Austrian Federal Economic Chamber, Division Bank & Insurance
The “natural bound of an instrument” as stated in Art. 3 (3) c should be specified.
The alternative application of the core or the simplified approach should be available for all institutions. As certain incentives to apply the core-approach exist, those institutions able to implement such an approach and to meet the administrative burdens and costs, should have the opportunity to do so. However mandatory application of an approach does not seem to be the ideal solution.
It should also be clarified, that in case the core approach is used on a parent institution level while the simplified approach is applied beneath, the parent institution is allowed to aggregate the data calculated with the simplified approach.
We support the idea that subsidiaries are allowed to use the simplified approach for their local reporting. Nevertheless differences in the amount of AVA can be substantial. Especially if the subsidiary has a high positive P&L the simplified approach may lead to a much higher AVA than the core approach.
The core approach does not have to be applicable for the subsidiary for consolidation.
We propose to solve this as follows: Each bank within the group should choose the method of AVA calculation based on its own discretion, not based on some threshold or based on the method chosen by the parent company also for consolidation purposes. The consolidation than would contain both simplified and core approach results. This would solve both ad 1.) and ad 2.)
In principal we do agree with a simplified approach, although double-counting has to be avoided. The disadvantage for institutions with a high amount of unrealized gains is substantial. Reducing the percentage from 25% and instead increasing the percentage of the overall fair valued assets and liabilities might be a less volatile alternative.
We do not understand the rationale for using also the unrealized profit as the base for the calculation (and presumably the more important one). It implies that instruments whose FV has decreased since their initial valuation actually lower the AVA costs (and vice versa, instruments with increasing FV increase AVA costs). In reality, there is no reason to think that the types of risk covered by this regulation have any connection to the upside/downside development of fair value of any individual instrument. We would agree with increasing the weight towards the total value of FV priced instruments. Further it might make the AVA provisions less volatile and unpredictable given the size of the portfolio.
If we consider PL as a source for AVA calculation, we do not comprehend why AVA is calculated from net realized profits and not from both net realized profits and losses.
In case the parameters used for calculation (25% and 0.1%) are outputs from a calibration exercise we think the RTS should describe how they were derived.
In our opinion we would exclude instruments which are central bank eligible. Those instruments can be used as collateral for immediate cash and therefore reduce the pressure to close the position. For the liability side we would differentiate between liabilities which are valued using the own funding curve and liabilities valued using a market implied CDS curve.
The provisions do not appear consistent from a systematic point of view.
This approach would lead to an extremely high additional value adjustment. If institutions are not able to calculate a single valuation exposure they should use a percentage of the aggregate absolute value of fair valued positions. The percentage should be retrieved from other banks which are able to calculate this AVA figure.
In principle the calculation is understandable, although a possible double-counting might occur. Furthermore the ongoing calculation on a regular basis implies a high amount of methodological, implementation, IT and validation efforts. It should be considered to apply already available tools, e.g. the VaR approach or the Bloomberg/Reuters scores for market depth.
Further the source of uncertainty mentioned in Art. 10 would already be covered by the Credit Valuation Adjustment.
For concentrated positions the information on the liquidity of a certain instrument is not always available. Those articles are very general, there is not much to comment on (especially the investing and funding AVA could me elaborated in more detail).
Neither the reason for including administrative costs in AVA is obvious, nor how it should be done.
1.) We propose to set AVA equal to 0 in case the asset is in Level 1 or Level 2 IFRS class.
2.) We propose to set administrative costs AVA equal to 0 for very standardized and liquid instruments.
The capital requirements for OpRisk are already specifically met, independent from any AMA.
We welcome the possibility to reduce the aggregated total amount of AVA to 50 %.
A precise answer cannot be delivered without knowing the exact methodology for the calculation of individual components nevertheless a simple sum might lead to double counting and should be avoided.
This will depend on the fact if the core approach will be extended to each subsidiary of a parent institution using the core approach or if the parent institution will have the possibility to aggregate data calculated on basis of the simplified approach (see Question 2). Otherwise the implementation costs would be substantial.