13. Traded Securities
- International Monetary Fund
- Published Date:
- June 2003
13.1 External debt in the form of traded securities corresponds to debt securities in the inward portfolio investment component of the balance of payments and IIP. In recent decades, the relaxation of restrictions on the foreign investment activities of banks and other institutional investors, combined with continued financial innovation, has resulted in a surge of cross-border investment in bonds (and equities). This has increased the interest of policymakers in data on this activity.
13.2 However, ensuring comprehensive coverage of traded securities is among the most difficult in the field of balance of payments and external debt statistics. In particular, the resident issuer is, in many cases, not in a position to identify the beneficial owner of their securities, and so may be unaware of whether the creditor is a resident or nonresident. Thus, almost inevitably, to compile position data, other than by accumulating flows on a previous position, the compiler needs to obtain information on the stock of traded securities of residents, and the owners of those securities, from a variety of sources. While it is relatively straightforward, but not a simple task, to obtain data on nontraded debt liabilities, for the following reasons, it is more difficult to identify the owner of a traded security.
Liberalization has facilitated the development of new channels through which investment can flow. In other words, compilers can no longer rely solely on traditional domestic data sources, such as banks or security dealers, because investors increasingly use foreign intermediaries, and security issuers may access foreign markets directly.
Unlike banks, which have a tradition of reporting to the central banks, as noted in the previous chapter, nonbank economic agents may be reluctant to report to the authorities on their ownership of traded securities, because, among others, of concern that data sent to the statistical agency may be passed on to other agencies. This, once again, highlights the need for the promotion of statistical integrity within the country.1 Noncompliance by respondents leads to gaps in coverage at a time when activity is rising.
The participation of various financial intermediaries in international transactions and the practice of registering of investment under nominee companies and in trusts can obscure the beneficial owner of the security.
International markets in certain “new” instruments have grown quickly in the past decade, causing difficulty in determining the “true” owner of the security. An example is the use of securities in reverse security transactions.
Rarely, if at all, is it possible for a government to have legal powers to require a nonresident investor to report on their ownership of securities issued by domestic residents.
Ways in which these difficulties might be overcome are examined in this chapter.2
13.3 In looking at ways to capture activity in securities for external debt purposes, countries should take into account any existing system they already have in place for the collection of data on portfolio investment and, more generally, balance of payments data, and also arguably, national accounts data. Respondents will know the existing system, and a considerable amount of human capital will have been invested in it at the compiling agency. Those concerned with external debt statistics should draw on this knowledge and expertise, not least because a detailed system of collecting data on inward and outward security investment can be resource intensive.
13.4 Also, there is a close linkage between cross-border securities activity and other data series such as direct investment. More important, inward, and outward, portfolio investment is directly affected by domestic activity. Whereas direct investment generally involves the establishment of a longer-term relationship between parent companies and their foreign affiliates, securities investment involves securities—both domestic and foreign—that potentially can be traded between residents and nonresidents. Depending on regulations and institutional arrangements, ownership of domestic and foreign securities can change quickly. Indeed, as exchange controls are lifted, inward or outward capital flows can arise from security transactions of both residents and non-residents. So, while the focus in the Guide is on foreign investment in securities issued by residents, when considering how to measure this activity due regard should be given to the measurement of residents’ investment in securities—issued by both residents and nonresidents.
13.5 This close relationship between data on traded securities in external debt, the balance of payments, and the national accounts means that it is important for agencies to cooperate. Otherwise potentially useful information may not be utilized, while at worst, respondents could end up reporting essentially the same information to two different statistical agencies. Cooperation need not only involve statistical agencies. In other government agencies there will be potential users of the data collected. For instance, information on nonresident ownership of government securities is likely to be of interest to finance ministries in helping to formulate government debt policy. Policy ministries can help the compiler in devising report forms, encouraging responses, and in evaluating the (aggregate) data.
13.6 Finally, any development of the data system to capture investment in domestic securities by nonresidents will inevitably lead to questions about the computer system on which data are to be stored and manipulated. Computer systems are obviously tools that help facilitate a more efficient statistical operation, but before a computer system is installed, it is necessary to consider the form of the data capture and manipulation; the data output required both in final form and from interrogation of the system; as well as any need to be compatible with data stored in other systems.
13.7 An important starting point in deciding how to measure positions (and flows) in traded securities is ascertaining how and through which channels security investment flows into and out of the country. This involves talking to market participants and generally gaining an understanding of the domestic security markets. The issues to explore are:
How do nonresidents invest in domestic securities?
Through which institutions do they invest?
Where do nonresidents arrange for the custody of their domestic securities? How are records held?
Where are trades settled?
Are security codes used in monitoring security positions?
Do residents issue securities directly abroad? Do residents invest in these securities?
13.8 The importance of preliminary research cannot be overstated because, once completed, the compiler can decide at which point or points in the “chain” of activity it is most appropriate to collect information. There is no one obvious answer for all compilers. For legal, institutional, and historical reasons, different countries have different market structures and practices, and so what suits one country may not suit another. Nonetheless, the pros and cons of collecting information from different types of market participants can be indicated, and these are set out in Table 13.1. The relevant importance of the various advantages and disadvantages will depend on individual economy circumstances. For different instruments and markets, different collection methods may be appropriate.
|Issuer of security||Will know about securities issued.||Unlikely to know beneficial owner of the security either at issue or during secondary market trading.|
|Banks (receipts/payments)||Transactions in domestic currency require settlement through resident banks. Transactions recorded could be cumulated on a previous position and, with appropriate valuation adjustments, provide new position data.||Nature of transaction may be difficult to establish. May have a problem in identifying direct investment transactions. Although a method for compiling position data in the short term, a more direct measure of the stock position might be required in the medium term, depending on the complexity of the reporting system. Also, only covers investment in securities issued in the domestic market.|
|Issuing agency (security house/bank)||Will know about securities issued.||May not know beneficial owners at issue and, unless a dealer, will not know about secondary trading.|
|Dealer (security house/bank)||Will have information on sales and purchases of securities. As with banks, transactions data could be used to compile position data.||May not cover all nonresident purchases of resident securities. May have a problem with nominees and identifying direct investment transactions.|
|Fund manager||Will have information on beneficial owners.||Unlikely to cover all nonresident purchases and holdings of resident securities.|
|Organized exchange||Will have a record of transactions on the exchange and perhaps positions. Data on positions might also be available via member firms.||May not cover all nonresident purchases and holdings of resident securities. May have a problem with nominee accounts.|
|Settlement agency||Will have a record of transactions.||May not cover all nonresident purchases and holdings of resident securities. May have a problem with nominee accounts and identifying direct investment transactions/positions. Records may not be kept in a form appropriate for external debt/balance of payments purposes.|
|Registrar||Will know who owns which securities.||Use of bearer securities undermines the use of a securities register. May have a problem with nominee accounts. May not cover transactions particularly well.|
|Custodian||Information on ownership available. Fewer in number than investors. Should know information on the outstanding value of holdings.||Coverage of nonresident purchases and holdings of resident securities is uncertain. May have a problem in identifying nonresidents, although tax status may help, and direct investment transactions. May not know exact details of transactions/may have difficulty extracting data in line with balance of payments methodology. Double counting a potential problem if subcustodians used.|
13.9 Before discussing the advantages and disadvantages of approaching different types of respondents, the relationship between the collection of transactions and position data needs to be considered. There are various ways transactions and position data can interact: (1) transactions data can be compiled separately from position data, and cross-checks introduced to validate both sets of data;3 or (2) transactions data can be added to a previous position and, with appropriate reevaluations and any other adjustments, a new estimated position calculated (although an independent benchmark position survey at periodic intervals is essential to check and improve the quality of the estimated position data)—see the appendix to Chapter 12; or (3) position data can be compiled on a security-by-security basis, supported by a database with information on individual securities issued by domestic residents (Box 13.1), using individual transactions data to update the individual holdings of securities (although even then periodic verification of the derived position data is recommended using alternative or additional inquiries). Whichever method is used, decisions on whom to approach and what to request in terms of position data are at least influenced by the approach taken to collecting transactions data. So, in the discussion below, both transactions and positions data are discussed.
Box 13.1.Security Databases
In measuring positions in traded securities, information may be collected from respondents at the level of the individual instrument. Such an approach potentially provides great flexibility in meeting requirements for external debt statistics. However, to utilize fully the potential of such information, the compiler is advised to develop or acquire a database that contains detailed information on individual securities—price, country of issuer, industrial sector of issuer, etc.—and that uniquely identifies securities through a security identification code.1 Through such a database, individual securities that are reported with an identification code can be located in the database, and the associated information can be drawn upon to compile information not only on outstanding positions but, depending on the scope of the associated information contained, statistics on the debt-service payments schedule, the currency composition of external debt, etc. Also, such an approach can enhance data quality by allowing the compiler to check the accuracy of submitted data and to resolve conflicting reports.
Sources of Information
Information on individual securities can be obtained from commercial sources, international organizations, and security numbering agencies.
By far the most comprehensive and complete databases are those available from commercial sources, usually at a commercial price. The best of these commercial sources supply high-quality, timely, comprehensive data to the international financial community to support investment activity. At the time of writing, some of the leading commercial vendors, in alphabetic order, are Bloomberg, Euromoney Bondware, Interactive Data, International Financing Review, International Securities Market Association, Reuters, and Telekurs.2
The Bank for International Settlements (BIS) maintains a database of international debt securities that is available to member central banks and perhaps other governmental organizations, as described in Chapter 17.
The Association of National Numbering Agencies has a database of individual securities that is commercially available. By linking the databases of national numbering agencies (NNAs)—the entities that assign the international security identification number (ISIN) in their own jurisdiction—this database provides key descriptive information on individual securities. Coverage of individual securities differs in completeness among NNAs, and information on market prices is not included. To understand more about the information available on this database, it is recommended that the compiler approach the NNA that allocates ISIN codes to securities issued within the domestic economy.
Role of the Security Identification Code
As noted above, in a compilation approach that uses a database of individual securities, the security identification code is of central importance—the respondent needs to provide a code so that the database can identify the security. However, different respondents could submit different security identifiers for the same security because any widely traded security could be allocated a domestic as well as an international security identifier. For instance, in the United States, a domestic security code (known as CUSIP) will be allocated to a domestic security. As a result, private investors have adopted a variety of different security identification systems as their primary identifier. National compilers should discuss the use of security identifiers with potential survey respondents. If national compilers can rely on survey respondents to use primarily one coding system—for instance, the ISIN—this enhances the efficiency of the compilation procedure. If not, then the agency is advised to acquire a database(s) that contains all the various identifier codes that a given security has been assigned by the different coding systems. These cross-reference databases may well be available from the same commercial firms mentioned above.
Nonresident Investment in Domestically Issued Securities: Potential Respondents
13.10 An obvious approach for compilers is to collect information on nonresident investment in securities issued domestically by residents from domestic financial intermediaries. This approach assumes that nonresidents will involve these intermediaries when undertaking transactions in the domestic market. For instance, for transactions and positions in government securities, the government might consider making such a reporting requirement a condition of any licensing approval that the domestic financial entity may need in order to have settlement accounts in domestic government securities.
13.11 Typically, banks are approached for data on external transactions and positions because of their role in the payments system; if domestic currency is used to settle transactions, a resident bank is likely to be involved. However, money flows through banks for a variety of reasons, and banks may have difficulty in establishing the specific nature of a transaction as a securities transaction. Also, it is important that transactions involving nonresidents are captured not only when money comes into the country but also whenever nonresidents transact in domestic currency, such as when a nonresident draws down a domestic bank account to purchase a resident security. This is the key issue: can banks identify and report in a comprehensive manner investment by nonresidents in domestic securities? The possible use of data from banks in their role as custodians is examined ahead.
13.12 Another method used is to gather data on securities from investment dealers, including banks, that conduct portfolio investment business on behalf of nonresidents. In other words, those who arrange and execute the deals. Dealers usually keep records of client transactions and may be better able to identify portfolio transactions than banks through their payments system activity. Invariably, the number of financial intermediaries are likely to be fewer in number than investors, and, legal circumstances permitting, should be approachable. This method of approach depends, of course, on nonresidents using domestic intermediaries. Also, these institutions will need to be able to identify residents and nonresidents and keep records in a manner that allows their use in external debt, as well as balance of payments and IIP compilation.
13.13 Canada has adopted a system of capturing foreign investment in securities using dealer reports, and this method is set out in Chapter 14. The dealers report individual transactions involving nonresidents and include the value of the deal and the unique code for the security (developed for settlement purposes).4 Information kept on a database of individual securities is used to confirm the residence of the issuer of the security and provide additional information. Canada’s detailed and complex statistical system also generates income data on an accrual basis.
13.14 Some countries carry out special surveys that are addressed to resident fund (investment) managers, and request information on own account and client account investments in resident and nonresident securities by resident and nonresident investors, thus providing the necessary data on residents’ debt liabilities to nonresidents. Data on the country and institutional sectoral distribution of ownership may also be requested. The information can provide good coverage of the household sector’s portfolio assets, provided that they use resident fund managers. However, such a survey will not provide comprehensive coverage of nonresident ownership of resident securities unless nonresidents use domestic fund managers extensively.
13.15 Another method is to capture nonresident investment in domestic securities at the point of the trade or settlement—for instance, using information on transactions from the stock market. At the least, the stock exchange usually has to keep a record of individual transactions, and at best may act in a settlement capacity and see the cash change hands. It may be possible for this information to be supplied to the compiling agency. Sometimes there may be a separate but similar market mechanism for bond trades. Through these markets, nonresident investment transactions and holdings may be obtained. For instance, the exchange might have or can obtain the authority to request that information be reported to it on who owns what securities. This might be undertaken not only for statistical but also for regulatory and policy purposes.
13.16 However, there may be reluctance for the stock exchange or settlement agency to release the information required by the compiling agency, and the prevalence of nominee accounts may lead to misidentification of the true investor (a common problem when “intermediaries” report). Other issues that could arise are whether the records kept can be readily utilized for external debt and balance of payments statistics purposes, and the comprehensiveness of the coverage of nonresident investment in resident securities.
13.17 Close links with the stock exchange may be important for the compiler in other regards. The stock exchange will be a source of information on market developments; it may well be the agency that needs to be kept informed by quoted corporations of new securities issues—helpful for information on security issues in foreign markets by residents (and domestic security issues by nonresidents); it may be the agency allocating code numbers to individual securities issued in the domestic market; and individual investors may need to inform the stock exchange of large equity holdings, thus helping the compiler to identify direct investment positions and transactions.
13.18 Another avenue is to approach registrars, who store information on the owners of securities (for example, to make coupon payments). For instance, details of ownership of debt securities issued by the government in domestic markets are frequently held on a computerized book-entry register, with change of ownership being evidenced by an entry on this computerized register, rather than the transfer of a physical certificate. Typically, these registers contain useful information such as, for each security, the outstanding balance for each investor, and the amount of accrued interest. Also, debt securities can be valued both at market as well as nominal value and can be classified by original as well as remaining maturity. However, problems arise in identifying ownership, given the frequent use of nominee accounts, not least for administrative efficiency.
13.19 Yet another method of measuring nonresident investment in securities issued by domestic residents is to collect data from custodians. Many countries use custodian surveys of one type or another, and an approach should be explored for compiling at least some element of the data on nonresident ownership of traded securities. Domestic securities owned by nonresidents may be deposited with local custodians for “safekeeping,” and these institutions, primarily banks, could be approached through a survey to report transactions and ownership of domestic securities by nonresidents. Such a survey can provide good coverage of resident securities denominated in the domestic currency and traded in national organized markets.
13.20 However, resident securities denominated in foreign currency, issued and usually traded in foreign organized markets (for example, international bonds, etc.), are unlikely to be captured by such a survey. Also, there are other possible drawbacks that the compiler needs to consider.
13.21 The custodian may have difficulty in distinguishing residents from nonresidents, although a possible different tax treatment from that applied to residents may be one way in which this distinction can be made.
13.22 A local custodian may be acting on the instructions of a “global” custodian, located in another economy, and so may not know the name of the beneficial owner of the security—the security might be registered in the name of a foreign global custodian. Resident custodians are likely to record security holdings in the name of the global custodians as nonresident holdings, but resident investors could subsequently purchase the securities but leave them entrusted with the global custodian, causing a mismeasurement of nonresident ownership. Periodic surveys to confirm the beneficial owner of securities may be warranted.
13.23 Another potential problem, and one that arises with all transactions and positions reported through financial intermediaries, is the difficulty in distinguishing securities related to direct investment activity from other cross-border security activity, leading to the possibility of double counting of investment activity if direct investment data are separately collected, which is usually the case.
13.24 Securities data from custodians can be reported on an individual or aggregate basis. As mentioned above, data reported on an individual basis is best supported by a database that records individual securities issued by domestic residents. This database can reduce the burden on the respondent and confirm the data reported. This is a method successfully employed by Austria, but, as with the Canadian system, there can be considerable resource cost for the compiling agency.
13.25 Alternatively, aggregate data can be requested. As always, checks are required: for instance, with aggregate data a custodian might report the number of securities owned rather than their value. It should also be recognized when requesting aggregate information that custodians may not hold their records of nonresident ownership of domestic securities in a way that is conducive to external debt reporting. Therefore preliminary discussions are essential to ascertain which data might be readily available.
Issues of Securities by Residents in Foreign Markets
13.26 Measuring foreign investment in securities issued abroad by residents can be difficult. Foreign intermediaries will not report to the domestic compiling agency. Swapping data with foreign compilers is one option, but this approach is difficult to implement because the compiler would need to know all the compilers to approach, and the nonresident compiler would need to have the requisite data. A more promising approach is to obtain information on gross issues and redemptions of international issues either from issuers themselves or from other sources, including the domestic stock exchange or other official bodies that should be informed of any new issues by quoted companies. International sources of information, such as the BIS international securities database, discussed in Chapter 17, could also prove useful.
13.27 If a database of individual securities is maintained—or aggregate information on foreign security issuance is reported by resident issuers—so that net new issues in foreign markets—gross issues and gross redemptions—are recorded, and the outstanding amounts of the securities issued by residents in foreign markets can be calculated, a reasonable assumption could be that these securities are purchased by nonresidents, excluding those known to be purchased by residents. In other words, external debt in the form of nonresident investment in debt securities issued in foreign markets by residents, including government, could be calculated by, all other things being equal, netting out domestic ownership of resident debt securities issued in foreign markets from the total outstanding. Information on resident holdings of these debt liabilities could come from domestic respondents, either the investors themselves or financial institutions involved in this activity. Besides being a method of calculating an element of external debt, the resultant information on resident and nonresident ownership of debt securities issued in foreign markets is of interest in its own right, as explained in Chapter 7.
13.28 This approach is a perfectly acceptable compilation technique and would require the compiler to liaise with the agency that complied data on domestic investment in domestic securities for the financial accounts. Indeed, some countries employ this technique to measure inward investment into all private sector traded securities.
Information on Securities Involved in Reverse Security Transactions
13.29 If the collateralized loan approach is employed to record reverse security transactions (such as repurchase agreements, repos), a memorandum table (Table 4.5) is provided in Chapter 4 for the presentation of data on securities issued by residents that residents acquire from or provide to nonresidents under these arrangements. It is expected that the majority of such transactions will occur in the domestic market, most likely in the government securities market. Most commonly, repos are transacted by financial institutions with other financial institutions, including central banks. Requiring domestic financial institutions to report domestic and nonresident securities sold to and purchased from nonresidents under reverse transactions, perhaps in their balance-sheet returns to central banks and/or statistical agencies, is likely to cover the bulk of the business. Other entities such as nonfinancial enterprises and governments may be involved in reverse security transactions, perhaps even in domestically issued securities in foreign markets. So the compiler is advised to investigate the significance of information from these institutional sectors as well.
13.30 However, in compiling data on securities issued by residents and traded under reverse security transactions by residents with nonresidents, care needs to be taken to avoid double counting. Experience indicates that where security registers are used to identify nonresident ownership and/or where custodians report, it is not always possible to identify securities subject to repos. So, if a custodian provides information on nonresident ownership of resident securities, it may be inclusive of securities purchased and sold under reverse security transactions; that is, the information might already include resident securities acquired by nonresidents from residents under reverse transactions, contrary to the collateralized loan approach. It is very important for the compiler to understand how securities involved in reverse security transactions are recorded in the position information provided.
13.31 Clearly, the more that transactions in domestic securities are concentrated in the domestic economy, the greater is the likelihood that domestic financial intermediaries can provide adequate coverage and, thus, a lower likelihood that there will be undercounting. The difficulty is then in ensuring that resident and nonresident owners are correctly identified and the concepts outlined in the Guide are adhered to.
13.32 On the other hand, overcounting is more possible where a number of methods are used to collect data. While more than one method may be needed to ensure comprehensive coverage—for instance, the measurement of foreign investment in government securities may differ from that for private securities—the compiler should be aware of the increased possibility for the double counting of activity when more than one method is used.
13.33 To reduce the possibility of mismeasurement, particular care needs to be taken in deciding on the respondent population; as noted in the previous chapter, a register of reporters, kept current, is essential and could be drawn from a centralized national register of reporting entities maintained for national accounts reporting purposes.
Periodic Position Surveys
13.34 As mentioned above, in the short term some economies might compile position data by accumulating transactions on a previous position. However, it is important to conduct periodic benchmark position surveys, perhaps once a year. The sources of position data could be different from those for transactions data. For instance, data on transactions might be compiled from information supplied by dealers or organized exchanges, whereas custodian information might be used for the position data. The results of the position survey can then be checked against the cumulative transactions data; in other words, reconciliation can be undertaken. This reconciliation is particularly important when financial intermediaries are reporting transactions because it can reveal inconsistencies and errors in reporting that might not otherwise be spotted. In some ways, independent verification of the data is helpful for the robustness of the compilation system. Alternatively, the same institutions could be approached for both transactions and position data so that any discrepancies can be rectified and improvements made for future years.
13.35 Because of the need to improve the coverage of portfolio investment assets globally, and also because of the difficulty of identifying nonresident ownership of resident securities, the IMF, in cooperation with other international organizations, has promoted the development of a Coordinated Portfolio Investment Survey (CPIS). This survey was conducted for the first time with a reference date of end-December 1997 and was designed to collect comprehensive information, with country attribution, of resident holdings of nonresident securities, both equity and long-term debt securities. It involved 29 countries using harmonized definitions and concepts based on BPM5. By exchanging bilateral data, the coverage and quality of portfolio debt liabilities was also improved. The results of the first survey were published in 1999.5
13.36 At the time of preparation of the Guide, the CPIS is about to embark on its second survey, with a reference date of end-December 2001, and thereafter could become a regular undertaking. As coverage is improved, such as more major investing countries participating and the weaker areas of measurement strengthened (for example, coverage of household investment), then over time these creditor-based data could gain in importance as a source of information for external debt compilers.
Integrity of disseminated data is one of the four dimensions of the IMF’s Special Data Dissemination Standard (SDDS) and General Data Dissemination System (GDDS). Among the type of actions that the SDDS and GDDS outline to promote integrity is the dissemination of the terms and conditions under which official statistics are produced, including those relating to the confidentiality of individually identifiable information.
Although a practical guide for the measurement of assets, a helpful source of information on compiling position data for traded securities is IMF (1996), Coordinated Portfolio Investment Survey Guide, and its second edition (IMF, 2002) is available on the Internet at the IMF’s website, http://www.imf.org/np/sta/pi/cpisgd.htm.
See IMF (1996), Coordinated Portfolio Investment Survey Guide, Appendix VIII, for an explanation of how to reconcile position and transactions data, and to estimate income from position data.
While national numbering agencies (NNAs) frequently issue their own code for securities issued in their jurisdiction, they also allocate a unique international security identification number (ISIN) code for each security. More information on ISIN codes is available in IMF (2002), Appendix VII, pp. 151–53.