7. Selected IMF and Other International Organizations’ Data Collection Initiatives
- Eduardo Valdivia-Velarde, and Tamara Razin
- Published Date:
- December 2014
7.1 This chapter discusses data sources provided by the International Monetary Fund (IMF) and other international organizations that may be relevant for balance of payments and international investment position (IIP) compilation. The main feature of the data sources discussed in “IMF’s Bilateral Data Collections” and “Bilateral Data from Other International Organizations” is that they provide detailed geographical breakdown that may be used by the counterpart economy for balance of payments and IIP statistical purposes. “Other Datasets” describes other datasets that should be consistent with data reported in the international accounts.
7.2 Data compiled directly by partner economies can also be relevant; however, their relevance depends on specific circumstances—for example, the coverage of data in partner economy’s statistics, the methodology applied by the partner economy, the accessibility to the data, confidentiality constraints, and so forth. The compiler should assess these factors on a case-by-case basis.
7.3 Datasets described in this chapter are maintained by the IMF, the Bank for International Settlements (BIS), the Organisation for Economic Co-operation and Development (OECD), and the World Bank, and relate to portfolio investment, direct investment, loans, deposits, development assistance, and trade in goods.
7.4 Initiatives that collect data by individual counterpart economy (bilateral data) include the IMF’s Coordinated Portfolio Investment Survey (CPIS) and Coordinated Direct Investment Survey (CDIS), the BIS international banking statistics (IBS), and the OECD development assistance data.
7.5 Data on partner economies compiled by international organizations can be used in two ways:
In a bilateral reconciliation exercise as a check of the compiler’s estimates; in recent years, there has been an increased use of cross-economy data comparisons to enhance the quality of individual economy statistics.
Directly in the compiling economy statistics in the absence of national data or to complement existing data sources; in this case, the compiler should make the necessary adjustments described ahead.
7.6 A bilateral reconciliation exercise of balance of payments or IIP statistics involves comparison of data that are provided by the reporting economy’s partner economies to international organizations with the reporting economy’s own data and that purport to measure the same set of transactions or positions. For example, economy A’s estimate of the value of direct investment in economy B could be compared with economy B’s estimate of direct investment from economy A disseminated by the IMF in the CDIS. As discrepancies are identified and explained, the compiler can improve the quality of the balance of payments or IIP data.
7.7 Direct use of data compiled by international organizations in the compiling economy’s balance of payments and IIP statistics can be relevant when the compiling economy does not collect the data itself or when collected data are partial or incomplete. For example, if economy A does not collect portfolio liabilities, then the CPIS can be used to estimate a liability position. Before using bilateral data it is important to assess the coverage of the data sources used by the counterpart economy.
7.8 While this chapter provides an overview of the various sources and their comparability to balance of payments and IIP needs and methodology, it is recommended to consult the data sources directly for a more detailed description of coverage and methodology.
IMF’s Bilateral Data Collections1
Coordinated Portfolio Investment Survey (CPIS)2
7.9 The CPIS provides information on individual economy year-end holdings of portfolio investment securities (short- and long-term debt securities and equities), valued at market prices, cross classified by the economy of issuer of the securities, and it is therefore most suited to be used for IIP compilation. The CPIS focuses on the geographical breakdown of portfolio assets of the participating economies. Some economies also provide data on portfolio liability positions. Cross border securities trading and settlement often uses the International Securities Identification Number (ISIN code) to identify securities. Data are valued at market prices. The CPIS follows definitions and classifications set out in the BPM6.3
7.10 The Survey of Securities Held as Foreign Exchange Reserves (SEFER) is conducted in parallel with the CPIS. Accordingly, participating economies also supply a geographical breakdown of the value of securities holdings that are included in reserve assets. Although the foreign securities component of reserve assets appear under a separate heading from portfolio investment in the BPM6, securities held as part of reserve assets are part of the portfolio investment liabilities of the issuers. From the perspective of the issuer, securities that are held as reserve assets are indistinguishable from those held in portfolio investment. Another survey is also conducted of large international organizations, the Survey of Securities Held by International Organizations, to obtain the value of their holdings of securities. In order to maintain confidentiality of individual returns, special procedures are used by the IMF to collect the data from these two surveys. Similarly, data from these two surveys are published only in aggregated form so that no economy’s or organization’s data are identifiable.
7.11 In the absence of a national data collection system, economies may use the CPIS derived liabilities data—for equity and investment fund shares, and short- and long-term debt securities—for identifying portfolio investment liability positions by partner economy and as an input for compiling the IIP. These data may understate an economy’s actual portfolio investment liabilities, because some economies do not report in the CPIS and others report holdings for only selected domestic sectors. Thus, the derived liabilities data should be viewed as a lower threshold that can be used to cross-check a given economy’s own estimates. The derived liabilities estimates should be evaluated for the extent of potential under coverage, and usually should be adjusted higher, before they are used in the IIP. However, the compiler should view the use of CPIS derived liabilities estimates as a temporary measure and work on developing a direct collection of data for compiling portfolio investment liabilities.
7.12 The CPIS derives portfolio investment liabilities positions for all economies, not just for participating economies from the reported data (see Tables 8 and 16 on the webpage indicated in footnote 2). The derived data can be used for checking a compiling economy’s portfolio investment liability positions; however, as mentioned in the previous paragraph, the compiler should consider it as a lower threshold for checking. The compiler should assess the extent that CPIS reporting economies included those economies that were most likely to hold securities issued by the compiling economies.4
7.13 The IMF plans to increase the frequency of the CPIS and SEFER surveys from annual to semiannual and to accelerate by three months the timeliness of reported data. The semiannual data will be released starting in 2014.
Adjustments to CPIS data for compiling the balance of payments and IIP
7.14 Although CPIS data are consistent with portfolio investment in the IIP, direct use of these data for partner economies could need some adjustments considering the absence of data from some counterpart economies either because they are not reporting economies or because they do not provide data with some partners to preserve confidentiality. Also, the compiler should keep in mind that the CPIS data for partner economies attributes security holdings to portfolio investment (by institutional sector) while portfolio investment liabilities in the IIP data includes also securities held by partner economies as reserve assets and securities held by international organizations. Grossing up techniques or models based on the available reported data could be used to estimate data for nonreporting economies. More details on grossing up techniques are presented in Chapter 2.
7.15 As mentioned earlier, the CPIS does not provide breakdown by sector of issuer, and cross border securities trading and settlement often use the ISIN code to identify securities. If the ISIN code is also used in a securities database, then this database can allow identifying characteristics of particular securities, including their issuer. Where feasible, the compiler is encouraged to classify portfolio investment by institutional sector of the issuer.
Coordinated Direct Investment Survey (CDIS)5
7.16 The CDIS collects data on direct investment positions for end-year by economy based on the location of the immediate counterpart to a direct investment position, with equity reported separately from debt investment. For inward direct investment, participating economies compile the value of outstanding end-year positions by immediate (first) counterpart economy (the economy from which the investment comes). For outward direct investment, participating economies provide information on the value of end-year outstanding positions by immediate (first) counterpart economy (the economy into which the investment is sent). Further breakdowns of information are encouraged showing gross debt instruments positions (total liabilities and assets separately identified), positions between fellow enterprises separately from those with direct investors/direct investment enterprises (DIENT), and positions of resident financial intermediaries separately from other direct investment positions. Also, as part of the overall reporting to the IMF, economies provide metadata that allow assessing whether data follow the standards. Direct investment relationships are described in paragraph 10.7.
7.17 Thus, CDIS contains data on direct investment positions, and is therefore suited to be used for IIP compilation. Partner economies can use the CDIS data to obtain information on their direct investment assets and, to some extent, their direct investment liabilities. Data collected through the CDIS are consistent in concepts, coverage, valuation, and classification with the BPM6 and the fourth edition of the OECD Benchmark Definition of Foreign Direct Investment.6 Data are available on an annual basis from 2009.
7.18 The CDIS does not contain enough information to calculate precisely transactions based on the difference between positions at the beginning and the end of the year. Therefore, it cannot be used without adjustment in the balance of payments accounts. (Box 10.2 of Chapter 10 provides an illustrative example of deriving transactions using data on positions and other price changes.)
7.19 The valuation method used for unlisted equity in CDIS is own funds at book value (OFBV), which is one of the methods of approximating market value of unlisted equity recommended in the BPM6 (paragraph 7.16). More details on different valuation methods including OFBV are contained in the Guide’s Appendix 4.
Adjustments to CDIS data for compiling the balance of payments and IIP accounts
7.20 The CDIS allows deriving from the reported data the outward and inward direct investment positions for all economies, not just for the participating economies (see Table 3 on the website indicated in footnote 5). These data can therefore be used either as a check of a compiling economy’s direct investment positions or, if direct investment data are not collected by the compiling economy, for use in preparing an IIP. As with the discussion of CPIS data, the compiler should assess the extent that CDIS reporting economies included those economies that were most likely to invest / receive direct investment in/from the compiling economy.
7.21 Derived data could be an important source of information to improve direct investment statistics of the compiling economy but should be treated with caution. Derived data could need some adjustments considering: (1) the absence of data from some economies, either because they are not reporting economies, or because they do not provide data with some partners to preserve confidentiality; (2) the degree of coverage of the standard components; and (3) consistency with international standards (e.g., exclusion of debt between selected affiliated financial companies, valuation method used for unlisted equity (OFBV), and inclusion of reverse investment and data on fellow enterprises). Data coverage can be assessed consulting the CDIS metadata questionnaire reported by individual participating economies. Grossing up techniques or models based on the available reported data could be used to estimate data for nonreporting economies when reported data cover most of the total value of the direct investment positions. Chapter 2 provides more details on grossing up techniques.
7.22 CDIS data are presented according to the directional principle, based on the direction of the direct investment relationship while the direct investment data in IIP are presented using assets/liabilities presentation (so that, for example, the netting of reverse investment is not built in). The standard components for direct investment positions may be rearranged to support both presentations. Table 7.1 shows how the assets/liabilities presentation can be rearranged to CDIS (directional principle) presentation.
|Asset/liability presentation in IIP||Directional principle presentation in CDIS|
|Direct investment assets||Total outward direct investment|
|Direct investment liabilities||Total inward direct investment|
Bilateral Data from Other International Organizations
International Banking Statistics
7.23 The Bank for International Settlements’ (BIS) international banking statistics (IBS) provides position data on cross border activities of banks in most major international banking centers. Data can be used to derive deposits assets and loans liabilities. The BIS statistics contain two datasets—locational and consolidated statistics. The locational data are based on the economy of location or residence of the creditor bank, and are consistent with the residence principle set out in the BPM6. These data can be used for balance of payments and IIP purposes. The consolidated data measure worldwide consolidated claims of banks headquartered in reporting economies, including claims of their own foreign affiliates but excluding interoffice positions.7 Since this classification principle differs from that used in the BPM6, the consolidated data are not a relevant data source for balance of payments and IIP compilation.
7.24 The locational statistics provide quarterly data on resident banks assets and liabilities, in the form of deposits and loans, with nonresident counterparts broken down by banks and nonbanks, and by economy.8 The data include both outstanding amounts and exchange-rate-adjusted changes, but no maturity breakdown of loans. Thus, the data can be used in compiling two components of IIP: (1) liabilities, other investment, loans, other sectors, by taking nonresident banks’ outstanding amount of loans to the compiling economy, and (2) assets, other investment, deposits, other sectors, by taking the compiling economy’s nonbanks’ deposits in nonresident banks.
7.25 The BIS debt securities can provide some information on an economy’s liabilities related to issuing international money market instruments, bonds, and notes. The methodology was revised in December 2012 for the full history of the statistics, to enhance the comparability of the statistics across economies. International issues were redefined as debt securities issued outside the market where the borrower resides—that is, offshore issues. These statistics cover only securities issued abroad by economies; they do not cover securities issued in the domestic market that are purchased by nonresidents. In addition, they make no allowance for international securities purchased by residents of the debtor economy. Caution should be used when using these data for IIP and balance of payments purposes.
Adjustments to BIS data for compiling the balance of payments and IIP
7.26 The nonbank sector in BIS statistics includes other sectors—other financial corporations, nonfinancial corporations, households, and NPISHs—as well as the general government sector.9 It is important not to double count if both data on nonbank sectors from the BIS and national sources for government sector data are used. The compiler should identify positions for general government included in the BIS dataset in order to exclude these positions from the nonbank sector data accordingly.
7.27 The information from the BIS statistics is partial but close to complete regarding the coverage of IIP components because international banking is concentrated mainly in the economies participating in the BIS IBS.10 Partial data can be adjusted using estimation techniques such as extrapolation or application of weights, deriving the weights of the reported data and grossing up to the universe. See Chapter 2 for more details on grossing up techniques.
7.28 This dataset does not provide a maturity breakdown for loans and deposits; thus the compiler will need to apply percentages based on alternative available data, to obtain these breakdowns.
7.29 The Development Assistance Committee (DAC) of the OECD is responsible for collecting internationally comparable data on official development assistance (ODA). This collection is referred to hereafter as the DAC Reporting System. The DAC has worked closely with the IMF’s Statistics Department, other international agencies, and member economies of the OECD to develop reporting directives and to align reported data to BPM6 concepts. Therefore the data are generally consistent with BPM6 guidelines for compiling data on development assistance flows (but see ahead). The compiler should be familiar with the reporting directives and work closely with the development assistance agency to ensure proper treatment of development assistance flows for the purpose of compiling the balance of payments and reporting to the DAC. It would be desirable to quantify any differences in treatment advocated by the reporting directives and the BPM6. Some donors, who are not DAC members, also report to the DAC.
7.30 The components of development assistance will be reflected in several items in the international accounts; the compiler should distinguish between current international cooperation (which is included in the secondary income account (see BPM6, paragraph 12.47)), investment grants (which are included in the capital account (see BPM6, paragraph 13.25)), and loans (which are recorded in the financial account and in IIP). For grants, it is necessary that offsets, such as exports of goods, provision of education services, other technical assistance, and provision of cash, be identified and included in goods, services, and currency and deposits accordingly.
7.31 Data in the DAC reporting system should, in practice, be compiled on a basis that closely parallels the concepts of timing and valuation advocated by the BPM6. The reporting directives comment specifically on the following:
Excluding official subsidies to private firms, since they support activities with a primarily commercial objective (the BPM6 includes these subsidies as primary income)
Including imputed costs of educating nonresident students when fees do not cover the costs of education programs (the BPM6 advocates the same treatment)
Including the administrative costs of ODA programs, subsidies to resident nongovernmental organizations, and refugee costs (the BPM6 does not include these costs because they involve resident-to-resident transactions)
Including interest subsidies paid to residents to provide “soft” financing as part of development assistance (the BPM6 does not advocate this treatment because the means of incorporating the impact within the economic accounts has not fully evolved); information on concessional debt could be provided through supplementary information (see BPM6, paragraph 12.51).
Including in development assistance the costs associated with technical cooperation, administration, education, and research incurred in the donor economy (The BPM6 does not advocate this treatment; however, paragraph 12.47(c) of the BPM6 indicates that payments associated with the salaries of technical assistance staff who are deemed to be resident in the economy in which they are working are included in current international cooperation.)
Excluding military equipment or services (In the BPM6, the supply of goods and services to military units or bases is included in government goods and services n.i.e. Transfers of military equipment are included in investment grants.)
Adjustments to DAC data for compiling the balance of payments and IIP
7.32 The coverage differences between ODA and the BPM6 requirements mentioned earlier should be considered by the compiler when DAC data are used for balance of payments and IIP compilation. For instance, adjustments should be made to include transfers to/from individuals or private companies.
7.33 In reporting forms submitted to the DAC, the major aggregates such as official development assistance should be classified by economy; however, transactions with international institutions are not classified by economy. Certain transactions, such as administrative expenditures in the donor economy are not—according to the reporting directives—allocated by economy and are shown unallocated. However, for the purpose of compiling partner economy balance of payments statistics described in Appendix 5, costs incurred in the donor economy should be classified by economy. If it is not possible to allocate specific costs to specific economies, the compiler could prorate costs across recipient economies by using aid flows that can be allocated by partner economy.
7.34 Reporting forms submitted to the DAC provide information on both commitments and disbursements of development assistance. The latter basis is relevant in compiling the balance of payments.
7.35 DAC data are compiled with a certain time lag. However, until actual data are available, extrapolations could be used.
External Debt Statistics
7.36 Two related databases on external debt statistics exist: Quarterly External Debt Statistics (QEDS) and Joint External Debt Hub (JEDH). An overview of these statistics is provided subsequently. A full description of the definitions, coverage, and methodology used in these statistics can be found on the home pages of the respective databases, although it is recommended to reference the original source data. Both the JEDH and QEDS are published on the World Bank’s webpage.
Quarterly External Debt Database13
7.37 The QEDS database brings together detailed external debt data that are published individually by economies that subscribe to the IMF’s Special Data Dissemination Standard (SDDS), as well as by some economies participating in the General Data Dissemination System (GDDS).14
7.38 Three main sets of data are available on a quarterly basis: (1) breakdowns of the total external debt position by sector, maturity, and instrument, (2) domestic-foreign currency breakdown of external debt, and (3) forward debt service schedule. The first two sets are relevant to IIP statistics. In addition, the SDDS QEDS includes six other sets of data that are of analytical value.
7.39 The definitions and criteria used regarding time of recording, concept of residence, and exchange rate conversion, along with classification by institutional sectors and financial instruments, will be fully aligned to the principles set out in the BPM6 in 2014. Accordingly, QEDS data should be consistent with IIP liabilities. However, it is important to be aware that the IIP components equity and investment fund shares, and financial derivatives and employee stock options, are excluded from the external debt statistics.
7.40 In some economies, QEDS data are compiled and reported by the balance of payments compiler that derived them from the IIP. In such cases, QEDS data are fully consistent with IIP. In other economies, QEDS statistics are compiled and reported by an institution other than the balance of payments compiling unit/institution. The QEDS compiling agency should assure that all IIP debt liability components are also covered in QEDS. This coverage consistency can be achieved by requesting from the balance of payments compiler data on IIP external debt instruments that usually are not monitored through the external debt monitoring system (e.g., currency and deposits, trade credit and advances, insurance technical reserves, and accounts payable). The IIP compiler also should assure that all data covered in external debt statistics are covered in IIP. A close cooperation and cross-checking of data between balance of payments and external debt compilation agencies should be undertaken each quarter. Generally, QEDS data are viewed as a source for cross-checking of IIP data and not as a data source for compiling IIP.
Joint External Debt Hub15
7.41 The JEDH jointly developed by the BIS, IMF, OECD, and World Bank—brings together external debt data and selected foreign assets from international creditor/market and national debtor sources. These sources include creditor data sources such as the African Development Bank, the Asian Development Bank, BIS, the Berne Union, the Inter-American Development Bank, the IMF, Paris Club Secretariat, the International Development Association, the OECD, and the World Bank—as well as information by various markets sources. The JEDH contains information on loans, other credits, including official trade credit and advances and bilateral loans, debt securities, and selected foreign assets and liabilities. Since not all sources used in the JEDH follow BPM6 requirements, it is important to consult the JEDH for detailed metadata description.
7.42 For comparison of the creditor/market data in the JEDH with the data reported by individual economies in the QEDS, the JEDH provides a comparison table for three broad instrument categories: loans plus deposits, debt securities, and trade credit and advances. In this way the compiler can compare national data with data from partner economies and market sources in an easy manner. The compiler should use JEDH data to check or review other data sources, but it is recommended to use the original data source to compile balance of payments and IIP data.
Data Template on International Reserves and Foreign Currency Liquidity16
7.43 The data template establishes standards for the provision of information on the amount and composition of official reserve assets, other foreign currency assets held by the monetary authorities and the general government, short-term foreign currency obligations, and related activities (such as financial derivatives positions and guarantees extended by the government for quasi-official and private sector borrowing) of the monetary authorities and the general government that can lead to drains on reserves and other foreign currency assets.
7.44 In principle, official reserve assets specified in Section I.A. of the International Reserves and Foreign Currency Liquidity: Guidelines for a Data Template (IRFCL) should correspond to the data on international reserves that economies compile for balance of payments and IIP purposes under the BPM6 guidelines. The definition of official reserves assets should be consistent across all macroeconomic statistics sets.17 If balance of payments/IIP data and the IRFCL are compiled by different institutions/units, a cross-check of data on reserve assets should be conducted each quarter for their full consistency.
Direction of Trade Statistics (DOTS)18
7.45 The Direction of Trade Statistics (DOTS) contains, for about 187 countries, current figures on the value of merchandise exports and imports disaggregated according to their trading partners. Data are available from 1980 for years, quarters, and months. Reported data are supplemented by estimates whenever such data are not current or are not available in monthly frequency. DOTS follows concepts and definitions set out in International Merchandise Trade Statistics: Compilers Manual (2004).
7.46 It is sometimes assumed that corresponding export and import data between partner economies should be consistent—that is, the exports from economy A to B should be equal to the imports of economy B from A, after taking into account the insurance and freight costs when economy B imports are valued on a c.i.f. basis. The DOTS estimation system uses this assumption in cases where one partner has not reported data.
7.47 However, different compilation practices may cause inconsistency between exports to a partner and the partner’s recorded imports. The principal reasons are differences in (1) classification concepts and detail (lack of uniformity in determining origin, transshipment, and destination economies), (2) time of recording, (3) valuation, (4) coverage (shipments to and from free-trade zones, exclusion of military and other confidential items, value thresholds for customs registration of shipments, returned goods, and other good (5) processing errors.
7.48 As most economies report data on exports and imports by counterpart economy, this dataset, in contrast to those included in “IMF’s Bilateral Data Collections” and “Bilateral Data from Other International Organizations,” is not used by the counterpart economy for balance of payments statistical purposes, but it could be used to check or verify the compiler’s own data, considering possible differences already explained. As discrepancies are identified and explained, the compiler can improve the quality of his or her trade data.
Users can access the detailed nonconfidential data as reported by each individual economy.
Methodology, data sources, and results from the survey are available at http://cpis.imf.org.
However, in consultation with the IMF’s Committee on Balance of Payments Statistics and CPIS participating economies, the IMF is increasing the frequency of the survey from annual to semiannual (beginning with end-June 2013 data) and to accelerate by three months the timeliness of reported data. More frequent and timely reporting of CPIS data could provide improved benchmarks for estimating quarterly portfolio position data. The scope of the data is also being enhanced to include the identification of short positions and sector of debtor. Further, with the aim of aligning the outputs with user demands, an additional disaggregation of the data on the institutional sector of resident holders cross classified by the institutional sector of the nonresident issuers of securities is also included in the revised CPIS reporting forms. However, in order to reduce the potential reporting burden, this disaggregation limits the sectoral detail for nonresident issuers to the 25 economies with systemically important financial sectors, the economies for which the IMF Executive Board agreed should have mandatory assessments every five years under the Financial Sector Assessment Program.
In 2013, with reference to data for 2012, 77 economies, including almost all the major economies and most large portfolio investment holders, reported their holdings of portfolio investment assets issued by nonresidents broken down by nonresident economies and instruments.
Methodology, data sources, and results from the survey are available at http://www.imf.org/external/np/sta/cdis/index.htm.
Around 100 economies, including almost all the major economies, but also many smaller economies, report annually their outstanding inward direct investment positions broken down by immediate counterpart economy. Most of these economies also report annually their outstanding outward direct investment positions in the same detail.
These statistics build on measures used by banks in their internal risk management systems.
BIS International Banking Statistics, Table 7A (all sectors) and 7B (nonbank sector) at http://www.bis.org/statistics/bankstats.htm.
The BIS Committee on the Global Financial System (CGFS) recommendations to enhance the BIS international banking statistics are being implemented. Stage two of the enhancements relevant to IIP and balance of payments purposes (collection of new data, including a finer sectoral breakdown in the locational banking statistics) will be reported from Q4 2013 data. See “Improving the BIS International Banking Statistics,” CGFS Papers, no. 47, November 2012, available at http://www.bis.org/publ/cgfs47.htm.
As of March 2013, 44 countries and financial centers report these data to the BIS.
See also measurement of development assistance in donor economies in Chapter 6.
The QEDS database, jointly developed by the World Bank and the IMF, brings together detailed external debt data of economies that subscribe to the SDDS and a selected number of economies that participate in the GDDS. As of June 2013, about 110 economies report to the QEDS (see http://www.worldbank.org/qeds).
GDDS countries’ participation in the QEDS does not require producing prescribed SDDS data.
The concordance between classifications of reserve assets in the BPM6 and the IRFCL is explained in Table 2.1 of the guidelines: http://www.imf.org/external/np/sta/ir/IRProcessWeb/dataguide.htm.