As the regulators consult on and the industry prepares for the latest incarnation of market risk rules — Fundamental Review of the Trading Book (FRTB) — we explore some of the implementation challenges the industry has ahead of itself.
- The Risk Factor Eligibility Test is being introduced to mitigate the possibility of internal models underestimating risk in illiquid markets.
- The first hurdle for any bank building a RFET compliance system will be in sourcing the ‘real’ observation data.
- The extensive scope of MiFID II trade data could mean it acts as an important source of ‘real’ observation data.
In March this year, the Basel Committee on Banking Supervision (BCBS) released a consultative document that proposed a number of amendments to the original Fundamental Review of the Trading Book (FRTB) standards published in January 2016.
This provided the industry with an opportunity to give feedback on the latest proposals.
With the consultation now closed and the responses available to the public, we have taken the opportunity to mull over some of the challenges the industry now faces.
This article is the first in a mini-series that explores some of the difficulties associated with acquiring ‘real’ observation data and seeks to understand the potential problems when attempting to conduct the Risk Factor Eligibility Test (RFET) assessment on portfolios.
This article looks at the criteria for the RFET and whether MiFID II could help banks with their sourcing of ‘real’ observation data.
Watch: Is FRTB a priority for banks?
What is the Risk Factor Eligibility Test?
RFET is being introduced to mitigate the possibility of internal models underestimating market risk in illiquid markets.
The logic goes like this:
Banks use historical time-series data as inputs into their internal market risk capital models.
For those risk factors where the associated instruments trade infrequently, the time-series data may become stale (since no new information is available), which causes the portfolio’s P&L to tend to zero.
The internal model mistakenly concludes the level of market risk is low, when what’s really being observed is a lack of activity in the market.
In order to protect against this risk, the BCBS is introducing RFET, which dictates the rules under which market risks can be modeled.
It mandates banks to collect executed trade and committed quote data, which they must map to risk factors and then conduct two checks:
- Ensure there are at least 24 observations per year.
- Ensure that there isn’t more than a month between any two consecutive observations.
Sourcing ‘real’ price observations
The first hurdle any bank building a RFET compliance system will come across is sourcing the ‘real’ observation data.
Irrespective of the bank’s portfolio profile, this will be challenging because it requires consolidation of data from so many sources.
Nevertheless, it will be substantially easier in highly transparent exchange traded markets, where instruments often trade frequently and executed trades are published, than in OTC markets that have often been characterized by limited transparency and individual instruments trading sporadically.
One of the key initiatives that could make sourcing ‘real’ observations for European OTC markets significantly more achievable is the introduction of MiFID II.
From January, MiFID II mandated that bond and derivatives trades conducted in Europe either executed on regulated platforms or bilaterally would be subject to post-trade transparency — i.e. the individual trades need to be published to the market.
For banks with European operations, the extensive scope of MiFID II trade data could mean it acts as the pre-eminent source of ‘real’ observation data and significantly reduces the burden of the RFET compliance obligations.
So with MiFID II data potentially acting as a key source of ‘real’ observations, we set out to get our hands on it in order to see whether it lived up to our expectations.
Bond data observations
To keep life simple, we limited our analysis in a number of ways. Firstly, we’ve only focused on bonds (we’ll leave derivatives for another time).
Secondly, the dataset is based on contributions from two of the more popular providers — Tradeweb and Trax.
These two sources only provide a partial view of the market. The obligation to publish trade data sits with a wide variety of regulated venues and reporting platforms.
Thirdly, the data covers trades executed during the first three months of 2018, since the regulation only came into effect at the start of this year.
What did we learn about FRTB?
The image above shows information on what is published and when. Ultimately, the data teaches us two lessons:
- There’s a lot of data. During Q1 2018 we observed over a million trades executed in 36,000 bonds (and remember this is only part of the market).
- Deferrals are significant and widely adopted: 87 percent of the bonds were subject to a maximum deferral length of four weeks or longer.
We find this initial review of MiFID II data very encouraging, as we know from the structure of MiFID II that our sample of data is only part of the market and that we need to consolidate many more sources before we have complete coverage.
However, we are cautious about the extensive use of deferrals. This is a systematic feature of the regulatory regime and not a feature specific to the first few months as firms get to grips with the MiFID II compliance requirements.
We encourage regulators to clarify the RFET rules and provide banks with sufficient time to collect data that’s subject to deferred publication prior to conducting RFET.
Read our full response to the consultative document, which is available on the BCBS website.
- In the next article, we explore some of the features and challenges by conducting the Risk Factor Eligibility Test on a selection of data sets covering bond and derivative markets.