFC banks grew faster than liabilities, and implicit net worth thus increased very fast, with implications for direct investment flows in the OFC balance of payments accounts. In industrial countries, the opposite occurred, and banks’ cross-border positions went into deficit between the two years.

Sources of Data

Estimates for offshore activities of the six OFCs in this section are based on comparisons between published balance of payments statements for the OFCs and data available from other sources. The statistics from outside sources are far from complete, but they constitute a starting point for judging the effects of OFC transactions on world capital flows.

In Tables 50 and 51 of this section:

Table 50.

Six Offshore Financial Centers: Adjustments of Balance of Payments Statements to Include Offshore Banking Activity, Net Annual Average Flows, 1986–89

(In billions of U.S. dollars; outflows ( –)) 1

article image

All figures are net balances of inflows and outflows on both assets and liabilities.

The balance of payments “other” category consists of all capital flows except direct investment, portfolio securities, and reserves/LCFAR. The amounts in the top half comprise all flows in that category for deposit money banks and for private nonbanks and all official “other capital” flows except loan asset and liability transactions.

No balance of payments accounts are included in the Yearbook for the Cayman Islands; the lower half shows the total IBS flows.

Table 51.

Total Effect of Adjustments to IBS Data, Direct Investment, Portfolio Investment, and Short-Term Securities Statistics, for Six Offshore Financial Centers, Net Annual Average Flows, 1986–89

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

article image

From Tables 12 and 14, based on partner country data. Excludes $3.9 billion residual derived from income estimates.

From Tables 18 and 20.

From Chapter 5.

Missing from these data are direct investment flows vis-à-vis countries other than those that provided the geographic detail used in Chapter 3, portfolio purchases by OFC residents other than those included in bank assets in OFCs, and OFC nonbank transactions other than those with international banks. These missing items are noted later in the chapter. The amounts and net balances in such flows are not known.

Adjustments to Capital Accounts

The relation of banking data for OFCs to their balance of payments data is illustrated in Table 50. Of the six OFCs in the table, the Cayman Islands is an important special case. The Cayman Islands is one of the most important of the OFCs, but it is omitted entirely in the Yearbook. International banking data for this OFC are an important addition to data on world capital flows, and Table 50 indicates the extent to which the Cayman Islands dominates OFC flows other than those that pass through Hong Kong.129

In order to keep the data concise, the comparisons in Table 50 are condensed into net flows—liability and asset flows combined into a form in which a positive number indicates a net inflow into the listed OFC. The top part of the table gives balance of payments bank-related data for the OFCs from the 1990 Yearbook. Most of the amounts are small, because they exclude most offshore activity by special-license firms and because they all are small economies. The bottom part gives the excess of corresponding IBS net flows—directly reported bank activity and net derived changes in OFC nonbank positions—over the Yearbook amounts. The IBS items are much larger than the amounts included in balance of payments statements. These residuals in the table are thus the amounts that directly affect world balances in this group of figures.

For both banks and nonbanks, the differences indicate a strong tendency toward net capital outflows from the OFCs in the form of credit extended. That is, the OFCs have had larger increases in banking-related assets than in liabilities. While the figures show clear general tendencies—as in Chapter 6 on derived non-bank flows, they indicate large flows missing from the Yearbook—and were used to adjust the global capital account, they are crude and lack the detail needed for use as regular balance of payments estimates. The flows are also highly volatile on an annual basis, and the data must be examined closely before they can be built into an annual series.

Table 51 combines these banking adjustments with those from Chapters 3, 4, and 5 in direct investment, portfolio investment, and short-term securities for the six OFCs. These additional adjustments combine to a $0.9 billion average increased inflow to the six OFCs for 1986–89. With this offsetting adjustment, the total net balance of offshore financial activity in the six centers averages to an $8.4 billion capital outflow for the group.

The data in these tables are, of course, incomplete. The direct investment amounts reflect reports for an incomplete set of partner countries. The portfolio figures include almost no transactions in assets, and resident nonbank “other capital” flows are only those with reporting international banks. Although the net effect of the unmeasured flows is unknown, the portfolio asset component could normally be expected to add to the net outflow.130 For example, although the six OFCs may have been important sources for the $6 billion unallocated portfolio outflows mentioned in connection with Table 48, these flows into securities markets are not included in Table 51. If the total $6 billion outflow were added to Table 51, the total net outflow for the six OFCs would average $14 billion.

Even an $8 billion additional net capital outflow for this group of OFCs immediately raises questions as to the sources of such a volume of funds. A large part of the answer is apparently in the net earnings of the financial operations in the OFCs. During 1986–89, banks in the six OFCs had average cross-border asset and liability positions in the range of $700 billion, and nonbanks in the group had net positions—assets less liabilities—of about $26 billion with foreign banks alone. If the OFC nonbanks were getting a reasonable yield on foreign net banking assets, and if the OFC banks had earnings from rate spreads and fees, then income for the six OFCs may have been $5 billion a year from banking activity alone.131 Beyond the bank-related income are direct investment and portfolio income that are also excluded from national balance of payments coverage. Income is just one aspect of the question, however, and the $8 billion outflow can be understood only in the broader context of a complete balance of payments statement.

A Balance of Payments Statement for Offshore Activities

Table 52 sketches a summary balance of payments statement for offshore activities in the six OFCs. The table combines the figures from Table 51 with others that can be deduced from them. It also indicates the major items that are missing for the full picture. The first column of the table is the full statement as it can be constructed from Table 51, and the second column shows the amounts of these flows that are mentioned in earlier chapters on direct investment, portfolio, and nonbank transactions with foreign banks. The third column contains the remaining amounts that emerge in this chapter.

Table 52.

Balance of Payments Statement for Offshore Financial Activity of Six Offshore Centers, Annual Average Flows, 1986–89

(In billions of U.S. dollars)

article image

Income and reinvested earnings in this column are based on the Yearbook, 1990, Part 2, Table 3. That table includes adjustments for OFC investment income that are not shown separately.

Included in the fifth line of Table 12. Derived in Chapter 3 from partner country data.

Included in the fifth line of Table 12 as an unallocated residual for OFCs.

Included in line 2b of Table 14.

Portfolio asset net flows by OFC nonbank residents. Portfolio adjustments in Chapter 4 include an average $6.4 billion of net purchases (outflows) by offshore centers, but these purchases could not be allocated to individual OFCs and are therefore omitted here. If all of this outflow were inserted into the table, errors and omissions would become $11.8 billion.

OFC nonbank “other capital” transactions with nonbanks.

The current account in Table 52 consists only of an estimate for investment income taken from Part 2 of the Yearbook, accompanied by an equal and offsetting outflow in the form of reinvested earnings. The earnings figure includes components for banks and non-banks separately. These earnings belong almost entirely to nonresidents, and in balance of payments accounting they are debited in the current account, as in the second line of Table 52, whether or not they are paid out to the owners. Earnings that are reinvested (that is, not distributed to parents) then appear again in the capital account as a credit. In balance of payments terms, it is the reinvestment of earnings as a capital flow that is a source of net asset growth in Table 52. In principle, the net current and capital account balances of offshore centers should reflect only the transactions of local residents other than the special-license firms.132

The large positive errors and omissions in Table 52 show the extent to which the available data are out of balance. Bank-related capital outflows—$9.3 billion on average—are larger than the income and other direct investment inflows that are included as possible sources of bank funds, even on the assumption that all investment income is retained in the OFCs. If the banking numbers are approximately correct for these OFCs, then Table 52 suggests strongly that the direct investment net inflows are still understated. This is a reasonable possibility. The four countries that separately report reinvested earnings show for OFCs an average total of just $1 billion (second column of the table) out of the notional $5 billion total, mentioned earlier, based on income estimates. On that very loose evidence, “other” direct investment inflows from all owners might average $5 billion to $10 billion in the table. Such inflows would reduce errors and omissions considerably.

Even more inflows are needed, however, to balance the accounts. Table 52 notes the two forms of capital flow that are missing altogether—portfolio assets and nonbank transactions with nonbanks. Both of these items can be expected to show net capital outflows in most years. OFC banking evidently is not a source of funds for such outflows, since in Table 52 banking appears to absorb funds from other OFC sources. The alternative in this context is still more inflows in the form of direct investment.

Table 52 thus implies that there may have been inflows of direct investment into these OFCs as large as $10 billion to $15 billion a year during 1986–89. These flows are entirely missing as credits in OFC capital accounts; roughly $5 billion are visible in other countries’ data as debits. Bringing these OFC direct investment flows into the world totals would improve the match of debits and credits in direct investment, but at present there is no basis in available data to support such a change. From the evidence in Chapter 3 and Table 52, proposals for OFC direct investment flows are limited to raising their net credits by an average of $8 billion for 1986–89.

Conclusions on Six OFCs

Table 52 illustrates that excluding offshore financial flows in OFC balance of payments results in, at least, an understatement of direct investment and banking flows in world balances. The excluded flows are large enough to create marked asymmetries in both income and capital flows. The problem in OFC data can be seen as the mirror image of the capital account problem for other countries. Most countries have better measurements of their cross-border inflows than of their outflows into foreign assets. The OFCs, on the other hand, show far more outflows than recorded inflows in Table 52. This difference is consistent with the nature and advantages of OFCs as investment havens. An expanded measure of nonbanking flows through OFCs, with or without geographic detail, undoubtedly would bring into the world picture large volumes of transactions that are now missing.

Host countries for offshore financial activity prefer to leave this activity out of their national balance of payments statements, and there are analytic reasons for that practice. World statistics on capital flows and positions suffer from the omission, however, and the Yearbook would be a richer mine for analysis with more information on offshore activities. This is an important objective. The Working Party recommends that OFC activities be brought into Yearbook statistics to the fullest extent possible. Countries hosting offshore activity should be encouraged to collect the data and to bring them into their balance of payments statements. The Fund should develop its own estimates for major offshore activities that are not included in national statements. Attention should be paid first to the smaller offshore centers where the flows can be identified most readily. Figures then should be introduced for centers, such as the Cayman Islands, that are not included in Yearbook statistics. Banking data reported by OFCs to the BIS and IMF can play a major role in such a compilation, and the derived nonbank figures discussed in Chapter 6 are also important. BIS information on international securities markets is another data source.

Hong Kong

Hong Kong is statistically in a different position from the six OFCs in the preceding section. Like the Cayman Islands, Hong Kong itself does not submit a balance of payments statement to the Fund. Unlike the Caymans, however, Hong Kong is included in Part 2 of the Yearbook, although it is not shown separately. The amounts included are compiled by the Fund, based in part on Hong Kong’s banking reports for BIS/IBS purposes. Hong Kong’s banking information is seriously out of balance, however, because it excludes bank liabilities vis-à-vis foreign nonbanks. This gap is consistent with the large positive errors and omissions found in the Hong Kong balance of payments.

Hong Kong banking positions grew rapidly in recent years—at an average annual rate of 35 percent from 1985 to 1989—and are now the largest in the offshore group that reports in the BIS/IBS program. By 1990, Hong Kong banks’ foreign assets were more than $460 billion, almost half of those of the six OFCs discussed previously combined. Hong Kong is thus a major part of world capital markets and it should be represented in the world capital accounts more completely than it has been.

Table 48 includes adjustments to Hong Kong data that oscillate widely and average –$1.0 billion a year. These are mainly a combination of portfolio inflows averaging almost $5 billion, $3.5 billion in outflows derived from international banking statistics for the years 1987–89, and about a $2 billion average increase in official reserves. These are major adjustments for an economy the size of Hong Kong. The Working Party recommends that the Fund use all of the information available on Hong Kong’s cross-border capital transactions in constructing a balance of payments statement for this important financial center.

OFC Activity in Other Countries

The six OFCs in the earlier section have the common characteristics that they are reporters for the BIS/IBS banking statistics and that they exclude most of their cross-border financial activities from balance of payments statements. For this group of countries, it appears possible to build a reasonable balance of payments statement for the excluded offshore activities that will help give a better picture of world capital flows.

A number of other countries also have offshore transactions that they exclude from balance of payments figures. Such activities have a variety of connections to the Yearbook world totals, and their effects on world balances are either accounted for or unknown. These transactions pass through resident institutions, but compilers in some cases exclude them as not relevant to the national economies.133 Two examples illustrate different forms of the practice.

Flows into custodial accounts may be largely missing in balance of payments statements for countries of origin. If custodial information were available from banks included in the BIS/IBS reporting system, the data could broaden substantially the scope of geographic information available to measure derived positions for nonbanks. Comparisons of such totals with balance of payments figures for nonbanks could then go well beyond the comparisons with direct bank relationships that appear in Chapter 6. Even without geographic detail the figures still suggest the size of missing flows in current as well as capital account statistics.

Without data on flows through custodial accounts, an information gap exists that may be large. The Working Party recommends inquiries in major countries to explore the size of the cross-border custodial positions held by banks for nonresidents.

Conclusions and Recommendations

A number of countries omit from their balance of payments statements large volumes of nonresident capital transactions that pass through local financial markets but that are unrelated to the domestic economies. A major part of these excluded flows goes through offshore financial centers that provide entrepot facilities for banking and other forms of financial business activities. Omission of these transactions causes imbalances in world capital accounts because counterpart entries are included, to some extent, in balance of payments statements of countries where the transactions originate and end. These omissions can be corrected in several ways. Within the framework of the Manual, analytic uses of the data probably would be best served by separate identification of offshore activities within host countries’ balance of payments statistics.

Two principal recommendations arise from the considerations given in this chapter: