Chapter

Annual Report of the IMF Committee on Balance of Payments Statistics

Author(s):
International Monetary Fund. Statistics Dept.
Published Date:
August 2004
Share
  • ShareShare
Show Summary Details

I. Introduction

The IMF Committee on Balance of Payments Statistics was established in 1992 for the following purposes: to oversee the implementation of the recommendations contained in the reports of two IMF working parties that investigated the principal sources of discrepancy in global balance of payments statistics published by the IMF;1 to advise the IMF on methodological and compilation issues in the context of balance of payments and international investment position statistics; and to foster greater coordination of data collection among countries. The membership of the Committee as of December 31, 2003 and its terms of reference are presented in Appendices 1 and 2, respectively. In 2003, the Committee held its sixteenth meeting in December, at the IMF headquarters in Washington, D.C.

This report is structured as follows: Section II presents the Executive Summary. Section III provides an overview of statistical discrepancies in the global balance of payments statistics published by the IMF’s Statistics Department. Section IV discusses the Committee’s work program during 2003, and Section V concludes with a discussion of the issues that the Committee plans to address in the coming year.

II. Executive Summary

Recent Trends in Global Balance of Payments Statistics

Balance of payments statistics reported to the IMF’s Statistics Department and published in the 2003 Balance of Payments Statistics Yearbook (BOPSY) continue to show large and variable discrepancies on the global summations of current, capital, and financial account transactions. Discrepancies in 2002 fell somewhat for both the current and financial accounts, but several of the components recorded larger imbalances. The provision of revised data resulted in the widening of the discrepancies for several earlier years. For goods, although the imbalance for 2002 rose from the previous two years, there was a noticeable improving trend. Total credits and debits in services were almost in balance in 2002, but there were large offsetting imbalances in several of the components. The imbalance on income remained high. In the financial account, the imbalance in direct investment for 2002 fell sharply, but, for portfolio investment, the difference between measured assets and liabilities deteriorated substantially.

Data Quality

Since the financial crises of the 1990s, policymakers and other users of statistics are giving increased attention to the availability of comprehensive, timely, and reliable financial and economic data. In view of the IMF’s responsibility for economic surveillance of its members’ economic policies, and to assist users of the data to evaluate data quality, the IMF’s Statistics Department developed the Data Quality Assessment Framework (DQAF) in 2001 for macroeconomic statistics. The DQAF provides a systematic approach and a common language for the assessment of data quality.

Following a request by the IMF’s Executive Board in 2002 that IMF staff continue to elucidate good practice for data revisions policy, the Committee considered several papers on revision policies and practices at its meeting in 2003. Included in these were papers from the Statistics Department, Hong Kong SAR, Chile, India, the European Central Bank (ECB), and the Statistical Office of the European Communities (Eurostat.) In addition, the Committee received a joint paper by Eurostat and IMF staff, comparing the IMF’s DQAF with Eurostat’s quality definition. These papers were well received, and the Committee endorsed proposals for further work on data quality, including efforts to describe international best practices in dealing with revisions.

Revision to BPM5

The Committee considered the draft Annotated Outline (AO) for the revision of the Balance of Payments Manual, fifth edition (BPM5), which was developed by IMF staff. To reflect the Committee’s comments, the draft will be revised and a final version of the AO sent to all balance of payments compilers and other interested parties in April 2004. Comments will be sought by July 2004.

The Committee also considered the IMF’s proposed process for the preparation and release of the new manual by 2008. To that end, three technical expert groups were set up (one for issues relating to direct investment, one for issues relating to currency unions, and one for issues regarding balance of payments and the international investment position (IIP) more generally) to advise the Committee and subsequently the IMF’s Statistics Department on input to the new balance of payments manual. These groups will draw from the specialized knowledge in these various areas of experts in all parts of the world. In addition, the Committee took note of the need to coordinate the rewriting of the new manual with the reviews of the System of National Accounts 1993 (1993 SNA) and the OECD Benchmark Definition of Foreign Direct Investment (third edition (BD3)), which are being conducted concurrently to the revision to BPM5.

Portfolio Investment

The Committee was brought up to date on the 2001 and 2002 Coordinated Portfolio Investment Surveys (CPIS), as well as work being undertaken to enhance the collection and compilation of CPIS data. These have wider application to other elements in the balance of payments and the IIP, as well as other parts of the suite of macro-economic statistics.

The Committee was informed that the results of the 2002 CPIS, and revised data for the 2001 CPIS, would be available in early 2004. The results were to be placed on the IMF’s website (http://www.imf.org). Efforts by IMF staff continued to increase the participation and coverage of the survey. The Committee also endorsed a proposal by IMF staff to undertake, resources permitting, a survey of the methodology used by those that produce positions data on portfolio investment liabilities.

The Committee was advised that work continues to assess the possibility of collecting data on third party holdings (TPH), that is, holdings of securities by nonresident custodians. A meeting to discuss these issues, and to take them forward, is to be hosted by the ECB in May 2004. Jurisdictions with major custodial business will be invited to attend. Obtaining information on TPH has the potential to fill a major gap in the present coverage of the CPIS.

The ECB’s central securities database (CSDB) has a potentially important role in improving the quality of portfolio investment statistics (flows, stocks, and income) by providing relevant information on individual issues of equity and debt securities. The database would not just assist the compilation of the balance of payments and IIP statistics, but would also represent a valuable input for other macroeconomic statistics, such as monetary and financial statistics and external debt. The ECB anticipates that the database will be operational toward the end of 2004 or early in 2005. Further enhancements are already planned thereafter.

The Committee also received the results of a survey exploring the information available on reverse transactions (repurchase agreements and securities lending without cash collateral) in the management information systems of financial institutions in nine countries. The information will provide important input for the CPIS and for the revised balance of payments manual.

International Financial Statistics

The Committee was informed that the Bank for International Settlements (BIS) continues to increase the number of participating central banks in its international banking statistics: by the end of 2003, Greece and Mexico began participating in the statistics on international bank activity on a locational basis and Australia, Greece, and Mexico joined the statistics on country risk exposures on a consolidated basis (Australia already participated in the locational statistics). It was noted that data quality has also improved, with data on loans, securities, and other assets and liabilities being provided separately in the locational statistics by nearly all participants, and information on currency and country breakdowns in the locational statistics and on an ultimate risk basis in the consolidated statistics having generally improved.

For data on foreign exchange and derivatives markets, the Committee was advised that a record 52 countries will participate in the next BIS triennial survey, which will be conducted in April 2004 covering turnover data on foreign exchange and over-the-counter (OTC) derivatives market activity on a locational basis and at end-June 2004 covering positions outstanding on OTC derivatives market activity on a consolidated basis. For the April part of the survey, nondeliverable forwards (NDFs) will be identified separately. The BIS also collects and publishes statistics on most OTC financial derivatives on a semiannual basis, covering stock data on notional and market values from approximately 60 major dealers in the Group of Ten countries on a consolidated basis. In addition, the BIS compiles and disseminates quarterly data on turnover and amounts outstanding for exchange-traded derivatives. The Committee was told that an increase in the frequency of the semiannual OTC derivatives statistics was not possible for the time being.

Direct Investment

The Committee’s views were sought on a proposal that the IMF’s Statistics Department had taken to the IMF’s Executive Board to undertake a feasibility study as a precursor to a possible internationally coordinated survey of direct investment petitions, that is, a Coordinated Direct Investment Survey (CDIS). Such a survey would build on the successes of the CPIS and the IMF-OECD Survey on the Implementation of the Methodological Standards for Direct Investment (SIMSDI). The Committee was advised that the proposal is for the feasibility study to be conducted in 2004 and 2005, and, if the results indicate that a full survey can be undertaken, that it would be conducted in 2007 at the earliest. The proposal, at the minimum, is for benchmark positions data, broken down by counterpart jurisdiction. The Committee gave in-principle support to the proposal. The Committee was also advised that the IMF and OECD were launching another SIMSDI in late 2003.

Income

Following on discussions on income at its meeting in 2002, the Committee received several papers on the subject at its 2003 meeting. These papers raised questions about the current treatment of reinvested earnings in balance of payments statistics; inconsistent treatment of the income between and among different types of equity investment; the treatment of employee stock options; the outcome of work by the ECB comparing credits and debits on income between partner countries in the European Union (EU); and how the United Kingdom presents to users information on income in the balance of payments.

These papers provoked considerable interest and discussion. The Committee agreed that many of the issues would need to be discussed within the context of the new balance of payments manual.

Uses of Balance of Payments and International Investment Position Statistics

The Committee welcomed a paper by Uganda on the uses of balance of payments and IIP statistics for highly indebted sub-Saharan African countries, using Uganda as an example. The paper gave a wide-ranging description of the value of these data. It noted the role of the link between the balance of payments and the national accounts, in particular, with regard to public versus private sector saving, and how, initially, any shortfall in saving is made up by borrowing from the rest of the world. In the short term, this will lead to an appreciation of the domestic currency, but as debt levels rise, this appreciation will reverse itself. The paper noted that the form of financial inflows was important, indicating that direct investment in entities involved in the production of traded commodities was likely to be the most sustainable. The paper also drew the link between balance of payments and monetary policy, especially when associated with a sharp increase in government borrowing that has been financed from abroad. This development prompts the need to sterilize this increase in liquidity domestically with sales of treasury bills to nonresidents. However, the paper notes that such a policy is not sustainable in the longer term.

Other Issues

The Committee also received papers on other issues, with potential far-reaching implications for the compilation of balance of payments and IIP statistics, and the development of the new manual, on the links between international accounting standards and macroeconomic statistics; the use of an Internet survey for estimates of business travel in France; and security measures for Internet reporting in Japan.

III. Recent Trends in Global Balance of Payments Statistics

The discrepancies in the global summations of current and financial account transactions for 2002, as published in the 2003 BOPSY, narrowed overall, but significant underlying discrepancies persisted. In principle, the combined surpluses and deficits arising from the current, capital, and financial account transactions for all countries and international organizations should equal zero, as the credits of one country or international organization are the debits of another. In practice, however, the data do not offset each other. Statistical discrepancies occur in the global statistics, reflecting the incomplete coverage of transactions, and the inaccurate and inconsistent recording of cross-border transactions by countries resulting from, for example, differences in classification and practices, or in the time of recording transactions. Further, many errors and omissions offset or cancel each other and are therefore not reflected in the data on global discrepancies shown in Tables 1 and 2 of this report.

Global Current Account

The global discrepancy for the current account declined in 2002, but this overall decline masked increases in the discrepancies of several subcomponents of the current account. The discrepancy on goods, at its highest level after 1999, is a positive discrepancy (more exports than imports) and as such offsets negative discrepancies (more payments than receipts) in services, income, and current transfers. Exports and imports of goods recovered from a slowdown in world trade in 2001. The 2002 levels of trade in goods were almost identical to the 2000 levels. The discrepancy in income was at its highest in the seven years under review, reflecting the increased discrepancy for portfolio and other investment income. On the other hand, the discrepancy for direct investment income declined for both reinvested earnings—which reached a record high in 2001—and other direct investment income.

Table 1.Global Balances on Current Account, 1996-2002(In billions of U.S. dollars)
Average
Imbalance
19961997199819992000200120021996-2002
Current account balance-32.329.9-48.7-80.0-102.7-117.8-76.0-61.1
Goods balance98.3114.070.545.716.31.744.155.8
Credit5,348.25,553.95,443.85,646.86,369.16,106.86,369.3
Debit5,249.95,439.95,373.35,601.16,352.86,105.16,325.2
Services balance-20.0-5.2-0.5-13.1-11.2-21.2-8.7-11.4
Credit1,315.41,363.41,385.91,431.81,517.41,518.91,622.4
Debit1,335.41,368.71,386.41,444.91,528.71,540.11,631.2
Transportation-55.7-55.7-52.2-50.3-62.1-53.9-49.4-54.2
Travel31.834.528.427.331.737.434.432.2
Government services-10.3-10.9-8.2-17.7-24.6-22.6-28.1-17.5
Other services14.126.931.527.543.917.934.328.0
Income balance-80.2-53.6-87.3-81.0-65.3-73.3-91.0-76.0
Credit1,049.11,112.01,203.31,248.61,431.11,345.51,263.3
Debit1,129.31,165.61,290.61,329.71,496.31,418.81,354.3
Compensation of employees-7.51.8-0.20.1-1.0-2.7-2.7-1.8
Reinvested earnings67.958.137.272.767.190.578.267.4
Other direct investment income-14.80.0-2.2-37.3-34.6-35.9-26.3-21.6
Portfolio and other investment income-125.7-113.5-122.1-116.5-96.6-125.2-140.2-120.0
Current transfers balance-30.4-25.2-31.5-31.6-42.5-24.9-20.4-29.5
Credit366.0355.4369.1377.5363.3380.0421.2
Debit396.4380.6400.5409.0405.8404.9441.6
Memorandum items
Current account balance as percent of
gross current account transactions0.20.20.30.50.50.60.4
Goods balance as percent of gross
goods transactions0.91.00.70.40.10.00.3
Services balance as percent of gross
services transactions0.80.20.00.50.40.70.3
Income balance as percent of gross
income transactions3.72.43.53.12.22.73.5
Current transfers balance as percent of
gross current transfer transactions4.03.44.14.05.53.22.4
Source: IMF, Balance of Payments Statistics Year book, Vol. 54, Part 2, 2003.
Source: IMF, Balance of Payments Statistics Year book, Vol. 54, Part 2, 2003.

Global Capital and Financial Accounts

The widening discrepancy in the capital account, which more than doubled in 2002, was largely due to an increase in debits (resource outflows) in capital transfers.

The discrepancy in the financial account, like that of the current account, declined by a third in 2002. Once again, this decline masked offsetting movements in the subcomponents’ discrepancies. After two years of large discrepancies, the discrepancy of direct investment decreased to pre-2000 levels. Increased awareness of methodology and practices for measuring foreign direct investment generated by SIMSDI might have been a factor causing a decline in the discrepancy but it is too early to be certain. Direct investment transactions (both inward and outward) were at their lowest level since 1998.

In 2002, portfolio investment recorded its highest discrepancy in the seven-year period under review (Table 2). Transactions in portfolio assets dropped dramatically to approximately half of the transactions recorded in each of the previous three years. The industrial countries, led by the United Kingdom and the United States, accounted for 94 percent of the decrease, which was spread among all portfolio instruments—equity (58 percent decline), bonds and notes (35 percent decline), and money market instruments (43 percent decline). This reported decline in portfolio asset transactions is consistent with financial market developments in 2002 as described in the Global Financial Stability Report (GFSR) March 2003. For instance, the GFSR noted that “falling prices in the major equity markets, sharply widening credit spreads, high levels of actual and implied volatility, and a withdrawal of banks from risk taking in response to loan and trading losses and falling share prices” negatively affected financial flows to emerging markets in 2002 (p. 34). On the other hand, reported portfolio liability transactions did not show the same large decrease, resulting in the large discrepancy even after adjustment was made by IMF staff for the sizable transactions in portfolio investment assets by reserve asset holders.

Table 2.Global Balances on Capital and Financial Accounts, 1996-20021(In billions of U.S. dollars)
Average
Imbalance
19961997199819992000200120021996-2002
Capital account balance-0.22.2-15.8-17.613.8-4.7-11.5-4.8
Credit55.950.644.348.371.551.958.0
Debit56.048.360.165.957.856.669.5
Financial account balance111.091.1-9.142.3214.5171.5112.8104.9
Direct investment6.523.512.015.6158.295.216.046.7
Abroad-368.9-442.8-682.1-1,077.8-1,351.0-702.6-634.1
In the reporting economy375.4466.3694.11,093.41,509.2797.8650.0
Portfolio investment118.0203.7-202.0148.272.732.9210.383.4
Assets-649.6-740.1-1,046.2-1,362.7-1,292.7-1,200.9-655.2
Liabilities excluding LCFAR2767.6943.8844.21,510.91,365.41,233.8865.4
Financial derivatives-10.1-6.6-13.014.8-9.16.42.9-2.1
Assets140.1148.1186.9195.4228.7234.6194.2
Liabilities-150.2-154.7-199.8-180.6-237.8-228.2-191.3
Other investment29.7-83.3205.8-122.445.677.6-87.29.4
Assets-761.3-1,330.5-345.6-540.0-1,253.1-718.8-647.3
Liabilities excluding LCFAR2791.01,247.3551.4417.61,298.7796.4560.1
Reserves plus LCFAR-33.1-46.2-12.0-13.9-52.8-40.6-29.2-32.5
Reserves-189.5-106.0-42.3-152.2-173.1-159.6-265.4
LCFAR156.359.830.3138.3120.4119.0236.2
Net errors and omissions-78.5-123.273.755.3-125.6-49.1-25.3
Source: IMF, Balance of Payments Statistics Yearbook, Vol. 54, Part 2, 2003.Note: In the financial account, a negative sign indicates an excess of recorded outflows: the absence of a sign in the balances indicates an excess of recorded inflows over outflows.

Table 2 also includes the global balance on net errors and omissions.

Liabilities constituting foreign authorities’ reserves. The data in liabilities constituting foreign authorities’ reserves were derived from information collected by the IMF from a sample of large reserve-holding countries. These data were used to adjust portfolio and other investment liabilities to align the data better with corresponding assets series.

Source: IMF, Balance of Payments Statistics Yearbook, Vol. 54, Part 2, 2003.Note: In the financial account, a negative sign indicates an excess of recorded outflows: the absence of a sign in the balances indicates an excess of recorded inflows over outflows.

Table 2 also includes the global balance on net errors and omissions.

Liabilities constituting foreign authorities’ reserves. The data in liabilities constituting foreign authorities’ reserves were derived from information collected by the IMF from a sample of large reserve-holding countries. These data were used to adjust portfolio and other investment liabilities to align the data better with corresponding assets series.

Future analysis and tracking of portfolio investment transactions data may be improved now that the CPIS is being conducted on an annual basis. The results of this position survey, for which data from selected countries are collected, may help in understanding the cause of the discrepancies in the portfolio transactions.

Figure 1.Impact of Revisions in 2002 on Global Current Account Imbalances, 1996-2001

(In billions of U.S.dollars)

Sources: 2002 BOPSY, Volume 53, Part 2 and 2003 BOPSY, Volume 54, Part 2.

Figure 2.Impact of Revisions in 2002 on Global Financial Account Imbalances, 1996-2001

(In billions of U.S.dollars)

Source: 2002 BOPSY, Volume 53, Part 2, and 2003 BOPSY, Volume 54, Part 2.

The discrepancies for other investment are more volatile, changing from positive to negative over the past seven years. The data for other investment, before adjustments for transactions by reserve asset holders, are presented in Tables B-31 to B-33 of Part 2 of the 2003 BOPSY, and include a relatively small discrepancy of -$24 billion in 2002. However, there are large underlying discrepancies for the subcomponents of loans (-$253 billion) and other financial assets and liabilities ($229 billion). The latter includes trade credits, currency and deposits, and miscellaneous items such as capital subscriptions to international nonmonetary institutions.

As noted above, in compiling the global aggregates, the IMF adjusts the data for portfolio investment and other investment liabilities to take account of those liabilities for which counterpart asset transactions are classified as reserve assets.2 Reserves increased substantially from $160 billion in 2001 to $265 billion in 2002, the highest net changes in transactions recorded in the past seven years.

Revisions

Revisions to global data for the 1996-2001 period appear in Figures 1 and 2. The largest revisions for the period, as reported in the 2002 and 2003 editions of the BOPSY, occurred in direct investment in 2001. Data for direct investment abroad transactions were revised upward by $82 billion and data for direct investment in the reporting country transactions were also revised upward by $69 billion. These offsetting revisions reduced the discrepancy on direct investment transactions for 2001 somewhat, but not significantly. However, the revisions from the United States’ Benchmark Survey of U.S. Direct Investment Abroad, 1999, which is undertaken every five years, contributed to a significant reduction in the discrepancy for the financial account in that year. It can be seen that, although the total portfolio investment holdings by residents of the United States fell in 2002, it retained its ranking as the largest direct investor abroad, and the largest recipient of direct investment from foreign countries.

Figure 1 shows that the overall discrepancy for the current account in 2001 was almost identical in the two editions of BOPSY. A closer look at the subcomponents shows that there were revisions to each category, although they were not large. Revisions in the current account tended to increase the discrepancy (except for 1997, when there was a positive current account balance). The financial account is more variable, with revised data increasing the discrepancy in three of the six years.

IV. Work Program Undertaken by the Committee in 2003

The work undertaken by the Committee in 2003 reflected the priorities established in the medium-term work program at the end of 2002. Top priority was given to issues regarding data quality within the DQAF; updating the BPM5; and the concept of residence, specifically with regard to the treatment of nonpermanent workers and of entities without a physical presence. Other high-priority topics included issues regarding portfolio investment (the CPIS, reverse transactions, third party holdings), direct investment (metadata, borderline issues with other types of investment, permanent debt, and the possibility of conducting a coordinated direct investment survey), investment vehicles (such as mutual funds, hedge funds, and other private investment vehicles), statistical treatment of insurance activities, and the relationships between international accounting standards and the concepts in macroeconomic statistics. The Committee reviewed issues relating to global imbalances in balance of payments data; the policy application of balance of payments statistics; the extent to which countries report balance of payments data to the IMF’s Statistics Department on the basis of the classification system of BPM5; and the harmonization of reporting of data by multinational corporations for balance of payments purposes. Inconsistencies in the treatment of returns on various types of equity were also considered, with a particular focus on reinvested earnings. Work undertaken by the BIS on international financial statistics was also reviewed by the Committee.

Other work undertaken during the year included methodological work on the classification for statistics on international trade in services. In addition, the Committee received papers on (i) an update on the way that France is meeting the challenge it faces in measuring its travel account, in light of the introduction of the euro, through the use of an Internet users survey; and (ii) the use of employee stock options in Japan.

Data Quality

Since the financial crises of the 1990s, policymakers and other users of statistics are giving increased attention to the availability of comprehensive, timely, and reliable financial and economic data. In view of the IMF’s responsibility for economic surveillance of its members’ economic policies, and to assist users of the data to evaluate data quality, the IMF’s Statistics Department developed the DQAF in 2001. The DQAF provides a systematic approach and a common language for the assessment of data quality.

The framework comprises a generic assessment framework, as well as specific assessment frameworks for the main aggregates used for macroeconomic analysis and policy, covering monetary and financial, government finance, balance of payments, national accounts, and consumer and producer price statistics. Thus there is an overarching, generic framework for assessing data quality as part of the IMF’s overall surveillance work, prompting more emphasis on accuracy and reliability of statistics.

In light of the increased importance placed on data quality, the Committee has taken a keen interest in issues related to quality. At its 2003 meeting, seven papers on quality-related issues were presented.

Papers were received from three countries (Chile, India, and Japan) on their revisions policies and practices. The paper from Chile noted that although no explicit revision policy exists in Chile, the revision practice is well accepted by users. The main objective of such practice has been to give stability to historical data and at the same time incorporate up-to-date information for the most recent quarters. The paper explained that the primary reason for major revisions is methodological changes. The paper also raised two important issues: (i) how to deal with high-frequency data (such as weekly and monthly dissemination of selected balance of payments components) and (ii) the need to harmonize revision policy with other data sets, particularly with national accounts and external debt.

The Indian paper explained that the main consideration for the choice of revision cycle in India is the time lag for receiving source data. Quarterly data undergo revision until the last quarter of financial year and revisions are not frequent for annual data. In general, no large changes occur after revision as the practice is to use comprehensive data when preparing the data released for the first time.

The paper from Japan distinguished four types of revision (statistical, methodological, definitional, and presentational) and elaborated on the Japanese practices. The paper explained that balance of payments data are revised only once and the IIP data are not revised, as the view is that frequent data revision may impair the credibility of statistics. Relying on comprehensive and timely international transactions reporting systems (ITRS) as the main data source has meant that there has been less need for revisions. The paper noted that Japan intends to review the revision policy, primarily because (i) there have been cases where large-value transactions have been omitted, and (ii) retained earnings are not recorded in the period they are earned. A move toward more frequent revisions is now being considered and issues involved in this process are being reviewed. The Committee found these practical papers interesting and stimulating, noting that there is often a difficult balance between accuracy and timeliness of data, and that users may need to be educated on the reasons for revisions.

The Committee was provided with a paper by IMF staff on revisions policy for official statistics that had been prepared for the International Statistical Institute meeting in Berlin in August 2003. The Committee found the paper to provide a very useful framework for revisions policy. Several suggested improvements were offered, including the need for consistency across related macroeconomic statistical data sets (for example, between the balance of payments and the national accounts, and monetary and financial statistics; between the balance of payments and the IIP; and between the IIP and external debt data) and with socio-demographic statistics (for example, between the national accounts and households surveys) in order to view macroeconomic statistics within a broader context; coordination where statistics are produced by different agencies; emphasis on the absence of political interference, so that pre-release of data to ministers is clearly stated and adhered to; and that causes of revisions, and their contributions, be provided. A range of future revisions was also considered to be valuable (though difficult) as a means of both informing users and defusing criticism of subsequent revisions. Although the Committee emphasized that revisions do not represent errors, it was noted that errors can arise, and that these may be beyond the control of the compiling agency. Under those circumstances, it was suggested that users be informed that revisions may be larger than usual. It was also suggested that options on back casting of newly created series might be included in the document, especially as it was noted that if the compiling statisticians do not make the estimates, users will, and that these data are likely to be weaker. It was also noted by some members of the Committee that it was important to bear materiality in mind, as trivial revisions add little to the analytical value of the data series, while potentially causing users difficulties.

The Committee also received a joint paper from IMF and Eurostat staff on their respective data quality approaches, namely the DQAF of the IMF and the quality definition of Eurostat. The paper pointed out that the two approaches were modified to harmonize common areas. The paper elaborated on the structures, purposes, and uses of both approaches, but noted that, while the IMF DQAF focuses on process-oriented indicators with qualitative measurements, the Eurostat approach focuses on output-oriented indicators with quantitative measures. The two approaches serve similar purposes and complement each other, although the DQAF provides a more holistic approach. The paper proposed that the way forward would be to further exploit the complementarity of the two approaches. The Committee was advised that the Eurostat framework does not pay as much attention to institutional arrangements as does the DQAF because these are set up by regulation in the EU countries, and therefore they do not need to be monitored in this framework. Moreover, as part of the balance of payments regulation, member states are required to make regular quantitative quality reports.

The Committee welcomed this paper and suggested that the IMF and Eurostat might try to bring the two approaches into a single document. However, several members cautioned about placing too much emphasis on quantitative measures, especially as such an emphasis might militate against revisions to data.

The ECB presented a paper to the Committee on the work of a joint ECB-Eurostat Task Force on Quality (TF-QA). The interim report had not yet been approved at the time of the Committee meeting, but had been extensively commented upon by member countries. The approach followed by the TF-QA involves several steps: from the selection of those dimensions/elements contained in the DQAF that were deemed as most relevant for quantitative measurement to identification of possible measures, as well as testing and analysis of indicators and outcomes. The paper noted that the quality dimensions that could, directly or indirectly (through elements), be measured quantitatively are accuracy and reliability, and serviceability. Methodological soundness is assessed in qualitative terms through the “European Union balance of payments and international investment positions statistical methods,” ECB, November 2003 (last update of the so-called “B.o.p. Book”) and by the Accession Countries Manual, ECB, May 2003, which set out the agreed standards and requirements and assesses remaining departures in the countries’ data. Eurostat complemented them with a questionnaire on services. Regarding integrity and accessibility, quantity measurement is generally not appropriate, but the paper noted that these dimensions are well covered by the EU countries. The TF-QA is working to further refine the use and dissemination of quality indicators in reports, in line with the needs of compilers and users and an appropriate balance between quantitative measures and qualitative assessments.

The Committee found that the paper raised some important issues. Some concern was expressed regarding the interpretation of the quantitative information. It was felt that some countries may be making improvements on a continual basis, but that these might not be reflected in the quantitative measures. At the same time, the Committee noted that some quantifications will not affect quality, whereas others will; it was suggested that some nuance might be required. It was also pointed out that one of the assumptions in the paper is that errors are normally distributed, but no evidence was provided to support this assumption. Equally, it was pointed out that where there have been revisions, analysis might be undertaken to explain the reasons. The ECB indicated that refinements will continue to be made. It was pointed out that the participants need experience in interpretation of revisions, but, in time, it was felt that the use of indicators will fall with the rise of the focus on analysis.

The Committee received a paper from Hong Kong SAR that described the experience of the Census and Statistics Department of Hong Kong SAR in applying the DQAF for the assessment of the quality of the balance of payments statistics. The Committee found the paper very interesting, and especially relevant for developing countries that are introducing more surveys because their ITRS are not as viable following the abolition of exchange controls. In this situation, it was noted, there is often interest from a wide range of institutions that want to collect data. In an attempt to overcome “survey fatigue,” statistical agencies may try to combine their requirements into one survey. As a consequence, there are frequently too many respondents. It was noted that countries in this situation need to develop their practices regarding transparency. In addition, it was noted that customs authorities are not often interested in collecting statistics; their focus is revenue collection, with the result that the quality of the c.i.f./f.o.b. adjustment suffers. A similar difficulty is in regard to the collection (and usefulness) of data at market value rather than book value when there is no readily observable market price.

It was explained that in Hong Kong SAR, the Census and Statistics Department identifies the concentration of respondents for direct investment, portfolio investment, and other investment. It was indicated that the marginal improvement in the data for small companies was not cost effective. The importance of both book value and market value information was emphasized. Fluctuations in market value may be very large, so there is benefit in also knowing book values. Regarding the question of conversion of data from a c.i.f. basis to an f.o.b. basis, it was suggested that it might be better for the balance of payments compiler to make the adjustment.

Revision to BPM5

At its 2002 meeting, the Committee considered a draft unannotated outline for revising BPM5. On the basis of this and other related documents, the Committee gave approval, in principle, to the proposal to revise BPM5 and asked the IMF’s Statistics Department to prepare a more detailed, annotated outline. Accordingly, IMF staff prepared a draft AO and circulated it in September 2003 to the Committee and other interested parties, such as the Inter-Secretariat Working Group on National Accounts (ISWGNA), for comment. These comments were taken into account and a revised version was presented to the Committee at its 2003 meeting.

The AO sets out, in considerable detail, a series of proposals for modifying both the structure and some of the concepts in BPM5. The document also asks a number of questions, seeking input on the proposed changes and possible alternatives. The Committee endorsed the IMF staff proposal that the AO be circulated, in early 2004, to all interested parties, including balance of payments compilers, for comment by mid-2004. See Box 1 for some of the major changes proposed in the AO.

Box 1.Some Major Changes Proposed in the Annotated Outline for the Revision of the Balance of Payments Manual, Fifth Edition

  • 1. Increased emphasis on the IIP and Other Changes in Financial Assets and Liabilities Account.
  • 2. Review of the concept of residence, with particular focus on the “12-month rule” and entities without a physical presence (such as special purpose entities (SPEs), trusts, and holding companies).
  • 3. Consideration of increased use of fair value, in light of possible changes in international accounting standards.
  • 4. Refinement of the definition of reserve assets.
  • 5. Consideration of inclusion of untraded securities in Other Investment.
  • 6. Adoption of the new treatment for the measurement of output of other depository corporations and insurance companies, in light of resolution of issues by OECD Task Forces.
  • 7. Treatment of employee stock options.
  • 8. Harmonization of institutional sectorization with the 1993 SNA and the Monetary and Financial Statistics Manual 2000 (with a particular focus on mutual funds and entities without a physical presence).
  • 9. Increased harmonization with the 1993 SNA, particularly in light of its concurrent review.

The Committee also considered proposals from the IMF’s Statistics Department for the process of consultation and review in the revision to BPM5. Included in these proposals was the creation of three technical expert groups, which would provide advice to the Committee and the IMF for input to the new manual. Membership of these technical groups will be drawn from experts in balance of payments and IIP methodology from all over the world. These technical groups will be for (i) issues related to direct investment (DITEG); (ii) issues related to currency unions (CUTEG); and (iii) balance of payments issues, more generally (BOPTEG).

It was agreed that DITEG will work with the OECD’s Workshop on International Investment Statistics (WHS), which has long had a role in direct investment methodology, and has been responsible for three editions of the OECD Benchmark Definition of Foreign Direct Investment (BD). In order to continue this function, the WHS has created the Benchmark Advisory Group to revise the third edition of the BD. Once DITEG’s work has been completed, it is planned that the Advisory Group will use the agreed outcome to prepare a fourth edition of the BD (BD4). DITEG will attempt to conclude its work as soon as possible. It is expected to meet only three times and to conclude its work in early 2005. DITEG will report to both the WIIS and to the Committee. A list of the issues that the Committee agreed should be considered by these expert groups is shown in Box 2.

Box 2.List of Topics Referred by the Committee to the Technical Expert Groups Set Up for the Revision of the BPM5

BOPTEG
TopicPriority1/
1. Activation of guaranteesHigh1
2. Accrual principles for debt arrearsHigh1
3. Nonperforming loans, valuation of loans and deposits, write-offs, valuations of traded loans in balance sheets, and related issuesHigh1
4. Rules for identification of branchesHigh1
5. Muliterritory enterprisesHigh1
6. Institutional sector classificationHigh1
7. Residence of householdsHigh1
8. SPEs and shell and holding companies sector (units, sectorizations, residence, and transactions)High1
9. Residence of corporations (see also SPEs and holding companiesMedium
10. Application of direct investment to government corporationsMedium
11. Untraded and formerly traded securities in portfolio investmentHigh2
12. Traded loans—criteria to become portfolio investmentHigh2
13. Goods for processingHigh1
14. Repairs on goodsMedium
15. Direct investment—reinvested (including negative reinvested earnings)High1
16. Retailed earnings of mutual funds and rerouting of investment income of technical reserves of life insurance enterprises and pension funds
17. Interest on debt securitiesHigh1
18. Income on securities lending and reversible gold transactionsLow
19. Treatment of technical assistanceHigh2
20. Mutual insuranceMedium
21. ReinsuranceHigh1
22. Appendices:
  • -Changes from BPM5
Ongoing
  • -Reconciliation with SNA
High2
  • -Process for amending the revised manual
High2
  • -Risk issues
High2
  • -Exceptional financing
High2
  • -Insurance and reinsurance
High2
  • -Financial services and income
High2
  • -Finance leasing
High2
  • -Standard components
High2
  • -Debt reorganization
High2
  • -Reverse transactions
High2
  • -Tourism
Low
23. Other possible appendicesLow
  • -Data dissemination and quality
High2
  • -Payments for use of assets
Medium
  • -Transactions with IMF
High2
  • -Research agenda
Ongoing
DITEG
TopicPriority1/
1. Valuation of direct investment equityHigh1
2. Valution of branch intangible assetsHigh1
3. Land owned by nonresidentsHigh2
4. Ultimate benefical owner/ultimate destinationHigh2
5. Direct investment—10 percent thresholdHigh2
6. Indirect investment—fully consolidated system, U.S. methodology, or 50 percent ownership.High1
7. Inclusion in investment of transactions between a nonfinancial direct investmentHigh1
8. Reverse investment—classificationHigh1
9. Round trippingHigh2
10. Permenent debt between affiliated financial intermediariesHigh2
11. Use of maturity and full instrument split for direct investmentMedium
12. Merges and acquisitionsHigh 1
13. Direct investment—reinvestment earnings (including negative reinvestment earnings)High1
14. Other possible appendices
  • -Bring together all direct investment issues (stock, flows, and income between affiliates)
High1
  • -Bring together all direct-investment-related issues (transactions in goods and services, income, financial flows, and stocks between affiliates)
Low
CUTEG
TopicPriority1/
1. Issues paper for research related to economic and currency unionsHigh1

Indicates the priority accorded to the topic: High1 indicates that the subject will be considered at the first meeting of the technical expert group, High2 indicates that the topic will be considered at the second meeting of the technical expert group. Should there be time, the other topics will be considered according to the priority indicated.

Indicates the priority accorded to the topic: High1 indicates that the subject will be considered at the first meeting of the technical expert group, High2 indicates that the topic will be considered at the second meeting of the technical expert group. Should there be time, the other topics will be considered according to the priority indicated.

In conjunction with these processes for consultation and discussion by, and within, the balance of payments community will be the link to the process to revise the 1993 SNA. The ISWGNA has created an Advisory Expert Group (AEG), with a similar remit to the roles of the technical expert groups for the balance of payments revision, that is, the AEG will examine proposed changes and report back to the ISWGNA. The IMF’s Statistics Department is represented on both the ISWGNA and the AEG, which will permit ready coordination of effort between the revision to the 1993 SNA and to BPM5.

Residence

The Committee also considered the question of residence, in particular, the treatment of nonpermanent workers within the balance of payments framework. Papers were presented by South Africa, Hong Kong SAR, and India. The paper from South Africa noted that non-permanent workers were an important issue in southern Africa, and the paper suggested that the strict application of the 12-month residence rule may be inappropriate in that region. The paper argued that the “center of economic interest” for many of the workers from Lesotho, Swaziland, and Mozambique in South Africa was their “home” country, not South Africa, even though they may be working in South Africa for much longer than 12 months. This was reflected in the institutional arrangements among the various countries, especially between South Africa and Lesotho and Mozambique. For example, it was noted that taxes that are levied on the incomes of workers from these countries while working in South Africa are payable to their “home” countries, not to the Government of South Africa. The paper from Hong Kong SAR noted that nonpermanent workers are a major element in the Hong Kong SAR’s balance of payments, especially with regard to movements with the People’s Republic of China (in both directions), but also for nationals of other countries who come to Hong Kong SAR for periods frequently well in excess of the usual 12-month rule, but who maintain considerable links with their former residence. For many such cases, the application of the concept of “center of economic interest” is not always easy to establish. The paper argued that the new manual should provide clear guidance on how residence should be determined for such cases. The paper from India echoed many of the concerns presented in the other two papers, especially with regard to Indians working abroad.

The Committee took note of these papers. Several members pointed out that the circumstances described had more general application. The Committee agreed that these issues need to be taken into account when the new balance of payments manual is written.

Portfolio Investment

The Committee was brought up to date on the 2001 and 2002 CPIS, as well as on work being undertaken to enhance the collection of CPIS data. This work has wider application to other elements in the balance of payments and the IIP, as well as other parts of the suite of macro-economic statistics. It covers investigations into how to measure holdings of securities by nonresident custodians (that is, TPH); the results of a questionnaire on the information systems of financial institutions that are involved in reverse transactions (repurchase agreements and securities lending without cash collateral); and the development of the ECB’s CSDB.

The Coordinated Portfolio Investment Survey

The CPIS, which is coordinated by the IMF’s Statistics Department, is an international survey of the holdings of portfolio investment assets. The survey provides data on holdings of equity, and long-term and short-term debt securities, by counterpart jurisdiction of issuer. The survey has been conducted on an annual basis since 2001 (following a limited survey in 1997).

Of the 67 jurisdictions that participated in the 2001 CPIS, all but one have agreed to continue to participate on an annual basis. One particularly noteworthy aspect of the CPIS is that about 20 small economies with international financial centers (SEIFiCs) now take part on a continuing basis. Overall, the willingness of so many countries to participate reflects the benefits that they anticipate will flow from the availability of data on creditor holdings of securities issues.

The data from the survey help to fill an important statistical gap by providing a database that counterpart debtor countries may use to construct estimates of their own outstanding securities liabilities, as well as assisting other data analysts in understanding this fast-growing cross-border exposure. Table 3 and 4 provide, in matrix form, summary results of the 2001 CPIS for the 10 largest holders and issuers of securities. More detailed results can be found on the IMF website (http://www.imf.org/bop). Total securities reported in the 2001 CPIS were $12,613 billion, and, in 2002, $13,731 billion. It can be seen that, although the total portfolio investment holdings by residents of the United States fell in 2002, it retained its ranking as the major investing country.

Table 3.Geographic Breakdown of Total Portfolio Investment: Top Ten Economies by Holders and Issuers, at Year-End 2001(In millions of U.S. dollars)
Investment from:12345678910Total
Investment In:United StatesUnited KingdomJapanLuxembourgGermanyFranceItalySwitzerlandNetherlandsIrelandOthervalue of investment
1 United States308,986490,200177,910108,168116,53074,00176,389137,334155,2811,433,5723,078,371
2 United Kingdom504,318110,35660,30972,96168,60333,73123,06640,78788,650277,0671,279,847
3 Germany131,442130,030111,307150,10186,35359,56656,61268,22132,637333,7201,159,988
4 France141,786130,57764,42259,18774,19143,91126,44545,19118,964170,152774,826
5 Netherlands146,35772,35946,13954,17486,57184,49353,15434,64613,141106,754697,787
6 Italy47,45591,06933,98646,99980,99173,5576,45244,04526,716127,847579,117
7 Japan196,35994,09027,34210,39917,27312,11510,27311,5089,653144,379533,392
8 Luxembourg12,54922,03945,864103,23530,643106,05266,3898,34912,988112,977521,084
9 Cayman Islands58,65945,965132,78520,23012,31921,48917,77611,7722,9756,61167,654398,235
10 Canada200,47415,18421,91011,0544,51310,3892,7375,6602,9194,87628,659308,374
Other809,120393,746232,786213,308238,267201,002148,979178,101124,34171,687670,9233,282,259
Total value of investment2,248,5191,304,0441,289,754820,614791,616710,330552,022495.804485,669441,2043,473,70412,613,281
Table 4.Geographic Breakdown of Total Portfolio Investment: Top Ten Economies by Holders and Issuers, at Year-End 2002(In millions of U.S. dollars)
Investment from:12345678910Total
Investment In:United StatesUnited KingdomJapanLuxembourgGermanyFranceItalySwitzerlandNetherlandsIrelandOthervalue of investment
1United States358,459499,048176,180103,462120,00970,609147,061185,98380,1581,537,4313,278,401
2United Kingdom452,10794,84665,53373,40575,83932,92549,758118,96725,748336,5561,325,683
3Germany94,761128,410121,007171,262113,76356,461106,09545,47861,691422,4591,321,387
4France123,37597,12766,55269,24180,33553,98453,53623,66433,661237,104838,579
5Netherlands124,49384,02841,95962,918104,498106,71254,15020,82443,122162,799805,503
6Italy40,919107,58445,21467,16598,245110,33852,75136,8438,444153,076720,580
7Luxembourg8,72020,56855,799123,97337,432146,7508,26315,50781,969132,520631,501
8Japan193,97586,43822,83217,54516,5208,63712,05411,13010,263138,006517,398
9Cayman Islands56,25351,109175,16624,44718,32826,56313,0024,3687,79815,423108,875501,330
10Spain36,45226,94316,06523,49049,05660,35811,72928,46319,2604,00453,162328,981
Other802,863434,063278,863240,332230,306213,338147,812107,65580,041158,509767,4763,461,258
Total value of investment1,933,9181,394,7291,394,520923,399899,152880,871598,056570,003565,494522,9924,049,46513,730,001

Among the more salient features of the results is the importance that some small economies play in international financial markets. For example, Luxembourg is both a major source and destination of cross-border portfolio investment, while the Cayman Islands is one of the principal destinations.

To assist jurisdictions to undertake the survey, the survey guide3 has been translated into French, Spanish, Arabic, Russian, and Chinese. These documents are available from the IMF on request. The guide sets out the purpose of the survey and the conceptual underpinnings, and offers compilers advice on how to undertake a survey of this nature (such as how to establish a survey frame; how to develop contacts with respondents; software requirements; to address low or nonresponse rates; what data validation procedures might be put in place; follow-ups with respondents; and data release).

The Committee also discussed the usefulness of the CPIS. Several members pointed out how the data are used to analyze the breakdown of their holdings. In addition, given that the geographical breakdown of liabilities are hard to estimate from more standard sources of data, the derived liabilities were felt to be particularly valuable. It was also felt that the data would become increasingly important as the time series develops.

In view of the importance of having a revisions policy (as part of data quality), IMF staff presented the Committee with a proposal for a revisions cycle for the CPIS. The proposal is that the first, preliminary release of the CPIS data for any particular year would be made toward the end of the year following the reference date, that the data would be revised about six months later, and that they would then be revised again at the time of the release of the next year’s preliminary release. The Committee agreed with the proposal.

As part of its ongoing attempts to involve as many jurisdictions as possible in the CPIS, the IMF’s Statistics Department held a workshop for members of the Arab League, the Islamic Republic of Iran, and India, in Beirut, in June. The Statistics Department also seeks to assist SEIFiCs to the maximum extent possible, in view of their limited resource base and their importance in the survey. Accordingly, a workshop for SEIFiCs was also held in Bermuda in May. Funding for the workshops was provided by Government of Japan through the Japan Administered Account for Selected Fund Activities. The workshops were designed to provide compilers with the opportunity to exchange their experiences in conducting the 2001 CPIS; to encourage those who have not yet decided to participate in the CPIS to do so; to see how the survey might be conducted more efficiently and effectively; to establish a network of contacts among the compilers; and to indicate what plans they had for the 2002 survey.

In order to obtain as comprehensive a picture as possible of cross-border asset holdings of securities, and so that counterpart liabilities can be constructed, comparable information to the CPIS is necessary on securities held as reserve assets and securities assets of international organizations, as these assets, by definition, are not included in countries’ holdings of portfolio investment assets. To this end, the IMF’s Statistics Department conducts two annual surveys, Survey of Securities Held as Foreign Exchange Reserves (SEFER) and Survey of Securities Held by International Organizations (SSIO). The results of these surveys are available on the IMF’s website (http://www.imf.org/bop).

In line with the IMF Statistics Department’s view that all data should be supported by metadata (to explain to users such elements of the survey as the methodology employed, the survey frame, the response rate, and other related issues), all jurisdictions that participated in the CPIS were asked to complete a metadata questionnaire. Sixty-four jurisdictions did so. The results of this questionnaire have also been placed on the IMF website. The information will play an important role in the analysis of the results and in the exchange of information between and among the participating countries and will be valuable for users and compilers. Knowledge of the metadata is an integral part of analyzing the data.

Third Party Holdings

At its 2003 meeting, the Committee considered a report from the Technical Group on Third Party Holdings (TGTPH). This group had been set up by the Committee in 2000, with a remit to examine how TPH might be captured in official statistics. TPH are securities that have been placed with a custodian, resident in one jurisdiction, directly by end-investors resident in another jurisdiction. It is likely that there are many instances of TPH, especially by households or small or medium-sized enterprises (SMEs), involving in total substantial amounts of securities. As such holdings lie outside the scope of balance of payments and IIP statistics of the economy of the custodians’ jurisdiction, and as households and SMEs are usually not covered in surveys of end-investors, these holdings are likely to be missed in countries’ IIPs and their CPISs.

In its initial work, the working group on TPH found that (i) there were many different types of business models used by custodians, so that there was no simple approach that would permit these holdings to be readily identified;

(ii) there were often multiple custodian chains, so that there were major risks of double counting; and, as a result

(iii) the focus of any further work should be on high-worth individuals, as in this area there would likely be less risk of double counting and custodian records might be more readily usable for statistical purposes. Following this earlier work, the TGTPH held discussions in 2003 with Euroclear (in Brussels) and Clearstream (in Luxembourg) to determine the extent to which these bodies, as depositories, could identify the ultimate beneficial owners (UBO). Except in a few, very limited, instances, neither of these bodies were able to do so.

In light of this outcome, the Committee agreed to the proposal by the TGTPH that the next step explore how holdings on TPH might be identified by those few jurisdictions with major international custodial business. In view of the highly concentrated nature of the custodial industry, it is expected that a very high proportion of the TPH of high worth individuals could be captured from custodians in only a few jurisdictions. To that end, the ECB offered to host such a meeting, which offer was accepted. The TGTPH will prepare for consideration at that meeting an issues paper, together with a draft questionnaire that might serve as the basis for further discussions with custodians in the jurisdictions that agree to participate in such a survey. Should such a survey proceed, the first step would be to find out the types of business models that custodians use, in order to avoid any double counting and to minimize the respondent burden. There are still several problems that have to be addressed before any data on the value of TPH cross-border holdings by high-worth individuals might be available, including the lack of clarity in some countries as to whether they have the legal authority to request the information.

Reverse Transactions

The Committee has discussed reverse transactions on many occasions, reflecting their importance, their complexity, and the difficulty in measuring them. Reverse transactions (repurchase agreements, securities lending without cash collateral, gold swaps, and gold loans or deposits) have grown rapidly in volume and complexity in recent years, as financial markets round the world have come to use them much more extensively and effectively. The nature of reverse transactions has also evolved since BPM5 was published. As a result, the Committee, at its 2001 meeting, took a number of decisions in principle: (i) to continue the current recommended treatment of recording repurchase agreements (involving cash collateral) and gold swaps as collateralized loans and to assess further the availability of supplementary information on the sector of the counterparty and the issuer of the security; (ii) not to record a transaction at all for securities lending (without cash collateral) and gold loans/deposits, but instead to provide supplementary information similar to that recommended for repurchase agreements and gold swaps; and (iii) in the event that the asset acquired under a reverse transaction is on-sold outright, the seller of that asset so acquired should record a negative asset (that is, a “short” position).4

Even so, the Committee felt that it was necessary to pursue the extent to which the supplementary information (on sector of the counterparty and the issuer of the security for repurchase agreements and securities lending) was available, and to determine how to classify the payment for the use of the lent asset in securities lending and gold loans/deposits. To address these issues, the technical group on reverse transactions (TGRT) was set up. The group comprised representatives of nine economies (Belgium, Hong Kong SAR, Japan, Portugal, South Africa, Spain, Switzerland, the United Kingdom, and the United States), the IMF, the ECB, and the BIS. The technical group’s mandate was to explore how the reporting of reverse transactions can be improved (for monetary statistics, for flow of funds accounts and for sectoral balance sheets, as well as for balance of payments and IIP statistics), bearing in mind the Committee’s decision on the appropriate treatment.

As a first step, in 2002 the TGRT undertook a survey of a selection of end-investors, custodians, brokers/dealers, and fund managers in the economies of the participants in the technical group, to determine whether these institutions’ information systems could readily identify when reverse transactions take place. The results of this survey were presented to the Committee at its meeting in 2003. Among the important results were: (i) the introduction in the United States of a new reporting system, by financial institutions, that allows the identification of reverse transactions; and (ii) a significant minority of respondents to the survey indicated that they record repurchase agreements as both a collateralized loan and a transaction in the underlying security (the so-called “four way reporting”). These are important developments as they open routes for improved measurement of reverse transactions. The new system in the United States is not compulsory (it is not a government sponsored exercise), but most of the custodians in New York use the software. At present, Treasury securities are not covered, but that will change in 2004. The Committee requested that the TGRT explore with the authorities in the United States whether this software might be available for use in other countries. At the same time, a significant minority of respondents to the survey indicated that they record reverse transactions as both a loan and a transaction in the underlying asset, which may mean that this approach could become more widespread and result in improved measurement.

As to the appropriate treatment of the fee on securities lending and gold loans, the TGRT was unable to come to a consensus. The issue will be explored more fully during the revision to the BPM5.

In its work on how reverse transactions are recorded, the TGRT examined the role of central clearing houses (CCHs) in tripartite repurchase agreements. These institutions are becoming increasingly important and may have a major impact not only on how the data are collected but also on what the data represent. At present, there is no standard reporting practice, especially for gross versus net reporting. Moreover, as these clearing houses are not other depository institutions, transactions by banks with other banks but channeled through the clearing houses, will have an impact on monetary aggregates. The Committee requested that the TGRT investigate CCHs in more detail.

Global Securities Databases

Several participants in the CPIS use a security-by-security approach to obtaining the information of portfolio investment asset holders. This approach can provide very high-quality data, provided that there are other sources of information against which the information can be checked. One of the major ingredients for high-quality data, using a security-by-security approach, is a securities database.

From previous work on the CPIS, it has become evident to the Committee that there is considerable interest in the development of a global securities database (GSDB). A GSDB, as so envisaged, would be a multidimensional database, with information on such variables as the name, sector, and jurisdiction of the issuer of each security; the amount issued and its date of issue; the currency and coupon (if any) of the issue; the maturity date(s); and, possibly, the sector of holder of the security. Price information would also be included. Such a GSDB has a potentially very wide application: in addition to the CPIS, it could be used for the construction of estimates of portfolio investment transactions in the balance of payments; for external debt estimates; for the international reserves and foreign currency liquidity template; for monetary statistics; for flow of funds tables; for the sectoral balance sheets in the national accounts; and for financial soundness indicators.

In response to its own needs and those of its members, the ECB has undertaken a comprehensive exercise to create its CSDB, which has many of the features of a GSDB. The Committee was advised of the progress that the ECB has made in this exercise. There are to be three “feeders,” viz. (i) commercial providers, including national numbering agencies; (ii) the internal sources of the European System of Central Banks (ESCB), including the member states’ national central banks’ databases; and other sources, such as the BIS, Euroclear, and Clearstream. The Committee welcomed these developments and requested that it be kept informed. It is anticipated that the CSDB will be operational about the end of 2004.

In addition to the ECB’s database, the BIS has maintained a securities database for several years, primarily for international securities but increasingly for domestic securities as well. The Committee was advised that the coverage of international securities is good (with data coming from commercial providers) but that the coverage of national securities on issue is currently limited to 42 countries. Data quality on securities has improved as the frequency of reconciliation of data from commercial providers is now on a quarterly basis. Sector classification has also improved, with public and private splits now available for banks, nonbank financial institutions, and other businesses. Use of commercial data providers on domestic issues is being assessed, and a decision was expected by the end of 2003. Time series on international securities have been improved, having been carried back to 1966.

International Financial Statistics

The Committee was informed that the BIS continues to increase the number of participating central banks in its international banking statistics: by the end of 2003, Greece and Mexico began participating in the statistics on international bank activity on a locational basis and Australia, Greece, and Mexico joined the statistics on country risk exposures on a consolidated basis (Australia already participated in the locational statistics). It was noted that data quality has also improved, with data on loans, securities, and other assets and liabilities being provided separately in the locational statistics by nearly all participants, and information on currency and country breakdowns in the locational statistics, and on an ultimate risk basis in the consolidated statistics having generally improved. Most participating central banks submit their data to the BIS within 12-13 weeks. As of end-2004, the BIS will be collecting enhanced consolidated banking data that will cover more detailed and comprehensive data on country risk exposures on an ultimate risk basis including exposures from derivative instruments.

For data on foreign exchange and derivatives markets, the Committee was advised that a record 52 countries will participate in the next BIS triennial survey, which will be conducted in April 2004 and will cover turnover data on foreign exchange and OTC derivatives market activity on a locational basis and at end-June 2004 covering positions outstanding on OTC derivatives market activity on a consolidated basis. For the April part of the survey, nondeliverable forwards (NDFs) will be identified separately. The BIS also collects and publishes statistics on most OTC financial derivatives on a semiannual basis, covering stock data on notional and market values from approximately 60 major dealers in the Group of Ten countries on a consolidated basis. In addition, the BIS compiles and disseminates quarterly data on turnover and amounts outstanding for exchange-traded derivatives. Data quality in the forthcoming triennial survey is expected to improve by the following: (i) the definition of “dealers” and “in-house” or related-party deals has been clarified; (ii) data will be collected consistently on a sales desk basis, as opposed to a trading desk basis; and (iii) there will be greater use of concrete examples for data reporting in the guidelines to the survey. It is hoped that the data will be reported earlier than in the previous survey, so that turnover data can be disseminated a month earlier in September and amounts outstanding data a month earlier in November 2004. The Committee was told that an increase in the frequency of the release of the semiannual OTC derivatives statistics was not possible for the time being.

Direct Investment and Direct-Investment-Related Issues

The Committee gave consideration to a number of issues related to direct investment, including borderline issues, investment vehicles, the SIMSDI, a proposal for a feasibility study of a possible internationally coordinated survey of direct investment position, that is, a CDIS, and harmonized reporting by multinational corporations.

Borderline Issues

At the Committee’s 2002 meeting, the United States had presented a paper that raised several points on borderline issues, to which the Committee had been asked to provide written comments. At the 2003 meeting, a summary of those responses was presented. There was general support for the position of the United States on the treatment of mutual funds, that, in the main, they should not be classified as direct investors, though there may be circumstances where they might be considered direct investment enterprises. However, Committee comments showed the need for a definition of what constitutes a mutual fund, and how mutual funds are distinguishable from other types of collective investment schemes, such as trusts, and other similar investment vehicles (such as real estate investment trusts). As to the treatment of trusts by the United States, there was no disagreement with the U.S. position. Regarding permanent debt, there was no clear view on what might be done, but, as with mutual funds, there was a general view that permanent debt needed to be defined more clearly. The Committee responses did not support the practice in the United States regarding the specific exemption from direct investment of transactions and positions between nonfinancial parents in the United States and their financial affiliates in the Netherlands Antilles. As to shell companies, there was no consensus, but several responses noted that the U.S. approach would lead to global asymmetries. There was general support of the approach by the United States regarding the treatment of agents, and support for the new manual to be more precise on this topic. The responses also supported the approach in the United States to the treatment of corporate inversions, and that the existing treatment of financial derivatives is appropriate.

Two specific issues were raised in the response of the United Kingdom. The first is with regard to a corporation with dual residence, that is, where it is incorporated in the United Kingdom and in another country: all the transactions, assets, and liabilities are allocated in a fixed proportion between the two countries. The second issue was with respect to “toll processing” within a corporate family. Goods are sent from the parent to a foreign affiliate for processing, and then returned without a change of ownership. A processing service charge is recorded, rather than exports and imports of merchandise, which is what BPM5 recommends.

In light of these responses, the Committee indicated support for a proposal that the new manual should include a recommendation for a supplementary ownership-based presentation. It also discussed whether the current 10 percent rule for the establishment of a direct investment relationship should be revisited in light of the international accounting standard that uses 20 percent. Opinions were mixed, ranging from emphatic support for the proposal to an equally strong disinclination to reopen the issue. Given this diversity of views, it was decided that there was a need to consider the issue further. It was also proposed that the IMF’s Statistics Department take up this issue with the WIIS.

Following on from the paper that was presented in 2002, the United States presented another paper on issues related to the borderlines between direct and portfolio investment. The paper explores issues related to whether mutual funds with “feeder/master” relationships and some types of hedge funds (especially those involving limited partnerships) might be considered to have direct investment relationships. The paper also raises questions about the appropriate classification of holding companies with cross-border relationships. It noted that, whereas, in the Standard Industrial Classification (SIC), trusts are treated as financial entities, and estates are not, in the North American Industry Classification System (NAICS), trusts and estates are both treated as financial entities. Moreover, in the SIC shell companies are not explicitly mentioned, but in NAICS they are not classified as financial entities. In the United States, all three types have been treated as financial entities, but, beginning in 2003, with the adoption of NAICS, the United States treats trusts and estates as SPEs, but not shell companies. The paper also discussed the question of the treatment of insurance intercompany/branch claims, noting that these should not be considered to be direct investment, if they represent technical reserves. On the other hand, the paper argued that private equity funds in which the ownership share meets the 10 percent threshold should be treated as direct investment.

The Committee found that the paper raised some very important questions. It was of the view that holding companies that hold only domestic companies be classified on the basis of the activities of their principal operating companies. However, the Committee was not clear whether this same principle should be applied in the case where the holding companies hold only foreign operating companies.

Treating SPEs as financial or nonfinancial entities would affect the treatment of transactions between them (as to whether or not nonequity and nonpermanent debt transactions were recorded as direct investment). The Committee felt that, for some uses, there were those who preferred to “look through” SPEs (though this is likely to lead to global imbalances as different compilers will treat them differently), while it was noted that they perform some function, which might suggest that they be shown in a supplementary presentation. As to the treatment of positions on insurance technical reserves of policyholders between related entities, the Committee felt that BPM5 clearly indicated that these were not direct investment. The Committee agreed that all these issues needed to be addressed in the new balance of payments manual.

Survey of Methodological Standards for Direct Investment

At the end of October 2003, the IMF published Foreign Direct Investment Statistics: How Countries Measure FDI, 2001, which contained the results of the joint IMF/OECD update of the 1997 SIMSDI.5 The 2001 SIMSDI update was restricted to 61 countries—a subset of the 114 countries that participated in the 1997 SIMSDI—and covered the 30 OECD countries and 31 other countries, including all subscribers to the Special Data Dissemination Standards.

The report on the results of the 2001 update showed that there have been marked improvements since 1997 in both the availability of direct investment statistics, particularly in position data, and in the application of a number of the recommendations set out in BPM5 and BD3.6 However, it also showed that there are gaps in the available data and there remain a number of important areas where the majority of countries surveyed have not yet implemented the international recommendations. The report provided a wealth of information on country compilation and dissemination practices and how these practices compared across countries and against the recommendations set out in the international statistical manuals.

At the request of the IMF’s Executive Board for a seminar on the topic of foreign direct investment, the IMF’s Statistics Department prepared a paper, “Foreign Direct Investment Trends and Statistics” (SM/03/363, October 28, 2003), and a summary paper (SM/03/363, Supplement 1, October 28, 2003). These papers drew heavily on the SIMSDI results.

The IMF and OECD are launching another joint SIMSDI survey in late 2003. The questionnaire for the 2003 SIMSDI has been modified to take into account the changes in the methodology introduced since 1997, and will be sent to significantly more countries than the 61 covered by the 2001 SIMSDI update. As with the 2001 update, summary metadata drafted in consultation with national compilers and cross-country comparison tables will be prepared and are expected to be posted on the IMF’s website in late 2004. A joint IMF/OECD report will also be prepared and is expected to be published in 2005. At the seminar, the IMF’s Executive Board supported continuing the SIMSDI exercise and extending it to cover a larger segment of the IMF’s membership.

Coordinated Direct Investment Survey

The Committee’s views were sought on a proposal that the IMF’s Statistics Department had taken to the IMF’s Executive Board to undertake a feasibility study as a possible precursor to a CDIS. Such a survey would build on the successes of the CPIS and SIMSDI. The proposal is for the feasibility study to be conducted in 2004 and early 2005, and, if the results indicate that a full survey can be undertaken, that it would be conducted, at the earliest, in 2007. The proposal is for benchmark positions data, broken down by counterpart jurisdiction. The Statistics Department had advised the Executive Board of the potential value of a CDIS, in the light of the key importance of direct investment data and the discrepancies in the existing data. It also advised the Board of the complexities of such an undertaking. The Department had stressed the importance of conducting a feasibility study before embarking on a much more ambitious exercise than either the CPIS or SIMSDI. The Board had supported the idea of a feasibility study, even while recognizing the difficulties of a fully fledged CDIS. It was noted that the feasibility study and the work on direct investment for the new manual would be undertaken in tandem. Work would need to be done on various methodological issues before the full survey could be carried out. The feasibility study would be carried out in conjunction with the OECD and other international organizations with an interest in direct investment statistics, with the work being undertaken at the same time as the DITEG, so that the CDIS, should it proceed, would be based on the new methodological standards in the new balance of payments manual and BD4.

The Committee generally supported the proposal, but recognized the difficulties that a CDIS would involve. Among these, the Committee felt that a common basis for valuing direct investment positions would be the most problematic.

The Committee was informed of progress on a joint project between the ECB and Eurostat on the harmonization of reporting by multinational corporations (MNCs). While offering the prospect of considerable synergies to reduce respondent burden, the process still required considerable further work. Most of the reactions from MNCs to an initial approach by the two agencies had indicated a general level of support, but, for many of the MNCs, much of the information that would be requested under a Uniform Reporting Model was not readily available and would require major investment in information technology. The Committee was advised that, if harmonized reporting for MNCs were to be adopted in the EU, there would be possible benefits for non-EU countries, leading to a unified reporting format for all national compilers.

Income

Following on discussions on income at its meeting in 2002, the Committee received several papers on the subject at its 2003 meeting. A paper from Canada noted that, during recessions, dividends have been paid in excess of the enterprise’s earnings. While this would, without the imputation of reinvested earnings, have led to a deterioration in the net lending of the enterprises concerned, with the imputation of reinvested earnings, such a deterioration is not observed. The Canadian paper argued that the imputation of reinvested earnings series tended to be counterintuitive. The paper also indicated there also are concerns in Canada about inconsistencies between the imputation of reinvested earnings in balance of payments statistics and the absence of a similar imputation between similar ownership relationships in resident: resident relationships where there is an equivalent ownership level (that is, 10 percent or more of ordinary shares or voting power). The paper also noted that, for the accounts for the rest of the world in the national accounts, reinvested earnings are excluded, with a reconciliation statement provided as a link to the balance of payments. At Statistics Canada, there is concern about imputing a flow when there has not been a transaction.

Related to the paper from Canada was one from IMF staff that elaborated on a similar paper presented to the Committee at its 2002 meeting, identifying apparent inconsistencies in the treatment of returns on holdings in equity. Four alternatives were proposed in the paper to address some of these inconsistencies: (i) treat dividends payable as the only distribution of earnings of corporations and unincorporated entities (thereby eliminating reinvested earnings and the imputation of premium supplements on the technical reserves of life insurance companies and pension funds); (ii) retain the present treatment, but extend to situations where one resident entity holds 10 percent or more in another resident entity (thereby creating symmetry between resident: nonresident relationships and resident: resident relationships); (iii) reroute all saving of corporations and unincorporated entities to the ultimate beneficial owners (thereby reducing all saving by the financial and nonfinancial business sectors to zero); and (iv) regard the payment of dividends as a withdrawal of equity (thereby eliminating any income on equity).

The Committee found that the paper raises fundamental questions on income and saving, issues that also need to be taken up in other forums (such as the review of the 1993 SNA and Eurostat’s European System of Accounts 1995 (ESA95). Given that the review of the 1993 SNA is not intended to involve major changes, it is very likely that a major project of this nature may lie beyond the purview of the present exercise. Even so, the Committee thought that it might be worthwhile raising the issues with national accountants to see whether they would be prepared to consider the issue in this current review of the 1993 SNA.

The Committee also received a paper from Japan on employee stock options, which also discussed the results of a survey conducted by the Bank of Japan on the use of stock options, payable to residents of Japan by nonresident corporations. Such arrangements are quite common, and have a long history (going back at least 20 years in some cases). The paper also pursued some of the classification issues, suggested that stock options be treated as financial derivatives, and recommended that the question of the appropriate treatment of stock options be considered in the revision to the 1993 SNA as well as the new balance of payments manual. The Committee welcomed the paper, found it an interesting issue, and agreed that the treatment of stock options is a matter for the new balance of payments manual and the revision to the 1993 SNA.

The Committee also was given papers by South Africa, the ECB, and the United Kingdom, on income-related topics. The paper from South Africa discussed the classification of mutual funds and the treatment of their earnings, and indicated that there is a case for showing share/units in mutual funds as a separate subclassification within equities, given their distinct nature and their growing cross-border importance. The paper also supported the position of the United States that mutual funds should not be direct investors under any circumstances, but that it is possible that they could be direct investment enterprises. The paper from the ECB noted that the approach by ESA95 (that all saving by mutual funds should be deemed to be distributed to their shareholders, and then the portion that is not distributed as dividends is shown as being reinvested back in the mutual funds through the financial account) ensures consistency between creditor and debtor. However, the paper recognized that the approach in ESA95 does not apply equally for interest of collective investment schemes and their income on equity. The United Kingdom provided the Committee with a presentation of what the United Kingdom provides to a group of its users, with regard to income in the balance of payments. It was noted that the issue of negative reinvested earnings had been discussed with this group, some members of whom doubt the usefulness of the artificiality introduced by these imputations. The Committee found these presentations very interesting and valuable input for the new manual.

The ECB also provided the Committee with a paper on the outcome of work that had been done comparing results on income credits and debits on portfolio investment between counterpart members of the EU. The paper pointed out that there are large differences, especially during periods of large changes in interest rates. The paper noted that there are different approaches to the calculation of income flows in different member states: from the use of market indicators for calculating accruals on a purely aggregate basis, to using aggregate data supplemented by security-by-security data, to pure security-by-security, and these different approaches may contribute to the differences in the results. The paper noted that, for example, the use of an aggregate approach, even when using the same conceptual basis, may produce different results from a security-by-security approach, as the weights of nonresident holdings is unlikely to be equal across all securities. The application of market rates may obscure these different weights. The paper also compared the advantages of these two approaches and concluded that the security-by-security approach, taken together with a central securities database, will tend to produce more accurate results. This approach, however, has to be set against the costs of setting up and maintaining a system to cope with the volume of transactions.

Uses of Balance of Payments and International Investment Position Statistics

The Committee was presented with a paper by Uganda on the uses of balance of payments and IIP statistics for highly indebted sub-Saharan African countries, using Uganda as an example. The paper gave a wide ranging description of the value of these data. It noted the role of the link between the balance of payments and the national accounts, in particular, with regard to public versus private sector saving, and how initially any shortfall in saving is made up by borrowing from the rest of the world. In the short term, this will lead to an appreciation of the domestic currency, but as debt levels rise, this appreciation will reverse itself. The paper noted that the form of financial inflows was important, indicating that direct investment in entities involved in the production of traded commodities was likely to be the most sustainable. The paper also drew the link between the balance of payments and monetary policy, especially when associated with a sharp increase in government borrowing that has been financed from abroad. This development prompts the need to sterilize this increase in liquidity domestically with sales of treasury bills to nonresidents. However, the paper notes that such a policy is not sustainable in the longer term.

The paper also raised data collection and quality consideration in an economy that has been liberalized quickly. In the Ugandan case, this liberalization had led to the removal of many restrictions, but, in so liberalizing, many of the data sources for the balance of payments and the IIP were lost. Uganda is endeavoring to recover these data sources, but it has proved very costly and a loss of time series.

The paper was very well received by the Committee, recognizing the importance that balance of payments compilers need to place on uses of their statistics, and the link to other data sets (such as the national accounts and the monetary and financial statistics). The Committee felt that the paper should serve as an important input on the chapter on uses in the new balance of payments manual, together with the South African paper that was presented to the Committee at its 2002 meeting.

Reporting Under BPM5

The Committee reviewed the progress countries were making in reporting balance of payments and IIP data to the IMF’s Statistics Department on the basis of the classification system of BPM5, as well as the use of electronic reporting. For the 2003 Balance of Payments Statistics Yearbook, 157 countries reported balance of payments data using the coding system of BPM5 (compared with 151 in 2002). One hundred and fifty-eight countries reported in electronic form (compared with 144 in 2002) Twenty countries are now reporting using the IMF Internet Correspondent System. One hundred countries now report quarterly balance of payments statistics to the IMF. A growing number of countries are reporting IIP data: for 2003, there were 83 comprehensive reporters, and a further 12 that provided partial IIP data.

Other Issues

The Committee also received papers on other issues, with potentially far-reaching implications for the compilation of balance of payments and IIP statistics, and the development of the new manual, on the links between international accounting standards and macroeconomic statistics; the use of an Internet survey for estimates of business travel estimates in France; and security measures for Internet reporting in Japan.

Two papers (one from the ECB, one from Eurostat) were presented on international accounting standards, and their links to macroeconomic statistics, specifically with regard to the International Accounting Standard (IAS) #39 (“The Recognition and Derecognition of Assets”). The Committee was informed that the European Commission has decided that for all listed companies the international accounting standards will become compulsory at the consolidated level, and that several member states have also required that individual entities will do so as well at the national level. As a result of these decisions, and because the IAS are close to the international statistical standards, there should be an increase in data quality at the national level. This was especially important for direct investment, for which EU member states use different accounting standards, and also given that EU member states are moving away from bank reporting systems to direct surveys to measure direct investment. Reinvested earnings would also likely be reported on a basis more in line with the current operating performing concept, as the IAS, if properly implemented, would also permit the separation of current operations from other holding gains and losses, for example. However, to the extent that some (unlisted) enterprises do not adopt the IAS there will be some inconsistencies in reporting compared with those that do. The Committee was also advised that the IAS is apparently moving to full application of market price principles for all assets and liabilities, including deposits and loans, which would be a major change from BPM5, which recommends the use of nominal value. To the extent that this change occurs, it would mean that the creditor approach to the measurement of interest would become more widely used. It was pointed out that the IAS for insurance, especially for reinsurance, would assist the implementation of reporting for balance of payments purposes.

The Committee also recognized the potential impact that the adoption of fair value accounting would have on Basel II capital reserve requirements, and the associated implication for financial soundness indicators. It was noted that this issue causes tension between those users interested in market valuations (or proxies thereto) and those who feel that nominal value is a better principle for deriving debt-service ratios and sustainability. It was also pointed out that the same issue arises with regard to tradable instruments for both market prices and nominal values. It was noted that, while there is concern about estimation (as evidenced by provisioning, for example), fair value is considered to be more realistic, and the impact on capital and reserves would be recognized incrementally. It was noted that nominal value adjusted for loan loss provisions might be closer to realizable value. Overall, the Committee saw the potential advantages from the general adoption of the IAS. Although not much impact can be expected soon, as these standards will be adopted progressively, there were some concerns on conceptual grounds about the adoption of fair value. It was noted that the impact on the data of the introduction of IAS would generally be negative in the short-to-medium term, but that there appeared to be genuine longer-term benefits.

The Committee was informed that new sources of information for travel in France had been operational since April 2003, and were consistent with the previous data. The Committee was advised that one of these new data sources is a survey of Internet users. Up to 10,000 respondents are contacted each month to a filter question about foreign travel; a maximum of 500 respondents are identified. These are then asked further questions about their foreign travel. It was noted that business travel constitutes an important part of France’s travel debits, and that 12 percent of all overnight stays outside France is for business reasons. It was also observed that two-thirds of business travel is outside the euro area, whereas personal travel by residents of France are more often within the euro area. Average business travel costs are substantially higher than personal trips, up to six times more. The Committee found this development to be very interesting, while recognizing that there were still some measurement issues that might need to be addressed.

The Committee received a paper on the security measures that the Bank of Japan is employing to secure information in an open network for electronic procedures on reporting data for balance of payments purposes. The procedures provide: (i) Bank of Japan server authentication; (ii) user authentication; (iii) data encryption; (iv) prevention of illegal access to the Bank of Japan server; (v) access control for electronic files at the Bank of Japan;(vi) access logging; (vii) virus checking; and (viii) backup. It was pointed out that there are many firewalls in sequence in the system, to maximize security and to ensure that the intranet system is secure. It was noted that many of the features in the new Bank of Japan system were modeled on those of the Bureau of Economic Analysis in the United States. Several Committee members thought that this paper had many useful features that could be considered for their own development of electronic reporting.

V. Future Work Program

Appendix 3 sets out in detail the medium-term work program agreed to by the Committee in 2003. Subjects are ranked by priority. The rankings are not intended to reflect the absolute importance of each topic but rather to reflect the relative priority assigned to each topic by the Committee, given the limited time and resources available for research and investigation.

Top priority for the Committee is (i) the updating of BPM5, (ii) direct investment issues, and (iii) data quality. For the updating of BPM5, the IMF staff will (a) prepare and disseminate the final version of the Annotated Outline to interested parties, especially balance of payments compilers, (b) draft several chapters of the new manual, and (c) prepare methodological documents for consideration by the Committee and technical expert groups. For issues relating to direct investment, IMF staff will provide progress reports on (i) the feasibility study to conducting a coordinated survey on direct investment positions, and (ii) the 2003 SIMSDI. As to data quality, Committee members will provide comments on the IMF staff paper on revisions policy for official statistics, and on the ECB’s quantitative indicators for assessing data quality that were presented to the 2003 meeting of the Committee.

High priority is accorded to portfolio investment, investment vehicles, reverse transactions, and heavily indebted poor countries. For portfolio investment, IMF staff will report on the 2002 and 2003 CPISs, the TGTPH will report back to the Committee on progress in developing an approach to the measurement of third party holdings in major financial centers, and the ECB will report on progress on the implementation of its CSDB. Members of the Committee have been asked to provide comments on the paper from the United States on investment vehicles. For reverse transactions, the TGRT will provide progress reports on (i) the extent to which the software used by financial institutions in the United States could be adopted by financial institutions in other countries; and (ii) provide a paper on the “four transactions” approach to measuring reverse transactions, including worked examples. Uganda and IMF staff will prepare a paper on how debt forgiveness, under the Heavily Indebted Poor Countries (HIPC) Initiative, is treated in the balance of payments and the IIP.

The Committee gave medium priority to global imbalances, implementation of BPM5, central clearing houses, international trade in services, external debt and the IIP, international financial statistics, financial derivatives, reporting systems, and data processing. IMF staff will prepare a paper that highlights imbalances, at a global level, in balance of payments statistics, and the TGRT will provide a report on central clearing houses. IMF staff will provide the Committee with a report on the activities of the Inter-Agency Task Force on Statistics on International Trade in Services. There will also be a report on the activities of Inter-Agency Task Force on Finance Statistics. The BIS will report on developments in international financial statistics, and Japan and France will present papers on changes to their reporting systems. Australia will provide a report on the restructuring and reengineering of its economic statistics program, while the United States will inform the Committee on the introduction of its survey on cross-border holdings and transactions in financial derivatives.

A paper by IMF staff on “short” positions was accorded a low priority.

VI. 2004 Meeting

The next meeting will be held at the South Africa Reserve Bank in Pretoria, in the week of October 25, 2004.

    Other Resources Citing This Publication