Discussion Topic: Swap Data Repository and Data Standards

Published: February 25, 2014

Based on the lessons of the recent financial crisis, the Dodd-Frank Act introduced a new requirement that derivatives trades be reported to centralized data warehouses called swap data repositories (SDRs). This requirement holds out the promise of bringing transparency to derivatives markets for regulators and market participants. Standardized data on derivatives exposures will also allow regulators and market participants to better measure exposures and risks throughout the financial system.

However, to date, the data collected in SDRs are fragmented and inconsistent. Data are reported differently across repositories (even within the U.S.) to the degree that data are difficult to add up and compare. These problems are even more pronounced across jurisdictions.

As noted by Mary Miller in her opening remarks at our recent conference, “We need to roll up our sleeves and address any obstacles to making these data useful for market participants and for regulators who are monitoring financial stability. We need to work with the industry and the repositories to establish standards for reporting, so that data can be aggregated and analyzed — in other words, so we can develop a more holistic picture.”

CFTC Commissioner Scott O’Malia has regularly noted the current shortcomings in SDR data and the current inability to compare data on cleared and uncleared derivatives in order to get a view of bilateral portfolio exposure, assets concentration and contagion, and correlated risk exposure.

OFR has begun to work with the CFTC to help resolve these problems. It is clear that using industry best practices can help ensure that data collections in SDRs provide high quality, consistent data that can be compared and aggregated. Data collection specifications that are comprehensive, thought out, and defined well form the bedrock of the data collection effort.

Questions for the Committee to consider in order to assist in this work:

  1. Regulators place high value on consolidating data through SDRs. What are the benefits for market participants? How can regulators help incentivize market participants to provide high quality data to SDRs?

  2. One definition of Data Quality is that ‘the data are fit for purpose.’ How complete, timely or accurate do SDR data need to be to be fit for financial stability analysis purposes? Is this different from risk management purposes?

  3. Are there incentives for the reporters to make the data ambiguous or inaccurate? If so, are there strategies, such as requiring granular data, which help avoid misreporting?

Additional Questions for consideration by the Data and Technology Subcommittee

  1. SDRs represent collection points with multiple entry points for trade data. What are the risks and mitigation strategies for collecting data through these Financial Market Utilities?

  2. There are several identification standards necessary for SDRs. The identifiers are detailed in Exhibit 1 and the industry best practices are listed in Exhibit 2.

  1. What is the best way to incentivize industry to build and maintain identifier and content standards?
  2. Do the best practices in Exhibit 2 all apply to each of the information standards noted in Exhibit 1?
  3. Are there additional best practices for all or any of the standards in Exhibit 1?
  4. Would using a compression algorithm (such as hashing) introduce any problems?
  5. Are you familiar with any existing identification standards that would meet the needs of these identification standards (Swap/Product/Transaction/Instrument)?
  6. Is it important to ensure, to the extent practicable, that the identifiers are compatible with financial market participants' existing automated systems?
Exhibit 1: Identification standards

Identification standards required by the CFTC’s rules include:

  • Unique Swap Identifier (USI) – § 45.5 - A unique identifier for every swap transaction. This allows for elimination of duplicate transactions where trades are reported by both counterparties.
  • Legal entity identification (LEI) – § 45.6 - This need is being met by the LEI initiative.
  • Unique Product Identifier (UPI) – § 45.67 - A unique identifier for categories of products such as stock, bond, convertible bond.

Identification standards not required by the CFTC’s rules:

  • Unique Transaction Identifier (UTI) – Modeled by ISDA after the USI to cover OTC derivative transactions not covered by USI.
  • Unique Instrument Identifier (UII) – Existing commercial and international standard identifiers do not cover the full universe of instruments.
Exhibit 2: Best practices for identification standards
  • Singularity and Uniqueness. Only one identifier should be assigned to any Entity/Swap/Product/Transaction/Instrument, and no identifier should ever be reused.
  • Persistence and Neutrality. An identifier should follow an identifier through its life regardless of corporate actions or other business or structural changes. Persistence is important not just because it reduces the need to research changes but also because it reduces errors in analysis. For an identifier to be persistent over time, it should be neutral and not contain any embedded intelligence.
  • Reliability and Interoperability. For an identifier to be widely accepted it must be reliable and interoperable. The identifier should be supported by a trusted and auditable method of verifying the identity of the Entity/Swap/Product/Transaction/ Instrument to which it is assigned, both initially and at appropriate intervals thereafter.
  • Open Source and Public Availability. The identifier should be an open standard and publicly available without any license restrictions.
  • Extensibility. The identifier should be capable of becoming the single international standard for unique identification of an Entity/Swap/Product/Transaction/Instrument across the financial sector on a global basis. Therefore, it should be sufficiently extensible to cover all existing and potential future Entities/Swaps/Products/ Transactions/Instruments.