INCREASING SURVEY PARTICIPATION
The twenty-nine countries that participated in the 1997 CPIS accounted for approximately 80 percent of estimated world holdings of portfolio investment in the form of equities and long-term securities. The major investing countries/territories that did not participate included Germany, Luxembourg, Switzerland, Hong Kong SAR, and a number of offshore financial centers. Nonparticipants indicated that their primary reasons for not taking part in the 1997 CPIS were the difficulty of obtaining, from nonresident custodians, data on portfolio investments of resident households; the lack of participation by important partner countries; the absence of plans to compile IIP statements; and concern that the release of confidential data to CPIS compilers might discourage investment. These issues must be successfully resolved prior to implementation of the 2001 CPIS. The participation of the major investing countries and OFCs is essential if the 2001 CPIS is to provide a reliable measure of world holdings of portfolio investment assets and an equally reliable measure, which will be based on estimates from partner countries, of each participating country’s portfolio investment liabilities. The participation of other countries is essential, as well, to achieve broader application of BPM5 methodology. Therefore, vigorous efforts should be made to secure the participation of more major investing countries in order to address the under-reporting of global portfolio investment assets and to confirm the reliability of the global data on portfolio investment liabilities.
As noted, about two-thirds of the countries that participated in the 1997 CPIS were already compiling IIP statements as integral parts of national balance of payments statements, and most of the other participating countries were planning to compile IIP statements. The 1997 CPIS facilitated this process by requiring geographic disaggregation of portfolio assets because, to provide the required data, many compilers (even those that were collecting stock data attributed geographically) introduced major changes and refinements. In some cases, the 1997 CPIS even provided an incentive to establish data collection systems to support compilation of IIP statements.30 The participation of more countries in the next CPIS depends, in part, on the degrees to which countries have implemented BPM5 recommendations regarding the balance of payments Financial Account and the IIP. As more countries take actions necessary for preparing annual IIP statements, the reporting of stocks and flows of portfolio investment should be improved, and global discrepancies should be reduced.
SECURING THE PARTICIPATION OF OFFSHORE CENTERS
Offshore financial centers play a variety of roles with respect to portfolio investment—for example, a banking role (accepting deposits from nonresidents, investing in portfolio assets abroad, issuing securities on behalf of nonresident clients); a mutual fund role (selling units in investment funds that invest abroad); or a custodial role (managing portfolios abroad on behalf of nonresident clients and assets under management by trust companies). Because residents are not involved in these transactions, statistical reporting requirements are likely to be minimal. Concern about protecting confidential information on customers from tax or other authorities in their countries of residence causes reluctance regarding the collection of data required for full participation in a CPIS—especially as there are potentially fewer benefits and greater costs of participation. Also, in many countries with offshore financial centers, existing legislation puts offshore companies beyond the reach of national statistical agencies. However, legislation governing the monetary authorities (or financial service centers) generally gives them the power to collect necessary statistics from offshore financial institutions. Plans for any statistical survey must include recognition of the limited resources available to the relevant authorities. To successfully conduct statistical surveys such as the CPIS, countries with offshore financial centers must establish appropriate statistical authorities empowered to protect confidential information and regulations governing the reporting of cross-border financial assets by offshore financial institutions. Therefore, a particular effort should be made to ensure the participation of offshore financial centers.
Definition of Offshore Financial Centers
The countries or territories recognized by the international community as OFCs have varied over time—not only because of the lack of a precise definition for an OFC but also because new centers continue to emerge and change the geography of financial activities.31 Offshore financial centers differ greatly in size, function, and structure; however, OFCs exhibit some or all of the following characteristics:
-
Entities operating in OFCs perform a broad spectrum of financial activities such as banking intermediation (taking deposits and lending), fund management, insurance services, trust services, asset protection, corporate planning, tax planning, and other financial services.
-
Many OFCs are located in small economies and represent the bulk of financial activities conducted by these countries.32
-
Entities operating in OFCs are often controlled by nonresidents, deal predominantly with nonresident counterparts, and engage in transactions in currencies that differ from those of the countries in which the OFCs are located.
-
OFCs are typically located in host countries that provide incentives to attract nonresident financial activity. Offshore entities are generally free of taxes and exchange controls imposed in domestic financial markets, and are sometimes granted additional “informal” incentives (banking secrecy and anonymity, for example).
For these reasons, OFCs are normally isolated from domestic financial systems, although a progressive integration of OFCs has been observed in a number of cases.
A useful taxonomy of OFCs has been suggested by Marcel Cassard in “The Role of Offshore Centers in International Financial Intermediation,” IMF Working Paper, WP/94/107, September 1994. Cassard distinguishes among international, regional, and booking centers.33 International or primary offshore centers are full-service centers that act as financial intermediaries for worldwide customers and provide both the sources and uses of funds in the centers’ market areas.34 Regional or secondary offshore centers are those that channel funds from outside the areas in which they are located to the regions in which they are located. Because of the low absorptive capacity of the regional economies, such centers may also channel regional surpluses to outside uses. The majority of regional OFCs require foreign institutions to maintain physical presence; employ staff; pay minimal capital requirements; and conduct financial transactions such as raising, lending, and investing funds within specific jurisdictions. (All of these activities involve initially high fixed costs.) In return, regional OFCs provide flexible regulatory environments, tax concessions, and attractive tax treaties to avoid double taxation of interest and dividends. Most regional OFCs specialize in a few financial activities that may range from interbank operations to trading of foreign currencies to fund management and leave markets for other financial services undeveloped. Booking offshore centers are those in which international financial institutions locate “shell branches” for the purpose of recording financial transactions.35 Funds come from outside the region and are used outside the region. Booking centers are primarily used by financial institutions as legal residences for registering financial transactions that are arranged and managed—for purposes of trust planning, tax evasion (particularly wealth and inheritance taxes), and anonymity— for individuals in other jurisdictions. In general, foreign institutions do not maintain substantial physical presences in these centers; do not have staff and facilities (only local representatives); and do not offer deposit-taking, withdrawal, or loan services.36
Regulation of Offshore Financial Centers
Over the past few years, concern about the migration of financial activities to unregulated offshore financial centers (regulatory arbitrage) has led to a growing awareness of the necessity for international regulatory standards applying to offshore financial centers and covering offshore banking and insurance activities as well as collective investment schemes. For example, such standards would require consolidated supervision of banks headquartered in third countries.37 Although these concerns are not directly linked with balance of payments issues, considerable attention has been paid to the role of OFCs in channeling external financing (both bank and nonbank) to emerging market economies in which the financial systems are potentially more vulnerable to reversals in capital flows, rapid accumulations of short-term debt, unhedged exposure to currency fluctuations, and selective capital account liberalization. More information about the activities taking place in OFCs is therefore needed to develop accurate assessments of the risks generated by or faced by OFCs. A better understanding of the nature of OFC activities and OFC links with the global financial system is even more pertinent if the financial system of the onshore economy has the potential for instability.
Availability of Data from Offshore Financial Centers
It is clearly stated in the BPM5 (see paragraphs 79, 365, and 381) that offshore entities should be treated as residents of the economies in which the centers are located. In accordance with this rule, the external transactions of such entities should be covered in the balance of payments and IIP statements of host countries. In many cases, however, OFCs are regarded as nonresidents in the countries where the centers are located. As a consequence, transactions between offshore entities and third economies are excluded from host countries’ balance of payments statistics and, somewhat perversely, transactions between host countries and offshore entities are included in such statistics.38 Many countries and territories classified as booking centers (the British Virgin Islands, the Cayman Islands, Guernsey, Jersey, the Isle of Man, the Marshall Islands, and Turks and Caicos) do not disseminate any balance of payments and IIP statements. When balance of payments statements are compiled, offshore transactions with third countries are regarded as occurring between nonresidents and excluded from the balance of payments statistics of host economies.39 More statistics are available for international and regional centers because of the larger size of the related local economies.40 With only a few exceptions, OFCs normally do not record any entries for portfolio investment, assets.
The major source of information on OFC banking activities is the BIS. Eighteen industrial countries and economic territories (Austria, Belgium, Canada, Denmark, Finland, France, Germany, Ireland, Italy, Japan, Luxembourg, the Netherlands, Norway, Spain, Sweden, Switzerland, the United Kingdom, and the United States) report data to the BIS, and six offshore centers (the Bahamas, Bahrain, the Cayman Islands, Hong Kong SAR, the Netherlands Antilles, Singapore) report as well. Data on banking assets are reported on a locational basis and a consolidated basis; data on banking liabilities are reported on a locational basis. Locational data are compiled in accordance with the balance of payments residency concept and are attributed by instrument (loans, deposits, securities, other assets and liabilities); currency (domestic and foreign—US dollar, euro, yen, British pound, and Swiss franc); sector (banks and nonbanks); and country (individual countries, international organizations, and monetary authorities). However, some OFCs are unable to distinguish holdings of securities from other types of assets such as loans.41 Therefore, it is not possible to derive, from geographical details on OFC banking assets, information on counterpart portfolio investment liabilities.
The largest offshore centers (those with end-1997 banking assets and liabilities of more than US$10 billion according to BIS reports) are: Hong Kong SAR, Singapore, the Cayman Islands, the Bahamas, the Netherlands Antilles, Panama, Bermuda, Bahrain, Liechtenstein, Cyprus, Monaco, and Macao.42 Some of these centers (the Cayman Islands, the Netherlands Antilles, Hong Kong SAR, Aruba, and Bermuda) are also known to have issued substantial amounts of debt securities in international markets. These debt securities were attributable to both bank and nonbank residents of OFCs. If balance of payments and IIP statements were compiled by the relevant OFCs and if the debt securities were acquired only by nonresidents, the securities should have been reported in the statements of the relevant OFCs. It appears that the debt securities were not reported because the (scarcely available) OFC balance of payments statistics for that period did not reflect a large proportion of securities outstanding. The omission possibly resulted from the non-coverage of offshore transactions in the balance of payments and IIP statistics compiled by countries in which OFCs were located. However, portfolio investment liabilities compiled from 1997 CPIS results were more in accord with the BIS measurement of international issues of debt securities, although the liabilities compiled from CPIS data exceeded, for some countries, those reported to the BIS. This anomaly may have occurred because the 1997 CPIS covered equity and debt securities issued domestically but the BIS did not.
Despite the limited availability of data on portfolio investment held by OFCs, there is reason to believe that the role of offshore financial centers as conduits for portfolio investment has grown significantly since the Godeaux Report was written. It is presumed that such centers account for a significant fraction of world portfolio investment holdings.43 However, at present, there is no comprehensive picture of OFC international investment positions. The lack of adequate data on OFCs has severely limited the study of their role in the international financial system. The availability of data from OFCs must be increased if the potential impact of OFC activity is to be effectively discerned.
Reporting by Offshore Financial Centers
Data-gathering efforts for OFCs are hampered by
the very limited resources available to the relevant statistical agencies
laws or regulations that, in some cases, restrict access to (or the release of) relevant information
the fact that the population of offshore companies to be covered is often large
the limited physical presence (requiring survey responses to be solicited from accounting centers located abroad)of many entities from the offshore financial centers.
For these reasons, improvements in statistical reporting by OFCs will be shaped by (a) the introduction of legislation empowering statistical agencies to collect data from offshore companies and (b) the re-design of data collection systems to make the best use of existing data sources.
Improved statistical reporting by the offshore financial sector will depend greatly on regulation. Most countries in which OFCs are located have established, or are in the process of establishing, regulatory authorities for oversight of offshore financial institutions (banks, insurance companies, collective investment schemes, hedge funds, trust companies, etc.). The implementation of regulatory standards would facilitate improved statistical reporting. Offshore companies could be treated (in a manner consistent with international statistical standards) as residents of the countries in which the OFCs are legally domiciled. The contribution of offshore enterprises to value added and the impact of an OFC on other sectors of a country’s economy could be assessed. The structure, historical evolution, and financing of the offshore economy could be studied. Comparisons of offshore centers could be made, and the collective impact of OFCs on the rest-of-the-world could be assessed.
The participation, in the BIS international system for reporting locational banking statistics, of countries that are hosts to OFCs was strongly encouraged in the report of the Financial Stability Forum’s Working Group on Offshore Financial Centers. Because of the importance of OFCs as channels for cross-border portfolio investment, the participation of OFCs in the 2001 CPIS is also essential to ensure that coverage for this second survey is global. In both instances, the quality of the data reported would be greatly enhanced by the implementation of regulations that establish standards for statistical reporting.
As there are no international regulatory standards for trust companies, the capacity of OFCs to report the value of cross-border portfolio investment assets held in trust for nonresidents is uncertain. Current practice in most countries is to attribute such assets to the sector of the beneficial owner, which is likely to be an investing country and not an OFC. Thus the interest of OFCs is essentially that of third parties. The collection of third-party data from custodians and trust companies is a potentially useful means of filling gaps in the 2001 CPIS now being planned by participating countries.
COVERAGE
Short-term Instruments
The 1997 CPIS focused on long-term portfolio investment instruments because the authors of the Godeaux Report considered the under-reporting of long-term portfolio investment assets to be the most likely cause of the global discrepancies in portfolio investment. However, it is probable that short-term portfolio investment flows were also substantial and also suffered from under-reporting. During liquidity crises, short-term instruments are particularly important because financial instruments are no longer effectively rolled over or refinanced. The reporting of data on short-term instruments would provide complete coverage of portfolio investment asset holdings, and most existing data collection systems could be adapted to obtain this information.44 More complete coverage of debt securities would be especially useful for developing partner country sources of data on external debt statistics. Expanded coverage should also permit the making of more accurate comparisons of total debt securities as some survey participants may have difficulty in correctly separating short- and long-term securities.
Therefore, as in the 1997 CPIS, coverage of portfolio investment assets in the form of equity and long-term debt securities will be mandatory for the 2001 CPIS; coverage of short-term debt securities (voluntary in 1997) will be required in 2001. Disaggregations of securities held by nonresident households and entrusted, for the purpose of investment abroad, to resident custodians (third-party holdings) and geographic attributions of long- and short-term portfolio investment liabilities will be voluntary.45 For the reasons stated in section IV, positions in financial derivatives will not be covered. Because existing collection systems cannot be readily adapted to support the reporting of data on a remaining maturity basis, classification by remaining maturity will not be required for the 2001 CPIS. To facilitate the analysis of survey results, all participating countries will be asked, although not required, to prepare an IIP statement for end-December 2001.
Portfolio Investments Held by Households
A number of countries that participated in the 1997 CPIS expressed concern that reliable estimates of portfolio investment made abroad by households could not be constructed from available sources of data. This was particularly true when households did not employ the services of resident intermediaries or custodians. Some countries did not participate in the 1997 CPIS because such deficiencies could not be resolved without the participation of offshore financial centers or of major investing countries with “onshore” financial centers. Therefore, efforts should be made to develop improved sources of data on households making portfolio investments abroad.
Some 1997 CPIS participants were doubtful about the effectiveness of any data collection strategy (end-investor or custodian, aggregate or security-by-security) for tracking portfolio investments made abroad by households using the services of nonresident agencies. In some industrial countries, the extent to which households used nonresident custodians had a significant impact on data collection, and the same was thought to be true in many countries that had exchange controls or underdeveloped financial infrastructures. Only a few countries attempted to collect portfolio investment data through household surveys; other countries approached nonresident custodians directly or requested that data on household portfolio investments be reported on a voluntary basis. Results were less than satisfactory in all cases.
An alternative approach that could be taken by countries already relying on data reported by resident custodians would be to request that these custodians also report information on portfolio investments made abroad on behalf of nonresident clients (third-party holdings). To prevent double-countings custodians would have to distinguish between corporate clients and households. Information on households could be useful; information on corporate clients would likely be reported elsewhere. Third-party reporting, which is a complex data collection strategy, covers of the nonresident household’s country of residency and the issuer’s country of residence. Problems to be addressed include the treatment of nominee accounts and, in the event that chains of custodians are involved, identification of the custodian serving as the first counterpart. Moreover, without the requisite legal authority, third-party reporting would be purely voluntary, and it could be difficult to persuade offshore financial centers to collect and report information on holdings of nonresident households.
Repos and Securities Lending
A number of countries found it difficult to implement the 1997 CPIS requirement that repos and securities lending should not be reported as portfolio investment but as collateralized loans. Such difficulties arose when these holdings were recorded as portfolio investment in source data (such as securities registers and, in some cases, transactor records). This practice may have resulted in inconsistent reporting by countries that were parties to the same transactions. However, it is believed that most repos and instances of securities lending consisted of short-term instruments issued as a means of providing liquidity to financial markets. If such remains the case, inconsistencies in the treatment of repos and securities lending will not affect the outcome of future surveys on long-term securities. Nonetheless, an effort should be made to determine whether the impact of the requirement for reporting repos and securities lending as portfolio investment should be evaluated.
Creditor/Debtor Reconciliations of External Debt Statistics
Reconciliations of creditor/debtor data for the traded elements of external debt are critical for improving the quality and timeliness of some countries’ external debt statistics. Therefore, the 2001 CPIS should facilitate such reconciliation by encouraging countries to report data on the institutional sectors of issuers, the institutional sectors of holders, securities issued in domestic and foreign currencies, and securities issued in domestic and foreign markets. The addition of supplementary data categories does, however, add to the complexity of the survey. Coverage of these data may have to be postponed until securities databases from which such additional information can be readily obtained are in place.
Disaggregations of portfolio investment assets by security issuers’ countries of residence have intrinsic value because the information is useful for assessing “country risk” and cross-border contagion.46 Data from the 2001 CPIS should be useful for identifying nonresident holdings of securities issued, both domestically and abroad, by residents because survey coverage of securities issued by residents should be broader than that available through market sources. Survey data should also comprise a useful supplement to available banking and market sources on the creditor side and to official external debt statistics on the debtor side. The value of external debt data from the 2001 CPIS would be enhanced by identification of the institutional sectors of the issuers (and possibly the holders) and by identification of domestic and foreign currency issues.
DATABASES FOR SECURITIES
In the opinion of many compilers, the security-by-security approach to data collection produces data of better quality than the aggregate approach because, for the most part, the security-by-security approach permits more rigorous checking of data. A key element of this approach is the creation of national databases containing information on securities. Therefore, efforts should be made through the 2001 CPIS to promote the creation of national and international databases containing information on securities.
Such databases would also facilitate the collection of data on remaining maturities, currency composition, nominal values, ultimate debtors, related accrued income, and future debt servicing. (Often, the data collection systems of countries that do not follow the security-by-security approach cannot readily be adapted to capture these additional data.) National databases for securities would also be useful for reconciling data on partner country assets with data from countries reporting liabilities and for developing flow of funds accounts that could provide information on the dynamics of security markets and permit conditions to be monitored from a systemic risk standpoint. To facilitate the creation of national databases for securities and exchanges of data among countries, the BIS, ECB, and IMF are investigating several possibilities:
the feasibility of designing a model securities database
the scope for improving the global international securities identification number access mechanism network (a means of access to the records of all national numbering agencies) and supporting metadata
the use of commercial databases for statistical purposes and ways of sharing information contained therein
the role of international organizations in facilitating data exchange.
THE CASE FOR AN ANNUAL CPIS
A number of factors support the conducting of annual coordinated portfolio investment surveys at dates subsequent to that of the 2001 CPIS. Many countries now conduct annual portfolio investment surveys to support the compilation of annual IIP statements. Countries that have relied on sample surveys in conjunction with less frequent (such as triennial) surveys have found that sample surveys have tended, at least in recent years, to result in significant underestimation of portfolio investment assets and liabilities. Moreover, some countries consider that conducting the CPIS on a triennial basis is equivalent to starting afresh each time and re-experiencing the attendant loss of staff, institutional memory, and set-up costs for computer systems. Countries that have established, or are planning to establish, securities databases to support the security-by-security approach to compiling portfolio investment statistics would certainly be able to participate in annual surveys. Thus, for both compilers and respondents, setting up a collection system that would be used on an ongoing basis would produce better data and be more cost-effective than periodic surveys.
The case for annual coordinated portfolio investment surveys is reinforced by the usefulness of time series data, collected on a country-by-country basis, on holdings of portfolio investment assets. These data could facilitate the assessment of contagion issues and other aspects of “country risk.” Likewise, the generation, on a country-by-country basis, of partner country data for portfolio investment liabilities would facilitate the assessment of countries’ vulnerability to reversals of capital flows. Data from annual coordinated portfolio investment surveys could also be used to perform more frequent reconciliations of external debt data from creditor and debtor sources.