French Banking Federation

We would like to put forward the following suggestions in order to improve clarity of the draft Regulation and transparency of the NCAs decisions:
Article 1 of the draft Regulation states that NCAs shall take into account three categories of conditions. The third category of conditions, labelled as “other conditions”, is defined by article 4, when considering the opportunity to set higher risk weights (“RW”), and article 6 as regards setting higher minimum LGD values.
Articles 4 and 6 mentioned above comprise a first set of “other conditions” that includes:
• an explanation as to why increasing the RW or minimum LGD is deemed to mitigate the financial stability considerations, as defined in the draft Regulation ;
• an explanation as to why the RW or minimum LGD are increased in a way that ensures correspondence between the higher risk weight and the loss expectation related to one or more property segments/the higher minimum LGD value and the LGD expectation of the concerned portfolio;
• an assessment of the potential pro-cyclical effects of setting higher RW or minimum LGD values in the current stage of the economic cycle on the financial stability considerations.

We consider that these conditions, or steps, are essential for an NCA in the process of determining whether it should increase or not the RW or minimum LGD value. Therefore we would suggest amending the draft Regulation to include these conditions in article 1 as mandatory steps the NCAs must complete before deciding to increase RW or minimum LGD values and not categorize them as « other conditions ».

Moreover, the explanations and assessment mentioned above would be most instrumental for credit institutions to understand the decision taken by the NCA, particularly when the NCA decides to set higher minimum LGD values since the draft Regulation doesn’t include any LGD indicative benchmarks. These three conditions could therefore be considered as « transparency measures » and we would strongly support their disclosure, although the CRR does not include any requirement regarding such disclosure. However since CRR requires the EBA, pursuant to articles 124 (2) and 164 (5), to publish RW and LGD values, these elements could be included in such publication.
Article 2 (2) of the draft Regulation states that when adjusting the loss experience to determine the loss expectation, NCAs shall take different criteria into account and in particular:
• historical and expected evolution of prices and volatility in these prices in the immovable property market;
• time horizon over which the forward-looking property market developments are expected to materialise;
• fundamental drivers of demand and supply;
• past and present structural and cyclical characteristics of the immovable property market (historical volatility in immovable property prices, the size of the immovable property market, national taxation systems and the national regulatory provisions for buying, holding or letting immovable property).

These criteria are also used to determine the LGD expectation, pursuant to article 5 of the draft Regulation and the last criterion listed above, past and present structural and cyclical characteristics, participates to the definition of financial stability considerations (article 3 (1) (c) of the draft Regulation).
We suggest amending the wording of point (e) of article 2 (2) of the draft Regulation to allow NCAs to take into account other characteristics which could impact developments in the immovable property markets and are related to specific practices of certain Member States. In France for instance, the risk of default on mortgage loans remains low since credit extensions and lending conditions are based on the borrower's creditworthiness and thus on the stability of the borrower's income. This practice of assessing the risk of default is independent of changes in real estate property prices unlike in other Member States such as, in particular, the UK, Spain, Denmark or the Netherlands where property-value based mortgage loans are the rule. In those Member States, where loans are secured by the value of the property asset, households can make mortgage equity withdrawals for purposes other than home purchase against the equity in their property. Due to these risk assessment methods, residential mortgage loss rates incurred by institutions are structurally low in France regardless of property prices.
Point (e) of article 2 (2) of the draft Regulation could therefore be drafted as follows:
e) the past and present structural and cyclical characteristics of the immovable property market, where whether structural characteristics which may relate to the historical volatility in immovable property prices, the size of the immovable property market, national taxation systems and the national regulatory provisions for buying, holding or letting immovable property, cyclical characteristics or other characteristics deemed relevant to determine forward-looking developments ;

We would like to emphasize that the historical evolution of prices may not provide an adequate representation of the French real estate sector. The decision to lend in France is based on the solvency of the borrower; therefore loss ratios are better indicators of the resilience of the French real estate market. We suggest that the EBA considers loss ratios as important criteria amongst its “other conditions” when setting conditions for higher RW / LGD floors.
Furthermore we would appreciate a feedback from the EBA on “IP losses” reports (Article 101 (1) c) & f) CRR). It could be useful :
- to ensure consistency in practices among banks and Member States on how these reports are filled in;
- to establish what conclusions could be drawn on financial stability.
We do not share the EBA’s methodology underpinning benchmarks calculations. The 0.3% level of losses mentioned in paragraph 3(a) in articles 125 and 126 of the CRR is a precondition for the assessment of the level of guarantee applied to residential and commercial property respectively. The 0.3% reference is therefore not appropriate; as a result the lower bounds of the indicative benchmarks proposed by the EBA are not relevant.
The data brackets proposed by the EBA as indicative benchmarks under the standardised approach for loss expectations are too low both for the residential and for the commercial real estate portfolios. We believe proposed values are calculated on a purely theoretical basis and are not an adequate reflection of the portfolios risk profiles. We therefore would like to suggest that the EBA reviews these values in light of loss expectations back-testing based on loss experience data widely available at national competent authority level (for example on the “IP losses” reports). An impact assessment by the EBA of proposed indicative benchmarks would be welcome.
See our answer to question 2 as regards structural characteristics determining forward-looking immovable property market developments (article 3 (1) (c) of the draft Regulation).
See our answer to Question 1.
See our answer to Question 2 regarding the determination of loss expectation and the elements NCAs should base their adjustments on.
If the three conditions (explanations as to why increasing RW or minimum LGD values and how this increase ensures correspondence with loss or LGD expectation and assessment of pro-cyclical effects) are included in article 1 of the draft Regulation, as suggested in our answer to Question 1, article 6 could be deleted.
In addition if an NCA decides to take a macro-prudential measure to increase minimum LGD values, it should have demonstrated beforehand that the decision to increase capital requirements stabilizes the market and will not generate any pro-cyclical effects.
As pointed out in our general comments to this paper, the EBA’s consultation provides for more conservative prudential treatment of mortgage exposures by national NCAs; these decisions must rely on high-quality, independent and reliable data.
As the EBA points out on page 25 of its consultation paper “[…] data indicators are unfortunately not always available in all national immovable property markets […]”. As already stated in our general comments, we would like to raise the EBA’s and the NCA’s level of awareness to this point, especially as far as commercial real estate exposures are concerned, and encourage them to use extra care in relation to data sources underlying their decisions to increase RW / LGD floors on real estate exposures.
Alexandra Merlino