# Risk.net podcast: DTCC’s Lind on FRTB, data pooling and NMRFs

## As many as 70 banks globally could adopt internal model approach for market risk capital

Reports of the death of internal models may have been greatly exaggerated.

Tim Lind, head of data services at the Depository Trust and Clearing Corporation, believes the largest banks with the most complex trading books will still use internal models to calculate their market risk capital requirements under the Fundamental Review of the Trading Book (FRTB).

“The number of banks that go [for the] internal model approach – the IMA – we’re thinking somewhere in the neighbourhood of 60 to 70 banks globally,” Lind says. “It is a strategy that we think the more sophisticated, more cross-asset-class type of trading banks will pursue.”

In this inaugural Risk.net podcast, Lind, who joined the DTCC in January, discusses the complexities of FRTB and the data pools firms such as DTCC are creating to help banks opting for the IMA.

Lind is leading the DTCC’s effort to create a shared data pool for banks struggling with FRTB’s costly capital add-ons for non-modellable risk factors (NMRFs). Markets where data is thin or patchy are in greater danger of falling into the NMRF category. Bloomberg and IHS Markit are working on similar projects.

The creation of these platforms has not been easy. Progress was initially slowed by dealers’ concerns over privacy, data standardisation and a lack of clarity from regulators over governance. Large banks with bigger trade datasets argue they should not have to pay to use the data they submit voluntarily to pooling services.

Some regional banks are also reluctant to cough up local market data. In March, Risk.net reported six of Canada’s largest banks were working on an internal project – known as the Canadian Data Utility – to pool their trade data in local markets. Nordic banks are mulling a similar option. Lind addresses the development of global data pools alongside smaller, regional counterparts, and how the two can coexist.

FRTB is due to come into force in January 2022.

Lind also talks about the likelihood of regulators loosening modellability requirements for risk factors, which currently specify that consecutive price observations cannot be more than a month apart. Banks have taken issue with this, citing the seasonality of some assets. Some have suggested a wider gap between consecutive observations would lower the number of risk factors that qualify as NMRFs.

Lind pours cold water on this theory. Doubling the time between real prices would not lead to a significant increase in modellability, he argues.

“When we went through data on asset classes we integrated, changing [the] gap period doesn’t have a massive or material impact in terms of the percentage of notional value that is now modellable that wasn’t previously modellable,” Lind says. “It’s not moving it by 50%, where 50% of my notional value was non-modellable before and you went to two [price observations] in 60 [days] instead of a one in 30, it is not that kind of magnitude of improvement.”

Interview by Dan DeFrancesco

Index

3:00 – Update on FRTB and data pooling

4:51 – The relationship between regional and large data pools

11:51 – Managing multiple data pools

17:05 – Landscape of vendors’ data pools

23:35 – Use of proxy data

25:38 – Potential changes to NMRF

31:00 – Internal versus standardised approaches

35:45 – Timeline for FRTB

To hear the full interview, listen in the player above, or download. Future podcasts in this series will be uploaded to Risk.net. You can also visit the main page here, and subscribe in iTunes here.