Source: https://www.prarulebook.co.uk/pra-rules/market-risk-internal-model-approach-crr/01-01-2027?p=1 · captured 2026-09-28 · checked 2026-09-28 · b65a8e4d

2201
22023.
2203
2204An institution shall not limit the validation of the risk measurement models of an institution to back-testing requirements and P&L attribution requirements, but shall, at a minimum, include the following:
2205
2206(a) tests to verify whether the assumptions made in the internal model are appropriate and do not underestimate or overestimate the risk;
2207
2208(b) own internal model validation tests, including back-testing in addition to the regulatory back-testing programmes, in relation to the risks and structures of their portfolios; and
2209
2210(c) the use of hypothetical portfolios to ensure that the risk measurement model is able to account for particular structural features that may arise, for example, material basis risks and concentration risk, or the risks associated with the use of proxies.
2211
2212[Note: This rule corresponds to Article 325bj of CRR as it applied immediately before revocation by the Treasury]
2213
221401/01/2027
2215
2216Article 325bk Calculation of Stress Scenario Risk Measure
2217
22181.
2219
2220For the purposes of this Article, the ‘stress scenario risk measure’ of a given non-modellable risk factor means the loss that is incurred in all trading book positions or non-trading book positions that are subject to foreign exchange or commodity risk of the portfolio which includes that non-modellable risk factor when an extreme scenario of future shock is applied to that risk factor.
2221
222201/01/2027
2223
22242.
2225
2226An institution shall develop appropriate extreme scenarios of future shock for all non-modellable risk factors.
2227
222801/01/2027
2229
2230Development of extreme scenarios of future shock for individual risk factors
2231
22323.
2233
2234An institution shall develop the extreme scenarios of future shock for a single non-modellable risk factor for the purposes of paragraph 2 such that the resulting stress scenario risk measure is at least as conservative as:
2235
2236(a) an expected shortfall measure calculated for that non-modellable risk factor alone;
2237
2238(b) for the stress period in accordance with paragraph 10;
2239
2240(c) at 97.5th percentile, one tailed confidence interval;
2241
2242(d) calculated with base time horizon of 10 days; and
2243
2244(e) scaled to a time horizon that is the greater of 20 days and the liquidity horizon of that non-modellable risk factor in accordance with the following formula:
2245
2246\[SS_j=SS_j(T)\sqrt{\frac{max{(20,LH_j})}{10}}\]
2247
2248Where:
2249
2250SSj=
2251
2252the standalone expected shortfall measure of non-modellable risk factor j
2253
2254LHj=
2255
2256the liquidity horizon of non-modellable risk factor j, as set out in Article 325bd
2257
2258T=
2259
2260the base time horizon, where T = 10 days
2261
2262SSj(T)=
2263
2264the expected shortfall measure that is determined with a 10-day time horizon for only the non-modellable risk factor j
2265
226601/01/2027
2267
22684.
2269
2270An institution may use a variety of methodologies for developing the extreme scenarios of future shock for different non-modellable risk factors and shall:
2271
2272(a) apply those methodologies in a consistent manner across similar non-modellable risk factors;
2273
2274(b) document a clear rationale for the methodology used for each non-modellable risk factor; and
2275
2276(c) validate that the methodologies meet the conditions in paragraph 3.
2277
227801/01/2027
2279
22805.
2281
2282In developing the extreme scenarios of future shocks in accordance with paragraph 3 an institution shall ensure that the extreme scenarios of future shock adequately consider any limitations to the methodologies used, including but not limited to:
2283
2284(a) any skewness or kurtosis in the distribution of returns on the non-modellable risk factor; and
2285
2286(b) any material non-linearity in the institution’s portfolio with respect to that non-modellable risk factor.
2287
2288Conceptually, an institution shall estimate the confidence interval around the extreme scenarios of future shocks produced by their methodologies due to the methodological limitations, and ensure that the extreme scenarios of future shocks used are at the conservative end of that confidence interval.
2289
229001/01/2027
2291
22926.
2293
2294Where an institution determines the extreme scenarios of future shock based on a proxy risk factor, the institution shall demonstrate that that proxy results in a stress scenario risk measure that meets the conditions in paragraph 3 with a high degree of confidence. Where an institution determines the extreme scenarios of future shock indirectly by scaling to the stress period a risk measure calibrated to another period of time, the institution shall demonstrate that the scalar is generally appropriate for the non-modellable risk factors to which it is applied and results in stress scenario risk measures that meet the conditions in paragraph 3 with a high degree of confidence.
2295
229601/01/2027
2297
2298Development of extreme scenarios of future shock at standardised bucket level
2299
23007.