APPLICATION OF THE COUNTERCYCLICAL CAPITAL BUFFER IN LITHUANIA

Transcription

1 TEMINIŲ STRAIPSNIŲ SERIJA No 5 / 215 TRANSLATION APPLICATION OF THE COUNTERCYCLICAL CAPITAL BUFFER IN LITHUANIA Bank of Lithuania, 215 Reproduction for educational and non-commercial purposes is permitted provided that the source is acknowledged. Address Gedimino pr. 6 LT-113 Vilnius info@lb.lt

2 Abbreviations BCBS Basel Committee on Banking Supervision CCB countercyclical capital buffer CDS credit default swap CRD IV Capital Requirements Directive IV CRR Capital Requirements Regulation DSTI debt-service-to-income ratio EBA European Banking Authority ECB European Central Bank ESA 21 European System of Accounts 21 ESRB European Systemic Risk Board EU European Union GDP gross domestic product HP Hodrick-Prescott (filter) IMF International Monetary Fund LTV loan-to-value ratio MFIs monetary financial institutions RE real estate RW risk weights RWA risk weighted assets 2

3 CONTENTS Summary Purpose of the countercyclical capital buffer New capital requirements Purpose of the countercyclical capital buffer Reciprocity of the CCB rate Impact of the countercyclical capital buffer on the economy Impact transmission channels Impact on the economy Undesirable consequences of the application Setting of the countercyclical capital buffer rate Setting of the CCB rate in periods of increasing systemic risk Reduction or release of the CCB rate Practical implementation of decisions and their publication References Annex. Information on data used to compute the main and complementary indicators... 21

4 Summary 4 The main purpose of this paper is to inform market participants and the public about the principles for the use of a countercyclical capital buffer one of the instruments of macroprudential policy in Lithuania. The application of the CCB is regulated by the Rules for Capital Buffer Formation, approved by the Board of the Bank of Lithuania on 9 April 215. The Bank of Lithuania takes a decision on setting the CCB rate on a quarterly basis. The decision and background information for it are published on the Bank of Lithuania s website. A number of changes in financial sector s supervisory and regulatory system have been implemented in response to the implications of the global financial crisis and in attempt to prevent the build-up of imbalances within the financial sector. At a global level, a huge importance is given to broad-based standards Basel III 1 as one of the most significant regulatory initiatives, as well as to active implementation of macroprudential policy. Along with other requirements, the new Basel III framework provides for tighter requirements for the level and quality of bank capital. Most of the components of Basel III to be applied at EU level are included in the CRD IV, which was approved on 26 June The provisions of this Directive have been transposed into the national legal systems of EU Member States. Apart from other supervisory requirements, the CRD IV regulates the application of the countercyclical capital buffer. Under the Law of the Republic of Lithuania on the Bank of Lithuania, as of 24 September 214, the Bank of Lithuania must implement macroprudential policy with the aim of contributing to the stability of the financial system, including strengthening of the financial system resilience and preventing the build-up of systemic risks, thereby ensuring sustainable contribution of the financial sector to economic growth. The Law instructs the Bank of Lithuania to continuously monitor and evaluate threats arising for the domestic financial system and apply necessary macroprudential policy measures. The Macroprudential Policy Strategy, approved by the Board of the Bank of Lithuania on 12 March 215 3, sets out the interim goals of macroprudential policy, its instruments, the procedures for decision-making and making decisions public, as well as the principles for cooperation with other institutions. One of the most significant interim goals of the macroprudential policy pursued by the Bank of Lithuania is limiting excessive credit growth and leverage, and taking efforts to avoid them. The CCB is one of the instruments to achieve this goal. In Lithuania, the CCB is to be increased if credit growth becomes unsustainable, in order to build up a sufficient capital buffer for covering potential bank losses during a crisis period, thereby reducing pro-cyclicality in credit supply and strengthening the financial system s resilience. The CCB rate above per cent will be set in the event of excessive growth of the domestic credit market, which could lead to the formation of systemic risks in the financial sector. The CCB is an additional requirement which must be met by using banks Common Equity Tier 1 capital. The decision on setting the CCB rate will be taken based on a comprehensive analysis of the credit and RE market situation, as well as various complementary quantitative and qualitative information of relevance. In the case of Lithuania, two main and four complementary indicators have been selected after carrying out an analysis. The CCB reference rates are computed by using the main indicators. The credit-to-gdp gap has been chosen as the main indicator because of its credibility in predicting financial crises. Reference rates are set above per cent where the credit-to-gdp ratio is significantly higher than its long-term development trend. Complementary indicators show sustainability within the most relevant financial and economic areas for the use of the CCB, and provide additional information concerning the soundness of the CCB rates. The final decision regarding the CCB rate for Lithuania will be based primarily on the main and complementary indicators, but expert assessment is to be taken into account as well. In view of the dynamic development of the macroprudential policy, there may be changes in the list of the main and complementary indicators used to set the CCB rates and the CCB rate setting methodology in the future. 1 Basel Committee on Banking Supervision (21; rev. 211), Basel III: A global regulatory framework for more resilient banks and banking systems : 2 Directive 213/36/EU of the European Parliament and of the Council of 26 June 213 on access of the activity of credit institutions and the prudential supervision of credit institutions and investment firms, amending Directive 22/87/EU and repealing Directives 26/48/EU and 26/49/EU. 3 Decision of the Board of the Bank of Lithuania of 12 March 215 on the approval of the macroprudential policy strategy (No. 3-31).

5 5 (exceptions are available) 1. Purpose of the countercyclical capital buffer 1.1. New capital requirements The global financial crisis has shown that excessive credit growth can lead to major losses and failures in the functioning of the financial sector, as well as have a negative impact on the development of the real economy. Economic recessions are considerably deeper and recoveries slower when the preceding growth is related to strong credit growth (Jorda et al., ). At the same time it should be noted that too strong credit growth tends to be followed by economic downturns. As suggested by Schularick and Taylor (212) 5, excessive credit growth is a sufficiently good indicator for forecasting financial crises. Chart 1 Changes in banks capital structure after implementing the CRD IV Minimum requirement 8% Larger of the two rates is applied Systemic risk buffer, 5*% Global systemically important and other systemically important institutions (G-SII and O-SII) buffer, 1 3.5% and 2% respectively Countercyclical capital buffer, 2.5%* Capital conservation buffer, 2.5% Tier 2 capital, 2% Additional Tier 1 capital, 1.5% Common Equity Tier 1 capital (CET1), 4.5% Note: *Member States may set higher requirements. The cycle of a build-up, fast growth of systemic risk and its consequences could also be observed in Lithuania in 24 to 21. This cycle, first, was characterised by an extensive development of the credit market (the highest annual credit growth rate recorded was 5%) and particularly strong real estate price increases (the annual price increase was 2 7% in some sectors). Overoptimistic future expectations, aggressive development of the banking sector, competition among banks, low real interest rates, and the prevailing favourable situation in the EU and global financial markets were the main drivers of the growth rate of RE prices and bank loan portfolio being a few times higher than that of the domestic economy. The downturn in the economic and financial cycle, which started at the end of 28, coincided with the most severe global financial crisis in the recent 8 years. All this led to sharp changes in the development of the economic and financial sectors: domestic real estate prices fell by half; the GDP fell by 15 per cent in 29, while unemployment rose to 18 per cent. Most likely, the downturn would have been less disastrous if macroprudential measures, limiting systemic risks and credit market imbalances during the upswing of the financial cycle, would have been applied both in Lithuania and other countries. Before 28, the regulation of the financial sector was based mainly on the supervision of individual financial institutions and market discipline, while insufficient attention was given to the analysis of systemic imbalances and use of various measures aimed at limiting systemic risks. Significant imbalances may build up in the domestic economy, which 4 Jorda, Ò., Schularick, M., Taylor, A. M. (213). When Credit Bites Back, Journal of Money, Credit and Banking 45, p Schularick, M., Taylor A. M. (212). Credit Booms Gone Bust: Monetary Policy, Leverage Cycles, and Financial Crises, , American Economic Review 12(2), p

6 will be difficult to notice by only analysing individual financial institutions. For instance, the true extent of loans issued by the financial sector for the purchase of real estate only shows up when analysing the aggregate data, while the assessment of individual financial institutions may not reveal the exceptional crediting of the real estate sector by these financial institutions. Thus, in the assessment of the stability of individual financial institutions, their potential losses can be underestimated should the adjustment of real estate prices occur. In view of this, significant changes have been made in the regulation of the financial sector following the recent financial crisis: founding of institutions responsible for the supervision of the entire financial system started. A number of initiatives at EU and global level have been launched to create an effective system for the use of macroprudential measures. Basel III, the new capital formation standards, was one of such initiatives. These standards have been approved at EU level; many of their proposals have been transposed into the CRR and CRD IV, while the provisions of the latter two have been transposed into the national legal systems of EU Member States. 6 The CRD IV, which is to be implemented at EU level, provides for new capital buffers: systemic risk buffer, systemically important institutions buffer, and countercyclical capital buffer, that change basically the capital structure of the banking sector (see Chart 1). These new capital buffers are applied taking into account country-specific (and their financial institutions-specific) factors; therefore, countries will have a possibility to reduce systemic risk at a national level. All of the new capital measures can be divided into two groups 6 : those for the management of structural risk (e.g., systemic risk buffer, systemically important institutions buffers) and those for the management of cyclical risks (e.g., countercyclical capital buffers). The CCB is the key measure for managing cyclical risks within the financial sector. This buffer is used to mitigate the impact of the financial cycle on the entire economy. To ensure the efficiency of the CCB requirement within the internationally integrated financial systems, this measure is to be applied to banks exposures in individual countries rather than to individual institutions. This will ensure the CCB efficiency as all credit institutions on the territory of the lending country, irrespective of their legal status, will be affected Purpose of the countercyclical capital buffer Chart 2. Use of the CCB: an illustrative example Chart 3. Impact of the CCB: an illustrative example Capital adequacy requirement Total capital requirement CCB accrual CCB reduction Financial cycle with a buffer used Additional buffer Fluctuations Minimum capital requirements Base level Financial cycle without a buffer used The main purpose of using the CCB is to protect the banking system against potential losses that may be incurred due to the pro-cyclical growth of systemic risk, thereby contributing to sustainable crediting of the real economy during a financial cycle. To reduce the pro-cyclicality of the banking sector development considering the development of systemic risk and credit developments, the size of this additional capital buffer may be changed. The CCB requirement is set above per cent in case of unsustainable credit development in the country (e.g., when it is significantly higher than overall economic growth) (see Charts 2 and 3). This measure has a direct impact on the sector s resilience as the capital buffer would be accumulated during the period of unsustainable credit growth (see Chart 3) and would be reduced at the time of economic recession. Moreover, this measure is important for reducing the cyclicality of the financial system. The 6 More information about the new measures under the CRD IV is available in the 213 and 214 Financial Stability Review of the Bank of Lithuania.

7 CCB rate may vary between and 2.5 per cent or be above 2.5 per cent if necessary 7, and is to be set by the decision of the Bank of Lithuania on a quarterly basis. The ability to both limit the causes of an emerging crisis (unsustainable credit growth) and mitigate the effects of the crisis (decreasing crediting) is one of the major advantages of setting this buffer. 7 Unlike other macroprudential policy measures, the CCB is applied to domestic exposures of banks in Lithuania rather than to individual institutions; therefore, this requirement also has a direct effect on bank branches and foreign banks granting loans to Lithuanian residents. Individual institutions will have to calculate their institution-specific countercyclical capital buffer according to the CCB rates applied in individual jurisdictions. 8 This is particularly important, as cyclical systemic risks mainly build up in individual countries where bank branches of other countries operate too. A possibility to use this measure for asset exposures ensures the effectiveness of the CCB use within the financial markets of EU Member States which are deeply integrated internationally Reciprocity of the CCB rate Harmonising legal regulation in individual countries is very important as jurisdictional differences may prompt banks to shift their operations into other jurisdictions. If the CCB rate were not recognised by other countries, banks operating in Lithuania as branches would have an advantage as they, complying with the capital requirements of another jurisdiction, will not have to comply with the CCB rate set in Lithuania. The principle of reciprocity has therefore been enforced at EU level; its purpose is to ensure the effectiveness of the CCB applied by individual Member States and equal competition conditions for domestic banks and foreign bank branches. The Directive provides for immediate recognition internationally, at EU level, of the CCB rate not above 2.5 per cent. If a Member State sets the CCB rate below 2.5 per cent, this rate will have to be applied to all exposures in that country, including the exposures of bank branches and foreign banks. The Directive also provides that the portion of the CCB rate exceeding 2.5 per cent can be recognised by Member States on a voluntary basis. In such a case, countries that find it relevant (e.g., their banks have exposures in a country that sets the CCB) will have to consider the recognition of the buffer portion that exceeds 2.5 per cent and publicly inform about their decision. 9 It should be noted that the reciprocity principle will only be used in full after the transition period set in the Directive (as of 31 December 218). 1 If Member States provide for a shorter transition period and decide to start using the CCB before the timeline set in the Directive, automatic recognition of the reciprocity will not be mandatory with the CCB rate below 2.5 per cent. At the same time, the Directive provides that national competent authorities should consider recognition of the CCB rates set by Member States which had decided on the early introduction of the CCB. The Law on Banks of Lithuania also regulates the recognition of the CCB rates of third countries. If a third country s authority responsible for setting the CCB has set and published the CCB rate to be applied in the third country, the Bank of Lithuania may set a different buffer rate to be applied to exposures in that country, if it has reasonable grounds to believe that the CCB rate set by a respective national authority of the third country is not sufficient to protect its institutions against the systemic risk posed by excessive credit growth in that country. 2. Impact of the countercyclical capital buffer on the economy 2.1. Impact transmission channels Additional CCB requirement has a multi-directional impact on the financial sector and the credit market (CGFS, 212; see Chart 4). After raising the capital adequacy ratio requirement banks may either: a) increase their capital, or b) reduce their risk-weighted assets. There are two ways to increase capital: by attracting additional capital and/or accumulating additional capital from retained profits (thereby reducing profitability and dividend pay-outs). Additional capital of a bank would limit systemic risk as financial institutions would be more resilient to adverse economic situations and have larger buffers for absorbing potential credit losses. Should commercial banks decide to keep their capital unchanged, but reduce their risk-weighted assets, this would likely lead to the reduction of credit volumes and/or tightening credit standards. 7 If the Bank of Lithuania assessed that the CCB rate of 2.5 per cent is insufficient considering emerging risks. 8 Institutions have to build up an institution-specific countercyclical capital buffer that is equal to the total sum of their risk-weighted exposures multiplied by the countercyclical capital buffer rates set in individual countries. 9 It should be noted that the need to have the CCB rate above 2.5 per cent should not be seen often. For comparison, at the time of the peak of a credit cycle in Lithuania, when annual credit growth rate was 5 per cent, the CCB rate, calculated based on the credit-to-gdp gap, would be about 5.1 per cent. It should be noted that these retrospective calculations are based on the assumption that crediting would have remained unchanged even with a higher CCB rate. 1 This is provided for in Article 16 of the CRD IV.

8 Higher capital requirements It should be noted that an additional CCB may affect market participants expectations. For instance, real estate prices may be expected to decelerate following an increase of the CCB rate. If economic agents rely on unreasonable expectations and anticipate strong growth in real estate prices or the economy in the future, they may have incentives to lend to non-profitable businesses or loss-bearing projects. With the change of an economic cycle, the possibilities for the repayment of such loans would decrease significantly. For instance, if expectations regarding further fast RE price increases are not based on fundamental factors, changed expectations may lead to a fall in RE prices. Such a situation would reduce the profitability of RE projects and the ability of RE purchasers to repay their loans. Such excessive nonproductive lending creates the conditions for the build-up of various systemic risks (e.g., unsustainable increase in RE prices). One of the objectives of macroprudential policy is to limit non-productive credit, thereby ensuring crediting of productive activities. 8 Chart 4. Channels for the CCB impact transmission Potential impact Loan market Voluntary capital buffers Cost of borrowing Spillover to non-banking sector Dividends New capital Assets, primarily with higher risk weight Loan repricing Credit demand Credit supply Asset prices Impact on credit cycle Expectation channel Loss covering possibilities Tighter risk management Higher stability of the financial system Potential bank response Potential market response Decreases Increases Source: CGFS (212) Impact on the economy As the CCB is a new instrument, the empirical research of its impact on credit and economic development is rather limited. Hence, indirect research can be relied upon to analyse the impact of a quantitative increase of capital adequacy ratios on bank lending. The results of such studies vary depending on the methodology and assumptions used (whether a temporary or permanent increase of capital requirements is analysed), but many of them have revealed that tightening of capital requirements have only a modest impact on an increase in the cost of credit and/or a decline in credit volume. Based on available research (MAG, 21) 11, increasing of capital requirements by 1 p.p. in the long-term leads to an increase in bank lending interest rates of approximately 12.2 basis points and a decrease in the amount of loans of 1.47 per cent. The MAG (21) suggests that raising the bank capital adequacy ratio by 1 p.p. will change the cost and volume of credit, which, depending on the assumptions, will reduce the GDP level by.1 per cent in the long term. 12 Despite the observed negative impact of increased capital requirements on GDP growth in the short term, the positive effect on long-term economic development as a result of the reduced impact of crises would be markedly higher 11 For more details, see Macroeconomic Assessment Group The long-term is perceived as the impact after 48 quarters (median of estimates).

9 (Arregui et al., 213). 13 It should be noted that potential losses from financial crises can be particularly high. Various studies suggest a correlation between a higher capital adequacy requirement and a lower probability of a bank s bankruptcy (and systemic risk) (see Miles et al., 212; Sveriges Riskbank, 211; Kragh-Sørensen, 212). The depth of an economic crisis also directly depends on the severity of the banking crisis and failures in credit operations. The literature referred to (e.g., ESRB, 214, and Aikman, D. et al., 213) often presents the economic losses of a banking crisis as the cumulative GDP loss compared with long-term trends of economic development observed before the crisis. For instance, the BCBS study (21), which presented an overview of various studies of the impact of crises, indicates that accumulated (and discounted) losses of banking crises account for an average 63 per cent of the highest pre-crisis GDP level. As a matter of fact, the estimate of crisis losses as presented by various authors varies a great deal because of the different assumptions about the duration of the impact of a crisis and the size of the long-term impact of the crisis. According to some authors (e.g., Haldane, 21), the losses of a crisis may account for even 35 per cent of the precrisis GDP. Considering the scale of these potential losses and the relatively modest negative impact from the use of the CCB, in the conduct of macroprudential policy, using the CCB quite actively is worthwhile even when there is uncertainty about unbalanced current credit growth (IMF, 213). 9 According to Jimanez et al. (212), since, in times of an economic upturn, there are quite many substitutes for credit (e.g., firms have a possibility to issue bonds), in times of economic growth the impact of macroprudential instruments (CCB) may be lower than expected because lower supply of bank credit could be replaced with credit supply from other institutions or other forms of financing could be used (e.g., debt securities). During a recession, unlike during a credit boom, access to credit substitutes decreases along with credit availability. In that case, the easing of macroprudential policy through the reduction of additional capital buffers may have a positive effect on crediting Undesirable consequences of the application While the primary objective of the CCB is to promote sustainable development of the credit market and strengthen the stability of the financial system, the use of this instrument may have certain undesirable consequences either. After the CCB rate is increased, banks may attempt to reduce their risk-weighted assets or change the structure of their lending in order to meet tighter capital requirements. To reduce the need for capital requirement, banks may tend to lend to sectors to which lower risk weights are applied (e.g., grant loans for house purchase) and suspend lending to sectors to which higher risk weights are applied (e.g., to lend to businesses 14 ). Moreover, banks may attempt to refocus their operations from regular lending to areas that are less regulated in macroprudential terms, such as loan securitisation or investment in non-bank financial institutions engaged in the lending activity. In case these effects were substantial, this could cause an undesirable impact on the real economy because of lending redistribution among sectors, the shift of risky operations or the development of shadow banking. However, Lithuania s financial system is based on a traditional banking model and the likelihood of these effects to be significant is low. The Bank of Lithuania is prepared to monitor and assess the response of financial institutions to the additional capital requirements that are set and to take extra measures to reduce undesirable consequences if necessary. It should be noted that the slowdown of lending due to the use of the CCB cannot be viewed only as a negative (undesirable) effect. This impact should be compared to benefits arising from the lower likelihood of a financial crisis and lower amplitude of the financial cycle. Furthermore, it is complicated to precisely project the start of the adjustment of accumulated imbalances, and therefore uncertainty is to be faced when reducing the CCB. Reducing the CCB may also have certain negative consequences. Should the CCB be reduced after a financial crisis has gained momentum, it might only modestly contribute to credit encouragement due to persistent negative expectations. On the other hand, there is the risk of reducing the CCB too early and thus prompting the growth of imbalances, for example, decelerating price growth may accelerate again. 3. Setting of the countercyclical capital buffer rate The Bank of Lithuania sets the CCB rate based on a comprehensive analysis of the situation in the financial and real estate sectors, for which various quantitative and qualitative data are used. By setting the CCB rate in Lithuania on a 13 These studies provide only an approximate estimate of the impact of the CCB as actual data on the application of the CCB are not available; they focus on changes in capital requirements for institutions, whereas the CCB rate is applied to the whole banking system s exposures in a certain jurisdiction. In addition, some of the studies (Aiyar et al., 212; Francis et al., 212) look into a permanent increase of capital requirements, in contrast to the cyclical one in case of the CCB. 14 For instance, in Switzerland, the CCB is linked to sector-specific capital reserves and is only applied to housing loans, thereby attempting to avoid undesirable changes in the composition of bank loan portfolio.

10 quarterly basis, the Bank of Lithuania follows the ESRB guidelines and recommendations on the application of the CCB. 15 The CCB rate cannot be below and is calibrated in steps of.25 p.p. or multiples of these steps. The CCB increase and reduction phases differ. The CCB would be increased in order to accumulate capital during the times of unsustainable credit growth. The CCB increase would be based on the principle of guided discretion, i.e., on the pre-set indicators and the CCB reference rates, as well as the analysis of all available and relevant information, while the decision will be taken by the Board of the Bank of Lithuania. The CCB rate would be reduced or fully released when the decreasing overall risk in the financial sector or heightened financial market tensions signal an upcoming downswing in the financial cycle. Since this phase is related to uncertainty, all information available and expert assessment is to be used Setting of the CCB rate in periods of increasing systemic risk The purpose of increasing the countercyclical capital buffer is to accumulate an additional capital buffer in times of unsustainable credit growth in the country, which may lead to systemic losses for the banking sector in the future. It is important to determine accumulating imbalances in advance and gradually increase the CCB in order to accumulate a sufficient capital buffer to be used when a crisis hits. In the case of Lithuania, the CCB is set based on indicators that were the most useful in providing early signals of the build-up of imbalances during the recent crisis. These indicators are discussed in more detail below. The selection of indicators used for setting the CCB has been analysed in more detail in an occasional paper of the Bank of Lithuania (Valinskytė, 215), and the data used to calculate them are explained in the annex to this paper Setting of the CCB reference rate The Bank of Lithuania calculates the CCB reference rate on a quarterly basis. The calculation of this rate is based on the deviation of the credit-to-gdp ratio from its long-term trend, taking into account, inter alia, credit growth in the country and existing ESRB recommendations. In Lithuania, two CCB reference rates are computed, using the selected main indicators. The size of the CCB reference rates is estimated based on the size of the credit-to-gdp gap from its longterm trend. These two CCB reference rates differ in that the long-term trend of the credit-to-gdp ratio and the gap from it are computed in two ways in accordance with the standardised method proposed by the BCBS and a method adapted to the Lithuanian data (by using a forecast of the credit-to-gdp ratio). Chart 1. Long-term trend of the credit-to-gdp ratio and the gap, computed in accordance to the standardised Basel method Chart 2. Long-term trend of the credit-to-gdp ratio and the gap, computed using a forecast (Q1 2 Q4 214) Percentage 1 Percentage points 6 (Q1 2 Q4 214) Percentage 1 Percentage points Crisis period Credit-to-GDP gap (right-hand scale) Credit-to-GDP ratio Average ratio (Q Q4 214) Long-term trend of credit-to-gdp ratio (based on the standardised Basel method) Sources: Statistics Lithuania and Bank of Lithuania calculations. Note: The long-term trend is computed by applying a one-sided HP filter with a smoothing parameter set to 4, Crisis period Credit-to-GDP gap (right-hand scale) Credit-to-GDP ratio Average ratio (Q Q4 214) Long-term trend of credit-to-gdp ratio (computed using a forecast) Sources: Statistics Lithuania and Bank of Lithuania calculations. Note: The long-term trend is computed by applying a one-sided HP filter with a smoothing parameter set to 4,; before applying the filter, the credit-to- GDP ratio is modelled for the next five-year period using a four-quarter weighted average. 15 It is provided for in Article 136(3) of the CRD IV.

11 When calculating the value of the CCB reference rate, the credit-to-gdp gap (the difference between the actual indicator and its estimated long-term trend) would be calculated first, and the derived value of the gap would be related to the CCB rate. This rate may vary due to different credit definitions or different methods used to calculate the long-term trend (e.g., for accuracy purposes, the trend used is augmented with a simple forecast or the long-term trend is calculated by using different methods). 11 Reference rates are set based on the broad credit definition, when all sectors (not only MFIs) are considered creditors, including non-residents. This is done in order to take into account the incentives of credit market participants to avoid the impact of the instrument by diverting lending via other (non-banking) institutions. The CCB reference rates are set above if the total credit-to-gdp gap is higher than 2 p.p., and the CCB rate is set at 2.5 per cent when the gap exceeds 1 p.p. When the gap interval is 2 1 p.p. 16, the CCB reference rate will vary linearly. Countries may set thresholds that suit their credit cycle the best. The evaluation of Lithuania s financial cycle 17 suggests that the thresholds proposed by the BCBS and the CRD IV (when the lower threshold is 2 and the higher one is 1 p.p.) would have been adequate during the crisis in Lithuania in 28, i.e., additional capital could have been accumulated a few years before the crisis started. The credit-to-gdp gap as recommended by the ESRB and BCBS has been chosen as the main indicator for calculation of the CCB reference rate due to several reasons. Firstly, the results of many studies (Drehmann et al., 21; Drehmann and Juselius, 213; Behn et al., 213) have shown that it is one of the most reliable indicators for early prediction of an upcoming financial crisis caused by excessive credit growth. Secondly, this indicator is directly related with the CCB goals, as it may be used to determine whether the growth of credit to the real economy is consistent with the development of the real economy (when calculating the credit-to-gdp ratio, the impact of credit demand, which is driven by general economic growth, is eliminated). The ratio s long-term trend may be increasing due to the deepening of the financial market, for instance, more efficient financial intermediation. If the indicator significantly exceeds its long-term trend, it is the first signal of credit growth being too strong and potentially reflecting the build-up of imbalances. The main indicators are defined as the gap from the long-term trend. A long-term trend is estimated by applying a one-sided HP filter with the smoothing parameter λ = 4, (according to the methodology presented in the BCBS, 21, and ESRB, 214, see Chart 5). While this method is good enough for Lithuania s historic data, higher uncertainty is faced due to the HP filter specifics when estimating the real-time gap. Such tendency is particularly pronounced during the change of a cycle phase. Although the uncertainty surrounding the long-term trend estimate does not necessarily influence the usefulness 18 of the credit-to-gdp gap as a leading indicator, estimates of the long-term trend and the gap should be as precise as possible to ensure credible macroprudential policy decisions. This issue (for more information, see Valinskytė, 215) has been discussed in literature analysing the assessment of long-term trends, while adjusted (e.g., augmented with a forecast) long-term trends are used also by other countries (such as Norway). The issue related to uncertainty may be reduced by using forecast-augmented indicators, where even a simple forecast reduce estimate errors (Gerdrup et al., 213). Comparison of various methods for the estimation of the longterm trend (without a forecast, and augmented with various simple forecasts) has showed that the most appropriate way to compute the gap for Lithuania s data would be to augment the credit-to-gdp ratio with forecasts calculated as a weighted average of observations for the last four quarters 19 (see Chart 6), these conclusions are in line with the results of studies by other authors (e.g., Gerdrup et al., 213). This is the method used to compute the second CCB reference rate that is based on the other main indicator. It differs from the method proposed by the BCBS as the HP filter is applied to the ratio augmented by its forecast for the next five-year period. The forecast has been computed automatically by extending the indicator with a four-quarter weighted average. 2 Such a procedure helps to reduce errors in the long-trend estimate at the endpoint of the data series. 16 AKR t = min {2,5%; 2,5% *(max{gap H-L t; L} L)} ; where H is the gap threshold that corresponds to the maximum CCB rate (2.5%); L is the gap threshold which has to be reached for starting to accumulate the CCB; gap t is the gap between the credit-to-gdp ratio and its long-term trend at period t. 17 For more information, see Valinskytė, According to Van Norden (211), Drehmann et al. (211), the gap calculated by applying a one-sided HP filter performs as well as (even possibly outperforms) the gap considered to be actual one and calculated by applying a two-sided HP filter. Empiric research suggests that it is true, for instance, for data in UK and Norway. 19 This conclusion is applicable to many other indicators, the gap between which and their long-term trend is computed. For more information, see an occasional paper on the methodology for the application of the CCB (Valinskyte, 215). 2 The forecast for the indicator over period T+p is calculated as X T+p =,4X T +,3X T-1 +,2X T-2 +,1X T-3, p = 1,,2, where T is the number of actual observations, which are available in real time. Applied weights (,4,,3,,2, and,1) correspond to the weights of the preceding four actual observations at the end of the forecast period when the forecast is computed as a moving four-quarter average with an undesirable fluctuation in the projected series at the beginning of the projection period avoided.

12 The use of complementary indicators to determine the validity of the CCB reference rate Due to the properties of the above-discussed credit-to-gdp ratio and its gap from the long-term trend as well as scarce international practice in the use of macroprudential instruments it would not be wise to rely on one indicator. Therefore, to increase the efficiency in setting the CCB, it is recommended to choose more indicators that would be closely monitored and would help adopt fully justified decisions on setting the CCB rate and its value. Complementary indicators would help assess the level, development and prevalence of financial imbalances, and estimate the adequate rate of CCB more precisely. 12 In the case of Lithuania, apart from the main indicators showing the sustainability of borrowing by the non-financial sector, four complementary indicators have been chosen, which inform about the situation in areas important for the application of the CCB, such as the RE market, and bank financing and liquidity: 1. MFI loans-to-gdp gap; 2. House price-to-household income gap; 3. MFI loan-to-deposit ratio; 4. Current account deficit-to-gdp ratio. Other countries that have published information about the CCB application principles, such as Norway and Switzerland, also use complementary indicators similar to those presented above. The ESRB recommends to monitor them; moreover, in the case of Lithuania (Valinskytė, 215), these indicators were among the most useful in revealing the build-up of financial imbalances in It should be noted that because of the inertia of the gap indicators, i.e., their change is slow after their development trend reverses, they are only suitable in analysing the buffer accumulation period during a gradual build-up of systemic risk. To determine the moment when the systemic risk materialises and when it would be appropriate to reduce the CCB rate, other, more dynamic indicators of the market situation will be used. The ratio of MFI loans to the private non-financial sector to GDP Chart 3. Long-term trend of the ratio of MFI loans to the private non-financial sector and GDP, and the gap (Q1 2 Q1 215) Chart 4. Dynamics of loans to the private non-financial sector (Q1 2 Q1 215) Percentage Percentage points Crisis period Loan-to-GDP gap (right-hand scale) Loan-to-GDP ratio Average ratio (Q Q1 215) Long-term trend of loan-to-gdp ratio Sources: Statistics Lithuania and Bank of Lithuania calculations. Note: The long-term trend is computed by applying a one-sided HP filter with a smoothing parameter set to 4,; before applying the filter, the loan-to-gdp ratio is modelled for the next five-year period using a four-quarter weighted average Percentage Crisis period Credit impulse* (right-hand scale) Annual change in nominal loan stock Loan-to-GDP ratio Long-term trend of loan-to-gdp ratio Percentage Sources: Statistics Lithuania and Bank of Lithuania calculations. Note: The long-term trend is computed by applying a one-sided HP filter with the smoothing parameter set to 4,; before applying the filter, the loan-to- GDP ratio is modelled for the next five-year period using a four-quarter weighted average. * The ratio of the difference between the annual change in the loan portfolio and the change in the loan portfolio in the preceeding year, and GDP The ratio of MFI loans to the private non-financial sector and GDP is a preliminary indicator for the development of the credit-to-gdp ratio. This indicator is useful as it may be computed on a monthly basis and with a small lag 21, which allows for more urgent response to the market situation. Credit granted by banks (or MFIs) is also referred to as the narrow credit. At the end of the fourth quarter of 214, this type of credit (loans) accounted for 72 per cent of the broad credit, with the historic trends of both credit series being similar. Moreover, this indicator allows for a more precise 21 The MFI balance statistics is published on a monthly basis 28 days after the end of the current month, while the Financial Accounts of Lithuania, used to calculate leading indicators, are published approximately 1 days after the end of a current quarter.

13 evaluation of the impact of bank lending on credit development. It should be noted that MFIs account for the largest share of credit in Lithuania; therefore, this indicator is reliable enough to forecast the development of the main indicators. If GDP declines faster than credit, the credit-to-gdp ratio may increase even during an economic recession (as it was the case in Lithuania in 29, see Chart 7). In this case, it is important also to take into account credit growth and the credit impulse (see Chart 8) in order to assess the overall credit market situation. 13 House price-to-household income ratio RE market has a huge effect on the credit market. It should be noted that housing loans account for a significant share of total bank lending, and their percentage has been growing in many industrial countries for a few recent decades (Jorda et al., 215). Growth in housing loans is often related to the increase in real estate prices, while the increase in property prices may prompt credit growth (Harvey, 21). In view of this, the RE market situation should also be analysed when setting the CCB. Lending to households for house purchase and developments in residential property prices are closely interrelated. Growth of house prices which relies too heavily on borrowed funds rather than fundamental factors, such as general economic growth, employment growth, etc., cannot be seen as sustainable. Too easy access to borrowing and unjustified anticipation of property price increases may lead to a build-up of self-fulfilling expectations, when house purchases are booming because of anticipated growth of house prices. Residential property is often used to secure bank loans; therefore, it is an attractive instrument for mitigating credit risk when house prices are increasing rapidly. In such cases, it becomes even easier to get a loan for house purchase, which creates the conditions for even faster build-up of imbalances. Moreover, trends in property market have a significant impact on the operational stability of MFIs, as more than half of the loan portfolio is directly related to this market. It is therefore important to assess trends within the property market and their impact on the loans market. The house price-to-household income ratio sums up information on the sustainability of real estate prices. The interaction between credit and house prices may prompt the build-up of imbalances in the credit market and real economy. With property prices going up, a possibility arises to get a larger loan because of the increasing value of the collateral. When RE prices go down, banks suffer both direct losses due to a decline in the collateral value and indirect losses due to unemployment growth, lower demand, and a slump in economic activity. To compute this complementary indicator, only residential property prices are taken into account, as data on commercial RE are less reliable. 22 The ratio gap from its long-term trend is also frequently mentioned as a useful early warning indicator of a crisis (ESRB, 214; Behn et al., 213; see Chart 9). Chart 5. Long-term trend of the house price-to-household income ratio and the gap (Q1 2 Q4 214) 21 = 1 16 Points Crisis period Ratio gap (right-hand scale) House price-to-household wages and salaries ratio Average ratio (Q Q4 214) Long-term trend Sources: Statistics Lithuania and Bank of Lithuania calculations. Note: The long-term trend is computed by applying a one-sided HP filter with a smoothing parameter set to 4,; before applying the filter, the ratio is modelled for the next five-year period using a four-quarter weighted average Due to a short data series, it is impossible to explore the usefulness of the indicator in forecasting a crisis in Lithuania based on historic data and determine its objective value that could show imbalances.

14 MFI loan-to-deposit ratio The MFI loan-to-deposit ratio shows the share of loans issued by banks in the domestic market deposit portfolio. The non-deposit asset backing is often cheaper and less stable, and related to foreign capital flows, while its unbalanced growth is quite a good indicator of excessive crediting (see Chart 1; a similar indicator is used by Behn et al., 213). The largest banks operating in Lithuania are subsidiaries and branches of foreign banks, and parent institutions in some cases can provide cheaper financing to them, compared with the cost of deposits in the domestic market, thereby encouraging unbalanced credit growth within the country. Excessive growth of the loan-to-deposit ratio may show unsustainable credit growth. 14 Chart 6. MFI loan-to-deposit ratio (Q Q1 215, seasonally adjusted) Percentage Crisis period Loan-to-deposit ratio Average ratio (from Q4 1993) Long-term average +/ 2 standard deviations Source: Bank of Lithuania calculations. Note: The development of the ratio is considered to be balanced if it does not deviate from its long-term average by more than two standard deviations. Standard deviation has been computed on the basis of Q Q1 26 data covering only the period of the ratio s moderate dynamics. Chart 7. Current account balance in Lithuania and the build-up of risk 23 (Q Q4 214, four-quarter moving sums) Percentage Ratio of current account balance to GDP Average (Q Q4 214) Sources: Statistics Lithuania and Bank of Lithuania calculations. Note: The risk level is identified based on Reinhart, S.M. and Reinhart, V.R. (28), Capital flow bonanzas: An encompassing of the past and present, NBER Working Paper, The MFI loan-to-deposit ratio shows the share of loans issued by banks in the domestic market deposit portfolio. The non-deposit asset backing is often cheaper and less stable, and related to foreign capital flows, while its unbalanced growth is quite a good indicator of excessive crediting (see Chart 1; a similar indicator is used by Behn et al., 213). The largest banks operating in Lithuania are subsidiaries and branches of foreign banks, and parent institutions in some cases can provide cheaper financing to them, compared with the cost of deposits in the domestic market, thereby encouraging unbalanced credit growth within the country. Excessive growth of the loan-to-deposit ratio may show unsustainable credit growth. Current account deficit The results of research of financial crises (e.g., Kaminsky and Reinhart, 1999; Kauko, 212) suggest that widening of a current account deficit can also be associated with the increase of systemic risk within the financial sector. However, unlike the credit-to-gdp gap, it is difficult to associate this indicator directly with specific values of the CCB. According to the ESRB (214), the monitoring of the current account deficit may help avoid false alarms arising from the credit-to-gdp gap. Moreover, the study performed by S.M. Reinhart and V.R. Reinhart in 28, which presents an analysis of 181 countries divided into groups by their income level (low, average, and high income), revealed that the current account deficit widens before financial crises in many countries. Based on this study, according to the current account deficit trends in crisis and pre-crisis (2 years before a crisis) periods, 6 reference thresholds 24 were determined, which would allow to assess the sustainability of the current account balance and materialisation of a potential crisis (see Chart 11). Estimate of the probability of a crisis To assess how various sets of indicators suggest the formation of financial imbalances, discrete choice models can be used. Combinations of indicators in discrete choice models (e.g., logit, probit models) are turned into an estimate of 23 In this case, thresholds have been established as the current account deficit across low- and medium-income countries during the crisis period and the pre-crisis period of 2 years. 24 These thresholds are reference ones, and in case the indicator comes close to an individual threshold, this would not necessarily mean a material risk increase.