The categories listed in Article 1 are correct, however they are not all the factors which should be taken into account. It is silently assumed that losses may only result from an insufficient collateral value. However, the demand to increase collateral will not solve the problem as the quality of loans, especially long-term loans, has been gradually deteriorating. This factor must also be reflected in the standard.
It is worth emphasizing that observations of long-term trends, both in immovable property markets and markets for loans secured by mortgages, are relatively short-term in the new European Union member states, which is due to the fact that free-market economies have a relatively short history in those member states.
As far as the Polish market is concerned, what is also important is the share of retail exposures secured by mortgages in banks’ assets, which is relatively high, and the share of exposures financing commercial immovable property in banks’ assets, which is relatively low. Our lending history is too short for adequate conclusions to be drawn as to the behaviour of portfolios at different stages of the economic cycle. Therefore, greater caution on the part of the competent authorities may be warranted.
The main categories of conditions specified for the purpose of determining of risk weights and LGD values are as expected. The level of generality, particularly of the provisions involving financial stability considerations and other conditions, is high and gives a high degree of discretion to competent authorities; however, that is warranted by the wide scope of application and the required appropriate level of generality of the guidance. This also means that the final assessment of the method of calibration of risk weights and the LGD parameter will depend on the practical solutions used by each competent authority, which are not yet known.
It seems that the high number of conditions which must be analyzed by the competent authorities of a given country in order to increase minimum risk weights and LGD floors will make it possible to address the issue comprehensively. However, it is important to have data on the basis of which it can be reliably stated that the risk weight and LGD floor increase criteria are met. Competent authorities should be required to create and support the development of inter-bank databases comprising such information.
It should be borne in mind, however, that for the Polish banking system, which is already applying stricter requirements for the financing of the immovable property market than is the case in other countries (a higher risk weight, Recommendation S requirements), any further increases of prudential requirements for the financing of immovable property in Poland will result in the inflow of competitive financing from abroad, giving rise to a deterioration in the financial situation of the Polish banking sector. An additional condition should be created, namely that banks which are not located in a given country but which invest in it should also be required to increase the risk weights or LGD floors (as is the case with the anti-cyclical buffer).
Setting higher risk weights is required both from banks which apply the Standardised Approach (STA) and from those which apply the Advanced Approach (IRB) The higher risk weights should not distort the competition between banks which apply different methods (STA or IRB). Banks which apply the IRB approach are expected to modify the LGD parameter. At the same time, capital requirements are also affected by other parameters (PD, CCF, etc.), which the institution may control to a certain extent to its own benefit. On the other hand, the LGD parameter is not constant over time, hence it may to a certain extent reflect the current market conditions, therefore its increase may multiply the effect of its introduction.
The abundance of factors which may cause risk weights to be increased, and their application in different countries by respective competent authorities, may give rise to differences in the assessment of the situation by the competent authorities in each country. This mechanism may also improve competitiveness of banks from specific countries (at the expense of their stability).
You cannot identify forward-looking residential immovable property market developments on the young Polish market with adequate probability, what with the unstable tax laws and frequently-changing residential property buyer support schemes. Forward-looking commercial immovable property market developments may be identified with higher probability and the answer may be affirmative in this respect.
Moreover, the adopted basis for the determination of the loss experience and the loss expectation is imprecise. In our opinion, the statistics set out in Article 101(1) CRR do not make it possible to reliably determine the economic loss experience. The loss experience (or rather its changes over time) is in turn supposed to be the basis for the determination of the related loss expectation. What is also relevant to the setting of risk weights is the potential future level of unexpected losses and not only current loss expectation forecasts. In particular, the ratio of the unexpected loss to the loss expectation also changes over time and depends on several factors, not all of which are indicated.
From the perspective of some countries (including Poland), what is missing is perhaps a factor involving exchange rate volatility where there is a significant proportion of loans denominated in foreign currencies.
The approach expressed in Questions 2 and 3 is incorrect also for another reason. For assets secured by mortgages, the relationship between losses on the mortgage portfolio and the risk weights is much more complex than Article 2 of the draft appears to suggest. Loans secured by mortgages are of high quality early in their tenure. Their quality begins to deteriorate after about five to seven years. The level of losses will be completely different on a stable market with a stable and small growth rate than on a market where the growth rate of new loans granted is high and variable. In particular, the introduction of loss limits as referred to in the explanation to Questions 2 and 3 is hard to accept as the draft RTS are to be qualitative in nature, indicating the necessary analysis factors and the procedure for the translation of the results into specific risk weights. However, they must not impose any values. No one should have the impression that the purpose of the RTS is to prevent excessive risk weight increases rather than setting them too low. Therefore, no additional limitations should be imposed on competent authorities, going beyond those specified explicitly in Articles 124, 126 and 164 CRR.
a. The differentiation of weights due to national specificities is by all means reasonable. Although the differentiation of the benchmarks across different countries is necessary, EBA should not interfere with the detailed determinations made by national competent authorities (unless there are dramatic deviations) as that diminishes the responsibility of each bank and local regulator for rational decisions in this area.
b. For loss expectation justifying the 35% risk weight of exposures, the lower limit of 0.1% is too low, whereas the upper limit of 1.5% is too high (it could be 0.75%). For the 50% risk weight, the limit could be set at 1.0%.
Yes, although the wording of Article 3(1)(a) gives preference to the old European Union member states where the majority of global systemically important institutions are located.
The entire Article 4 appears somewhat unclear – by definition, competent authorities are supposed to be able to take account of other factors" on the basis thereof, while the provisions in fact boil down again to the setting of benchmarks for the loss expectation."
a. The average LGD should be statistical in nature and it should not be the benchmark for the specification of indicators in national markets.
b. The other qualitative conditions seem to be selected correctly. The quantitative recommendations, however, should be removed. The market for immovable property and loans secured by mortgages on immovable property is very complex, strongly dependent on several national macroeconomic factors. For that reason, unconditionally binding quantitative recommendations will be absolutely inadequate to the actual problem and they may do more harm than good. Moreover, if the solution was so simple as the draft RTS suggest, the legislator would place it in the CRR instead of introducing a national option in the Regulation.
It should be borne in mind that the historical data is often incomparable in the new European Union member states due to the system transformation which they have undergone; therefore, any developments used as the basis for adjustments have lasted not more than about a dozen years.
c. An immovable property market is a local market, with all the consequences thereof. We believe that the specificities of national immovable property markets should be taken into account.
Yes, although the sentence in (2)(c) should probably be verified – the setting of higher minimum LGD values should have anti-cyclical and not pro-cyclical effects.
It is a very good thing that the regulator invites suggestions on the effects of the regulation as part of the consultation process. A relevant expert opinion is definitely worth preparing. This would, however, be both time- and work-consuming. This is a task to be performed over a longer period of time, by a larger team of experts.
This data set does not seem to be necessary, yet it will do no harm if it is treated as an illustration. It does not seem to be necessary to point competent authorities to the essential variables as the authorities usually perform more complex analyses and they are also familiar with financial stability considerations and their assessment. However, although it is redundant, it may stay if this will improve the quality of the RTS.
It is worth adding the following indicators:
• time needed to complete a sale transaction on the market for a specific type of immovable property;
• the average immovable property debt collection period.