I am trying to generate/prove the magnitude of the over-conservativeness of the regulatory VaR (internal models) under Basel III against what a more accurate VaR would be. However, I can't seem to find what the Basel III standard way to add RNIVs to VaR anywhere on google.

Basically, I have a basic Historical simulation-based VaR (which has the usual yield curve and FX rates). I have generated RNIV (risks-not-in-VaR) that are the (a) tenor basis, (b) cross-currency basis and (c) OIS-Libor spreads.

Now, I know and can estimate the correlations of these RNIV market-data to the yield curve, and I believe an accurate way is to add these RNIVs to the base VaR using these correlations.

However, I have been challenged that I need to do simple additive of these RNIVs to the base VaR, because Basel III rules states so. I can't seem to find these explicit rules anywhere.

Any regulatory capital rules expert out there?

Kind regards

  • $\begingroup$ It is my understanding that internal models have to be approved by the regulator and usually that involves some discussion with them on your approach. If your methods satisfies some basic checks then I imagine you could convince them of its validity, But playing devils advocate, what is the risk to your correlations failing to hold true and what is the proof of its validity. What about stress tests for bad scenarios. Also what about other risks such as reset (fixing) risks? $\endgroup$
    – Attack68
    Commented May 22, 2019 at 16:26

2 Answers 2


I assume this is UK specific as RNIV is a PRA concept. You can’t recognise diversification as per the requirements which are detailed in the ss13/13: see section 2.


The EBA adapted the above framework for the banks under its supervision and called it Risk not in Model, again diversification is not allowed- please see section 7 of the below:


  • $\begingroup$ I thought RNIV is a Basel, and hence EBA concept? PRA, or FCA takes directives from EBA and hence applies RNIV to UK firm/financial institutions as well? But yes, this was what I was trying to confirm........... we (unfortunately) cannot recognize diversification even if our risk (VaR) models have been approved for IMA (internal models approach) $\endgroup$
    – Kiann
    Commented May 24, 2019 at 7:42

I think that you might also want to take a look at Basel Committee’s paper “Minimum capital requirements for market risk” that includes the new Basel III FRTB SA (in particular RRAO part) and IMA (in particular NMRF part) charge calculation methods.


  • $\begingroup$ thanks @ir7, would you be able to specify which particular section relating to NMRFs? $\endgroup$
    – Kiann
    Commented May 24, 2019 at 7:44
  • $\begingroup$ Calculation of capital requirement for non-modellable risk factors, MAR33.16. Standardised approach: residual risk add-on, MAR23. (The whole paper is a masterpiece. :) ) $\endgroup$
    – ir7
    Commented May 24, 2019 at 20:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.