While the manner in which the national survey is to be conducted is left to the national compiler, the concepts and principles underlying the national survey are to be applied in conformity with the fifth edition of the International Monetary Fund’s Balance of Payments Manual. To assist compilers in meeting this task, the Task Force identified the following areas where practical guidance is required:


While the manner in which the national survey is to be conducted is left to the national compiler, the concepts and principles underlying the national survey are to be applied in conformity with the fifth edition of the International Monetary Fund’s Balance of Payments Manual. To assist compilers in meeting this task, the Task Force identified the following areas where practical guidance is required:

  • nomenclature;

  • country attribution;

  • valuation;

  • distinction between direct and portfolio investment; and

  • securities where there is a potential to double count.

I. Introduction

64. The concepts and principles underlying the Coordinated Survey are those contained in the fifth edition of the IMF’s Balance of Payments Manual (BPM5). The first section of the Survey Guide’s glossary of security terms (appendix V) explains the most important of these. As national compilers apply these concepts and principles, some practical reporting issues may arise. To guide compilers, the IMF published in 1995 the Balance of Payments Compilation Guide, which describes how the conceptual framework in BPM5 can be implemented in practice. To further ensure both consistency and quality of reporting across the participating countries in the Coordinated Survey, this chapter covers certain practical issues that go beyond those covered in the Compilation Guide. This chapter draws on the work of the Organisation for Economic Cooperation and Development’s (OECD’s) Working Group on Harmonized Investment Definitions, which met in 1993/94.13

II. Nomenclature

65. As discussed in chapter 1, the Coordinated Survey is confined to nonresident equity securities and long-term debt securities in the form of bonds and notes owned by residents. BPM5 defines these securities in chapter XIX, paragraphs 388 and 390 (see also appendix V—the glossary of security terms) and discusses the methodological treatment of selected instruments in paragraphs 395 to 408.

66. Nonetheless, financial markets continually evolve. Consequently, national compilers face changing classification and/or valuation or other balance of payments accounting problems.

67. The Bank of England took the following approach to address these problems: To provide national compilers in the European Union with a reference tool and to encourage a consistent approach in classifying financial instruments, the Bank of England, in response to a request from the European Monetary Institute (EMI), created a financial terminology database (FTD) of financial instruments. The database, which is periodically updated, provides (for each financial instrument listed) its “economic rationale” and the balance of payments treatment. Where appropriate, the relevant BPM5 reference is provided; if the instrument is not directly covered in BPM5, a treatment is proposed by the Bank, in line with the IMF methodology, which is then discussed and agreed to by the balance of payments experts of the EMI. More details on the FTD are provided in appendix VII.

68. For the Coordinated Survey, national compilers should endeavor to use the methodology of BPM5 to classify financial instruments. However, if compilers desire further guidance, they may find the FTD to be a valuable reference.

III. Country Attribution

69. A principal requirement of participation in the Coordinated Survey is the provision of a country attribution of domestic residents’ investment in nonresident securities. This information is required so that bilateral information can be exchanged. This section addresses the issues raised by this requirement.

70. BPM5 recommends that holdings of securities should be attributed according to the debtor/creditor principle: “Financial claims of the compiling economy are allocated to the country of residence of the nonresident debtor, and liabilities are allocated to the country of residence of the nonresident creditor” (paragraph 484).14

71. Ensuring the correct country attribution of securities is crucial to the success of the Coordinated Survey. For data to be consistent, each reporter/compiler must attribute the same security issue to the same country, regardless whether the national survey is being conducted on a security-by-security basis or whether it is on an aggregated basis. Therefore, for the conduct of the Coordinated Survey, the country of residence of any enterprise is the country where the enterprise is legally incorporated or, in the absence of legal incorporation, where it is legally domiciled. For instance, securities issued by offshore entities should be attributed to the offshore country in which the issuing entity is legally incorporated, as opposed to the country of the parent enterprise. Securities issued by international organizations should be recorded as such, and not attributed to the country in which they are located. Other characteristics of a security, such as the currency of denomination, the market where it is issued, listed, or traded, or the country of the guarantor or underwriter of the issue, should not be used as a substitute for the resident country of the issuer.

72. The coding systems used by the securities industry to identify securities can help ensure consistency of geographic attribution of securities by compilers across countries. For example, International Securities Identification Number (ISIN) code numbers are now allocated to virtually all internationally traded long-term securities by national numbering agencies (NNAs), who have sole right to allocate numbers within their jurisdiction. Each security has its own unique identifier. For equity securities (but not debt securities), the first two digits can be used to identify the country of the issuer. It is recommended that national compilers, when discussing their national survey with potential survey respondents, evaluate the use of the ISIN code system in their own country and gauge its potential use in classifying securities in their national survey. More information on the coding system and the associated security databases is provided in appendix VII.

IV. Valuation

73. In paragraph 468, BPM5 recommends that stocks of assets (and liabilities) should be valued at current market prices16 at the appropriate reference date; it also provides advice on how to value securities:

For equities that are listed on organized markets or are readily tradable, the value of outstanding stocks should be based on actual (market) prices. The value of equities that are not quoted on stock exchanges or otherwise traded regularly should be estimated by using the prices of quoted shares that are comparable as to past, current, and prospective earnings and dividends. Alternatively the net asset values of enterprises to which the equities relate could be used to estimate market values if the balance sheets of enterprises are available on a current value basis. For debt securities that are listed in organized markets or are readily tradable, the outstanding value of stocks also should be determined on the basis of current market values. For debt securities that are not readily tradable, net present value of the expected stream of future payments/receipts associated with the securities could be used to estimate market value. (The net present value of any future receipt is equal to the value of that receipt when discounted at an appropriate interest rate.)

Security Databases

Countries conducting a security-by-security survey have found it extremely useful to acquire a security database that contains detailed information on securities, for instance, price, country of issuer, industrial sector of issuer, etc. If survey respondents report a securities identifier for each security held, then compilers can use these databases to provide missing information, check for accuracy of submitted data, and resolve conflicting reports. Appendix VII provides more information on security databases.

Before acquiring a database, however, national compilers should be aware that there is a potential difficulty: different respondents could submit different security identifiers for the same security because any widely traded security could be allocated an identifier by more than one coding system. Indeed, in the model survey form in appendix II, survey respondents are given the option of reporting either an ISIN code or a code allocated by a number of coding systems. National compilers should discuss this issue with potential survey respondents; the allocation exercise becomes more straightforward and efficient if national compilers can rely on survey respondents to primarily use one coding system, for instance, ISIN. 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, and then build links between all these data and the reported information.15

15Information on how to store data in a set of linked two-dimensional tables (relational databases) its provided in chapter 4, box 4.

74. It is essential that, as far as possible, there be consistency of reporting across countries. For those securities that are traded, which should by far be the majority, prices should be commonly available from stock exchanges, commercial vendors, for instance, Reuters, Bloomberg, etc., or organizations that maintain security databases, for instance, the International Securities Market Association (ISMA), Telekurs, etc. (see appendix VII). However, compilers should be aware that local practices vary. For instance, in some markets the market price quoted for securities excludes interest that has accrued but that has not yet been paid. In other markets the reverse is true. The recommended treatment in BPM5 is to regard interest accrued but not paid as part of the financial liability for the debtor and, hence, to be included in the market price of the security (see BPM5 paragraphs 121 and 396).

75. Where market prices are not available for securities (e.g., in private placements and in organized exchanges where they are not actively traded), valuing securities may create difficulties for national compilers. One approach is to request respondents to provide a market value. For instance, in France—which has a security-by-security approach—custodians reporting security holdings of clients are instructed to value listed securities at stock market prices on the reference date or at the price chosen for the annual valuation of the respondent’s portfolio for unlisted securities. In Australia—which has an aggregated approach—respondents are given a range of possibilities with which to value both equities and debt; the order of preference is based on the recommendations of BPM5 (including the method of valuing securities issued at a discount, which is outlined in paragraph 396).

76. Of course, survey respondents may face the same problems that compilers face in obtaining good information on market valuation. Therefore, when compilers are relying on respondents to provide information on the market value of securities, it is important that some method of quality control be introduced. When survey respondents report on a security-by-security basis, compilers might be advised to request additional information: the quantity of each security held and the face value of the security and/or the price used to value it. With this information, compilers could confirm that the price used looks appropriate, either by using their own knowledge or by comparing prices quoted for comparable or the same securities by different or the same respondents.

77. Alternatively, survey respondents could be asked to report individual securities at the nominal value, allowing the compiler to convert to market value. For infrequently traded securities, the compiler could estimate a market price using information on comparable securities. For example, the compiler could use the market price of a security that closely approximates the tenor, risk, and payments characteristics of the unpriced security. Or the compiler could use a price index. The index could vary in complexity between that based on similar maturities in the same currency or that which identifies individual issues that mirror the characteristics of the unpriced security. One approach, using the price index, could be to use the broad-based indices for a number of major currencies (U.S. dollar, yen, deutsche mark, French franc, sterling, and Swiss franc) available from organizations such as the ISMA. If national compilers intend to create their own price index from information already available, then they need to carefully consider technical complexities.

78. Regarding those respondents who report on an aggregate basis, the compilers’ scope for independently verifying that securities have been correctly valued is more restricted. If the data are being collected within an integrated system, the compiler could, as in Australia, request that the reporters reconcile transactions and position data on the same survey form; alternatively, the compilers themselves could, as in the United Kingdom, reconcile transactions and position data (see chapter 4, section VI and appendix VIII for more information on quality control).

Valuation of Asset-Backed Securities

The vast majority of asset-backed securities are U.S. securities backed by pools of residential mortgages. At year-end 1995, there were approximately 800,000 such securities outstanding, with nonresident ownership estimated at approximately $100 billion.

These securities present a special problem, given that there can be partial repayments of principal at any time. Because the market price is a function of market interest rates and perceived creditworthiness, and not the amount of principal outstanding, then simply revaluing the original face value to end-period market prices will cause an overvaluation of position data if there has been a partial repayment. The compiler need only revalue the face value of the remaining amount outstanding—the so-called “factor value.” If the original face value of a security was $1,000,000, and $700,000 of the principal has been repaid, the “factor value” is 300.

Problems have arisen—at least when collecting data from U.S. custodians—because some custodians keep track of only the original face amount of principal outstanding on the security, whereas other custodians keep track of only the remaining principal outstanding. Thus, if collecting data on a security-by-security basis from custodians, compilers are advised to ask for both the original and remaining principal outstanding; if only one is available, the compiler should at least be clear as to which has been reported, so that the appropriate adjustments can be made. If data are collected on an aggregate basis, prior consultation with major respondents is recommended in order to decide upon the appropriate course of action.

79. Countries that intend to conduct an aggregate survey, particularly a first-time survey targeted at custodians, are advised to gain some idea of the reported value of securities not revalued to end-period market prices. One approach is to ask respondents to report this information on the survey form. For instance, in the model survey forms in appendices III and IV, this information could be collected at the end of forms 2a and 2b, as an “of which” item of the total value of all equity and long-term debt securities reported. These figures might provide only a rather rough indication of the size of the problem in value terms—experience suggests that for rarely traded equities the number rather than the value owned could be reported—but they would help compilers identify, contact, and, perhaps, directly assist those survey respondents who have the most difficulty revaluing securities.

V. Distinguishing Between Direct and Portfolio Investment in Securities

80. The objective of the Coordinated Survey is to collect data on domestic residents’ investment in nonresident securities for use in the compilation of portfolio investment data. If the securities held are considered to be direct investment securities, then they should be excluded. Hence, not all holdings of nonresident securities come under the auspices of the Coordinated Survey.

81. The definition of direct investment is in Chapter XVIII of BPM5 and the OECD Detailed Benchmark Definition of Foreign Direct Investment. A direct investment relationship is established when a resident in one economy owns 10 percent or more of the ordinary shares or voting power (for an incorporated enterprise) or the equivalent (for an unincorporated enterprise). Direct investment comprises not only the initial transaction establishing the relationship between the investor and the enterprise but also all subsequent transactions between them and among affiliated enterprises, both incorporated and unincorporated (BPM5 paragraphs 359 and 362).

82. Therefore, both the equity held that establishes the direct investment relationship, and any other equity or debt that is issued by the direct investment enterprise or the direct investor or its affiliates and owned by them, should be excluded from the Coordinated Survey.17 In practice, the method by which they are excluded may well depend on the general collection method adopted.

83. In an end-investor survey, the reporting instructions should be clear as to what should be included and what should be excluded. Because end-investors are best informed about the composition of their investments, they should be able to distinguish portfolio from direct investment assets. Indeed, as direct investment statistics themselves are commonly collected through a survey of investors, the reporting instructions can make clear to respondents that securities reported as direct investment should not be reported as portfolio investments. The separate reporting of direct and portfolio investment data is the approach in Japan. The compiler can also compare the direct and portfolio investment returns for each respondent as an additional check.

84. In a custodial survey, the instructions should also clearly state that direct investment securities should be excluded. However, the applicability of these instructions might well vary among countries. The experience among Task Force members is that, in some markets, investors tend to hold all securities with custodians, whereas in others they hold direct investment securities separate from their portfolio investments and away from custodians. It is recommended that national compilers conducting a custodian survey evaluate market practice in their own country.

85. In markets where investors tend to hold all securities with custodians, compilers collecting data for the Coordinated Survey from custodians will need to be particularly vigilant in ensuring that securities that should be classified as direct investments are excluded. This could take the form of monitoring particularly large holdings of securities and comparing data from direct investment survey returns with those from the Coordinated Survey. Alternatively, as a quality check, direct investment securities could be identified separately as a memorandum item. Box 3 illustrates how Austria ensures that direct investment securities are excluded from its survey of portfolio investment.

86. Finally, national compilers—whether they approach end-investors or custodians—need to be careful when classifying securities issued by entities based in offshore centers. If held by the parent enterprise or associated affiliate, these securities should be classified under direct investment, unless the sole purpose of the offshore entity is to serve in a financial-intermediary capacity. In that case, only those securities associated with permanent debt and equity are to be regarded as direct investment securities (see BPM5 paragraph 365).

VI. Treatment of Securities Where There is Potential for Double Counting

87. In this section, four types of transactions of growing importance, which could potentially lead to significant under- or double counting of securities holdings between countries, are identified. These types of transactions are (1) repurchase agreements, (2) securities lending agreements, (3) depositary receipts, and (4) stripped securities.

Austria’s Experience in Excluding Direct Investment Securities from the National Survey Results

In Austria, custodians provide the Oesterreichis-che Nationalbank (ONB) with information on individual securities at par value (see chapter 2, section III). The custodians are not able to distinguish between portfolio investment securities and securities held for direct investment purposes. Hence, all securities are included in the report by custodians,18 and the data need to be adjusted by the ONB using an estimate of direct investment securities held with domestic custodians.

As in many countries, the data on outward direct investment in Austria are based on reports by the domestic direct investors. The ONB, using these direct investment reports, chooses a sample of domestic companies—based on their size, the type of industry, and the geographical breakdown of their investment abroad, etc.—and requests that they provide information on the custody of their direct investment equity securities. When this exercise was last conducted in 1992, domestic companies accounting for about 65 percent of the total outward direct investment, reported. The results, as a percentage of the total nominal direct investment capital, were:

article image

It is assumed that the sample is representative of all domestic companies that held direct investment securities. Consequently, 28.3 percent of the total amount of the nominal capital of outward direct investment—the amount of equities held in custody with domestic banks—is deducted from the custodian reports on domestic residents’ holdings of nonresident equity securities.

18The same procedure has been adopted by France.

A. Repurchase Agreements


88. A repurchase agreement (repo) is an arrangement involving the sale of securities at a specified price with a commitment to repurchase the same or similar securities at a fixed price on a specified future date. A reverse repo is the same transaction seen from the other side, that is, an agreement whereby a security is purchased at a specified price with a commitment to resell the same or similar securities at a fixed price on a specified future date. There is more than one type of repo. In some repo transactions, ownership of the underlying security remains with the seller; and in some others, the ownership of the underlying security changes hands. The underlying security may be held by a third party for safekeeping.

Potential for double counting

89. Repurchase agreements were developed primarily as a method of financing security positions, but they have also become a method by which investors can borrow securities to cover “short positions” (and so are equivalent in economic terms to securities lending arrangements). Regardless of the type of repo transaction, the market risks of owning the security, and all significant benefits, remain with the “seller.” Hence, accounting practice in many countries is to regard the securities involved in repo transactions as remaining on the books of the seller and to record the cash received as a collateralized loan. On the other hand, there are countries that, for institutional, legal, or other reasons, treat repos, where ownership changes hands, as transactions in the underlying security, that is, as an outright sale of the security for cash. The subsequent repurchase is regarded as a separate and independent security transaction.

90. The intention in BPM5 and the System of National Accounts (SNA) is to distinguish repos by their nature: if ownership does not change hands, then a collateralized loan should be recorded; if ownership of the underlying security changes hands, then a sale and purchase of securities should be recorded (BPM5 paragraph 418).19 However, BPM5 goes on to state that in some instances, because of legal, institutional, and other considerations, national compilers may be required to adopt an alternative treatment of repos; in such instances, this information should, if feasible, be separately identified and reported to the IMF.

91. In practice, national compilers in some countries find that they have no realistic choice other than to require data on security holdings (and transactions) to be reported in line with the accounting and/or regulatory conventions prevailing in their countries. It appears that countries rarely treat repos exactly according to the intention of BPM5: repos tend to be treated either as outright sales (and purchases) of securities (in which case the “buyer” rather than the “seller” of the security should report the security holding) or more commonly, particularly among those countries with the largest repo markets, as collateralized loans (in which case, the “seller,” not the “buyer,” should report the holding).20

Treatment of repos in the Coordinated Survey

92. Against this background, the Task Force recognized that a practical approach should be adopted regarding the reporting of securities holdings underlying repos.

93. Because many repos are conducted between residents within the same (creditor) country, even if the underlying security is the liability of a resident of another (debtor) country, it is vital that countries participating in the Coordinated Survey be internally consistent in their reporting approach. In this way, countries’ claims on the debtor country will be accurately measured regardless of the approach taken to repo holdings. If the treatment of repo holdings is not consistently applied among all respondents within a creditor country, then claims on the debtor country will be inaccurately measured. Box 4 explains the reasoning behind these conclusions. However, ensuring internally consistent instructions is not necessarily straightforward, particularly for those countries that conduct a custodian survey but treat repos as collateralized loans. They will, for instance, need to capture “short positions”; otherwise, the security holdings could be double counted (again see box 4).

94. As a first step, national compilers are encouraged to obtain some idea of the importance of the repurchase market in their own economy. If it is significant, then national compilers should gain an understanding of the market and discuss possible reporting approaches with potential respondents. For those countries that collect data through an aggregate survey, the respondents should be required to report in line with the instructions provided, for instance, usually in line with accounting and legal requirements. Countries that collect data from custodians will need to discover whether their custodians are able to separately identify securities held under repo agreements when reporting clients’ holdings. (This should not be an issue with regard to their own account holdings.)

Statistical Treatment of Repurchase Agreements21

Explaining the possible statistical entries arising from repurchase agreements can soon become immensely complicated, not least because there are frequently three variables to consider—the two counterparties to the repo transaction and the issuer of the underlying security—or four variables if the security is deposited with a custodian for safekeeping.

Provided both the counterparties are domestic residents and provided their statistical treatment of repos is identical (that is, national compilers ensure that all respondents have the same instructions), then claims vis-à-vis the country of issuance of the underlying security remain unchanged. The reason for this conclusion is as follows:

For example, if resident X of country A repos 100 security liabilities of country B to resident Y of country A, then regardless whether X maintains this holding on its books, and Y does not, or vice versa, country A has a claim of 100 on country B, that is, only one entity in country A has a reported claim on country B. If, however, resident X treats repos as a collateralized loan, and resident Y treats repos as a purchase of securities, then country A’s claim on B will be incorrectly measured at 200, that is, both X and Y report a claim on country B. Similarly, if resident X treats repos as a sale of securities, and resident Y treats repos as a collateralized loan, then country A’s claim on B will be incorrectly measured at zero, that is, neither X nor Y report a claim on country B.

However, holdings data may not necessarily be “correct” even if both X and Y treat repo transactions in a consistent manner. Assume that X and Y both treat repos as collateralized loans, but Y onsells the securities to fellow resident Z, not as a repo. Y’s sale of the security may be accurately reported in the transactions data, but given that Y has sold a security that was not on its balance sheet, then Y has a “short” or negative position. If country A’s collection system cannot record negative positions, then country A’s claim on country B will again be incorrectly measured at 200, because both X and Z will report claims on country B, even though only one claim exists.

The potential reporting possibilities become more complex if one of the counterparties is a nonresident (see table that follows). From the viewpoint of the Coordinated Survey, the residency of the counterparty to the repo transaction is not directly relevant. It is rather the residency of the issuer of the security used for the repo that is relevant, because the Coordinated Survey is measuring security claims on nonresidents. However, the residency of the counterparty is of relevance to the extent that the treatment of the securities underlying repos differs between countries. To the extent that national compilers adopt different treatments, there is potential for under- and over recording at the global level for the same reasons as explained above in the domestic context.

The same considerations as apply to repos also apply to securities lending.

21This box draws on work produced by Statistics Canada.

95. Second, it is recommended that national compilers investigate the extent to which residents conduct repo transactions with nonresidents in nonresident securities. If the extent is significant, then national compilers should, if feasible, consult their counterparts in the appropriate foreign country(s) to discover whether both countries are treating repos in a consistent manner. Different approaches can lead to asymmetric reporting of securities holdings; given the size and growth of the repo market in the 1990s, this could cause a problem when data are exchanged. Bilateral swaps of additional information might be required to meet both countries’ needs.

96. Nonetheless, national compilers should be aware that the clear majority of the countries represented on the Task Force, and, in particular, most of those with the largest markets, in practice treat repo transactions as collateralized loans. In the model survey forms—appendices II to IV—repo transactions are, consequently, treated as collateralized loans.

97. Finally, information on how countries are intending to treat “repo” holdings should be supplied to the IMF (along with other documentation regarding the conduct of the survey at the national level [see paragraph 18]), so that inconsistencies in approach between countries can be monitored and, if need be, brought to the attention of the relevant national compilers.

Table. Repurchase Agreements: Examples of International Position Data Reporting Possibilities

article image

B. Securities Lending


98. Security lending is an arrangement whereby the ownership of a security is transferred, usually in return for collateral such as another security, under the condition that the security will revert to its original owner on a specified future date.

Potential for double counting

99. Securities lending is an activity usually associated with dealers needing to meet commitments to deliver securities that they have sold but do not own, that is, covering a “short” position. Basically, a securities loan, in the absence of cash collateral, involves swapping securities: the borrowed securities for other acceptable securities. As with repos, the market risks of owning the securities involved in a securities loan (and all significant benefits) remain with the original owners. (Indeed, repos and securities loans are equivalent in economic terms.) In all cases, the ownership of the securities changes hands.

100. The development of systems to collect data on securities lending is less advanced than for repurchase agreements, and the size of the market is considerably smaller. Unlike for repos, the methodological treatment of securities lending is not covered in either BPM5 or the SNA.

101. In some countries, the statistical practice is to regard the securities (both those lent and those provided in return as collateral) as remaining on the books of the original owner. If cash is given as collateral, the transaction is regarded as a “collateralized loan.” If a borrowed security is on-sold (which is usually the case), then both the original owner and the third party will regard itself as owning the security, while the “borrower” will regard itself as having a negative holding. In other countries, the practice is to regard securities lending as a transaction in its own right: the lent security becomes owned by the “borrower,” and the collateral, either cash or security, becomes owned by the “lender.” If a borrowed security is on-sold, then the borrower no longer holds it and the sole owner is the purchaser. As with repos, there is the possibility of asymmetric reporting among countries.

Treatment of securities lending in the Coordinated Survey

102. As with repurchase agreements, a practical approach should be adopted regarding the reporting of securities holdings underlying securities lending. The issues are similar, and the importance of internal consistency of approach equally valid. Therefore, the recommendations with regard to repos apply equally to securities lending: the importance of the market should be assessed; internally consistent reporting instructions, developed after consultation with potential survey respondents; the scope for the lending of nonresident securities to nonresidents, investigated; bilateral contacts, established with national compilers in the appropriate countries; and information on the intended treatment, sent to the IMF.

C. Depositary Receipts


103. Depositary receipts are securities that represent ownership of securities held by a depositary.22 Usually the country of residency of the issuer of the underlying securities is different from the country in which the receipts are issued.

Potential for double counting

104. The issuance of depositary receipts is becoming increasingly prevalent: issuers and intermediaries are finding that investors frequently prefer to acquire securities in the financial markets where payment and settlement systems, registration procedures, etc. are familiar, rather than in the home market of the issuer, where systems and procedures may not be so familiar or as efficient. The potential for double counting lies in the existence of both the underlying security, held by the depositary, and the depositary receipts. That is, two securities could be reported as held, but only one liability exists. Clear guidance to survey respondents is required in order to avoid double counting.

Treatment of depositary receipts in the Coordinated Survey

105. The reporting should follow the guidelines below:

  • Depositary receipts are certificates that represent ownership of securities held by a depositary. These receipts should be allocated to the country of residence of the issuer of the original (or underlying) security and not to the residency of the financial intermediary that issues the receipts. In other words, American depositary receipts (ADR) are liabilities of the non-U.S. enterprise whose securities underlie the ADR issue and not of the U.S. financial institution that “issued” the ADRs.

  • Financial intermediaries should not report holdings of any nonresident securities against which depositary receipts have been issued and sold.

  • If a depositary receipt has been issued before the financial institution arranging the issue has acquired the original (or underlying) securities, then that financial institution should report a negative holding in the original (or underlying) securities.

D. Stripped Securities


106. Stripped securities (strips) are securities that have been transformed from a principal amount with periodic interest coupons into a series of zero coupon bonds, with the range of maturities matching the coupon payment dates and the redemption date of the principal amount. Strips can be created in two ways. Either the owner of the original security can ask the settlement or clearing house in which the security is registered to “create” strips from the original security, in which case the strips replace the original security and remain the direct obligation of the issuer of the original security; or the owner of the original security can issue strips in its own name, “backed” by the original security, in which case the strips represent new liabilities and are not the direct obligation of the issuer of the original security.

Potential for double counting

107. The issuance of stripped securities is becoming more prevalent: the creation through strips of zero coupon bonds eliminates the reinvestment risk inherent in a bond with a coupon and allows investors greater leverage; that is, less money needs to be invested to gain the same market exposure as with an equivalent bond with a coupon. When the entity issuing the strips is creating new liabilities, double counting does not arise. The potential for double counting arises when the strips have replaced the original security although the latter has not been redeemed. In essence, the original security is “dormant” in the settlement or clearing house, until such time as it is reconstituted or redeemed.

Treatment of stripped securities in the Coordinated Survey

108. The reporting instructions should follow the guidelines below:

  • If strips have been issued by an entity in its own name, then the residency of the issuer is that of the entity that has issued the strips, and the issuing entity should report its holdings of the existing nonresident securities.

  • If strips have been created from a nonresident security and remain the direct obligation of the original issuer, then the residency of the issuer remains the same as for the original security.

  • Dealers who request that a settlement or clearing house create strips from an existing nonresident security should not report their holdings of the underlying nonresident security once the strips have been created.

  • Strips with an original maturity of less than one year are money market instruments. If identifiable, they should not be reported.