The Basel Committee has issued a long awaited consultation paper with proposals to revise some elements of the January 2016 market risk framework.

The main elements are:

  • A recalibration of the standardised approach (SA) to market risk;
  • Major changes to the profit and loss attribution (PLA) framework, which would make adoption of the internal models approach (IMA) a more attractive option for more banks as fewer trading desks would be likely to suffer the cliff effect of dropping to the SA;
  • No significant changes proposed to Non-modellable Risk Factors (NMRF); and
  • The Basel Committee is seeking comments by 20 June 2018.

Implications for banks

Overall, banks are likely to welcome these sensible and measured proposals. There are no major surprises, and the proposals continue to provide an incentive for banks to adopt an internal models approach.

Banks on the standardised approach could see a reduction in capital requirements for market risk of 15-30 percent.

Banks using the IMA will benefit from the proposed revisions to the PLA framework, whereby less serious failures of the PLA test will not require trading desks to switch from the IMA to the SA.

Banks that have focused on foundational aspects first (such as data alignment, data granularity, migration from Monte Carlo to historical simulation etc.) will not have to re-engineer much as a result of these proposals. However, banks that have a prototype SA/IMA calculator will have to tweak their calculators in order to align to the new requirements.

NMRF will remain an area of concern for some banks, because the Basel Committee has not proposed any significant revisions here.

Detail of the proposals

1. Revisions to the standardised approach

The Basel Committee proposals would:

  • Reduce the risk weights applied for interest rate risk (~20 - 40 percent), equity risk and foreign currency risk (~25 - 50 percent). No specific revisions to risk weights are proposed for credit spread and commodity risk classes. The final calibration will be based on further analysis of impact data and responses to the consultation. The Basel Committee may also consider changes to the risk weights used in the standardised approach to credit valuation adjustment risk (SA-CVA), since these were based on the January 2016 standards.
  • Permit liquid currency pair triangulation and treat the resulting new FX pair as liquid (this revision will also apply to the internal models approach) and therefore subject to lower risk weights.
  • Update the low correlation scenario in order to address the current overly conservative treatment.
  • Amend the treatment of non-linear financial instruments such as options:
    1. Apply consistent scenarios to risk factors defined in the same bucket for credit spread, equity and commodity risk classes (the consultation is also seeking feedback on an alternative approach that defines sectors as a subset of each bucket). The current approach results in two closely related financial instruments being treated differently due to the worst result being taken from two shocked scenarios.
    2. Introduce a floor to address circumstances where the aggregate curvature risk creates a cliff effect. The current approach can result in an abrupt increase in capital requirements, especially when the curvature risk positions are negative.
    3. Seek further feedback from the industry on the potential double-counting of FX curvature risk when neither of the FX currency pairs is the same as the bank's reporting currency. If this is found to be material then it is proposed to divide the curvature sensitivities by a scalar for all instruments where none of the underlying currencies is the reporting currency. 

All of these proposals would reduce SA capital requirements, bringing them closer into line with the Basel Committee's original intentions. The reduction in risk weights will translate directly into similar reductions in the associated risk capital charges, while KPMG experts reckon that allowing liquid currency pair triangulation will significantly increase the number of liquid pairs from the original 23 to upwards of 190. This will reduce the corresponding risk capital charge because the risk weight for liquid pairs is approximately 30 percent lower than for illiquid pairs.

However, some issues are not addressed, such as the over -capitalisation of a delta-hedged portfolio that is in the money with long gamma.

2. Revisions to the internal models approach

(i)  Profit and Loss Attribution (PLA)

  • Banks may align their input data for the Hypothetical and Risk-theoretical profit and loss calculations, subject to conditions that must be met in order to ensure that the risks are adequately captured.
  • Amend the PLA test metric by replacing the variance and mean ratio tests with the use of the Spearman correlation coefficient and the Kolmogorov-Smirnov (KS) or Chi-squared test respectively. The test would be conducted on a quarterly basis instead of monthly, and time series data for 12 months would be used rather than using just the previous month.
  • A modified traffic light approach to mitigate the cliff effect of trading desk failure. An `amber' category would be introduced whereby less serious failures of the PLA test would result in banks still being able to use an internal model approach for the respective trading desk but with the application of a capital surcharge the application of a capital surcharge based on the trading desk's relative materiality and a multiplier of 50 percent. Banks would still have to pass the 10 percent minimum capitalisation test. The proposed thresholds are subject to monitoring before finalisation to ensure they are robust and appropriate.

Based on the proposed modifications to the test, it is likely that more trading desks will pass this test compared to before, increasing the incentive to adopt an internal model approach.

(ii) Non-modellable Risk Factors (NMRF)

The Basel Committee has clarified the definition of `representative' real price observations in the context of non-modellable risk factors and the use of data for internal model calibration. There are no `hard' methodology changes proposed, although the Committee seeks further compelling evidence from the industry on various items such as:

  • Seasonality factors impacting the modellability criteria;
  • Recognition of diversification benefits for idiosyncratic equity risk; and
  • Any design flaw in the NMRF approach that results in disproportionately high capital impact for some risk factors.

In the absence of compelling evidence, the Committee will not make any changes to the treatment of NMRF.

Since the industry has already had two years to lobby on the methodology for NMRF it seems unlikely that any changes will result from this further consultation. Given the flexibility with the choice of data and considering data-pooling schemes, it may be better for banks to look into data vendor solutions to progress further on this front.

3. Introduction of a simplified standardised approach

The Basel Committee published a consultative document in June 2017 to propose a simplified alternative to the standardised requirements based on two approaches:

  1.  A reduced form of the January 2016 standard's sensitivities-based method; or
  2. A recalibrated version of the Basel 2 standardised approach.

The Committee now proposes to adopt the second approach, by recalibrating (applying multipliers to) the Basel 2 standardised approach capital requirements for each risk class of market risk. This is intended to deliver slightly higher capital requirements than under the January 2016 `full' standardised approach. Finalisation of these parameters are subject to further analysis.

4. Other revisions

1. Treatment of structural FX positions

Scope revisions to include the FX exemption on the FX risk stemming from the investment rather than the amount of the investment itself, and to include structural FX positions in foreign branches in the exemption.

2. Boundary between trading and banking book

Clarification of some details of the boundary between the banking book and the trading book, including equity investments in funds; and clarification that banks should use the mandatory banking book list as the starting point to classify positions.

3. Trading desk requirements

Some revisions to assigning a single head trader per trading desk, since this may not be appropriate for all banks. A trader may be assigned to a maximum of two trading desks subject to supervisory approval but this must not be done with the intention of optimising the likelihood of success in the PLA test. This will provide more flexibility in the way trading desks are established and result in enhanced resource allocation within the bank.

5. Frequently Asked Questions (FAQs)

The FAQs published by the Basel Committee provide further clarification on points raised by the industry but do not propose any significant changes. The table below summarises these clarifications.

No. Topic Summary
1 IMA - Expected Shortfall
  • P&L can be simulated from full revaluation or sensitivity-based valuation.
  • Banks must determine policies to assess material changes in risk factors of the portfolio. Regulators will assess the concepts developed and implemented by banks.
2 IMA – Liquidity Horizon (LH)

LH for a risk factor that has maturity lower than prescribed is assigned the LH floor that is equal to or longer than the maturity (10, 20, 40, 60, or 120). E.g. if maturity is 30 days, and prescribed LH is 60, then apply LH 40 (LH ≥ maturity).

With multi-asset class indices (i.e. equity-credit):

  1. Compute weighted average LH
    • Multiply LH of each instrument/component by its weight in index
    • Sum across all instruments
  2. Use weighted average LH to determine final LH value, using equal to or longer rule. E.g. if weighted average LH is 12, final LH is 20.
3 IMA- Default Risk Charge (DRC) Bank may not use simplified modelling approach for credit derivative positions with multiple underlyings (but may use them for equity multiple underlyings).
4 Backtesting Volatility scaling that results in shorter observation periods is not allowed, except to reflect recent stress periods. Further use of the scaled data requires notification to the supervisor.
5 Sensitivity-based Approach
  • Banks must use the prescribed vertices for calculation, instead of internal ones.
  • For an option with no specified maturity, banks must use the longest prescribed maturity vertex and apply RRAO.
6 Residual Risk Add-on (RRAO) Bonds with multiple call dates are believed to bear other residual risks for RRAO.
7 FX Risk Non-deliverable currencies are treated in the same way as the deliverable currencies.
8 SA – DRC
  • When computing DRC, convertible bonds cannot be the same way as vanilla bonds, due to the optionality. Otherwise, the jump-to-default risk would be under-estimated.
  • Bond B guaranteed by a different issuer A is permitted to offset with another bond issued by issuer A only if the guaranteed bond satisfies the credit risk mitigation (CRM) requirements in the Basel 2 framework.
9 Trading Book Boundary
  • Instruments cannot switch between books even though they are reclassified for accounting purpose. Thus delisting of an equity cannot enable it to switch books.
  • Continuous monitoring and managing of net short credit or equity positions in banking book (BB) - frequent reporting will be required in order to monitor and manage.
  • Instruments having embedded derivatives issued from BB are required to assign the embedded derivatives portion to the trading book (TB). For example, Bond options, Cap/Floor etc.
  • External hedges can be made up of multiple counterparties as long as that exactly matches the internal risk transfer.

To discuss the implications further, please contact Rob Smith (, Jalpa Dodhia ( or Clive Briault (

Further insights can be found in the article series ‘Basel 4: The way ahead’, which delves into the impact of the recently finalised Basel 4 standards.

Visit Risk and regulatory challenges for regulatory insights.

Connect with us

Also on