Annex 4: Security-by-Security Databases
- International Monetary Fund
- Published Date:
- November 2009
What is a security-by-security database?
A4.1 A security-by-security (SBS) database is a micro database that stores statistics at an individual debt security level.48 Statistics are stored according to a range of attributes or characteristics that may vary depending on the purpose of the database. For statistical uses, the attributes may include the international securities identification number (ISIN), name of the issuer, residence of the issuer, sector and sub-sector, issue date, redemption date, type of security, currency of denomination, issue price, redemption price, outstanding amount or market capitalisation, and the coupon payments and dates (see Diagram A4.1).
Diagram A4.1Attributes of statistics stored in SBS databases
A4.2 The production of statistics from SBS databases can be presented as a three-stage process (see Diagram A4.2). The first stage typically involves the inputs by collecting and purchasing individual security statistics from a range of sources, such as central banks, government agencies, commercial data providers and securities exchanges. The second stage covers data quality management. The individual security data collected from different sources are received into the database, merged, and stored. Checks for completeness, plausibility and consistency are then performed, and where errors are detected, observations are corrected. The third stage involves storing individual security data according to various classification criteria, as discussed in Section 6.
Diagram A4.2Stages in the development of SBS databases
Benefits and costs
A4.3 When deciding whether to construct a SBS database, the full range of benefits and costs should be considered. Most of the arguments for and against SBS databases relate to the compilers of securities statistics, although respondents and users are also affected.
A4.4 One of the main advantages of SBS databases is that compilers, rather than respondents, are responsible for the statistical classification of securities. This promotes accuracy and consistency of the data, and adherence to international statistical standards. For statistical purposes, particularly in cases of statistics on securities issues, government finance statistics and institutional sector accounts, individual SBS issues data are usually aggregated according to various statistical categories. SBS databases offer the flexibility to produce different aggregates based on SBS data, with no need for any further data collection. Moreover, SBS databases allow the derivation of data on positions, transactions and other flows. SBS databases also allow quality checks at a very detailed level to detect outlier observations within specific statistical categories. Outliers may indicate a misclassification but they can also be caused by financial innovation, which would require further investigation and potentially an amendment to the statistical aggregation categories.
A4.5 SBS databases benefit respondents by reducing the amount of detailed breakdowns to be reported to compilers. Respondents no longer need to map their internal data into statistical reports and instead provide relevant information for each individual debt security in their database. A drawback for respondents is the necessity to meet data standards agreed with compilers.
A4.6 SBS databases can be extended to include information on securities holdings for resident holders grouped by sector and sub-sector, as well as for non-resident holders. For that purpose, information provided by respondents is linked at the individual security level to the data stored in the SBS database. The link is often done by using ISIN.
A4.7 From the user’s perspective, there may be interest in detailed disaggregated data or in combining different classifications, particularly as debt securities markets become more complex and globalised. SBS databases allow this decomposition in debt securities statistics. Sometimes a panel of individual securities data may be set up to analyse common developments. SBS databases also permit an analysis of the financing of different sectors, the size of different market segments or the importance of different debt securities. The databases allow users to track changes in the credit ratings, prices and liquidity of individual securities and issuers.
A4.8 At the same time, there are significant costs for compilers to set up and maintain SBS databases and adapt them to changes in users’ needs. SBS databases are largely sourced from commercial database providers. The acquisition of this information is expensive and can often be incomplete. Information technology costs for database storage and processing are significant. This reflects the complexity of SBS databases from an operational and methodological perspective, large volumes of statistics stored in them, and database management costs that shift to compilers from respondents. A minimum level of quality of data is needed, such as a full coverage of specific categories of securities. There are administrative expenses related to setting up contacts with data providers for regular reporting or to conduct surveys. Some manual intervention is also needed to cross-check corresponding data received from various data providers. Furthermore, there can be legal obstacles preventing data exchange between central banks, statistical agencies, and other authorities. There are also expenses related to finding available human, financial and information technology resources.