German Banking Industry Committee - GBIC

The three main categories (loss experience, forward-looking market developments, financial stability considerations) under which higher risk weights or higher LGD floors may be imposed result from Article 124(2) and Article 164(5) CRR. Insofar, the draft RTS follows the CRR specifications. However, what is of crucial importance is the methodology applied to initiate a prudential assessment of the appropriateness of the capital requirements on the basis of the joint analysis of the three main categories mentioned.
As we understand the CRR specifications, loss experience is the starting point and hence the basis of the analysis. And this makes sense. As this is experience gathered in the past, it is necessary to make quantitative adjustments for anticipated market price and volatility trends on the basis of historical values where market expectations are deviating. Increased loss expectations for the adjusted historical values should be a necessary precondition to get to increased risk-weights or increased LGD floors. Financial stability considerations should come into play only if increased loss expectations exist as a necessary precondition. The capital requirements assessment process should be initiated only when these considerations as well show systemic risks which are not already reflected in the adjusted loss expectations.
We welcome a possible wide analysis basis provided by the other conditions" which can cover numerous national conditions."
We basically find that recourse to the data of the surveys pursuant to Article 101 CRR is reasonable to determine the loss experience. With that, a uniform data basis is applied and the traceability of the results is strengthened. However, the CRR purposely does not directly link the loss experience values to the privileged risk-weighting.
With respect to the adjustments of the loss experience to anticipated market developments, Article 2(2) of the draft RTS mentions substantial influencing factors for adjustment. Such forecasts – in particular if they include forecast horizons for the fulfilment of the expectations - are highly complex. It is important to state first that the market expectations need not necessarily result in the loss experience having to be adjusted upward. It must still be possible in principle to expect lower losses than predicted by the historical data. We are critical of the adjustment of the loss experience also because as far as we know the supervisory authorities as well have no experience with the handling of such a procedure. We, therefore, suggest that the competent supervisory authority create and consult its own procedure for deriving loss expectation from loss experience based on the influencing factors mentioned in Article 2(2) of the draft RTS.
Moreover, it appears to be unclear whether the EBA has intended that only a selection of criteria needs to be fulfilled for the risk-weights in the standardised approach to be appropriate (Article 2(2) of the draft RTS: such adjustments shall be based on any of the following"), while all criteria need to be fulfilled for the minimum LGD values to be appropriate (Article 5(2) of the draft RTS: "such adjustments shall be based on all of the following"). We ask to clarify this.
In our opinion, the criteria mentioned for the assessment of the appropriateness of risk-weights and LGD floors give the competent authorities in the EU Member States a very wide scope of discretion. The definition of further conditions in Article 4 and Article 6 of the draft RTS, which are to be taken into account in addition if the risk-weights or the LGD floors are to be increased, even further widen the authorities' scope of discretion. Our experience is that forward projection and assessment of real estate market prices and volatilities often causes difficulties. To enable the prediction of future trends with a higher degree of certainty, we believe it is necessary to determine the strived-for level playing field for the competent authorities within the Eurozone in more detail by means of test criteria. We see a danger that more favourable competitive conditions for the financing of real estate investments can be created arbitrarily within the EU Member States."
We generally support the use of benchmarks, as long as they bare an indicative character and as long as there is now an automatic increase in risk weights if the limit is exceeded. A breach of the benchmark should only trigger a screening process without a predetermined conclusion.
We propose to take the initial risk weights of 35% or 50% as starting point and base the calibration on the level of loss expectations for setting higher risk weights that takes into account these initial risk weights.
As a matter of fact the maximum risk weight of 1250% used in the CRR is calibrated to cover a loss up to 100% of the exposure value. Therefore, a risk weight of 100% serves for covering up to 8% losses, thus the risk weights of 35% and 50% serve for covering losses up to 2.8% or 4% respectively.
We therefore believe that the level for the indicative benchmarks for increasing risk weights has to be significantly higher than the proposed 0.3%. The level of 0.3% mentioned in the consultation paper (see also point 5.1, page 24) is referring to the benchmark taken into account for the so called “hard test” for the preferential risk weights for residential and commercial real estate according to Article 125(3)(b) CRR and Article 126(3)(b) CRR. The aim of this benchmark has to be distincted from the target of the RTS.
This suggestion implies that in general the average loss implied by the initial preferential risk weights of 35% and 50% is the “normal” level to which the average losses reported under Article 101 CRR need to be compared. Given the 2.8% or 4% maximum loss coverage, the initial risk weights of 35% and 50% can be understood as in general sufficient for an average loss expectation of 1.4% or 2%, respectively.
By this means loss levels and risk weights would be combined: A level of losses of 2% is observed or expected for the portfolio of exposures secured by residential real estate. The level of 1.4% is exceeded, and therefore the risk weight of 35% may be increased by a ratio of 2% /1.4% (= 143%), which leads to a new risk weight of approximately 50%.
The value of 1.4% should therefore be inserted in point (a) of Article 2(4) and point (a) of Article 4(3), and the value of 2% should be inserted in point (b) of Article 2(4) and point (a) of Article 4(4) of the RTS. In addition this should also be reflected within the specification of the intervals.
The higher risk weight should in any case only be applied if this is deemed necessary under financial stability considerations.
We find that the considerations regarding the assessment of the financial stability are not specific enough. They build on (spill-over) effects on the resilience of the financial system or the lending activity which could result from SIFIs or other banks being impacted when higher loss rates occur. We consider it critical that the fact of SIFIs being impacted might also lead to higher requirements on many small institutions although they may serve other areas of the immovable property market.
The competent authorities have a wide scope of discretion here that we believe is appropriate. We doubt whether it is reasonable to impose higher capital requirements out of financial stability considerations when disturbances of the lending process are observed. This will normally be counterproductive and should, therefore, not be further pursued as a criterion.
We refer to our comments on the questions Q2 and Q3.
As already explained for the standardised approach risk-weights we basically believe it is proper that adjustments shall be possible also with respect to the minimum LGD. However, clearly increased adjusted LGD average estimates should be available first before higher floor requirements are imposed.
We support the view that there shall be no indicative benchmarks. This enables more flexible handling and also takes account of the existing relationship of the LGD parameters with other IRB parameters, in particular the PD, and the very specific national conditions.
On the other hand, we believe the scope of discretion of the respective competent authorities of the EU Member States is too wide as regards the potential increase of the floor because it is not limited. In our opinion, it is necessary here as well to determine the strived-for level playing field for the competent authorities of the EU Member States in more detail by means of test criteria. A cap for the LGD floor increase should be defined (max. 100% increase).
We refer to our comments on the question Q6.
Michael Engelhard