Minutes of the CNB Bank Board meeting on financial stability issues on 23 May 2019

Present at the meeting: Jiří Rusnok, Marek Mora, Tomáš Nidetzký, Vojtěch Benda, Oldřich Dědek, Tomáš Holub, Aleš Michl.

The meeting opened with a presentation of Financial Stability Report 2018/2019, which focused on financial cycle indicators, the assessment of systemic risks and the evaluation of compliance with the CNB Recommendation concerning the provision of mortgage loans following the introduction of upper limits on the DTI and DSTI ratios with effect from October 2018. Following the opening presentation, the Bank Board discussed the setting of the countercyclical capital buffer (CCyB) rate and potential changes to macroprudential measures targeted at risks connected with mortgage loan provision.

The presentation by the Financial Stability Department on CCyB issues emphasised that, according to the aggregate Financial Cycle Indicator as well as other indicators, the domestic economy had shifted further in the growth phase of the financial cycle during the second half of 2018. In the first few months of this year, the trends from the second half of last year had continued with the exception of the volume of new loans to households for house purchase, which had recorded a year-on-year decline in connection with the effect of the limits on the DTI and DSTI ratios recommended by the CNB from October 2018. At the same time, according to the results of the analyses presented, the partial signals of vulnerability of the domestic banking sector to a potential adverse change in conditions had intensified in the second half of 2018. These signals included the provision of a relatively high volume of loans with riskier characteristics, a decline in risk weights for house purchase loans and low provisioning, which may not be sustainable from the longer-term perspective. The amendment to the CRD V/CRR II regulatory package approved by the European Council and the European Parliament in December 2018 would also act towards higher vulnerability of the domestic banking sector in the future, reducing room for the CNB to set capital buffers for some systemically important banks.

In the discussion that followed the presentation, the board members agreed that cyclical risks to which the CNB should respond in a preventive manner had emerged in the domestic banking sector in the second half of 2018. In connection with the year-on-year decrease in the volume of new mortgage loans at the end of last year and in the first few months of this year, doubts were expressed whether the evolution of cyclical indicators in the second half of 2018 was a sufficient argument for increasing the CCyB rate. Signs of slowing growth in loans to households and empirical evidence of weakening growth in housing prices were cited as the grounds for those doubts. In this context, it was pointed out that there was significant uncertainty regarding the possibility to reliably quantify the risks identified from the perspective of the appropriate setting of the CCyB rate. One source of uncertainty was the market reaction to the DTI and DSTI limits introduced in October 2018, more specifically whether the observed year-on-year decrease in new mortgage loans had been due to frontloading during last year or may represent a longer-lasting effect. Additionally, it was discussed what weight should be attached in decisions on setting the CCyB rate to factors linked with banking sector vulnerability as opposed to potential credit losses related to cyclical risks, for whose coverage the CCyB is primarily intended. The prevailing view in the discussion was that potential credit losses and some sources of vulnerability were cyclical to a significant extent and that it was therefore necessary to take them into account when setting macroprudential policy. It was also said that the methodology for setting the CCyB should take into account to a greater extent the question whether funds generated by the existing CCyB rate were already sufficient to cover the monitored cyclical risks.

The Bank Board also discussed whether the financial cycle in the domestic economy was already at its peak or even past it. A majority of the board members were of the opinion that the financial cycle was probably close to its peak and therefore the increase in the CCyB rate under discussion could be the last one in this cycle. Furthermore, the discussion concerned potential costs of increasing the CCyB rate on the price and availability of credit. The board members agreed that, given the existing capital surpluses and favourable outlook for bank profitability, an increase in the CCyB rate would not be an impulse for an additional tightening of credit conditions or a reduction in credit supply. In this context, it was emphasised that an increase in the CCyB rate would take effect at the one-year horizon and that it would be possible to revise it should the economic situation change. The Bank Board also agreed that in the period ahead attention would have to be paid to the preparation of a methodology for lowering the CCyB rate in the event of a turnaround in the financial cycle. It would also be necessary to analyse the procyclical elements embedded in the IFRS 9 accounting standard and assess the necessary form and degree of a reaction to the associated risks. Given the risk of regulatory arbitrage, it would also be necessary to monitor developments in the CCyB rate in other EU countries.

The second key part of the meeting focused on risks connected with the provision of loans for house purchase and the residential property market. According to the analyses presented, apartment price overvaluation had grown further in the second half of 2018 and currently stood at around 15% according to both methods used by the CNB. The high volumes of new mortgage loans provided in the second half of 2018 had to some extent reflected efforts to obtain loans in the months just before the recommended DTI and DSTI limits had been introduced. As expected, new mortgage loans provided had dropped at the end of 2018 and in the first few months of 2019. Banks were mostly compliant with the recommended LTV limits. The share of loans with LTVs of 80%–90%, which can account for a maximum of 15% of new loans, had decreased slightly further, reaching 9% in December 2018. Banks had provided some loans with LTVs above the individual limit of 90%. The share of these loans had remained low, falling below 2% of new loans in 2018 Q4.

With regard to compliance with the recommended DTI and DSTI limits, in 2018 Q3 – immediately before the recommended DTI and DSTI limits took effect – credit institutions had been lending to a significant extent to clients who had high additional debt service. It was mentioned repeatedly that was increasing the potential vulnerability of domestic banks. Following the introduction of the recommended limits in October 2018, this trend had been reversed and the shares of loans with DTIs and DSTIs above the recommended limits had started to head towards the 5% exemption. The adjustment process had not yet been completed and institutions had been non-compliant with the recommended limits overall in 2018 Q4. This mainly pertained to the DSTI ratio. The share of loans with DSTIs of over 45% had exceeded 24% of new loans in 2018 Q3, falling to less than 12% in Q4.

In the following discussion, the Bank Board agreed that, given the above estimate of housing price overvaluation, the current upper LTV limits could be regarded as upper bounds and it currently was not necessary to tighten them. However, continued growth in house price overvaluation could necessitate a reassessment of the sufficiency of the current limits. The Bank Board agreed with an adjustment of the reference base for the calculation of the volume of admissible exemptions from the individual LTV limit and the DTI and DSTI limits, which is included in the CNB Recommendation. This harmonisation of the base for all the indicators at one-half of the volume of loans provided over the previous two quarters will facilitate lenders’ planning of their business activities. The Bank Board also supported a clarification of the provision on the assessment of compliance with the DTI and DSTI limits in cases of refinancing of an unsecured loan for consumption involving a loan increase for clients who already have a loan for consumption secured by residential property.

In addition, the Bank Board discussed whether it was desirable to respond to the rise in interest rates on mortgage loans from exceptionally low levels by changing the upper limit on the DSTI ratio. This limit is more restrictive for mortgage applicants and providers than the upper limit on the DTI ratio. Some of the board members agreed that the potential room for an interest rate shock had been reduced by interest rates on mortgage loans approaching 3% at the end of last year and that an increase in the upper limit on the DSTI ratio to 50% would not make new mortgage loans riskier. However, the opinion clearly prevailed that conditions providing sufficient support for an increase in this upper limit had not been fulfilled so far. The rise in interest rates on mortgage loans since the decision on setting the DSTI limit had been relatively small. At the same time, the limit of 45% had not been met sufficiently during the last quarter of 2018. The fact that, according to the stress tests of households, loans with DSTIs of over 40% could be regarded as highly risky, especially for households with relatively low income, was a supporting argument against increasing the upper limit on the DSTI ratio. The main source of risk is a deterioration in borrowers’ income situation due to a weakening of economic activity. Additionally, some of the board members regarded the upper limit on the DSTI as a structural stabiliser that should not be subject to frequent changes. The issue of potential implementation in the CNB Recommendation of less strict limits for young applicants envisaged by an amendment to the Act on the CNB was also mentioned. There was a consensus that with regard to uncertainty about the outcome of the legislative process, this principle could not be implemented at the moment. The topics of the upper limit on the DSTI ratio and other potential adjustments to the CNB Recommendation would be discussed at the next meeting on financial stability issues on 28 November 2019, depending on market developments and the outcome of the discussion of the amendment to the Act on the CNB.

Following the presentation of the Financial Stability Report and the subsequent discussion, the Bank Board decided to increase the countercyclical capital buffer rate for exposures located in the Czech Republic by 25 basis points to 2.0% with effect from 1 July 2020. At the same time, it agreed that the likelihood of a further increase in the countercyclical capital buffer rate had decreased significantly given that the domestic economy was probably close to the peak of the financial cycle.

Author of the minutes: Jan Frait, Executive Director, Financial Stability Department