Primary tabs

Deutsche Bank

The minimum list is reasonable; however, the relevant data sources utilised should be those which are most appropriate given current market conditions. They should utilise consistent sources of expertise, technology and process.

Additionally, clarity is needed within Article 3(2) and Article 3(3). This clarity can be achieved by amending the text as suggested below (emphasis added to represent the amendments):

Article 3(2)
“The market data used to determine a prudent value shall consider available and reliable data sources, including the following, where relevant:”

Article 3(3)
“For cases where an expert based approach is applied for the purpose of Articles 8 to 10, alternative methods and sources of information shall be considered, including the following, where relevant.”

Moreover in both Articles the list of information sources should be flexible so that they are not all required for every valuation but are an indicative list of sources that can be used individually or collectively.
In our opinion the use of a threshold to determine when to use the simplified approach is appropriate and the proposed level is reasonable. However, this approach would be the most useful if it takes into account only on-balance sheet fair-valued assets. The inclusion of off-balance sheet fair-valued assets could cause confusion and incomparability across banks positions. Furthermore, we believe that the inclusion of liabilities in the threshold is not useful and is not aligned with the scope in Article 34 of the Capital Requirements Regulation (CRR) which only pertains to assets. Thus, we would request that the provisions contained in the text be revised to reflect this change.

DB also supports the concept that the AVA calculation take into account multipliers for different levels of the Fair Value Hierarchy (FVH) and amendments to the RTS text which includes the definitions of FVH taken from GAAP.
In DB’s view this provision does not add any additional operational cost or burden beyond that which would be placed upon the parent institution’s core requirements. However, we do question the benefit of potentially pursuing differing approaches within a single bank.
We recommend the utilisation of the FVH as a measure of observability and liquidity, which adheres to the spirit of the regulation, whilst also providing a simplified approach.

In our opinion, life-to-date unrealised P&L is not a fair or consistent measure of valuation uncertainty and would result in unanticipated outcomes given timing differences, the use of FIFO, LIFO or average cost, and the direction of the position held. Additionally, this would be very costly to implement for very limited benefit as a means of managing the valuation uncertainty of the organisation.
Please see our response to Question 4.
In our view, the approach detailed in the Articles is overly punitive and does not take into account the unintended impact of their implementation as proposed. Three examples are:

First, unrealised profits and losses are typically not stored within a bank’s systems. A change to this infrastructure would require significant development and cost.

Second, the 25% charge related to balance sheet market value is approximately 250 times as large as the charge proposed under the simplified method.

Third, the charge in Article 7 is pro-cyclical and in distressed market conditions estimating the charge could be impracticable.

Thus, a phased-in approach to the requirements would be beneficial to allow for all necessary changes to be made to systems and processes.
In relation to the detailed approach supplied, we agree with AFME that any surface generated to facilitate the calculation of Market Price AVA should not be unrealistic i.e. should not create discontinuities, for example, via the use of points directionally biased by specific risk exposures. This could, for instance, be an unintended consequence of a literal interpretation of the example provided, where we envisage problems producing realistic forwards that can be used by the bank’s analytic models and also when trying to extend the logic to a surface (e.g. for volatility) in practice.

The back-testing requirement as proposed is impractical for most circumstances (and extremely expensive where possible). We propose a less frequent and more informative test be derived that would fulfill the intended requirement, and make better use of firms' resources. Such a regime could include sampling of transactions, combined with information gathered from the independent price verification processes (i.e. leveraging existing practices).
Article 11: It is appropriate to consider Model Risk uncertainty alongside both Market Price and Close-out Cost uncertainty and therefore qualifying for the same aggregated AVA diversification treatment.

Article 12: This approach is reasonable. Notwithstanding this view, when defining liquidity, one should consider all potential market activity.

Article 13: The text should be amended to highlight that it is the expected contractual life of the trade from which to calculate the costs and benefits.

Article 14: The future administrative cost text is incongruent with our understanding of the AVA’s. From our perspective the AVAs are designed to estimate a prudent exit value. Therefore, in our opinion, a bank should be allowed to calculate this AVA by taking into account the lower of the costs incurred managing the portfolio or the incremental costs charged by a market participant managing the portfolio.

Article 15: The text is reasonable and we do not have any recommended changes.

Article 16: In our opinion operational risk additional valuation adjustments (AVAs) should not be included in a prudent valuation regime. Operational risk is already captured through the Advanced Measurement Approach (AMA).
We do not believe that a further definition is required beyond the guidance provided in the RTS.
We agree with the approach but reiterate the point we made in response to Question 8 where we discussed our view of model risk under Article 11.
It should be recognised that the outcome will tend to be more conservative through the simple aggregation of the constituent AVA.
Please also refer to our responses to Questions 8 and 10.
The approach as detailed offers minimal benefit in return for the material operational expense that would be incurred. As a result, we propose that a more stream-lined approach which leverages existing industry-wide processes be considered.
As detailed in our previous responses, we believe that there are areas within the consultation paper that would lead to a material increase in operational expenses without discernible benefits.

In particular we would propose that a review of requirements such as life-to-date unrealised P&L, the 100 days volatility measures and prescribed monitoring tools be revisited.
Marta Gaska