ESRB European Systemic Risk Board : Macroprudential approaches to non-performing loans

Macroprudential approaches to non-performing loans

January 2019

Contents

Executive Summary 2

1 Introduction 7

2 Drivers of system-wide increases in NPLs: lessons from the recent

crisis in Europe 10

2.1 Business cycle and asset price shocks 11

2.2 Vulnerabilities that built up before the crisis 12

2.3 Structural factors: the legal and judicial system 18

3 The role of macroprudential policy 20

3.1 Existing macroprudential toolkits 22

3.2 Early warning systems for signalling risks related to a system-wide

build-up of NPLs 23

3.3 Borrower-based measures 26

3.4 Capital measures 29

3.5 Dealing with vulnerabilities and structural factors falling outside the

scope of macroprudential policy 37

4 Concluding remarks

Annex 1 - Measures falling outside the scope of macroprudential policy

40 44

Annex 2 - The use of the SyRB to address structural NPL problems - the case of Romania

46

Annex 3 - The use of the SyRB to manage risks arising from problem CRE project loans - the case of Hungary

47

Annex 4 - The use of large exposure limits to address the over-indebtedness of NFCs - the case of France

References

Imprint and acknowlegements

49 50 56

Macroprudential approaches to non performing loans / Contents

Executive Summary

This report presents an analysis carried out by the ESRB in response to a Council of the European Union request to develop "macroprudential approaches to prevent the emergence

of system-wide NPL problems, while taking due consideration of procyclical effects of measures addressing NPLs' stocks and potential effects on financial stability". Relying on the experience and expertise of ESRB members, especially those from Member States in which system-wide increases in non-performing loans (NPLs) were observed in the aftermath of the recent crisis, the report begins by identifying the main triggers, vulnerabilities and amplifiers that can drive system-wide increases in NPLs. With these drivers in mind, the report then focuses on the role that macroprudential policy can play in preventing system-wide increases in NPLs and/or in increasing banks' resilience in the face of such increases.

The emergence and accumulation of NPLs can become a systemic problem when this affects a considerable part of the financial system, threatening its stability and/or impairing its core function of facilitating financial intermediation. A significant increase in NPLs

throughout the system can have a negative impact on the resilience of the banking sector to shocks, thus increasing systemic risk. NPLs may also be associated with higher funding costs and a lower supply of credit to the real economy. This may result from negative market sentiment towards banks with high levels of NPLs, which decreases banks' ability to access liquidity and capital markets (potentially leading to credit supply constraints). In the European Union, systemic concerns arose from the abnormally high proportion of NPLs which accumulated on banks' balance

sheets during the crisis, and their persistence after the crisis.

The report identifies business cycle and asset price shocks as two of the main drivers of system-wide increases in NPLs. A downturn of the business cycle and/or negative asset price shocks, particularly in sectors to which the banking sector is significantly exposed (e.g. residential real estate (RRE) and commercial real estate (CRE)), may trigger a system-wide increase in NPLs. In some cases such increases may also be associated with instances of significant resource reallocation within the economy.

The role played by vulnerabilities that built up before the crisis, i.e. (i) high indebtedness and excessive credit growth and (ii) underlying bank practices and governance, is also highlighted, as well as the role played by the judicial and legal systems. The increase in debt

levels seen in some EU Member States before the crisis made non-financial corporations (NFCs)

and households particularly vulnerable to negative shocks, especially when accompanied by a low capacity to generate internal capital and/or by low savings rates. Banks' internal incentives, culture

and practices also played an important role in explaining their exposure to imbalances that had built up before the crisis, as well as the impact of the crisis on their balance sheets.

Macroprudential policy messages

No fundamental changes to the existing macroprudential toolkits seem to be required, although a number of refinements should be considered. In particular, further work is needed

in areas relating to the use of sectoral capital buffers and the development of borrower-based measures (which are not harmonised at the European level).

Macroprudential approaches to non-performing loans

Executive Summary

•  Macroprudential authorities should develop early warning systems (EWSs) to monitor the risks of credit portfolio deterioration from a macroprudential perspective. Significant advances have been made in recent years in the development of EWSs for financial crises, although their focus is not specifically on signalling, at an early stage, potential system-wide increases in NPLs. The use of crisis prediction indicators might not, therefore, be the most appropriate way to monitor the risks of credit portfolio deterioration. Further research is thus warranted, both at EU and Member States level, to identify the systemic risk signals relating to a potential build-up of NPLs. The use of micro-datasets, at both bank and borrower level, can be a particularly useful way to identify vulnerabilities building up in specific sectors, or subsets of borrowers, that might become problematic in the future. The setting up of an EWS will also be crucial in that it will allow macroprudential authorities to communicate their views regarding the risks underlying potential system-wide increases in NPLs, and the need to address these at an early stage.

•  All Members States should include borrower-based measures in their national macroprudential toolkits, given the important role these instruments play in mitigating the vulnerabilities underlying the first stage of the lifecycle of a potential NPL, and their potential to lessen the adverse effects associated with credit misallocation.

  Nonetheless, so that borrower-based measures can mitigate different types of risk, and despite the need to guarantee a certain degree of harmonisation of definitions, decision-making should be kept at national macroprudential authority level. This will ensure flexibility with regard to the design, calibration and implementation of the most suitable set of tools.

•  Further exploratory work should be carried out on borrower-based measures for NFCs, notably at ESRB and European Central Bank (ECB) levels. The development of the aforementioned EWSs should be a good starting point for assessing whether borrower-based measures for NFCs can actually be developed further, given the underlying challenges in designing, calibrating and implementing them. This draws attention to the need to collect reliable data on borrowers' income (and sources of income), and to guarantee adequate collateral valuation. Given the absence of borrower-based measures for NFCs, macroprudential responses may, for the moment, make use of capital-based measures that target NFC exposures.

•  Capital-based instruments should also be considered for addressing vulnerabilities that might later result in system-wide increases in NPLs. The choice among existing capital tools will depend on to the degree to which the latent vulnerability is system-wide and accompanied by excessive credit growth, or affects only specific borrowing sectors or groups of lending institutions.

•  Macroprudential authorities should use the countercyclical capital buffer (CCyB) to prevent the systemic build-up of macro-financial imbalances and/or increase banks' resilience when dealing with NPL-related vulnerabilities. Additionally, releasing the buffer during a downturn phase, in order to ensure that lending continues to flow to the economy, is consistent with the need for banks to have more room for manoeuvre to clean up potential NPLs at an early stage. The financial cycle should be taken into consideration when choosing between capital and borrower-based instruments: there is an ongoing discussion with regard to the tightening of borrower-based measures earlier in the cycle (as their focus is mainly on preventing the origination of high-risk loans) and the use of the CCyB when there is a shift in

the financial cycle to a phase of stronger credit growth accompanied by an easing of credit standards (thus supporting banks' resilience and the cleaning up of NPLs at an early stage).

•  Macroprudential authorities should consider using the systemic risk buffer (SyRB) when the potential systemic increase in NPL flows is associated with developments in specific market segments or types of debtors as opposed to situations of generalised excessive credit growth. The SyRB can be tailored to become a targeted instrument, better suited to dealing with the cross-section structural nature of systemic risk which constitutes the main sources of system-wide increases in NPLs. The implementation of the so-called "banking package" is expected to further increase the flexibility of the SyRB, notably through the introduction of the sectoral SyRB and the removal of the reference to its "long-term, non- cyclical" nature. Sufficient flexibility should be provided - notably in EU-related regulation - in

  the definition of credit segments that can be targeted using the sectoral SyRB, as the source of a system-wide NPL build-up can change over time.

•  Macroprudential authorities can also use capital measures aimed at addressing excessive exposure concentrations when systemic risk appears to be building up in specific sectors/asset classes. Tightening large exposure requirements may mitigate concentration risk and the risk of shock propagation through the financial system, although it can also have an indirect impact by mitigating and preventing excessive credit growth.

  Additionally, higher own funds requirements can be applied by the designated authority in order to target asset bubbles in the residential and commercial property sector, and higher risk weights for RRE and CRE, or stricter loss given default (LGD) parameters, can be applied by national competent authorities on the basis of financial stability considerations.

•  When macroprudential authorities apply more targeted measures, they should follow a prudent approach in order to avoid procyclical effects and negative spillovers. Some capital-based measures may have (i) procyclical features, since they depend on the level of own funds and (ii) a significant procyclical impact if applied when risks have already materialised. Borrower-based measures may allow some potential procyclical effects to be avoided, as they do not apply to the stock of existing loans while, at the same time, they contribute to preventing a system-wide increase in NPLs. Borrower-based measures may, in fact, be particularly suitable for addressing vulnerabilities related to credit misallocation at an early stage of development (or in a preventive manner). In addition, since they do not have an impact on the stock of existing loans (they only target new loans) they may allow potential procyclical effects to be avoided. Macroprudential authorities should seek to follow a comprehensive approach by assessing and avoiding this kind of procyclical feature and, if necessary, by combining different measures with different activation and release timings. With regard to the potential negative spillover effects underlying the more targeted macroprudential measures, special attention should be paid to: (i) whether the heterogeneity underlying NFCs does or does not hinder the design of effective borrower-based measures applicable to this sector; (ii) whether adopting such measures may, in the end, simply shift the risk to other sectors not targeted by the measure and; (iii) whether these measures may lead to spillovers to non-bank financial institutions, notably by substituting bank-based financial intermediation with non-bank intermediation, and which measures should be adopted to avoid such spillovers. A more targeted approach may even lead to the reallocation of lending across