concerning supervisory back-testing of internal market risk models Guidance notice Content 31 July 2014

Transcription

1 (Please note that this is a non-binding English translation of the Merkblatt zu aufsichtlichen Rückvergleichen bei internen Marktrisikomodellen as of 31 July 2014) 31 July 2014 Guidance notice concerning supervisory back-testing of internal market risk models Content 1 Preamble 2 Determining the hypothetical and actual change in value 3 Calculation of the supervisory addend 4 Business days for the supervisory back-testing 5 Quarterly delivery of time series 6 Reporting process for the overshootings 6.1 Recipients of reports 6.2 Report deadlines 7 Notification and analysis of overshooting 7.1 Overshooting notification 7.2 Overshooting analysis Portfolio analysis Market analysis Model analysis 8 Annex 8.1. Overshooting notification template 8.2. Data fields for the quarterly time series delivery

2 1 Preamble Articles 362 to 377 of Regulation (EU) No 575/ of 26 June 2013 on prudential requirements for credit institutions and investment firms (hereinafter Regulation) govern the use of internal models for calculating own funds requirements for market risk positions (hereinafter market risk models). Regular back-testing is conducted to test the forecasting quality of the market risk models. These are used to establish multiplication factors pursuant to Article 366 of the Regulation. If the back-testing reveals that the forecasting accuracy of a market risk model is insufficient, pursuant to Article 366 of the Regulation it will lead to a higher multiplication factor and thus a higher own funds requirement for market risk. This notice provides guidance on the practical use of Article 366 of the Regulation to those institutions with a market risk model. The guidelines set out in this notice refer solely to back-testing for the analysis of the supervisory own funds requirement for market risks with the internal model. By contrast, the use of back-testing for validating the market risk model pursuant to Article 369 of the Regulation is intended, in particular, to determine the suitability of the model for managing market risks. The supervisory requirements for the use of back-testing for validation are significantly more comprehensive and thus not automatically fulfilled through the implementation of the guidelines set down below. 2 Determining the hypothetical and actual change in value Article 366 of the Regulation stipulates daily back-testing. The institution compares the value-at-risk (VaR) number calculated by the market risk model based on a holding period of one business day with the one-day hypothetical and actual change in the portfolio value. To determine the hypothetical change in value, the portfolio is revalued assuming unchanged end-of-day positions. The actual change in value is based on a comparison of the actual end-of-day value of the portfolio and the end-of-day value of the previous day, excluding fees, commission and net interest income. If a loss in excess of VaR occurs, BaFin and the Deutsche Bundesbank are notified of this overshooting. To establish the actual and hypothetical change in the portfolio s value, portfolio positions included in the back-testing must be revalued daily. In general, the positions shall be marked to market or to model pursuant to Article 105 (3) to (6) of the Regulation. Actual change in value In the actual change of value, accounting valuation adjustments, which are created for fair value measurement, are generally considered, even if these are not conducted daily (weekly, monthly etc). 2 Additional valuation adjustments resulting from the requirements for prudent valuation pursuant to Article 105 of the Regulation shall not be taken into consideration in the actual change in value. An adjustment for fees, commissions and net interest income shall be carried out. 1 Version published in the Official Journal of the European Union on 30 November For example, valuation adjustments for derivatives due to changes in the credit quality of the counterparties shall also be included.

3 The adjustment for net interest income is deemed to be accurate if, pursuant to BTR 2.2 No 3 of MaRisk, 3 funding costs are taken into consideration in the calculation of the economic result, so that no material profit is generated due to the holding of positions alone. The adjustment can be carried out at an appropriate trading-book portfolio level; a calculation on the basis of individual financial instruments is not required. The actual change in value of a portfolio also contains market risk events that are not attributable to market risk and thus cannot be captured by a market risk model. However, the actual change in value allows an assessment of the extent to which actual changes in the value of the portfolio are covered by VaR. Supervisors are aware that, besides the own funds requirements derived from VaR, the portfolio is subject to additional own funds requirements. 4 Hypothetical change in value An institution is permitted to calculate the hypothetical change in value through backcalculation from the economic result calculation required by MaRisk if compliance with the requirements of Article 366 of the Regulation and the interpretation used in this notice are assured. In order to calculate the hypothetical change in value, an institution may also conduct a separate revaluation of financial instruments. The calculation of the hypothetical change in value shall include changes and replacements to valuation models or parameters of valuation models to calculate model prices. By contrast, accounting and supervisory valuation adjustments are generally not to be considered. The hypothetical change in value thus only contains market risk events that should generally be forecasted by a market risk model. The hypothetical change in value can be used without limitation to assess the forecasting quality of a market risk model. Underlying portfolio An institution s trading book is basically the relevant portfolio for all supervisory backtesting. In addition to the position risk of the trading book instruments, foreign exchange and commodities risk are the other market categories pursuant to Article 363 of the Regulation. Financial instruments with foreign exchange and commodities risk may be outside the trading book. If the market risk model is used for the foreign exchange or commodities risk categories, the corresponding non-trading-book financial instruments exposed to foreign exchange or commodities risk are likewise to be included in the supervisory back-testing, but only with the change of value attributable to foreign exchange or commodities risk. 5 For financial instruments which are excluded from the model for specific risk pursuant to Article 371 (1) of the Regulation, only the part of the change in their value which falls under the general position risk shall be considered in the hypothetical change in value. These financial instruments shall be recognised in full in the actual change in value. If institutions do not use the internal market risk model for all risk categories pursuant to Article 363 (1) letters a to f of the Regulation to calculate their own funds requirements (hereinafter partial use), this can result in market risks from financial instruments in the trading book not being taken into consideration, or not fully, when calculating VaR. However, for supervisory back-testing, no (partial) exclusion of individual financial instruments is allowed for the changes in portfolio value to be calculated pursuant to 3 Last updated on 14 December For example, for the credit valuation adjustment risk or, if applicable, using the standard methodology for market risk. 5 This approach shall be used for both the actual and the hypothetical change in value.

4 Article 366 (3) of the Regulation, even if these do not contribute to VaR due to partial use of the market risk model. 6 For portfolios whose market risks are mainly determined by risk categories outside the regulatory confirmation of suitability, institutions partially using the market risk model may, with BaFin s permission, only include the change in value which falls under the risk categories of regulatory confirmation of suitability. In order to ensure that the meaningfulness of the supervisory back-testing is not impaired, this process will only be approved to a limited degree, in cooperation with supervisors on a case-by-case basis. 7 3 Calculation of the supervisory addend Pursuant to Article 366 (2) of the Regulation, the supervisory multiplication factor is the sum of at least three and an addend. To calculate the addend, the number of overshootings in the supervisory back-testing (within the meaning of Article 366 of the Regulation) for the past 250 business days based on hypothetical and actual change in value is calculated separately and the higher of the two numbers used. If the number of overshootings is five or more, the multiplication factor shall be increased in accordance with Table 1 in Article 366 of the Regulation. The institution shall calculate and apply the addend independently for every reporting date (= end of the quarter). The date of the revaluation is relevant in recognising the overshootings on a reporting date. Unless otherwise instructed by BaFin, all overshootings that have occurred shall enter into the calculation of the number of overshootings. Pursuant to Article 366 (4) of the Regulation, BaFin may, in justified individual cases, limit the added to that resulting from overshootings under hypothetical changes. For this purpose, the institution shall demonstrate that the overshootings under the actual change in value are not due to insufficient accuracy of the risk measurement model. This is reviewed by supervisors using the analysis of each overshooting provided by the institution. A separate application by the institution is not required, but an institution shall provide further clarification if it believes the overshooting is not due to the risk model s insufficient accuracy. When making their assessment, supervisors also take into consideration in particular that the actual change in value of a portfolio is also influenced by non-market risk events which thus cannot be captured in a market risk model. If the addend is measured using overshootings under the hypothetical change in value exclusively, BaFin notifies the institution of this decision in writing as appropriate, additionally defining the length and details of the procedure. Omission of individual overshootings is not possible pursuant to Article 366. An institution can only demonstrate that an overshooting was notified even though no overshooting occurred. Such notification may have been triggered by an incorrect process, for example. Using analysis provided by the institution, supervisors examine whether a notification of an overshooting thus does not represent an overshooting within the meaning of Article 6 Even if there is the possibility of just including the change in value which falls under the risk categories of partial use when calculating model prices for individual financial instruments, the complete changes in value shall be included, however. 7 BaFin will only approve this process if the portfolio is also used internally with the same composition. The institution is not permitted to assemble a dedicated portfolio for such sensitivity-based backtesting.

5 366 of the Regulation and may be cancelled. Application by the institution is not required in this case, either. Overshootings which were mainly caused by changes and replacements of valuation models or parameters of valuation models for calculation of model prices are explicitly exempted from this option. In such cases, supervisors assume that the adjustments were necessitated by a changed market environment and thus are to be included in the change in value of the portfolio. The timing for material adjustments shall result from a process that is stipulated by the institution in advance using internal guidelines and working instructions. The procedure shall be documented, so as to ensure that the overshootings cannot be concealed by suitable choice of timing. 4 Business days for supervisory back-testing If the bank conducts trading on a given day (ie active change in existing risk positions, which clearly goes beyond hedge adjustment) in a given location, this constitutes a business day for the bank. When conducting the supervisory back-testing, an institution shall define and document local and global business days. Global business days shall be geared towards the most important location for trade and risk management. The adequate determination of nonbusiness days shall be demonstrated by largely unchanged risk positions. Non-business days at multiple locations are not allowed to be merged for defining global business days. An institution is not required to revalue any global portfolios or calculate VaR on global non-business days. If a portfolio revaluation is conducted on a non-business day, the portfolio revaluation for the previous business day may be waived in justified exceptional cases. Supervisory back-testing shall be conducted for every revaluation. 8 The resulting change in value shall then be compared with the most recent available VaR of a previous business day. An institution shall document how it implements the requirements above. 5 Quarterly delivery of time series As part of the regular inspection of the risk models forecasting accuracy pursuant to section 44 (1) sentence 2 of the German Banking Act (Kreditwesengesetz), the backtesting time series shall be delivered to supervisors irrespective of whether overshootings have occurred. The transmission shall take place exclusively in electronic form. The backtesting time series and the respective quarter and year (in YYYY format) shall be entered in the subject line (eg back-testing time series 2013-Q3). The time series shall be sent to the following addresses (see 6.1). Delivery of the time series within four weeks after the end of the respective quarter shall be ensured. Delays shall be communicated immediately, prior to the fixed deadline, and the reasons for the delay shall be explained. The process for creation and delivery of the time series shall be documented in an appropriate manner. The time series shall include the previous quarter in each case (eg ). The date of value-at-risk is of key importance here. The data shall be presented exclusively as the first worksheet of an Excel file. Leading spaces or columns shall be deleted; the first row thus contains the names of the data fields and the actual data begin in row 2, column 1. The last row is populated by the data of the last day of the quarter. The pre-defined formats must be adhered to when entering data. One line contains all the data related to one day, eg the change in value matching the value-at-risk. Dates shall be entered uniformly using the ISO 8601: This ensures that the supervisory back-testing includes all changes in value of financial instruments, as each revaluation day contributes to a change in value.

6 format (YYYY-MM-DD, eg ). Figures of money amounts shall be represented in EUR million using signs, a comma as the decimal separator and without a thousands separator (±####,##, eg -1345,98 ). Missing data are not denoted with blank spaces but expressly indicated as NA. This also applies to data fields which, due to the construction of the respective market risk model, do not contain any information, eg the time series of value-at-risk calculated with a weighted review period. The size of the time series delivery may be adjusted if required. The institutions affected will be notified in writing. At present, the time series delivery includes the data fields shown in the annex. Any retrospectively required corrections to data once a time series has been submitted shall be transmitted to supervisors in a timely fashion. For this purpose, a further complete time series with the corrected data shall be delivered for the relevant quarter, ie data that have not been corrected are to be re-entered, subtraction (new value less the old value) shall not be carried out and the format specification shall be adhered to fully. If data corrections are necessary for several previous quarters, a separate complete file shall be delivered for each quarter in question, ie data shall not be combined in a single file. Notes on any corrections undertaken shall be included. These shall not be recorded with the data in the Excel file, but shall be communicated in a separate document. The notes shall present the corrections in a meaningful manner and enable supervisors to assess the adequacy and quality of the daily back-testing. In addition, ways to remedy the weakness in the underlying processes shall be stated. With the quarterly delivery of the time series, institutions shall also provide supervisors with a meaningful description of the major changes in the methodology used to calculate the change in value. If no revaluation or determination of value-at-risk can be undertaken for technical reasons, this shall likewise be documented in the quarterly delivery of the time series to supervisors. Missing days shall be stated in the cover letter in order to distinguish them from non-business days. The notes shall contain a description of measures to prevent such gaps in future. All additional information shall likewise be included in a separate document. 6 Reporting process for overshootings The objective is to report overshootings promptly and analyse them in sufficient depth. The reporting process as described here enables the implementation of these objectives. The reporting of overshooting shall be conducted in two separate steps. First the overshooting shall be notified pursuant to Section 366 (5) of the Regulation. In addition, a more detailed analysis of the overshootings shall be conducted. This is an important aspect of the supervision of market risk models. For this purpose, the institution shall test the suitability of its risk measurement model to find possible model shortcomings in the light of the overshooting and to render them transparent to supervisors. Transmission of the institution s internal analyses render the supervisory process more efficient; inspection measures can be dispensed with entirely or based on these analyses. 6.1 Recipients of reports Both the overshooting notification and analysis shall be sent to each of the following addresses:

7 Recipient Via Deutsche Bundesbank Bundesanstalt für Finanzdienstleistungsaufsicht Responsible Deutsche Bundesbank Regional Office By post Reference Bundesanstalt für Finanzdienstleistungsaufsicht - Referat Q RM 2- Postfach Bonn <Name of bank>, Overshooting as at <date or revaluation> Supervisors will circulate the information as appropriate. The institutions are free to decide whether to send the information electronically or by postal mail. 6.2 Reporting deadlines Listed below are the final deadlines for transmitting the information to supervisors. Notification of overshooting Analysis of overshooting No later than 5 business days after the overshooting occurs (beginning on the revaluation date) No later than 3 weeks after the overshooting occurs (beginning on the revaluation date) If an institution is unable to meet the 5-day overshooting notification deadline, it must submit the overshooting notification late, stating the principal reasons for the delay. If an institution cannot produce the overshooting analysis within 3 weeks, it shall report the delay promptly to supervisors and specify a probable delivery date. 7 Notification and analysis of overshooting The institution s reports should cover at least the following points. No details need be supplied on individual points which are irrelevant to the respective back-testing overshooting. If the institution cannot furnish details on individual relevant points concerning the overshooting, it shall state the main reasons for these omissions. 7.1 Back-testing overshooting notification Initial notification may be limited to data which the institution already possesses in order to identify the back-testing overshooting. No analysis is expected at this stage. In addition, the institution presents the most recent stressed value-at-risk (stressed VaR) on the forecast date. The details to be provided are:

8 Forecast date (VaR date) VaR Date of most recent stressed VaR Stressed VaR 9 DD.MM.YYYY DD.MM.YYYY Business day (revaluation date) Actual change in value Hypothetical change in value DD.MM.YYYY Pursuant to Article 366 (5) of the Regulation, the notification serves as a supervisory analysis of the multiplication factors. The addend used in the last reporting of own funds shall be provided. In addition, the numbers of overshootings relevant on the next reporting date according to the aforementioned procedure shall be listed separately according to actual and hypothetical change in value. Addend on the last reporting date Number of back-testing overshootings on the following reporting date Value Actual change in value Hypothetical change in value Number Number The form annexed to this notice shall be used to report an overshooting. 7.2 Overshooting analysis The detailed analysis shall cover the topics of portfolio analysis, market and model analysis. Each institution is free to devise its own analysis, but the structure should be comprehensible and commensurate to the overshooting. The tabulated, listed questions that have been inserted into the following text passages serve as guidelines as to what is to be covered by the detailed analysis. Each institution shall exercise its own discretion in selecting the focus of its analysis within the topic areas. Individual topics may be omitted if they are not relevant for the analysis of a specific overshooting. The analysis of overshootings under the actual change in value shall generally be conducted in a similar manner to the analysis of the overshootings under hypothetical change in value. Individual topics may often be omitted in the analysis of overshootings under the actual change in value, as they are not relevant for a specific overshooting. The model analysis may be omitted, for example, for overshootings which clearly result from the adjustment of valuation parameters which are not covered by the market risk model. There should still be appropriate portfolio and market analyses, however. This reduced analysis can, if already available, also be submitted with the notification of the overshooting (see 7.1), simplifying the two-stage reporting process. In this case, the usual reporting deadline of five business days for the notification of the overshooting shall be observed, too, and, where applicable, a reference made that no additional analysis will be conducted Portfolio analysis The detailed overshooting analysis shall contain the trading portfolio report from the day the overshooting occurred. This entails a comparison between the VaR, actual and hypothetical changes in value for suitable lower portfolio levels. This serves to identify 9 The stressed VaR is specified for a one-business-day holding period. If required, the same scaling method shall be used as for the VaR.

9 the portfolios that primarily caused the overshooting. If the corresponding portfolios can be identified, they shall be specified appropriately. Is it possible to identify the portfolios/portfolio positions which primarily caused the overshooting? Comparison of VaR, actual and hypothetical changes in value at lower portfolio level. Identification of significant overshootings (on the basis of actual or hypothetical change in value, depending which change in value caused the overshooting) at the suitable portfolio level. What trading strategies are being pursued with the identified portfolio positions? What financial instruments are included? What value-determining factors are relevant for these financial instruments? What sensitivities do the relevant financial instruments have to these valuedetermining factors? Analysis of relevant differences between actual and hypothetical change in value for these portfolio positions. What proportion of the deviation can be attributed to new transactions (purchases and sales)? (for the overshooting under the actual change in value) What proportion of the changes in value can be attributed to positions which are opened and closed again on the same day? (for the overshooting under actual changes in value) What proportion can be attributed to the formation of valuation adjustments (eg reserves)? (for the overshooting under actual changes in value) Are different valuations of financial instruments carried out to calculate hypothetical and actual changes in value, and what is their impact? Market analysis If significant market movements have been identified as the cause of the overshooting, the market movements should be described using valuation parameters that can be observed on the market (index levels, interest rates, securities prices, exchange rates, implied correlations, implied volatilities, etc). If the co-movement of several market parameters primarily caused the overshooting (eg in the case of correlation and hedging positions), these shall also be described. Is it possible to identify significant market movements which primarily caused the overshooting? Description of market movements using observable valuation parameters. In order to assess the significance of the market movements, current market movements must be embedded in a historical context. For example, a market movement to be analysed can be given as a quantile in the historical distribution (parametric or non-parametric). The appropriate time series (eg 250 business days) or the distribution of the risk factors in question shall also be provided.

10 Classification of the market movements using the time series of historical market movements Do these valuation parameters show movements which exceed the 99% confidence level (significant market movements)? Do these valuation parameters show unexpected interrelated movements in comparison with history? If economic factors can be identified as having triggered the significant market movements, these are to be explained. An example of such an economic cause would be an unexpected interest rate shift by the European Central Bank, triggering a shift in the yield curve. The relationship between economic causes and the observed market movements should be illustrated. If no economic causes can be identified, a qualitative description of the current market environment has to be provided. Description of economic causes Can individual causes be identified? Are the markets in a phase of high/low volatility? Are uncertainties or rumours causing strong market movements? Model analysis The aim of the model analysis is to assess the model s suitability in the light of the overshooting. If portfolio positions can be identified as having contributed significantly to the overshooting, the mapping of these positions in the market risk model is to be analysed. The aim is to assess whether the risk factors are appropriate for mapping these positions. Choice of risk factors In order to demonstrate the appropriateness of the chosen risk factors, a plausibility test should be carried out on hypothetical changes in value using the sensitivities of the positions to the risk factors and the realised movement in these risk factors. A substantial part of this P/L explain is the calculation of the share of the total hypothetical P/L accounted for by changes in value that can be explained by risk factor movements and the proportion which cannot be explained in this way ( unexplained P/L ). If a material share is unexplained P/L, an assessment of the suitability of the risk factors used in the market risk model shall be included. There is a particular need to examine whether the requirements with regard to selecting the risk factors pursuant to Article 367 (1) letters 1 to b of the Regulation have been implemented adequately. This P/L explain analysis shall be performed on the portfolios and risk positions which contributed significantly to the overshooting. Risk factor mapping The institution shall also determine whether the integration of the relevant risk factors into the risk model is appropriate. In addition, the distribution assumptions used in the risk model for these risk factors must be validated using historical data. The risk model parameters shall also be reviewed.

11 Analysis of the internal risk model s ability to capture the market risks of the relevant portfolio positions By what risk factors are the relevant positions mapped in the internal risk model? Can the hypothetical change in value be explained using the sensitivities to these risk factors and the realised movement? Which of the risk drivers that are relevant for the overshooting are possibly not captured by the internal risk model? Is the mapping of relevant risk factors in the risk model appropriate compared with the historical distributions? Aggregation of the risk factors If no individual portfolio position which primarily caused the overshooting has been identified, the overshooting might have been caused because the diversification which originates through the aggregation of risk factors in the model is larger than observed in the change in value. Then the correlations used and the model assumptions for collating risks are to be reviewed. Is the overshooting primarily caused by diversification in the aggregation of risk factors? Are the correlations used in the market risk model appropriate? What are the corresponding model assumptions with regard to the aggregation of risks? Process analysis The processes for calculating VaR, the actual and the hypothetical change in value are to be analysed. For example, overshootings can result from a failure to update market prices of financial instruments daily, due either to insufficient market liquidity or to process shortcomings. Analysis of the processes for calculating VaR, actual and hypothetical changes in value Is the underlying portfolio position in determining the VaR, actual and hypothetical changes in value identical? For which securities are no valid daily prices available, and which change in value is attributable to this? (under hypothetical or actual changes in value, depending on which P/L caused the overshooting). Did process errors or changes trigger/contribute to the overshooting? Stress tests and stressed VaR The qualitative requirements to be met by a market risk model include the regular calculation of stress tests. When calculating stress tests, large losses in value caused by unusual changes in market parameters and their interaction are identified. In addition, the stress tests serve to identify possible value losses which are not caused by risk factors included in the market risk model.

12 Moreover, a stressed VaR which shows the risk of current market risk positions in a stress period needs to be identified. The occurrence of an overshooting is therefore a good starting point for a review of the stress tests and the chosen stress period for the stressed VaR. If the overshooting can be attributed to unusual market movements, it shall be examined whether these types of market movements are factored into the stress tests and the chosen stress period for the stressed VaR. Furthermore, an analysis shall be conducted as to whether the extent of the hypothetical value loss that occurred in the overshooting has been captured by the stress tests and the stressed VaR. Review of the scenarios of the stress tests and the stressed VaR under the overshooting Are the observed market movements included in the stress tests? Is the extent of the incurred losses captured by the stress tests? Will the stress tests be adjusted as a result of the overshooting? How high was the stressed VaR on the day of the overshooting or the two days nearest the overshooting (date of VaR)? Is the chosen observation period for the calculation of the stressed VaR still appropriate? How large are the market movements which led to the overshooting in comparison with the market movements in the observation period for the stressed VaR? 8. Annex 8.1 Overshooting notification form

13 Form for the notification of back-testing overshootings in internal market risk models pursuant to Article 366 of the CRR 1 General Information Name of the institution Name Reporting date DD.MM.YYYY 2 Overshooting notification Forecast date (VaR date) DD.MM.YYYY Business day (revaluation date) DD.MM.YYYY Actual change in value VaR Hypothetical change in value Date of most recent stressed VaR DD.MM.YYYY Stressed VaR 3 Details on the current addend Addend on the last reporting date Number of overshootings on the next reporting date Value Actual changes in value Hypothetical changes in value Number Number

14 8.2 Data fields for the quarterly times series delivery Column Identifier Field description Format 1 VaR date Date of market risk positions on which VaR is based YYYY-MM-DD 2 VaR1d VaR forecast based on a 99% confidence interval, a one-day holding period and the observation period used for the calculation of VaR pursuant to Article 365 (1) of the Regulation, on the date in column 1 ####,## Figure with negative sign 3 StressVaR1d Stressed VaR pursuant to Article 365 (2) of the Regulation on the date in column 1 on a one-day basis For days without current calculation, the most recent stressed VaR (on a one-day basis) shall be updated ####,## Figure with negative sign or NA If the stressed VaR is calculated directly for 10 days, the cells of this column shall be identified with NA. 4 VaR10d VaR forecast based on a 99% confidence interval, a 10-day holding period, following the calculation method used by yourself, and the observation period used for the calculation of VaR pursuant to Article 365 (1) of the Regulation, using the date in column 1 ####,## Figure with negative sign 5 StressVaR10d Stressed VaR pursuant to Article 365 (2) of the Regulation for a 10-day holding period on the date in column 1 For days without current calculation, the most recent stressed VaR shall be updated ####,## Figure with negative sign 6 P/L date Date of next revaluation after the date of the VaR YYYY-MM-DD

15 7 Clean P/L Hypothetical change in value between the date in column 1 and the date in column 6 ±####,## Figure is entered with the sign of its result (Loss = negative sign) 8 Dirty P/L Actual change in value between the date in column 1 and the date in column 6 ±####,## Figure is entered with the sign of its result (Loss = negative sign), profit = no sign). 9 VaR-Limit VaR limit ####,## Figure with negative sign 10 VaR-Limit internal If the VaR limit in column 9 refers to an alternative VaR metric (eg other confidence interval), information on this size on date in column 1 ####,## Figure with negative sign or NA 11 VaR1d-nw VaR forecast based on a 99% confidence interval, a one-day holding period and an effective observation period of at least one year on the date in column 1 ####,## Figure with negative sign or NA 12 VaR1d-w VaR forecast based on a 99% confidence interval, a one-day holding period and an effective observation period of less than one year on the date in column 1 ####,## Figure with negative sign or NA