Basel Committee drops fixed correlations in new trading book proposals


The Basel Committee on Banking Supervision has dropped plans to work out trading book capital requirements using regulator-set correlations – a step that is already being welcomed by some bank capital and modelling experts. The u-turn appears in revised proposals for the committee's Fundamental review of the trading book, which was published on October 31.

The idea of using fixed correlations to aggregate modelled, desk-level capital numbers was one of the most controversial elements of the first consultation paper, which appeared in May 2012. Dealers argued regulators would not be able to determine appropriate correlations and feared the result would be punitive capital requirements that were not sensitive to risk.

In the new proposals, the approach is replaced by a formula that instead tries to counteract overly generous diversification benefits. The formula appears to be almost identical to one suggested by banks in their response to the original paper.

The new approach calculates capital as the weighted sum of a bank-wide risk measure and a linear calculation by risk factor, with the latter term effectively assuming zero diversification. The weights would be determined by a parameter picked by regulators for all regions and business lines, in contrast to industry proposals that would have linked it to the performance of models at individual firms.

"I'm actually pretty happy with that. It's a shame they didn't go for tying it to model performance, I think that would have been good, but it was fairly complicated so I can see why," says Gary Dunn, senior manager of wholesale credit and market risk at HSBC in London.

The big thing is that internal model banks will get to set their own correlations. That's a relief

The revised proposals also bow to industry calls to reduce the threshold at which the expected shortfall risk measure – selected as a replacement for the current value-at-risk approach – is calculated, from 99% to 97.5%. Banks had argued that a 99% confidence level would be too volatile, as the shorter tail would be more sensitive to new data.

But other industry gripes have been ignored. Model approval will still be calculated at desk level, for example – a proposal banks fear will result in a sudden jump in capital for any desks that lose approval and are forced to switch to the standardised approach instead. The industry had suggested a way of smoothing out the resulting cliff effect, but it appears to have been discarded.

The shape of the standardised approach itself is a little clearer. In the original document, regulators had outlined two alternatives – the so-called partial risk factor approach and a more complex fuller risk factor approach. The committee has come down on the side of the former, a relatively simple risk-weighted asset-style calculation that is in keeping with the existing standardised approach, but is intended to be more risk sensitive.

It splits the trading book into buckets of assets with "similar risk characteristics", applies regulator-set risk weights to their market value, and adds them up using long-term correlation parameters that in some cases vary depending on netting requirements. All banks, including those that use internal models, will be required to calculate their standardised capital requirements to enable comparison between institutions and to ensure modelled capital is not too low.

"I think it's the sensible decision. The fuller factor model was basically a kind of internal model. And in some ways it was harder to use than our real internal model. The partial factor approach should be better for the smaller banks. But the big thing is that internal model banks will get to set their own correlations. That's a relief," says a senior risk manager at a European bank.

As expected, the committee avoided growing questions over Basel III's controversial credit valuation adjustment, and the paper also notes that no decision has been made on whether to floor modelled capital at some percentage of the standardised approach – another topic that has raised industry hackles.

The document is open for comment until January 31, 2014.

  • LinkedIn  
  • Save this article
  • Print this page  

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact [email protected] or view our subscription options here:

You are currently unable to copy this content. Please contact [email protected] to find out more.

You need to sign in to use this feature. If you don’t have a account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an individual account here: