Macroprudential Diagnostics No. 8

Published: 12/7/2019

Introductory Remarks

The macroprudential diagnostic process consists of assessing any macroeconomic and financial relations and developments that might result in the disruption of financial stability. In the process, individual signals indicating an increased level of risk are detected based on calibrations using statistical methods, regulatory standards or expert estimates. They are then synthesised in a risk map indicating the level and dynamics of vulnerability, thus facilitating the identification of systemic risk, which includes the definition of its nature (structural or cyclical), location (segment of the system in which it is developing) and source (for instance, identifying whether the risk reflects disruptions on the demand or on the supply side). With regard to such diagnostics, instruments are optimised and the intensity of measures is calibrated in order to address the risks as efficiently as possible, reduce regulatory risk, including that of inaction bias, and minimise potential negative spillovers to other sectors as well as unexpected cross-border effects. What is more, market participants are thus informed of identified vulnerabilities and risks that might materialise and jeopardise financial stability.

Glossary

Financial stability is characterised by the smooth and efficient functioning of the entire financial system with regard to the financial resource allocation process, risk assessment and management, payments execution, resilience of the financial system to sudden shocks and its contribution to sustainable long-term economic growth.

Systemic risk is defined as the risk of an event that might, through various channels, disrupt the provision of financial services or result in a surge in their prices, as well as jeopardise the smooth functioning of a larger part of the financial system, thus negatively affecting real economic activity.

Vulnerability, within the context of financial stability, refers to structural characteristics or weaknesses of the domestic economy that may either make it less resilient to possible shocks or intensify the negative consequences of such shocks. This publication analyses risks related to events or developments that, if materialised, may result in the disruption of financial stability. For instance, due to the high ratios of public and external debt to GDP and the high demand for debt (re)financing, Croatia is very vulnerable to possible changes in financial conditions and is exposed to interest rate and exchange rate change risks.

Macroprudential policy measures imply the use of economic policy instruments that, depending on the specific features of risk and the characteristics of its materialisation, may be standard macroprudential policy measures. In addition, monetary, microprudential, fiscal and other policy measures may also be used for macroprudential purposes, if necessary. Because the evolution of systemic risk and its consequences, despite certain regularities, may be difficult to predict in all of their manifestations, the successful safeguarding of financial stability requires not only cross-institutional cooperation within the field of their coordination but also the development of additional measures and approaches, when needed.

1. Identification of systemic risks

Early 2019 was marked by a considerable acceleration in economic activity due to investment and goods exports growth accompanied by a continued increase in personal consumption. The available monthly indicators for the second quarter point to further favourable economic developments (see CNB Bulletin, 252), and according to recent projections of the CNB (see Macroeconomic Developments and Outlook No. 6) real economic activity is expected to grow in the remaining part of the year. However, the intensity of this growth could lessen slowly in the light of the anticipated unfavourable effect of the slowdown in economic activity in the European Union and major Croatian foreign trade partners. Further growth in economic activity over the medium term should lead to a further decline in the previously accumulated structural imbalances, primarily the high level of general government debt (74.5% of GDP at the end of March 2019) and external debt (75.9% of GDP at the end of March 2019) and heighten resilience of the domestic economy.

Figure 1 Risk map, second quarter of 2019

Note: The arrows indicate changes in relation to the risk map for the first quarter of 2019, published in Financial Stability No. 20 (May 2019).
Source: CNB.

Favourable macroeconomic developments and an improvement in the fiscal position, coupled with a fall in the general government debt-to-GDP ratio present for several years, led to a decreased risk perception of Croatia, with the CDS spread currently standing at an all-time low. Favourable developments in CDS spreads were observed in all Central and Eastern European countries, but in the last two years Croatia's CDS spread fell twice as fast as the CDS spreads in these countries, which reflects the continuous improvement in economic fundamentals in Croatia over the past few years. In the light of the improved economic fundamentals, in February 2019 the European Commission published its view that macroeconomic imbalances in Croatia were no longer excessive and in the context of their regular credit rating evaluations in March and June the credit rating agencies S&P and Fitch upgraded Croatia’s credit rating to investment grade, while Moody’s upheld its speculative credit rating but changed the outlook on Croatia to positive from stable in April.

Favourable developments were also observed in the private non-financial sector although its structural vulnerabilities continue to hold steady at a moderate level, primarily as a result of an unfavourable debt structure. Non-financial corporations saw an improvement in the interest rate and currency structure of debt in the first quarter of 2019, which has a favourable effect on the reduction of the sector’s structural vulnerabilities. Nevertheless, their debt level is still relatively high compared to the enterprises of peer countries (new EU member states). However, some segments of that sector, such as high-technology manufacturing corporations, have reduced considerably the use of bank loans as a form of financing and rely more on equity investment and non-bank creditors. Due to their specific business processes (for example a longer research and development horizon), the cost of financing of such corporations depends much more on the country’s risk premium (Analytical annex: Financing of high-technology manufacturing firms in EU countries). As regards structural vulnerabilities of the non-financial corporations sector, unfavourable demographic developments have led to a decline in qualified workforce that could have a negative impact on future business performance of corporations.

As regards short-term vulnerabilities of non-financial corporations, favourable developments continued, reflecting good business results in 2018 and the reduced burden of debt and total indebtedness. Capital and gross operating surplus continued to grow, exerting favourable effects on solvency and liquidity risk indicators.

In the first quarter of 2019, the household sector continued to witness an improved interest and currency structure of debt as a result of a drop in the share of loans with a variable interest rate and those indexed to a foreign currency, and a small growth of debt notwithstanding, the household-debt-to disposable income ratio reached the lowest level since the outbreak of the crisis in 2008.

The potential cyclic vulnerability source in the household sector lies in a noticeable growth in general-purpose cash lending of banks to the household sector with a prevalence of non-collateralised loans. Such loans were, already in 2018, granted according to much more lenient creditworthiness assessment standards in comparison with the increasingly stringent lending standards in housing loans. Combined with the increased average initial maturity and amount, this led to an accumulation of credit risk that could materialise in the case of less favourable economic developments. To prevent any associated undesired effects on financial stability and to protect consumers, at the end of February 2019 the Croatian National Bank issued a Recommendation on actions in granting non-housing consumer loans (for more details see Chapter 3 Recent macroprudential activities). Data for April and May 2019 point to a further growth in bank loans to households at a similar intensity, with only small changes in their structure: a small slowdown in the growth of general-purpose cash loans following the publication of the Recommendation and a faster growth of housing loans. Total housing loans were 4% higher year-on-year in May 2019 (measured by transactions), which is still a much slower growth rate than the two-digit growth of non-housing cash-loans.

The banking sector remains highly capitalised and liquid and bank exposure to currency and interest rate induced credit risk, although still moderately high, continued to decline due to a further increase in the share of kuna loans and loans with a fixed interest rate in the first quarter of 2019. Very low interest rates on time deposits led to a faster growth of funds in transaction accounts. This, coupled with longer loan maturities, leads to maturity mismatches in the balance sheets of banks. Structural vulnerabilities of the financial sector also reflect high banking market concentration and a concentration of exposures to the government sector and groups of connected persons, which is falling slightly with increased household lending.

The sources of risks arising from current developments in the financial sector have diminished from the previous report from a low to a very low level (Figure 1), as reflected in favourable developments in the financial stress index for Croatia and the euro area (Figure 2). Although the financial markets suggest that the risk of recession remains relatively high, the trend of yield curve flattening between 10-year and 2-year US bonds came to a halt in May 2019; however the markets are still marked by uncertainties regarding future financial and economic developments.

Figure 2 Croatian index of financial stress and individual markets’ contributions

Source: CNB.

Overall, the analysis of system exposure to systemic risks shows that system exposure has remained unchanged from the previous analysis (see Financial Stability No. 20), and has held steady at a moderate level (Figure 1).

2. Potential triggers for risk materialisation

The analysis of structural vulnerabilities of the domestic economy suggests that potential triggers for risk materialisation in Croatia lie in external developments, most notably the rise in protectionism and a possible greater than expected slowdown in global economic activity. While a final agreement on Brexit is still pending, Italy continues to be marked by great political uncertainty.

Rising protectionism in major global economies has resulted in a further growth in geopolitical uncertainties and a deterioration in economic expectations that may ultimately lead to a slowdown in global economic activity and a decline in the volume of global trade. Such developments might also lead to a deterioration in global financing conditions.

Against the backdrop of the expected slowdown of economic growth in the euro area and the keeping of the inflation target below the level of 2%, the European Central Bank (ECB) continued to pursue an expansionary monetary policy while the USA postponed its monetary tightening. The resulting prolonged period of low interest rates might lead to excessive risk-taking and additionally increase the unfavourable effect of the possible repricing of global risk premia over a medium term.

On account of the further Brexit delay granted by the European Council in April 2019, no final agreement has been achieved, leaving a hard Brexit a possibility. Such a scenario would have a negative impact on economic developments in the countries of the European Union (for more details see ECB’s Financial Stability Review), particularly the countries with strong foreign trade and financial links with the UK and on the conditions of financing on the international markets. Although the direct effects of Brexit on Croatia are estimated as negligible, its indirect effects stemming from the decline in economic growth in other countries of the European Union and increased financing costs on international markets would have an unfavourable impact on macroeconomic and financial developments in Croatia.

Political uncertainty in Italy has risen additionally following European Parliament elections and with new significant risks arising from the inability of the Italian government and the European Commission to reach an agreement about the planned budget for 2019. In the context of the 2019 European Semester, the European Commission issued in June 2019 a Recommendation for a Council Recommendation on the 2019 National Reform and Stability Programme for Italy stating that due to the general government debt reaching 132% of GDP in 2018 and further expectations of debt growth, it recommends a reopening of the excessive deficit procedure. Italy’s CDS soared after parliamentary elections in March 2018 and has since then been highly volatile, holding steady at considerably higher levels than those in peer euro area countries and Central and Eastern European countries. Any considerable decline in the economic activity in Italy could have an unfavourable impact on Croatia since Italy is Croatia’s most important trading partner. However, the unfavourable effect of a rise in the price of borrowing on Italian parent banks is not expected to spill over to domestic subsidiaries that are primarily funded by domestic deposits.

As regards the domestic environment, the main sources of risks are not estimated as significant. Nevertheless, if the continuous and fast growth of general-purpose cash lending of banks to households persists, this could lead to an excessive accumulation of credit risk that may materialise in the case of economic activity contraction and a growth in the unemployment rate. Particularly vulnerable are households earning below-average incomes, whose loan repayments, as shown by the Household Finance and Consumption Survey in the majority of households surveyed, exceeded the level that would enable them to dispose of the remaining part of income in the amount or above the minimum amount of living expenses as determined by the Foreclosure Act, thus restricting their borrowing activity in the future (for more details see Box 1 Indirect limit on the amount of loan repayment relative to debtor’s income). As regards non-financial corporations, there are still some uncertainties in connection with Agrokor; these are economic, linked with the future business performance of Agrokor, and also legal, related to the implementation of the settlement.

Other domestic vulnerabilities arise from a potential additional accumulation of arrears in the health sector. In addition, it is not possible to exclude the potential costs of law suits associated with the conversion of loans denominated in Swiss francs and the application of collective agreements in public services. And lastly, the planned further growth in government expenditure to GDP ratio in 2019 (under the Convergence Programme of the Government of the Republic of Croatia) might weigh down on domestic imbalances and impede government ability to adjust in the event of economic activity contraction.

3. Recent macroprudential activities

3.1 Continued application of the countercyclical capital buffer rate for the Republic of Croatia for the third quarter of 2020

A quarterly analytical assessment of the development of cyclical systemic risks has shown that there is still no pressure that would necessitate any corrective interventions on the part of the Croatian National Bank. Specifically, an increase in the stock of total domestic and foreign placements to the household sector and the non-financial sector in the first quarter of 2019 was accompanied by a relatively faster nominal GDP growth, with the result that the standardised relative debt indicator (i.e. total placements to nominal GDP ratio) decreased further. Since this ratio remained below its long-time trend, the credit gap calculated on the basis of this standardised ratio remained negative. A similar trend was also observed in the specific indicator of relative indebtedness, which is the ratio of domestic credit institutions’ loans to the seasonally adjusted quarterly GDP. Neither do other important indicators, such as developments in credit growth, growth in real estate prices or current account balance point to risks of excessive credit growth. In line with the results of the analysis, the Croatian National Bank issued in June 2019 the Announcement of the continued application of the countercyclical capital buffer rate for the Republic of Croatia for the third quarter of 2020

3.2 Recommendation ESRB/2015/1 on recognising and setting countercyclical buffer rates for exposures to third countries

Pursuant to Recommendation of the ESRB of 11 December 2015 on recognising and setting countercyclical buffer rates for exposures to third countries (ESRB/2015/1), the ESRB has to be submitted in the second quarter of each year a list of material third countries and, if necessary, notified of recognising and setting countercyclical buffer rates for exposures to identified third countries. Deciding on countercyclical buffer rates for third country exposures is also laid down in the Credit Institutions Act.

In line with the predefined analytical framework and schedule as well as the established criteria, at the end of the second quarter of 2018, the CNB reassessed the material exposures of Croatian banks to third countries, according to data available by the end of 2018. The analysis showed that, as in the previous year, only Bosnia and Herzegovina can be identified as a material third country for the Croatian banking sector. The analytical assessment also shows that, despite further positive growth rates of lending to households and non-financial corporations in Bosnia and Herzegovina, there is still no cyclical pressure requiring regulatory response. In June 2019, the ESRB was notified of the identified material third country.

3.3 Recommendation on actions in granting non-housing consumer loans

At the end of February 2019, the Croatian National Bank issued a Recommendation on actions in granting non-housing consumer loans recommending credit institutions to take into account minimum costs of living in accordance with the portion of salary exempt from seizure as laid down by the Foreclosure Act when assessing consumer creditworthiness for non-housing loans with initial maturity equal to or exceeding five years (for more details see Box 1). This recommendation aims to level out the conditions for assessing consumer creditworthiness for housing and non-housing loans with longer maturities and thus avoid the possibility of arbitrage between different types of loans. In addition, within its supervisory powers, the CNB asked banks to include in their internal assessments of capital requirements potential losses arising from general-purpose cash loans and to provide in their internal by-laws clear mechanisms for the repayment of banking bonuses in the event of excessive losses arising from such placements.

To assess credit institutions’ compliance with the Recommendation and to make adjustments to macroprudential policy instruments where necessary, the Croatian National Bank will require credit institutions to provide all relevant information on loans to consumers. Therefore it has recommended that credit institutions establish unique records of all non-housing consumer loans with information on loan user, loan, type and value of collateral and, for the purpose of monitoring the terms and conditions for granting all consumer housing and non-housing consumer loans, it has recommended credit institutions to establish records on various debt and debt service ratios in relation to the incomes of individual loan users.

Box 1 Indirect limit on the amount of loan repayment relative to debtor’s income

Since the beginning of 2018, credit institutions in Croatia have been applying tighter standards for the assessment of creditworthiness when granting housing loans, as the clients cannot be granted a loan greater than the amount that may be repaid from the part of their income eligible for seizure (by the bank or other creditors). To be specific, implementing EBA Guidelines on creditworthiness assessment (EBA/GL/2015/11) and EBA Guidelines on arrears and foreclosure (EBA/GL/2015/12) and based on the Act on Consumer Housing Loans, towards the end of 2017, the Croatian National Bank issued the Decision on the additional criteria for the assessment of consumer creditworthiness and on the procedure for the collection of arrears and voluntary foreclosure (OG, 107/2017, hereinafter: Decision). This Decision specifies the requirements for credit risk management of housing consumer loans and prescribes that in the process of granting housing consumer loans, credit institutions are obligated to determine the minimum costs of living that may not be less than the amount of salary exempted from seizure, as defined by the Foreclosure Act.

It should be noted that shortly before that, in mid-2017, amendments to the Foreclosure Act (OG73/2017) were adopted; they provided that the amount of salary exempt from seizure be increased for debtors with a net salary below the average net salary in the Republic of Croatia. These debtors have three-quarters of their net salary exempt from seizure (i.e. they have to be left three-quarters of their net salary at their disposal to cover living expenses), provided that the exempt part does not exceed two-thirds of the average net salary in the Republic of Croatia. For all other debtors the amount of salary exempt from seizure equals two-thirds of the average net salary in the Republic of Croatia (HRK 3,990 in 2017).

Credit institutions have complied with the CNB Decision and the provision that the amounts required for the legally determined minimum costs of living cannot be used for loan repayments. This introduced an indirect limit on the amount of loan repayments for housing loans relative to consumer income since creditworthiness assessment for granting housing consumer loans was aligned with the amount that can be seized in case of default. As a result, from 1 January 2018, the conditions for granting consumer housing loans of credit institutions for debtors in the Republic of Croatia with below average net salary have tightened considerably, as the highest permitted DSTI ratio for these debtors has been indirectly limited to a maximum of one-fourth of their net salary. For debtors with above average net salary, the part of the salary exempt from seizure is fixed at two-thirds of the average salary in the Republic of Croatia (calculated and published annually by the Central Bureau of Statistics) so the maximum permitted DSTI for such debtors rises as their net salary rises.

The introduction of new standards of housing lending in early 2018 coincided with a fast acceleration in the growth of consumer cash loans the granting of which was subject to more lenient criteria for creditworthiness assessment. The analysis of the conditions for granting cash general-purpose loans to households made by the CNB on a sample of banks towards the end of 2018 suggests that tightened conditions for granting housing loans resulted in a channelling of credit activities to cash loans and that credit institutions tended to grant more expensive and less favourable general-purpose cash loans to consumers who were not creditworthy for a housing loan. This led to an increase in maturity and average amounts of granted general-purpose cash loans and prompted the CNB to respond to the rising credit risk associated with the fast growth in cash loans by issuing the Recommendation on actions in granting non-housing consumer loans. Credit institutions were thus recommended, when determining a consumer’s creditworthiness for all non-housing loans
to consumers with original maturity equal to or longer than 5 years, to take into account the minimum costs of living in accordance with the part of salary exempt from seizure as prescribed by the Foreclosure Act. The Recommendation aims to level out the conditions for granting longer maturity housing and non-housing loans to consumers and to avoid the possibility of arbitrage (unfavourable for consumers in terms of the price) between different types of loans. Shorter maturity loans (up to 5 years) were not included in the Recommendation since they are considered to be real consumer credits which are generally less risky as they are predominantly loans of smaller amounts than longer-term non-housing loans and it is less likely that the debtor's financial situation will worsen over a shorter period of time.

The exemption of salary from seizure in the amount of minimum costs of living and the limit on the maximum amount of loan repayments relative to debtors’ income protect consumers from excessive borrowing and enable the banks to collect their claims in case of foreclosure but at the same time it weighs down on a consumer’s ability to take on new loans. In comparison with other EU member states, the implicit limit on the DSTI ratio in the amount of 25% for debtors with a below average income puts Croatia in the group of countries with more restrictive standards. The maximum permitted DSTI ratio in countries that introduced borrower-based measures and explicitly limited the DSTI ratio, generally ranges between 40% and 50%, although some countries allow for certain departures from this ratio for some categories of debtors and loans (Table 1).

Table 1 Limits on DSTI ratios for housing loans in EU member states

In Slovakia and Cyprus, the amount of repayments is limited relative to income minus living expenses. As a result, the rate shown is much lower in relation to original debtor income.
Note: In Romania and Slovenia the limits also apply to non-housing consumer loans.
Source: A review of macroprudential policy in the EU in 2018, ESRB, April 2019.

The analysis of the findings obtained by the Household Finance and Consumption Survey, conducted in the second half of 2017 with respect to household assets, liabilities, income and consumption in 2016 (that is, before the Foreclosure Act was amended and the CNB Decision entered into force) shows that most of the households included in the sample earn below-average incomes (Figure 1).

Figure 1 Characteristics of households included in the sample with regard to income level

Note: Loans include all loans to the household sector. The difference between the number of households that have loans and households that have loan repayments is mainly due to households that have overdraft loans and credit card loans without regular periodical payments. Vulnerable households are defined as households that have loan repayments and are indebted above the implicit limit based on the ratio of monthly debt service expenses to net salary, as set out in the Foreclosure Act.
Sources: Household Finance and Consumption Survey, CNB calculation and CBS.

Furthermore, almost one half of heads of households with loan repayments in Croatia included in the survey were indebted significantly above the limit defined by the amount of income legally exempt from seizure. This particularly refers to the group of debtors with below-average income, of whom 60%, according to survey findings, are not eligible for any form of new financing by loan under the provisions of the Foreclosure Act of 2017 and the CNB Decision as they had been overburdened by the repayment of existing debt even before the aforementioned regulations entered into force (Figure 2).

The ratio of loan repayment (existing loans) to the income of the head of household (the person earning the highest income in the household) and of the entire household, calculated according to the information on the income and debt of households from the Survey is shown in Figure 2. The orange line shows the implicit limit of the ratio of monthly debt service expenses to net salary set out by the Foreclosure Act. All ratios in Figure 2 above the implicit limit represent heads of households and households that are over-indebted according to applicable regulations and do not have the level of creditworthiness necessary to be granted new loans. Such households are mostly found in income groups with monthly salaries under and around the average, while the share of over-indebted households decreases with higher net salaries.

Figure 2 Debt service to monthly net income ratio (DSTI) of debtors or households and the implicit limit of the ratio under the Foreclosure Act relative to the monthly net income of debtors

Note: Gross salaries from the Household Finance and Consumption Survey have been converted into net salaries so as to enable their comparison with the part of net salary exempt from seizure under the Foreclosure Act. The average monthly net salary in 2017 was HRK 5,985.
Sources: Household Finance and Consumption Survey (carried out in 2017 with data collected for 2016), CNB calculation and CBS.

3.4 Action taken at the recommendation of the European Systemic Risk Board

The European Systemic Risk Board (ESRB) amended and extended its Recommendation on the assessment of cross-border effects of and voluntary reciprocity for macroprudential policy measures (ESRB/2015/2), recommending the reciprocation of the macroprudential policy measures adopted by the macroprudential authorities of Belgium (ESRB/2018/5), France (ESRB/2018/8) and Sweden (ESRB/2019/1).

In April 2019, Croatia reciprocated the measure adopted by the macroprudential authority of Belgium (OG, 41/2019.), which consists of a risk weight add-on for exposures of credit institutions using the internal ratings-based approach to calculate own funds based on retail loans secured by residential immovable property located in Belgium. Credit institutions that do not exceed the recommended materiality threshold may be exempt from the application of the measure, and the exemption currently applies to all domestic credit institutions. Furthermore, at the request of the macroprudential authority of Estonia, the ESRB recommended that, in relation to the systemic risk buffer for exposures in Estonia (the reciprocity of which was prescribed by the CNB in 2017, OG, 73/2017), credit institutions apply an institution-specific materiality threshold of EUR 250 million to steer the application of the de minimis principle. Accordingly, the CNB amended the previously adopted decision on the reciprocation of the Estonian measure (OG, 66/2019). The exemption continues to apply to all domestic credit institutions.

Considering the recent increase in the number of recommendations related to the reciprocation of macroprudential policy measures adopted by the macroprudential authorities of other EU member states and the very low exposures of domestic credit institutions to countries adopting the measures, the CNB will be more conservative when recognising the recommended macroprudential measures of other member states in the future. Specifically, the CNB will reciprocate only the recommended macroprudential policy measures of countries to which domestic credit institutions have material exposures (above the materiality threshold prescribed for the application of the de minimis principle). Such an approach is in line with the practice of the majority of other EU member states and complies with Recommendation ESRB/2015/2 provided that, once a year, exposures to other countries are reviewed and that relevant measures are reciprocated where the recommended materiality threshold is exceeded. Accordingly, the CNB decided not to reciprocate the macroprudential measures adopted by Sweden and France.

3.5 Implementation of macroprudential policy in other European Union member states[1]

In the first six months of 2019, the macroprudential policy instruments most frequently used by EU member states were the measures to mitigate risks associated with the upward phase of the credit cycle. A non-zero countercyclical capital buffer rate was applied by nine EU countries, with rates ranging from 0.5% to 2%. This includes two countries (Sweden and Norway) that will raise the applicable rate to 2.5% by the end of the year, which is the highest countercyclical capital buffer rate that other EU member states must automatically reciprocate and apply (pursuant to Article 137 of CRD IV, other EU member states may also recognise countercyclical capital buffer rates above 2.5%, but are not obligated to do so). An increase in the rate to be applied in 2020 has also been announced by the Czech Republic, Iceland and Denmark.

By the end of the year, four more countries will begin applying the non-zero countercyclical capital buffer rate, as announced 12 months earlier: Bulgaria (0.5% starting from October 2019 and 1% from April 2020), France (0.25% from July 2019 and 0.50% from April 2020), Ireland (1% from July 2019) and Luxembourg (0.25% from January 2020).

In January 2019, amendments to the regulations on the lending conditions for the household sector were adopted in Romania, according to which the ratio between the borrower's monthly loan repayment expenses and income was limited to 40% for loans in the domestic currency and to 20% for loans denominated in a foreign currency where the borrower is not hedged. For first-time home buyers, the ratios may be 5 percentage points higher, and 15% of newly-granted loans may deviate from the prescribed limits. The measures are applied to all household loans and all credit institutions, non-bank lenders and electronic money issuers.

The Hungarian central bank tightened the existing regulations related to the funding adequacy ratio for housing loans in order to reduce the maturity mismatch arising from the predominantly short-term sources of funding in the domestic currency and the long remaining maturity of housing loans in the domestic currency. The aim of the regulation was to stimulate the issue of domestic-currency mortgage bonds with longer maturities and the secondary mortgage bond market, and the amendments were made to increase the required share of long-term sources of funds.

The Central Bank of Malta adopted the legal basis for borrower-based measures, to be applied as of July 2019. Measures refer to housing loans and include the maximum allowed loan-to-value ratio ranging from 75%-90% for various borrower categories, the maximum allowed debt service-to-income ratio of 40% and a maturity cap of 25 or 40 years, depending on the category of the borrower (requirements are more relaxed for first-time buyers).

Table 1 Overview of macroprudential measures in EU countries

Note: The measures listed are in line with Regulation (EU) No 575/2013 on prudential requirements for credit institutions and investment firms (CRR) and Directive 2013/36/EU on access to the activity of credit institutions and the prudential supervision of credit institutions and investment firms (CRD IV). The definitions of abbreviations are provided in the List of Abbreviations at the end of the publication. Green indicates measures that have been added since the last version of the table.
Disclaimer: of which the CNB is aware.
Sources: ESRB, CNB and notifications from central banks and websites of central banks as at 15 June 2019.
For more details see: https://www.esrb.europa.eu/national_policy/html/index.en.html.

Table 2 Implementation of macroprudential policy and overview of macroprudential measures in Croatia

Note: The definitions of abbreviations are provided in the List of Abbreviations at the end of the publication. Green indicates measures added since the last version of the table.
Source: CNB.

Analytical annex: Financing of high-technology manufacturing firms[2] in EU countries[3]

High-technology manufacturing firms have an important role in strengthening a country’s competitiveness and are therefore a significant driver of economic growth.[4] Given their good and less volatile performance, which is less dependent on the economic cycle (like the energy sector or the pharmaceutical industry), such firms contribute to economic stability.[5] Nevertheless, better business performance than that of firms with lower technology intensity does not mean that high-technology manufacturing firms have access to financing at a lower cost. The reason for this is the uncertainty in the valuation of their business, which makes them vulnerable to changes in the risk premium and places less emphasis on their individual performance. Furthermore, the significantly smaller share of fixed assets and higher expenses also distances high-technology firms from banks.

According to the data available at EU level thus far (see footnote 3), high-technology firms pay higher costs of labour and fixed assets (depreciation), while their borrowing costs do not deviate significantly from other manufacturing firms. Higher costs of labour and fixed assets may be accounted for by the fact that the labour force they employ is highly qualified and the equipment they use is highly specialised, which is reflected in the higher level of technological progress seen in such firms. On the other hand, despite better business performance, and particularly better solvency, these firms do not have lower borrowing costs than other firms in the manufacturing industry (Table 1).

In order to empirically verify the determinants of borrowing costs, an adjusted model developed by Sakai et al. (2010) was used, according to which the financing cost of a firm depends on the country’s risk premium, the firm’s microindicators and the effects of time and age group. Bearing in mind the specific nature of the database used to obtain performance indicators for firms from various EU countries, shown according to activity, the effects of time and country were used in this annex, while groups were defined as groups with higher and lower technology intensity. The model used may be shown as follows:

where CDS is the risk premium of the country where the firm is resident, Y the vector of the microindicator of the firm’s business performance in the manufacturing sub-sector i, in country j, in the year t-1, while and are fixed effects for the country and year.

Results show that (besides time and country effects), the cost of borrowing of manufacturing firms is generally determined by the country’s risk premium and their margin and solvency. The risk premium expectedly raises the borrowing cost, but the correlation is more pronounced in firms with higher technology intensity. The margin also increases the cost of borrowing, reflecting the positive correlation between yield and risk, which is more pronounced in firms with lower technology intensity. The solvency of firms predictably lowers the borrowing cost in all groups of firms. As regards other significant determinants, it is important to note that a more intensified bank relationship (measured by the ratio of loans with banks and total liabilities) positively correlates to the borrowing cost in all firms, which is more pronounced in firms that are technologically more advanced. On the other hand, relationships with banks enable firms to save on other bank products (such as payment transactions). Although this may be less significant from the economic point of view, a greater asset turnover in firms with lower technology intensity is linked to lower borrowing costs. At the same time, the analysis does not suggest that lower expenses, aggressive growth (capital expenditures and growth in sales) and liquidity have any significant influence on the cost of borrowing (Table 2).

The stronger correlation between the cost of borrowing and the country’s risk premium in firms with higher technology intensity arises from the specific nature of their business process. Developing new technologies is a long process that is frequently perceived as risky by investors, particularly when they lack insight in the process itself. In addition to the fact that the longer investment horizon makes the present value of cash flows sensitive to changes in the discount rate, due to the asymmetry of information, valuators are often forced to take a conservative approach in determining the required yield for such firms. For that reason, significant investments in the research and development of high technology with longer and sometimes uncertain periods of return relativise the effect of the good present solvency of high-technology firms.

Table 1 Higher profitability and solvency of firms with higher technology intensity are not reflected in their borrowing cost

Note: The borrowing cost refers to all creditors and is calculated as the ratio of paid interest to the stock of financial debt. The cost of fixed assets is the ratio between depreciation and the stock of fixed assets. The average cost is the ratio of the cost of labour, fixed assets and debt to the total product (measured by income). Technological progress is calculated as the border effect of time on the given production function:

lny=lnk+lnl+1/2*t2+1/2*lnk2+1/2lnl2+1/2*lnk*lnl+lnk*t+lnl*t+ɛ,

where y represents the total production, l the cost of labour, k the cost of fixed assets, and t the time trend. Solvency is the ratio of earnings before interest and interest expenses. Asset turnover is the ratio of sales to total assets. Potential collateral is the share of tangible fixed assets in total assets.
Source: BACH.

Table 2 Determinants of borrowing costs of EU manufacturing firms

Note: Model specification has been made according to Sakai, Uesugi and Watanabe (2010): Firm age and the evolution of the borrowing cost: Evidence from Japanese small firms, Journal of Banking and Finance 34 (2010) 1970-1981. Margin is the ratio of net operating profit to total assets. Capital expenditures are the ratio of the sum of amortisation and the change in fixed assets to fixed assets. Liquidity is the ratio of the sum of cash and balances with banks to total assets. Bank relationship is approximated by the share of loans with banks in total liabilities. All regression equations include fixed effects for the given year and country, while independent variables are included with a lag of one year. Robust standard errors are clustered according to sub-sectors and shown in brackets. *,**,*** indicate significance at the level of 10%, 5% and 1% respectively. Extreme indicator values have been moved to the 5th and the 95th percentile of their distribution on an annual basis. The results are robust to a change in the sample of countries.
Source: BACH.

The aforementioned specific features of firms with higher technology intensity make it difficult for them to borrow from banks. The international comparison suggests that, as the technology intensity of a firm increases, financing with banks decreases, with the share of banks in all creditors of high-technology firms around 17%. To compare, almost one half of total liabilities to creditors in firms with lower technology intensity are liabilities to banks. Since they own fewer tangible fixed assets, which reflects their orientation towards know-how and larger investments in non-physical forms of assets (such as software, patents, royalties, etc.), firms with higher technology intensity have at their disposal less potential “classic” collateral. Furthermore, considering their specific production process they have a significantly lower asset turnover and lower liquidity, which is why ultimately, when their creditworthiness is determined, they seem less desirable as bank clients due to the lack of "hard” indicators (Figure 1).

Figure 1 EU firms with higher technology intensity use bank services less

Note: The weighted average is shown for all available countries.
Source: BACH.

Although the presented results do not suggest that by increasing borrowing from banks, the cost of debt of firms with higher technology intensity would drop, there remains the question whether banks could additionally diversify their portfolios by expanding their base of clients to such firms. To answer this question, it is necessary to determine what characteristics of firms are important determinants of their relationship with banks. For that purpose, the previously described model was adjusted so that the share of financing with banks depends on: a) macroeconomic conditions, b) firm microindicators and c) effects of country and year. The model used may be shown as follows:

where X and Y are vectors of macroindicators or microindicators for country j or firms from the manufacturing sub-sector i, at moment t-1, whileandare fixed effects for the country and year.

Results show that economic growth is negatively correlated to the share of financing with banks, which is particularly pronounced in firms with lower technology intensity. In such firms, the share of financing with banks decreases even as their profitability improves (Table 3). The conservative approach in the use of banking services is also reflected in the negative sign before the growth in sales, which is particularly noticeable in firms with higher technology intensity. As regards microindicators, the “hard” indicators of client creditworthiness are the most significant determinant of financing with banks; this refers to the cost-to-income ratio (in practice often used as an approximation for efficiency), liquidity and the share of tangible fixed assets that are more suitable as collateral. However, due to the specific nature of their business model, firms with higher technology intensity have, in traditional terms, lower efficiency (due to the costlier labour and pricier fixed assets they use), lower liquidity, and, finally, less eligible collateral (due to the fact that they more frequently rent than buy fixed assets, particularly real estate, and the fact that they invest in intangible forms of fixed assets).

Table 3 Determinants of the share of financing of EU manufacturing firms with banks

Note: The share of financing with banks is the ratio between debt to banks and debt to all creditors. Economic growth is the real annual GDP growth. Bank profitability is the ROE of the entire sector in a particular country. Operating profitability is the ratio of EBITDA to sales. All regression equations include fixed effects for the given year and country, while independent variables are included with a lag of one year. Robust standard errors are clustered according to sub-sectors and shown in brackets. *,**,*** indicate significance at the level of 10%, 5% and 1% respectively. Extreme indicator values have been moved to the 5th and the 95th percentile of their distribution on an annual basis. The results are robust to a change in the sample of countries.
Source: BACH.

To conclude, the borrowing cost of firms with higher technology intensity is higher than what would be expected based on their good performance indicators due to the specific nature of their business process which makes them vulnerable to changes in the risk premium and relativises their good solvency. Although no evidence was found that financing with banks would lead to lower total borrowing costs for such firms, their inclusion in the banks’ loan portfolios could have favourable effects on portfolio diversification. Even though such firms seem riskier from the aspect of traditional banking, by discriminating against them, banks are missing the opportunity to attract potentially good clients that, as a result, turn to venture capital funds, angel investors and crowdfunding platforms.

In order to stimulate the operation of firms with higher technology intensity, which are expected to contribute the most to development, some EU countries offer support through various economic policy measures, e. g. tax measures (tax credits, research and development tax incentives, options of hyper-accelerated amortisation, tax stimuli for patents, etc.), government investment co-financing, interest subsidies or government guarantees for loans, ensuring a legal framework for securitisation, etc. In that way, they reduce the cost of financing of firms with higher technology intensity and support investment and economic activity.

  1. Source: ESRB (https://www.esrb.europa.eu/national_policy/html/index.en.html) as at 15 June 2019

  2. The level of technology intensity is defined according to the Organisation for Economic Cooperation and Development (OECD) and the EUROSTAT classification of research and development intensity of individual industries as follows:a) high technology (HT): C21 – manufacture of basic pharmaceutical products and preparations, C26 – manufacture of computer, electronic and optical products,b) medium-high technology (MHT): C20 – manufacture of chemicals and chemical products, C27 – manufacture of electrical equipment, C28 – manufacture of machinery and equipment, C29 – manufacture of motor vehicles, trailers and semi-trailers, C30 – manufacture of other transport equipmentc) medium-low technology (MLT): C19, C22-C25, C33 – manufacture of coke and refined petroleum products, manufacture of rubber and plastic products, mineral products and metal products, repair and installation of machinery and equipmentd) low technology (LT): C10-C18, C31-C32 – manufacture of food products, tobacco products, beverages, textiles and wearing apparel, leather and related products, wood and paper and paper products, printing and reproduction of recorded media, manufacture of furniture and other manufacturing

  3. The BACH database currently includes non-financial corporations from 11 European Union countries: Austria, Belgium, the Czech Republic, Croatia, Germany, Spain, France, Italy, Luxembourg, Poland, Portugal and Slovakia. In this analytical annex, the data for the period from 2006 to 2017 was used for the aforementioned countries, except Luxembourg.

  4. The contribution of high-technology and medium-high-technology firms to gross value added in Croatia is relatively modest at 26%, which is noticeably lower than around 72% in Germany, 52% in the Czech Republic and 46% in Belgium.

  5. https://articles.marketrealist.com/2014/02/investors-guide-cyclical-counter-cyclical-industries/