Response to consultation on draft Regulatory Technical Standards (RTS) on prudent valuation
Go back
The language would imply that the bank would have to evaluate each and any of the sources listed for all valuation inputs. Hence, as a binding requirement, we feel that this list is excessive. In our view, the list as is can rather present a maximum list of possible sources. Whenever reliable data are being used, there should be no obligation to additionally draw upon less reliable data. From our point of view, whenever there is a liquid price, the use of consensus data or broker quotes for calculating the prudent value is inappropriate.
Under Article 3.2., the wording “a full range of available and reliable data sources” should be specified in greater detail. The language suggests that the bank has to draw upon data from all market data providers. Even if only the “reliable” sources were used, there would be redundancies in terms of the data which generate costs and tie up resources without improving the results' quality.
Especially sources which the bank does not use for valuation purposes at all (particularly consensus price services which are not being used for cost-benefit reasons) should not specifically have to be procured for this purpose. In terms of material costs and personnel costs, the integration of other sources or, moreover, also the participation in consensus pricing services would generally lead to a considerable additional burden. At this point, the largest part of the additional cost would be incurred by the data procurement and data storage. As far as we can see, this bears no relation to the poten-tial additional insights gained. From our point of view, there is a considerable danger of spending ma-jor amounts of money on the generation of spurious accuracy.
In its IPV or its trade-independent valuation, every bank typically has a process in place which decides on the integration of market data sources on the basis of their quality, availability and appropriate-ness. Depending on the respective valuation input, this may be one or several sources. In practice, it is extremely unlikely that all sources listed under paragraph 2 will be used. The decision which sources shall be used should remain incumbent upon the bank. Based on the principle of proportion-ality, during the selection of the information sources we therefore advocate a greater degree of flexi-bility. In Germany, it is a common market practice, that auditors also audit the appropriateness of the sources used as a part of their audit of the financial statement.
Regarding Art. 3 (3) we propose that central bank eligibility of bonds and corresponding haircuts de-fined by central banks in their collateral policy should be explicitly included.
A 25% haircut has to be applied to the unrealised profit of the fair value positions. The term “unreal-ised profit” lacks any national and international standard definition; it also lacks a harmonised under-standing. Whilst there is a quick consensus regarding an individual securities position (buying price 80, market value 90) it becomes difficult with regard to entire trading books featuring the most di-verse products.
Example 1: A closed swap book with valuation gains and valuation losses amounting to EUR1 billion each. The trading profit equals zero. Are the valuation profits unrealised and would a 25% haircut or, moreover, a haircut of EUR250 million have to be applied? In our understanding, in this case there is no need to create AVAs.
Example 2: An open futures position (e.g. bund futures bought at 140, market value 144) with a profit of EUR1 million. Has the profit been realised because there has been an inflow in the form of variation margin or is it an unrealised profit because the position is still open? In our view this case does not require any AVAs.
Example 3: A security (bought at 90, market value 95) was hedged either partly or in its entirety by an opposite position (swap, short holdings, future or similar). The security includes an unrealised profit of 5, the hedging position includes a loss of 5). The trading result equals zero. Would it be necessary to apply a 25% haircut on this 5 and would this have to be deducted from capital? In our understanding, also in this case there will be no need for AVAs.
Furthermore, valuation uncertainties exist not only in the field of unrealised profits but also in the field of losses. Hence, there should be no differentiation between profits and losses in return for the appli-cation of a lower percentage rate. This would also mitigate the arbitrary market driven volatility of cap-ital deductions. In our view, the percentages for capital deductions are clearly too high and should be reviewed during the QIS.
Furthermore, in the Simplified Approach there should be a clarification to the effect that also hedging relations shall be subsumed under the wording “matching, offsetting assets and liabilities” (Art. 5). Hence, when it comes to hedged transactions, neither the underlying transaction nor the hedging transaction should have to be included. In order to avoid any misunderstanding, we suggest a more detailed definition of this wording.
The Simplified Approach should not be conceived of as a sanction rule. This is due to the fact that the EBA itself assumes that “institutions with small fair value portfolios will typically be subject to limited valuation uncertainty”. The Simplified Approach should envisage the same logic and the same netting options as the Core Approach.
We suggest to use a categorization of Level 1/2/3 fair value hierarchy and corresponding different methods to derive AVA. Level 1 instruments could be excluded from AVA determination due to its inherent characteristic of highly liquid market prices. Level 2 and 3 instruments may be charged with different AVA corresponding to their market price reliability. Fair value hierarchy and the classification of instruments in the LCR calculation should be used in a consistent manner and should be consequently applicable in the classification of AVA requirements. In this context we again see a need to ensure a consistent treatment of different regulatory items.
Generally, we hold the view that the fallback solution under Art. 7 is acceptable. However, we would like to question the use of the extremely high values. In our opinion, the approach specified under Art. 7(4)(b), i.e. 10% of the nominal value is inappropriately high. As an alternative, we suggest an adjustment as a percentage rate only (e.g. 25%), of the fair value adjustments that have already been carried out.
In our preliminary understanding, a bank that is incapable of determining merely one single AVA cate-gory for a position will automatically have to apply the fallback approach described in Article 7(4). We hold the view that this is excessive. A regulatory alternative choice the EBA might consider is allowing a combination of the Core Approach and the fallback solution if and when merely individual AVA cate-gories cannot be determined.
We therefore suggest that institutions should not be required to calculate AVAs for those valuation inputs that according to a cost-benefit-analysis (immateriality) give rise to valuation exposures that account for less than 1% of the total valuation exposure or less than EUR1 billion in absolute terms.
Article 8 (2a) and article 9 (2) refer to “firm evidence of a tradable price”. We kindly ask to specify what we should subsume under “firm evidence of a tradable price” (e.g. all sources of market data from article 3 (2a-c)?).
Article 8 (5a1) specifies that market price uncertainty AVAs can be computed on the basis of exit prices. In our opinion, in doing so, this also implicitly yields a close-out cost AVA. This should be clarified in the RTS.
Article 9 (2) implies that where there is no uncertainty of market mid prices and the corresponding AVA has been set to zero according to Article 8 (2), the close-out cost AVA may be assessed to have zero value. We think that even if mid prices may be found in a very tight range, that does not imply the absence of bid and/or ask prices different from mid prices. Therefore we would like to question the connection made between Articles 9 (2) and 8 (2).
In the assessment of market price uncertainties or close-out costs (Art. 8 and 9), we assume that the spectrum of a valuation input (e.g. by observation of the data from Composite Reuters Instrument Codes (RICs)) may be estimated within a reasonable period of time/intra-day; this is due to the fact that, more likely than not, the various contributors represent independent data sources.
In article 9 there is no clear statement whether close-out cost AVA are assumed to have zero value when the position’s price already reflects bid-offer close-out spreads. We would welcome an additional statement, that based on a fair value approach with exit prices (Bid- or Offer-side dependent on the position) in the IFRS P&L there is no further need to consider close-out cost AVA.
The hurdles for a reduction of the number of parameters for market price uncertainty and close-out costs are inappropriately high. Along with Articles 8 and 9, this particularly also applies to the provi-sions under Article 20 which have to be seen as preconditions for a reduction.
Under Indent 21, the Discussion Paper explicitly clarifies that no prudent values would have to be cal-culated for DVAs due to the fact that these do not enter the own funds calculation. This provision was fit for purpose. Yet, this clause is absent from the Consultation Paper. Notwithstanding that the Con-sultation Paper only refers to CVAs (Art. 10), we suggest including this clause in the forthcoming standard in the form of a footnote.
By way of analogy to the DVAs, under the provisions of Article 33(1) CRR the results from hedging transactions for cash flows (from so-called cash flows hedges) posted in the IFRS equity shall be elim-inated from own funds. In our view, by way of analogy to the DVAs, to this end it will neither be nec-essary to calculate any AVAs nor to eliminate them from own funds. This is another aspect which should be reflected in the final RTS.
We suggest clarifying that (with regard to the CVA) no double counting takes place. In our preliminary understanding, provided it was not already taken into account in the CVA itself, it is merely necessary to recognise the uncertainty of the CVA. Furthermore, we assume that the adjustments in the proba-bility of default do not have to be taken into account.
In addition to this, we wonder whether the market price uncertainty and the model risk have to be taken into account during the calculation of the expected future exposures.
Regarding article 13 we strongly believe that the general requirement to calculate an AVA for funding and investment cost is misguided. The prudent price of a derivative is already accounted for via a fair value credit valuation adjustment plus the requirements in article 10 of the RTS for an ACVA. Investing and funding cost (or funding valuation adjustment) is not part of a (prudent) market price. Rather it is an internal profit hurdle, that institutions use in deciding whether or not to enter a trade. It also relies on the business strategy of an institution, whether it intends to hedge a derivatives trade and if so, if it chooses a counterparty with which a collateral agreement is agreed or not. Take for example a German government bond with a short term, where the market price corresponds to a yield of zero percent. Institutions will not be able to fully finance this position in the market so that they can achieve profitability by holding the government bond to maturity. Still the market price is the market price and a prudent value of the bond position will not take account of article 13 of the RTS, since it is not a derivative (however, this mismatch will be accounted for in funds transfer pricing systems). In the same manner, then, the value of a derivative which an institution enters with the German government should not be written down due to funding costs, which the institution incurs in its financing operations (accounting: liabilities). Institutions should be aware whether they incur a profit or a loss in entering such a trade, but the asset side of the balance sheet should not be valued according to positions an institution may enter or not on the liability side. Allowing an institutions funding spread to determine the valuation of its assets could potentially lead to a self-reinforcing vicious cycle, spiralling an institution into insolvency. Still, in entering a running derivatives trade with positive market value (possibly against a less refined counterparty), some institutions, instead of using market credit spreads, use their theoretical funding costs alongside historical PDs for the counterparty in order to calculate a price, treating the derivative effectively like a loan. However, this is just a different pricing approach (production costs in absence of a transparent and liquid market) and there is no overlap to the fair valued market value of the derivative. Therefore, in the RTS there should be incorporated a clause, which qualifies the requirement to calculate an investing and funding AVA. The AVA should only be calculated for derivatives positions, for which institutions have not already calculated an (additional) credit value adjustment (based on market credit spread information).
Concerning Article 14 calculation of future administrative costs, we believe that article 14 (1) is confusing. Where a close-out cost AVA is calculated based on bid-ask-spreads for example, this already covers transaction and administrative costs and is indicative of the existence of a market for the valuation instrument. Therefore, we believe that a future administrative cost AVA should only be calculated, where a close-out cost AVA cannot be calculated (since there is no market or other relevant valuation input) and the valuation instrument is likely to remain on an institution’s books until maturity. We suggest that EBA include a qualification concerning this matter into Article 14. A possible way to go forward would be to tie the calculation of this AVA for administrative costs to instruments that are marked-to-model using unobservable inputs (level 3 in what concerns IFRS). Additionally, we would like EBA to clarify on which market data other than exit prices a close-out cost AVA could be calculated.
Furthermore, in order to avoid any misunderstanding, we would welcome further explanations as to the meaning of model calibrations “other than calibration from market derived parameters” (Art. 11(1)).
Regarding the determination of relevant instruments for concentrated positions AVA we suggest to exclude at least Level 1 positions of the Basel III Liquidity Coverage Ratio classification as these instruments are deemed to be of highest liquidity. Consequently no material differences between the fair value of these positions and the potential exit price are expected. This is also expected for certain non-Level 1 assets with comparable market liquidity assumptions due to comparable risk and background (e.g. business development banks).
We assume the observable size of Bid-Offer spreads in combination with the average daily traded volume of a financial instrument may also serve as an indicator of the concentration level. Resulting we would suggest to include the “average daily traded volume” additionally in Article 3 – Sources of market data. In cases of a consequent application of wider Bid-Offer spreads in the calculation of close-out cost AVA would imply a certain part of the concentrated positions AVA already included in another AVA category (Article 9 – Calculation of close-out cost AVA). Further guidance on how to avoid these risks of double-counting some parts in different AVA categories would be welcomed in the RTS.
Regarding the estimation of a prudent exit period, a distinction between trading book positions and banking book positions should be allowed. Presumably the exit period in the banking book could be longer than in tha trading book as positions are held for different reasons.
The Funding Valuation adjustments (FVA) are being discussed in a controversial manner within the industry and the scientific community. However, Article 13 fails to achieve a uniform consideration of the funding costs and funding benefits for uncollateralised derivatives. This is due to the fact that the application of different approaches continues and that still no standard has emerged. Furthermore, it remains unclear how expected funding costs shall be taken into account as AVA.
We would appreciate a clarification: Does Article 13 (1) of the Consultation Papers refer to the inap-propriate valuation (valuation framework) of collateralised derivatives (e.g. non-application of an OIS-based yield curve hierarchy in the valuation of collateralised positions)?
From Art. 14 (1) it becomes apparent that the AVA for future administrative cost shall not become applicable if close-out costs AVA have already been taken into account on the basis of Bid/Ask-Spreads – i.e. exit prices – of the yield curve instruments. In order to avoid any misunderstanding, we would appreciate a confirmation of this preliminary understanding.
We believe that not every AVA will be relevant for every product and hence would welcome a clause in the RTS saying that institutions should define which AVAs are applicable and will be calculated for the products they hold and include this in their auditable documentation. Additionally, please also see our answers to Q8 regarding articles 13 and 14.
Furthermore, concerning the possibility to disaggregate market price uncertainty AVA and close-out cost AVA it is worth noting a peculiarity in the proposed calculation method (Article 8 (5a No. 1). Dur-ing the application, cases have occurred where the calculation of the prudent values in two stages using the components 1) market price uncertainty AVA and 2) close-out cost AVA led to a lower value than the value of the lowest underlying bid price used. We are of the opinion that, in these cases, it would be more constructive to calculate market price uncertainty and close-out costs in one compu-tational step as a 90% quantile directly to the bid prices. Yet, at the same time this begs the question how the respective result should be allocated to the factors market price uncertainty and / or close-out costs. We advocate in favour of a solution where, based on assessments made by experts, the amount may respectively be allocated on a pro rata basis to said two AVA categories.
Bond example:
AVA market price uncertainty:
BID ASK MID
A 87,693 93,794 90,744
B 88,000 90,000 89,000
C 85,000 87,000 86,000
D 88,000 92,000 90,000
Average 88,936
90% Quantile 86,900
AVA close-out cost:
BID ASK Bid-Ask
A 87,693 93,794 6,101
B 88,000 90,000 2,000
C 85,000 87,000 2,000
D 88,000 92,000 4,000
Average 3,525 90% Quantile 5,471
Market price uncertainty (90% Quantile) = 86.9
Close-out costs (90% Quantile) = 5,471
Prudent Value = 86.9 – (5.471 / 2) = 84,165 versus lowest bid price of 85,000
Furthermore, paragraphs 2 and 3 should be specified in greater detail. Unless they have already been defined under Article 2, we would welcome definitions of the terms used in Article 20(2) and 20(3). It is also unclear what is meant by “valuation inputs that match the contractual price”.
We have difficulties in comprehending the rationale for the interpolation proposed under paragraph 3(c) and (d). We would welcome an explanatory example.
From the analysis of the “impact of the proposals” in 5.2.14-18 in our view an important issue is missing. EBA did not address the future enlarged volatility of bank capital. To date bank capital varies with the variation in the fair values of its assets. When crisis looms, on account of the target level of certainty, prudent value will be far more volatile than fair value given ever more disparate data for valuation inputs (pro-cyclicality). At the same time, even the problem definition in 5.2.4 is somewhat flawed, since EBA recurs to a notion of “true realisable value”: A bank may opt NOT to realise value. It may particularly opt so in cases where asset prices fall, but sufficient funding is available to hold the concerned assets on its books, acting in fact like a long-term investor who looks “through-the-cycle”. In this context we would like to encourage EBA to rethink its definition and its understanding of the essence of prudence in valuation. The dictionary gives a good definition of the meaning of the word “prudent” [see http://dictionary.reference.com/browse/prudent]: “wise or judicious in practical affairs; sagacious; discreet or circumspect; sober; carful in providing for the future; provident”. Thus, the requirement in CRR for a “prudent valuation” also encompasses the requirement for a “sober and circumspect valuation”.
We strongly believe that bank capital should not be too tightly tied to the whims of the market, which is driven not only by (intrinsic) value, but also by (market) liquidity. When liquidity is pushed into markets and prices go up, valuations should be conservative, but on the other hand, when liquidity evaporates, valuations should be allowed to become more “aggressive” in the sense that they should more accurately reflect intrinsic value instead of market doom. Applying prudent valuation in its proposed form bears the risk of aggravating a deteriorating situation. In the recent financial crisis, prudent valuation requirements would have led to significant intensification of stress in the banking sector. In a steep market downturn, therefore, prudent valuation requirement should be enriched with a countercyclical measure, allowing for mid or even ask prices to be used in prudent, i.e. sober and circumspect valuation. The conception of a prudent value then would entail distrusting the markets “animal spirits” where herd-like behaviour prevails and liquidity dissipates. Measures that could be incorporated into the rules in order to install a “relief valve” could be, for example:
• Bloomberg’s Relative Strength Index which measures the momentum of a security’s price to determine if it is an overbought or an oversold condition and connecting this to valuation;
• comparing the current level of some volatility index (e.g. VIX) to its long-run average and relaxing prudent valuation requirements once certain thresholds are breached;
• indices calculated by central banks, such as the “St. Louis Federal Reserve Bank Financial Stress Index” could be used to set “dynamic target levels of certainty”.
We believe that our proposed notion of prudence is fully in line with Capital Requirements Regulation articles 105 (1) and (5). The benefits of such a notion of prudence would be a better recognition of the intrinsic virtues of fair valued financial instruments as a liquidity resource (in contrast to loans, for example), counter-cyclical regulatory impulses and increased bank capital stability. Indeed, to address the procyclicality of the financial system by, for example, stipulating the accumulation of buffers in “good times” so that these can be drawn down in “bad times” is a crucial component of the macroprudential approach. Tools which are already used in this regard include countercyclical capital buffers or dynamic loan loss provisioning.
An issue we miss from the impact analysis is the complexity of the RTS and the complexity the RTS adds to management reporting, bank controlling and auditing. Interpretation of the rules governing prudent valuation, implementation of systems and processes and auditing a prudent valuation approach will potentially and likely increase the differences across banks and will also add complexity to regulators and auditors alike.
Do you agree with the minimum list of alternative methods and sources of information defined above for expert based approaches? If not, what others could be included, or which points from the current list should be removed? State your reasons.
Article 3(1) sets out that, where institutions calculate AVAs on market data, they shall use the same market data used in independent price verifications (“IPV”). However, in general, only larger banks with active trading are able to draw upon an established IPV process. Hence, its existence cannot be taken for granted. The IPV process could serve as a means of orientation for the implementation – but this implementation should not be specified in an absolute and final manner. Paragraph 2 extends the scope of the market data used to the “full range of available and reliable data sources” and lists sources which should all be included. From our point of view, this definition is a source for concern on several counts.The language would imply that the bank would have to evaluate each and any of the sources listed for all valuation inputs. Hence, as a binding requirement, we feel that this list is excessive. In our view, the list as is can rather present a maximum list of possible sources. Whenever reliable data are being used, there should be no obligation to additionally draw upon less reliable data. From our point of view, whenever there is a liquid price, the use of consensus data or broker quotes for calculating the prudent value is inappropriate.
Under Article 3.2., the wording “a full range of available and reliable data sources” should be specified in greater detail. The language suggests that the bank has to draw upon data from all market data providers. Even if only the “reliable” sources were used, there would be redundancies in terms of the data which generate costs and tie up resources without improving the results' quality.
Especially sources which the bank does not use for valuation purposes at all (particularly consensus price services which are not being used for cost-benefit reasons) should not specifically have to be procured for this purpose. In terms of material costs and personnel costs, the integration of other sources or, moreover, also the participation in consensus pricing services would generally lead to a considerable additional burden. At this point, the largest part of the additional cost would be incurred by the data procurement and data storage. As far as we can see, this bears no relation to the poten-tial additional insights gained. From our point of view, there is a considerable danger of spending ma-jor amounts of money on the generation of spurious accuracy.
In its IPV or its trade-independent valuation, every bank typically has a process in place which decides on the integration of market data sources on the basis of their quality, availability and appropriate-ness. Depending on the respective valuation input, this may be one or several sources. In practice, it is extremely unlikely that all sources listed under paragraph 2 will be used. The decision which sources shall be used should remain incumbent upon the bank. Based on the principle of proportion-ality, during the selection of the information sources we therefore advocate a greater degree of flexi-bility. In Germany, it is a common market practice, that auditors also audit the appropriateness of the sources used as a part of their audit of the financial statement.
Regarding Art. 3 (3) we propose that central bank eligibility of bonds and corresponding haircuts de-fined by central banks in their collateral policy should be explicitly included.
Do you agree with the proposed simplified approach? Do you think the risk sensitiveness of the approach is appropriate? Are there alternative approaches that you believe would be more appropriate? State your reasons.
Naturally, the Simplified Approach can hardly be deemed risk sensitive. Particularly the assumed cor-relation between “net unrealised profits” and AVAs is not immediately plausible in our view: The risk sensitivity of the unrealized gains is less obvious since it is a historical number. For instance, the unrealized profit could be locked-in by counter-balancing trades. In addition, the lifetime unrealized gain of a fair valued financial instrument is not required to be collected for other purposes in case fair value changes are recognized in profit or loss. Hence, institutions may incur large implementation costs for deriving a number that is not risk sensitive.A 25% haircut has to be applied to the unrealised profit of the fair value positions. The term “unreal-ised profit” lacks any national and international standard definition; it also lacks a harmonised under-standing. Whilst there is a quick consensus regarding an individual securities position (buying price 80, market value 90) it becomes difficult with regard to entire trading books featuring the most di-verse products.
Example 1: A closed swap book with valuation gains and valuation losses amounting to EUR1 billion each. The trading profit equals zero. Are the valuation profits unrealised and would a 25% haircut or, moreover, a haircut of EUR250 million have to be applied? In our understanding, in this case there is no need to create AVAs.
Example 2: An open futures position (e.g. bund futures bought at 140, market value 144) with a profit of EUR1 million. Has the profit been realised because there has been an inflow in the form of variation margin or is it an unrealised profit because the position is still open? In our view this case does not require any AVAs.
Example 3: A security (bought at 90, market value 95) was hedged either partly or in its entirety by an opposite position (swap, short holdings, future or similar). The security includes an unrealised profit of 5, the hedging position includes a loss of 5). The trading result equals zero. Would it be necessary to apply a 25% haircut on this 5 and would this have to be deducted from capital? In our understanding, also in this case there will be no need for AVAs.
Furthermore, valuation uncertainties exist not only in the field of unrealised profits but also in the field of losses. Hence, there should be no differentiation between profits and losses in return for the appli-cation of a lower percentage rate. This would also mitigate the arbitrary market driven volatility of cap-ital deductions. In our view, the percentages for capital deductions are clearly too high and should be reviewed during the QIS.
Furthermore, in the Simplified Approach there should be a clarification to the effect that also hedging relations shall be subsumed under the wording “matching, offsetting assets and liabilities” (Art. 5). Hence, when it comes to hedged transactions, neither the underlying transaction nor the hedging transaction should have to be included. In order to avoid any misunderstanding, we suggest a more detailed definition of this wording.
The Simplified Approach should not be conceived of as a sanction rule. This is due to the fact that the EBA itself assumes that “institutions with small fair value portfolios will typically be subject to limited valuation uncertainty”. The Simplified Approach should envisage the same logic and the same netting options as the Core Approach.
We suggest to use a categorization of Level 1/2/3 fair value hierarchy and corresponding different methods to derive AVA. Level 1 instruments could be excluded from AVA determination due to its inherent characteristic of highly liquid market prices. Level 2 and 3 instruments may be charged with different AVA corresponding to their market price reliability. Fair value hierarchy and the classification of instruments in the LCR calculation should be used in a consistent manner and should be consequently applicable in the classification of AVA requirements. In this context we again see a need to ensure a consistent treatment of different regulatory items.
Could a differentiated treatment for some asset/liability classes be considered, for example with regard to their liquidity? Please state the pros and cons of such a differentiation. How would you define the degree of liquidity of an asset/liability class (e.g. fair value hierarchy, eligibility for the LCR, other)?
see Q4Do you agree with the approach defined above to calculate an AVA where the approaches in Article 8 and 9 are not possible for a valuation exposure? If not, what other approach could be prescribed? Explain your reasoning.
For us, this generally begs the question how valuation adjustments should be handled which cannot be subsumed under the value influencing factors set out under Articles 8 to 16. In order to avoid any misunderstanding, we suggest clarifying the policy which will have to be adopted in such cases.Generally, we hold the view that the fallback solution under Art. 7 is acceptable. However, we would like to question the use of the extremely high values. In our opinion, the approach specified under Art. 7(4)(b), i.e. 10% of the nominal value is inappropriately high. As an alternative, we suggest an adjustment as a percentage rate only (e.g. 25%), of the fair value adjustments that have already been carried out.
In our preliminary understanding, a bank that is incapable of determining merely one single AVA cate-gory for a position will automatically have to apply the fallback approach described in Article 7(4). We hold the view that this is excessive. A regulatory alternative choice the EBA might consider is allowing a combination of the Core Approach and the fallback solution if and when merely individual AVA cate-gories cannot be determined.
We therefore suggest that institutions should not be required to calculate AVAs for those valuation inputs that according to a cost-benefit-analysis (immateriality) give rise to valuation exposures that account for less than 1% of the total valuation exposure or less than EUR1 billion in absolute terms.
Do you agree with the approaches defined above to calculate AVAs for market price uncertainty, close-out costs, and unearned credit spreads? If not, what other approach could be prescribed? State your reasons.
If there is firm evidence of a tradable price or if a price can be determined from reliable data based on a liquid, two-way market and if there is no indication of any material market valuation uncertainty, the market price uncertainty AVA shall be assessed to have zero value (Article 8). There should be a clari-fication that these requirements are in any case fulfilled on a permanent basis in the event of an ac-tive market / level 1 valuation in line with the IFRS fair value hierarchy.Article 8 (2a) and article 9 (2) refer to “firm evidence of a tradable price”. We kindly ask to specify what we should subsume under “firm evidence of a tradable price” (e.g. all sources of market data from article 3 (2a-c)?).
Article 8 (5a1) specifies that market price uncertainty AVAs can be computed on the basis of exit prices. In our opinion, in doing so, this also implicitly yields a close-out cost AVA. This should be clarified in the RTS.
Article 9 (2) implies that where there is no uncertainty of market mid prices and the corresponding AVA has been set to zero according to Article 8 (2), the close-out cost AVA may be assessed to have zero value. We think that even if mid prices may be found in a very tight range, that does not imply the absence of bid and/or ask prices different from mid prices. Therefore we would like to question the connection made between Articles 9 (2) and 8 (2).
In the assessment of market price uncertainties or close-out costs (Art. 8 and 9), we assume that the spectrum of a valuation input (e.g. by observation of the data from Composite Reuters Instrument Codes (RICs)) may be estimated within a reasonable period of time/intra-day; this is due to the fact that, more likely than not, the various contributors represent independent data sources.
In article 9 there is no clear statement whether close-out cost AVA are assumed to have zero value when the position’s price already reflects bid-offer close-out spreads. We would welcome an additional statement, that based on a fair value approach with exit prices (Bid- or Offer-side dependent on the position) in the IFRS P&L there is no further need to consider close-out cost AVA.
The hurdles for a reduction of the number of parameters for market price uncertainty and close-out costs are inappropriately high. Along with Articles 8 and 9, this particularly also applies to the provi-sions under Article 20 which have to be seen as preconditions for a reduction.
Under Indent 21, the Discussion Paper explicitly clarifies that no prudent values would have to be cal-culated for DVAs due to the fact that these do not enter the own funds calculation. This provision was fit for purpose. Yet, this clause is absent from the Consultation Paper. Notwithstanding that the Con-sultation Paper only refers to CVAs (Art. 10), we suggest including this clause in the forthcoming standard in the form of a footnote.
By way of analogy to the DVAs, under the provisions of Article 33(1) CRR the results from hedging transactions for cash flows (from so-called cash flows hedges) posted in the IFRS equity shall be elim-inated from own funds. In our view, by way of analogy to the DVAs, to this end it will neither be nec-essary to calculate any AVAs nor to eliminate them from own funds. This is another aspect which should be reflected in the final RTS.
We suggest clarifying that (with regard to the CVA) no double counting takes place. In our preliminary understanding, provided it was not already taken into account in the CVA itself, it is merely necessary to recognise the uncertainty of the CVA. Furthermore, we assume that the adjustments in the proba-bility of default do not have to be taken into account.
In addition to this, we wonder whether the market price uncertainty and the model risk have to be taken into account during the calculation of the expected future exposures.
Do you agree with the approaches defined in Articles 11 to 16 to calculate the various categories of AVAs? If not, what other approach could be prescribed for each AVA? State your reasons.
In Article 12 (2) the RTS specifies that the time horizon for the market risk measure used to calcu-late own funds requirements (in the trading book) should be used in the calculation of the concen-trated position AVA. The requirement not only applies to the trading book but also to all fair valued positions in the banking book. Positions in the banking book are not held to be traded within a 10 day horizon. In our view, therefore, the time horizon for calculation of the AVA for banking book fair valued positions should be considerably enlarged. (Please see also our comment in relation to the LCR in our response to Q13.). Otherwise, we would also like to point out that there would be additional discrimi-nation of fair-valued assets held in the banking book vis-à-vis (illiquid) loans.Regarding article 13 we strongly believe that the general requirement to calculate an AVA for funding and investment cost is misguided. The prudent price of a derivative is already accounted for via a fair value credit valuation adjustment plus the requirements in article 10 of the RTS for an ACVA. Investing and funding cost (or funding valuation adjustment) is not part of a (prudent) market price. Rather it is an internal profit hurdle, that institutions use in deciding whether or not to enter a trade. It also relies on the business strategy of an institution, whether it intends to hedge a derivatives trade and if so, if it chooses a counterparty with which a collateral agreement is agreed or not. Take for example a German government bond with a short term, where the market price corresponds to a yield of zero percent. Institutions will not be able to fully finance this position in the market so that they can achieve profitability by holding the government bond to maturity. Still the market price is the market price and a prudent value of the bond position will not take account of article 13 of the RTS, since it is not a derivative (however, this mismatch will be accounted for in funds transfer pricing systems). In the same manner, then, the value of a derivative which an institution enters with the German government should not be written down due to funding costs, which the institution incurs in its financing operations (accounting: liabilities). Institutions should be aware whether they incur a profit or a loss in entering such a trade, but the asset side of the balance sheet should not be valued according to positions an institution may enter or not on the liability side. Allowing an institutions funding spread to determine the valuation of its assets could potentially lead to a self-reinforcing vicious cycle, spiralling an institution into insolvency. Still, in entering a running derivatives trade with positive market value (possibly against a less refined counterparty), some institutions, instead of using market credit spreads, use their theoretical funding costs alongside historical PDs for the counterparty in order to calculate a price, treating the derivative effectively like a loan. However, this is just a different pricing approach (production costs in absence of a transparent and liquid market) and there is no overlap to the fair valued market value of the derivative. Therefore, in the RTS there should be incorporated a clause, which qualifies the requirement to calculate an investing and funding AVA. The AVA should only be calculated for derivatives positions, for which institutions have not already calculated an (additional) credit value adjustment (based on market credit spread information).
Concerning Article 14 calculation of future administrative costs, we believe that article 14 (1) is confusing. Where a close-out cost AVA is calculated based on bid-ask-spreads for example, this already covers transaction and administrative costs and is indicative of the existence of a market for the valuation instrument. Therefore, we believe that a future administrative cost AVA should only be calculated, where a close-out cost AVA cannot be calculated (since there is no market or other relevant valuation input) and the valuation instrument is likely to remain on an institution’s books until maturity. We suggest that EBA include a qualification concerning this matter into Article 14. A possible way to go forward would be to tie the calculation of this AVA for administrative costs to instruments that are marked-to-model using unobservable inputs (level 3 in what concerns IFRS). Additionally, we would like EBA to clarify on which market data other than exit prices a close-out cost AVA could be calculated.
Are there cases where the above AVAs may have a zero value that could be defined in the RTS? If yes, please specify.
Under the provisions of Art. 11, model risks shall be taken into account as AVA. Depending on the position in need of valuation, model risk can manifest itself in various shapes and forms. By way of example, at this point we would like to mention: correlation risk, basis risks, recovery risks, dividend risks. For us, this begs the question how this diversity of model risks can be quantified within the meaning of the AVA. This also applies to the option of extracting a series of plausible parameters from an alternative, suitable model, a calibration approach or an expert estimate.Furthermore, in order to avoid any misunderstanding, we would welcome further explanations as to the meaning of model calibrations “other than calibration from market derived parameters” (Art. 11(1)).
Regarding the determination of relevant instruments for concentrated positions AVA we suggest to exclude at least Level 1 positions of the Basel III Liquidity Coverage Ratio classification as these instruments are deemed to be of highest liquidity. Consequently no material differences between the fair value of these positions and the potential exit price are expected. This is also expected for certain non-Level 1 assets with comparable market liquidity assumptions due to comparable risk and background (e.g. business development banks).
We assume the observable size of Bid-Offer spreads in combination with the average daily traded volume of a financial instrument may also serve as an indicator of the concentration level. Resulting we would suggest to include the “average daily traded volume” additionally in Article 3 – Sources of market data. In cases of a consequent application of wider Bid-Offer spreads in the calculation of close-out cost AVA would imply a certain part of the concentrated positions AVA already included in another AVA category (Article 9 – Calculation of close-out cost AVA). Further guidance on how to avoid these risks of double-counting some parts in different AVA categories would be welcomed in the RTS.
Regarding the estimation of a prudent exit period, a distinction between trading book positions and banking book positions should be allowed. Presumably the exit period in the banking book could be longer than in tha trading book as positions are held for different reasons.
The Funding Valuation adjustments (FVA) are being discussed in a controversial manner within the industry and the scientific community. However, Article 13 fails to achieve a uniform consideration of the funding costs and funding benefits for uncollateralised derivatives. This is due to the fact that the application of different approaches continues and that still no standard has emerged. Furthermore, it remains unclear how expected funding costs shall be taken into account as AVA.
We would appreciate a clarification: Does Article 13 (1) of the Consultation Papers refer to the inap-propriate valuation (valuation framework) of collateralised derivatives (e.g. non-application of an OIS-based yield curve hierarchy in the valuation of collateralised positions)?
From Art. 14 (1) it becomes apparent that the AVA for future administrative cost shall not become applicable if close-out costs AVA have already been taken into account on the basis of Bid/Ask-Spreads – i.e. exit prices – of the yield curve instruments. In order to avoid any misunderstanding, we would appreciate a confirmation of this preliminary understanding.
We believe that not every AVA will be relevant for every product and hence would welcome a clause in the RTS saying that institutions should define which AVAs are applicable and will be calculated for the products they hold and include this in their auditable documentation. Additionally, please also see our answers to Q8 regarding articles 13 and 14.
Furthermore, concerning the possibility to disaggregate market price uncertainty AVA and close-out cost AVA it is worth noting a peculiarity in the proposed calculation method (Article 8 (5a No. 1). Dur-ing the application, cases have occurred where the calculation of the prudent values in two stages using the components 1) market price uncertainty AVA and 2) close-out cost AVA led to a lower value than the value of the lowest underlying bid price used. We are of the opinion that, in these cases, it would be more constructive to calculate market price uncertainty and close-out costs in one compu-tational step as a 90% quantile directly to the bid prices. Yet, at the same time this begs the question how the respective result should be allocated to the factors market price uncertainty and / or close-out costs. We advocate in favour of a solution where, based on assessments made by experts, the amount may respectively be allocated on a pro rata basis to said two AVA categories.
Bond example:
AVA market price uncertainty:
BID ASK MID
A 87,693 93,794 90,744
B 88,000 90,000 89,000
C 85,000 87,000 86,000
D 88,000 92,000 90,000
Average 88,936
90% Quantile 86,900
AVA close-out cost:
BID ASK Bid-Ask
A 87,693 93,794 6,101
B 88,000 90,000 2,000
C 85,000 87,000 2,000
D 88,000 92,000 4,000
Average 3,525 90% Quantile 5,471
Market price uncertainty (90% Quantile) = 86.9
Close-out costs (90% Quantile) = 5,471
Prudent Value = 86.9 – (5.471 / 2) = 84,165 versus lowest bid price of 85,000
Do you agree with the approach defined above for the aggregation of valuation exposure level AVAs within the market price uncertainty and close-out cost AVA categories? If not, what other approach could be prescribed? State your reasons.
We do not agree with Article 17 (3b), since valuation uncertainty will be greatest exactly where there is more than one valuation input, due to complexity. The 50 %-number seems incompre-hensible. We would appreciate an adjustment of the percentage rate or, at least, a cogent expla-nation for the level at which this percentage rate is set.Do you agree that category level AVAs described in Articles 11 to 16 within the core approach should be aggregated as a simple sum? If not, what other approach could be prescribed? State your reasons.
Whilst we welcome the consideration of diversification effects (Art. 17) it is, however, difficult to work out empirically. Notwithstanding the foregoing, banks should be granted the right to calculate these effects on the basis of internal approaches that are fit for purpose and subject to regular reviews.Do you agree with the requirement for institutions using the core approach to implement the above ongoing monitoring tool as an indicator of the adequacy of data sources of valuation inputs used to calculate the AVAs described in Articles 8 to 10? If not, what other approach could be prescribed? State your reasons.
Article 20(2) and (3) merely refer to reduced valuation inputs. In order to avoid any misunderstand-ing, we would appreciate an explicit clarification that banks which do not reduce the number of pa-rameters of the valuation input shall be exempt from the requirements set out thereunder.Furthermore, paragraphs 2 and 3 should be specified in greater detail. Unless they have already been defined under Article 2, we would welcome definitions of the terms used in Article 20(2) and 20(3). It is also unclear what is meant by “valuation inputs that match the contractual price”.
We have difficulties in comprehending the rationale for the interpolation proposed under paragraph 3(c) and (d). We would welcome an explanatory example.
Do you agree with our analysis of the impact of the proposals in this CP? If not, can you provide any evidence or data that would explain why you disagree or might further inform our analysis of the likely impacts of the proposals?
Referring to the speech by Andrew Haldane, Bank of England, from 19th December 2011, the accounting of assets should not be separated from the source of their funding or “holding ability”, respectively. Given forthcoming regulatory initiatives regarding Liquidity Coverage Ratio, Net Stable Funding Ratio and the trading-book review (in particular the proposed new rules concerning separation of banking vs. trading book and the introduction of “liquidity horizons”), we strongly believe that Prudent valuation rules should be fine-tuned to the other initiatives so that bank regulation in its entirety would be consistent. In particular, once the new rules will be applied, “holding ability” will be dramatically strengthened and the realisation of only less beneficial outcomes (e.g. “fire-sale prices”) will be less likely. Therefore we also think that in cases where institutions can demonstrate that their assets are fully funded (to maturity), there should not be a requirement for prudent valuation. We also see the danger that the proposed rules in fact push institutions to classify their assets in a “loans and receivables” category where it is possible to evade the requirement to value them prudently as well as to forego the added information that comes with fair value. Historical cost, in this regard, is less informative and insensitive to the signals market prices emit, which hampers managements’ decision-making.From the analysis of the “impact of the proposals” in 5.2.14-18 in our view an important issue is missing. EBA did not address the future enlarged volatility of bank capital. To date bank capital varies with the variation in the fair values of its assets. When crisis looms, on account of the target level of certainty, prudent value will be far more volatile than fair value given ever more disparate data for valuation inputs (pro-cyclicality). At the same time, even the problem definition in 5.2.4 is somewhat flawed, since EBA recurs to a notion of “true realisable value”: A bank may opt NOT to realise value. It may particularly opt so in cases where asset prices fall, but sufficient funding is available to hold the concerned assets on its books, acting in fact like a long-term investor who looks “through-the-cycle”. In this context we would like to encourage EBA to rethink its definition and its understanding of the essence of prudence in valuation. The dictionary gives a good definition of the meaning of the word “prudent” [see http://dictionary.reference.com/browse/prudent]: “wise or judicious in practical affairs; sagacious; discreet or circumspect; sober; carful in providing for the future; provident”. Thus, the requirement in CRR for a “prudent valuation” also encompasses the requirement for a “sober and circumspect valuation”.
We strongly believe that bank capital should not be too tightly tied to the whims of the market, which is driven not only by (intrinsic) value, but also by (market) liquidity. When liquidity is pushed into markets and prices go up, valuations should be conservative, but on the other hand, when liquidity evaporates, valuations should be allowed to become more “aggressive” in the sense that they should more accurately reflect intrinsic value instead of market doom. Applying prudent valuation in its proposed form bears the risk of aggravating a deteriorating situation. In the recent financial crisis, prudent valuation requirements would have led to significant intensification of stress in the banking sector. In a steep market downturn, therefore, prudent valuation requirement should be enriched with a countercyclical measure, allowing for mid or even ask prices to be used in prudent, i.e. sober and circumspect valuation. The conception of a prudent value then would entail distrusting the markets “animal spirits” where herd-like behaviour prevails and liquidity dissipates. Measures that could be incorporated into the rules in order to install a “relief valve” could be, for example:
• Bloomberg’s Relative Strength Index which measures the momentum of a security’s price to determine if it is an overbought or an oversold condition and connecting this to valuation;
• comparing the current level of some volatility index (e.g. VIX) to its long-run average and relaxing prudent valuation requirements once certain thresholds are breached;
• indices calculated by central banks, such as the “St. Louis Federal Reserve Bank Financial Stress Index” could be used to set “dynamic target levels of certainty”.
We believe that our proposed notion of prudence is fully in line with Capital Requirements Regulation articles 105 (1) and (5). The benefits of such a notion of prudence would be a better recognition of the intrinsic virtues of fair valued financial instruments as a liquidity resource (in contrast to loans, for example), counter-cyclical regulatory impulses and increased bank capital stability. Indeed, to address the procyclicality of the financial system by, for example, stipulating the accumulation of buffers in “good times” so that these can be drawn down in “bad times” is a crucial component of the macroprudential approach. Tools which are already used in this regard include countercyclical capital buffers or dynamic loan loss provisioning.
An issue we miss from the impact analysis is the complexity of the RTS and the complexity the RTS adds to management reporting, bank controlling and auditing. Interpretation of the rules governing prudent valuation, implementation of systems and processes and auditing a prudent valuation approach will potentially and likely increase the differences across banks and will also add complexity to regulators and auditors alike.