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, according to 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 events 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 consequentially 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

Strong economic growth is supporting the continued decrease in the total exposure of the Croatian financial system to systemic risks, which remain moderate (Figure 1). The financial position of households has benefited from the strong increase in income and the robust labour market, while the debt repayment burden has not increased, thanks to the high share of loans granted at fixed interest rates and the relatively low statutory cap on the variable interest rate on housing loans. The corporate sector has performed well despite a significant increase in interest expenses, which led to an increase in capital. Against such a backdrop, the credit risk of banks has decreased further, high profitability and liquidity continuing to boost their resilience. However, as interest rates grew and the loan servicing burden increased, the share of housing loans with an elevated debt service-to-income ratio increased in the segment of new housing loans as well as in the segment of general-purpose cash loans. In addition, real estate prices are continuing to grow strongly, although the number of transactions is still trending downwards.

Figure 1 Risks to financial stability continue to decline

Notes: The arrows indicate changes from the risk map published in Macroprudential Diagnostics No. 22 (February 2024).
Source: CNB.

In early 2024, the euro area economy began showing noticeable signs of recovery as inflationary pressures eased. Real economic activity in the euro area surpassed expectations, with an annual growth rate of 1.2% in the first quarter. At the same time, the Croatian economy grew by 3.9% in real terms thanks to strong domestic demand and thus continued to expand significantly faster than the EU and euro area average. In addition, in the first five months of 2024, inflationary pressures continued to ease in Croatia, primarily under the influence of a slowdown in the growth in the prices of food and industrial products.

Figure 2 The repayment cost will grow slightly for housing loans with interest rates linked to the NRR

a) Housing loan interest rate developments according to reference parameter

Notes: Figure shows the interest rate on the stock of housing loans with a variable interest rate according to reference parameter (Figure 2a) and the interest rate on the stock of household time deposits in euro and the 6-month NRR (right). EURIBOR6M refers to the monthly market value of the 6M EURIBOR, while EURIBOR (reference parameter) refers to the variable interest rate linked to the EURIBOR.
Source: CNB.

b) Developments in the NRR and interest rates on deposits

Notes: Figure shows the interest rate on the stock of housing loans with a variable interest rate according to reference parameter (left) and the interest rate on the stock of household time deposits in euro and the 6-month NRR (Figure 2b). EURIBOR6M refers to the monthly market value of the 6M EURIBOR, while EURIBOR (reference parameter) refers to the variable interest rate linked to the EURIBOR.
Source: CNB.

As key interest rates are still high even after their slight decrease in June, the cost of repayment could increase slightly for debtors who have loans linked to the NRR, while for households with loans linked to the EURIBOR, the cost will remain unchanged.During the period of key interest rate increase, interest rates did not increase for most housing loans with a variable interest rate; in fact, on average they decreased slightly (Figure 2). Since interest rates on existing loans linked to the EURIBOR did not grow in the period of its strong increase due to the application of the statutory cap on the interest rate level^[1], they are not expected to decline as a result of a EURIBOR decrease. Although the NRR, the most common parameter in existing household loans, was stable, as time deposits began to increase more rapidly in late 2023 following the increase in interest rates on time deposits, it increased slightly (by slightly less than 0.3 p. p. by the end of the first quarter of 2024). Since banks periodically adjust variable interest rates linked to reference parameters, in the second half of 2024, this could lead to a slight increase in the cost of repayment, and, depending on the developments in interest rates on deposits, this could continue in the upcoming year as well. However, as the NRR is not expected to rise by any great extent, the cost of loan repayment should not increase significantly either (see Box 2, Financial Stability No. 25).

The rise in general-purpose cash loans accelerated in the first half of the year, and some of these loans may be considered risky. The annual growth in cash loans accelerated to 14.4% in May, and their momentum (quarterly growth rate on an annual basis) rose to 21% (Figure 3). According to the bank lending survey, the main drivers of growth in general-purpose cash loans are increased consumer confidence and durable goods consumption, particularly reflected in the increase seen in the import and registration of cars under the positive influence of a strong labour market and increasing real wages. However, a significant portion of new cash loans may be considered risky, as indicated by the fact that around a half of such loans were granted with a debt service-to-income (DSTI) ratio above 30%, and slightly over a quarter of them with a DSTI ratio exceeding 40% (Figure 4). Therefore, in the event of unfavourable macroeconomic developments, repayment difficulties could occur rapidly with such loans, leading to a rise in non-performing loans.

Figure 3 Strong growth in cash loans continues

Source: CNB.

Figure 4 A significant portion of cash loans is potentially risky

Source: CNB (Lending conditions).

In the first half of the year, housing loans continued to grow moderately as lending standards eased. Current lending activity (momentum) ranged between 6% and 8% in the first half of the year, while the annual growth rate was slightly higher, reflecting higher loan amounts granted under last year’s housing loan subsidy programme. The growth in housing loans reflects increasing amounts of individual loans amid growing real estate prices and elevated interest rates, while the number of newly-granted loans is stagnant. The rise in the average amount of new housing loans is accompanied by income growth, so that the debt-to-income ratio (DTI) remained stable in the first quarter of 2024 (standing at 6.2, which means that the average housing loan equals 6.2 annual incomes of the debtor). However, as the rise in interest rates reduced the ability of purchasers to finance residential real estate, the average DSTI ratio is noticeably increasing (having reached 41% at the end of March 2024) (Figure 5), while loans with an elevated DSTI ratio (above 40%) constituted almost a half of new loans at the end of March.

Activity on the residential real estate market continued to slow down gradually in the first quarter of 2024, but price growth remains pronounced. In 2023 and in early 2024, the number of purchase and sale transactions on the market continued to decline, primarily on the Adriatic coast and in Zagreb. The growth in the prices of residential real estate slowed down somewhat amid reduced investment demand, which primarily refers to the demand of non-residents. However, the growth at the beginning of 2024, with an annual rate of 9.1% in the first quarter, was still among the largest in the EU.

Low activity and insufficient depth remain the main vulnerabilities on the commercial real estate market. Commercial real estate prices continued to grow in the first quarter, and as the value of pledged commercial real estate increased, the average LTV ratio in loans linked to commercial real estate shrank further in the first quarter of 2024. However, low liquidity still poses a significant risk on the market, which, in the case of major disruptions, could lead to a rapid fall in prices, and, consequently, to a rise in the LTV ratio.

Figure 5 Lending standards remained somewhat more relaxed for housing loans

Source: CNB (Lending conditions).

Figure 6 Increase in residential real estate prices remains high

Notes: For details on the structure of the nominal index, see Kunovac and Kotarac (2019): Residential property prices in Croatia.
Source: CNB.

Corporate lending remained subdued in the first half of 2024 amid tightened financing conditions and reduced need for the funding of working capital. Following the rapid growth from mid-2022, interest rates on newly-granted corporate loans stabilised at the end of 2023. The transmission of tightened financing conditions to the servicing costs of existing loans was very fast considering the relatively short average loan maturity and the prevalence of loans with variable interest rates linked to the EURIBOR. On the other hand, the increase in interest rates has been slowed down somewhat by existing loans with fixed interest rates and variable interest rates linked to the NRR. In such an environment, corporate lending remained subdued, and the increase in loans was mainly driven by construction companies and companies engaged in real estate activities and, to a lesser extent, companies in manufacturing and trade.

Favourable corporate performance seen in 2023 continued in the first five months of 2024, supporting corporate resilience to potential shocks. In 2023, the income of non-financial corporations increased by some 10.5%, or 1.5 p. p. more than overall costs. Consequently, corporate net earnings grew strongly (38%), which supported the increase of capital by almost one fifth. Moreover, good business performance continued in the first five months of 2024 (according to data on fiscalisation), particularly in IT/communication and tourism-related activities. Although good business results and capital increase have had a favourable effect on the resilience of the non-financial corporate sector, to monitor corporate vulnerability, in addition to the current materialisation of risks, it is also necessary to observe risk-taking (Box 1 Corporate vulnerability indicators).

Banks’ loan portfolio quality remains high and non-performing loans continue to decline. In the first quarter of 2024, the share of non-performing loans in total loans remained at the historically low level recorded in late 2023 (2.6%), while their nominal amount dropped somewhat, reflecting the more successful collection of corporate non-performing loans and, to a lesser extent, household loans. At the same time, the share of loans with increased credit risk since initial recognition (so-called stage 2 loans) in total loans decreased from 15.6% at the end of 2023 to 13.6% in March 2024 (Figure 8), which mainly reflects a more favourable risk assessment of loans in manufacturing and energy sectors.

Banking system resilience is supported by very high levels of liquidity and capital and favourable expectations regarding its profitability (Figure 7). Own funds stagnated as a result of ample disbursements of last year’s profit, while the total capital ratio dropped to 0.6 p. p. in the first quarter of 2024, but nevertheless remained high at 23.3%. The decline reflects an increase in risk-weighted assets, i.e., the average risk weight, which is linked to stronger lending activity in the segment of general-purpose cash loans and repo transactions with other financial institutions accompanied by a decline in the level of assets held with the central bank. The liquidity coverage ratio and the stable funding ratio also decreased slightly, but remain significantly above regulatory requirements.

Banks’ financial results continued to improve, reaching historically high levels, from which they could begin to gradually decline. The net profit of banks in the first quarter was almost two fifths higher than in the same period in the year before. The increase in profit was primarily brought about by higher net interest income, while the contribution of increase in net non-interest income and release of provisions for credit losses formed in the preceding years was smaller. Profitability probably peaked at the beginning of the year and could decrease gradually as net interest spread narrows.

Figure 7 Bank profitability is at its peak and could decrease slightly in the upcoming part of the year

Source: CNB.

Figure 8 Credit risk continues to decline, although the increased share of stage 2 loans points to a possible change of trend

Source: CNB.

2 Potential risk materialisation triggers

The deteriorating geopolitical environment could trigger the materialisation of risks to financial stability, with further escalation of Russian aggression against Ukraine possibly having a considerable impact on European economies (Figure 9). Over the past two years, the war in Ukraine significantly affected international food and energy markets; however, over time, they stabilised. Still, a new potential cycle of increases in the prices of energy and other raw materials, and, consequently, food, and the growing numbers of refugees could burden European Union economies which are already seeing increased defence costs as a result of changes in the geopolitical environment. Therefore, in addition to increased risks of escalation of the conflict, the prolonged duration of the war could increase public expenditures and hamper fiscal adjustment.

At the same time, the escalation of the military conflict in the Middle East and the worsening geopolitical situation in the Far East are significant triggers of risk for the global economy. The spreading of the conflict in the Middle East could drive oil prices upwards (around 30% of global oil is produced in the Middle East^[2]) and lead to a further decline in the traffic through the Suez Canal. Furthermore, an escalation in the relations of Taiwan and China could result in supply chain bottlenecks and thus slow down global economic growth^[3].

The political situation in the Western Balkans could also affect financial stability in Croatia. A potential rise in political instability in the countries neighbouring Croatia could slow down growth and potentially trigger a recession in Croatia by reducing the perception of safety, which would negatively affect tourism and real estate demand. A drop in tourist arrivals would lead to a decline in revenues from tourism and make the debt repayment of indebted corporations and households for which tourism is a significant source of income difficult.

Growing uncertainty regarding policies that will be implemented after elections in numerous large economies in 2024 could have a negative effect on financial markets. Specifically, in many countries where elections are to be held in 2024 there is a significant level of social polarisation as political options advocating rapid changes in the domestic and global order grow increasingly popular. Should such political options be in a position to implement the policies they have announced, uncertainty will grow on the financial markets, which could lead to increased volatility and have a negative effect on market liquidity. In such an environment, a change in investor sentiment could primarily have a negative effect on the valuation of firms whose operations are based on artificial intelligence and which grew strongly over the past few years. Considering the large share of such firms in market capitalisations on major global markets, a drop in their valuation could result in increased uncertainty and volatility on global financial markets. Ultimately, deglobalisation could intensify, i.e., international trade could become more fragmented (as a result of, for instance, reshoring), which could, in the longer term, lead to increased manufacturing costs and a slowdown in global economic activity (Figure 10).

Figure 9 Geopolitical risks remain elevated, with changes in their geographical distribution

Notes: Indicators (except the geopolitical risk index) track the relative frequency of brokerage reports (via the Refinitiv platform) and financial news (Dow Jones News) related to certain risks. In the formation of the indicator, greater weight is assigned to data from brokerage reports. Indicators reflect the level of market attention for each risk in relation to the preceding five-year period.
Source: Black Rock Institute, Caldara, Dario and Matteo Iacoviello (2022), “Measuring Geopolitical Risk,” American Economic Review, April, 112(4), pp. 1194–1225.

Figure 10 Political uncertainty in Europe after the elections for the European Parliament has led to increased risk premiums

Note: The figure shows the interest spread between 10-year government bonds.
Source: ECB.

3 Recent macroprudential activities

In the first half of 2024, the mature phase of the financial cycle continued amid the strong expansion of domestic economy, as strong lending to households continued and residential real estate prices began to slow down gradually, albeit from a relatively high level, so that cyclical risks remained elevated. Consequently, in late June, a higher countercyclical capital buffer rate began to apply, as announced 12 months beforehand, which additionally boosted the resilience of the banking system. The regular annual review of the systemic materiality of third countries for the banking system of the Republic of Croatia was performed. Bosnia and Herzegovina and Montenegro were once again identified as material countries, with no cyclical risks being identified in those countries that would require a macroprudential response of the CNB in relation to domestic banks.

3.1 The announced countercyclical capital buffer rate to remain at 1.5%

The countercyclical capital buffer rate of 1.5%, which entered into force in late June, is currently considered adequate, having in mind the elevated level of cyclical vulnerabilities. A decline in the number of transactions continued in the residential real estate market. At the same time, the increase in residential real estate prices has been slowing down for several consecutive quarters, so that in the first quarter of 2024, these prices grew at an annual rate of 9.1%. In this phase of the financial cycle, the accelerated growth in household loans contributes the most to the level of cyclical risks, with the increase being particularly pronounced in general-purpose consumer lending. Total loans to households increased by 11.3% in annual terms (based on transactions) in May 2024, with particularly rapid growth in general-purpose cash loans (14.3%), while growth in housing loans remained stable (10.5%). Lending to non-financial corporations remained subdued until the end of May, with the annual rate of growth in these loans standing at 4.1%. Although the CNB assessed the countercyclical capital buffer rate of 1.5% as adequate with regard to the level of accumulated cyclical risks in the mature phase of the financial cycle, if the deepening of cyclical vulnerabilities continues, particularly in terms of the continued strong growth in cash loans in a favourable macrofinancial environment, a further increase of this buffer rate could be required.

3.2 Annual analysis of third-country materiality for the banking system of the Republic of Croatia

Bosnia and Herzegovina and Montenegro remain material third countries for the banking sector of the Republic of Croatia. This is confirmed by the results of the regular annual analysis carried out by the CNB in the second quarter of 2023. Material third countries have been identified based on three metrics: original exposure, defaulted exposure and risk-weighted exposure, as laid down by the European Systemic Risk Board^[10]. The exposure of the domestic banking sector to these countries on a consolidated basis refers mainly to the exposure of three domestic other systemically important institutions through their subsidiaries.

In addition to identifying Bosnia and Herzegovina and Montenegro as material third countries, systemic risks of a cyclical nature were also analysed in these countries. The analysis shows low risks of excessive credit growth that would require the implementation of macroprudential measures for Croatian banks exposed to these markets. However, due to signs of the gradual strengthening of cyclical risks in both countries, it is necessary to monitor their development regularly to enable appropriate measures to be taken in a timely manner if need arises.

3.3 Implementation of macroprudential policy in other European Economic Area countries

Against the backdrop of an orderly contraction of the financial cycle in most EEA countries, in the second quarter of 2024, they mostly continued to apply the macroprudential instruments already in force with a view to maintaining the resilience of the financial system. At the same time, as earlier decisions on the increase of the countercyclical buffer rate entered into force, the share of releasable capital buffers increased in many countries, while a smaller number of countries also introduced new measures.

Italy introduced a sectoral systemic risk buffer for credit exposures and counterparty risk-weighted exposures to Italian residents for all Italian credit institutions. The target rate of 1% will be achieved by raising it in two stages: to 0.5% as of 31 December 2024 and to 1% as of 30 June 2025. The measure should strengthen the resilience of the banking system in case of the materialisation of systemic risks, irrespective of their source. To ensure that all credit institutions operating in Italy operate under the same conditions, the ESRB will be asked to issue a recommendation on reciprocation.

Denmark activated a sectoral systemic risk buffer of 7% that will, as of 30 June, apply for all Danish credit institutions with regard to exposures to non-financial corporations operating in real estate and in the development of building projects in Denmark. Credit institutions in Denmark have substantial and growing exposures to companies operating in real estate, the activities of which are very sensitive to the risk of change in interest rates and other macrofinancial conditions. By introducing this instrument, the capitalisation of credit institutions is reinforced, which will enable them to better withstand possible losses related to exposures to the real estate sector. The Danish central bank requested the ESRB to recommend the reciprocation of the measure.

Greece introduced borrower-based measures for loans secured by immovable property, applicable as of 1 January 2025^[11]. The measures consist of a cap on the debt service-to-income (DSTI) ratio of 50% for first-time buyers and 40% for other buyers and a cap on the loan-to-value (LTV) ratio at origination of 90% for first-time buyers and 80% for other buyers. A quota of allowed exemptions was adopted with the measure, so that banks may exempt 10% of the total number of new loans granted in each quarter from the application of the aforementioned caps. The measure does not apply to loans granted under national housing policy or green transition programmes.

Table 1 Overview of macroprudential measures in EEA member countries and the United Kingdom

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

1. The legally prescribed cap on variable interest rates on housing loans even decreased due to the drop in the average variable interest rate on the stock of housing loans, as loans linked to the NRR, which reflects the banks’ financing costs, constitute the largest share (around 60%) of housing loans. ↑

2. https://www.statista.com/statistics/265200/middle-eastern-oil-production-in-barrels-per-day/ ↑

3. With a share of around 50%, Taiwan is the world’s largest manufacturer of semiconductors which are an indispensable component in producing most consumer electronics (source: https://www.cnbc.com/2024/01/12/taiwans-strength-in-semiconductors-could-be-its-achilles-heel-economist-says.html). ↑

4. Distance-to-default indicators are based on regression equations which determine thresholds for company default in a particular sample. The thresholds indicate the point of default for a company, and the distance from point of default (or points, if the indicator in question contains multiple components) becomes the measure of risk for that company. The Altman model (1968) is estimated according to the ratios of working capital, retained earnings, current earnings and equity to assets. An indicator value below 2.8 is considered company default. Zmijewski (1984) used the ratio of current earnings and total liabilities to assets and the ratio of current assets and current liabilities. An indicator value below 0 is considered company default. Both indicators were originally developed for manufacturing companies in the USA. ↑

5. Altman, E. I. (1968): A Further Empirical Investigation of the Bankruptcy Cost Question, The Journal of Finance, Volume 39, Issue 4, September 1984 ↑

6. Zmijewski, M. E. (1984): Methodological Issues Related to the Estimation of Financial Distress Prediction Models, Journal of Accounting Research, Vol. 22, Studies on Current Econometric Issues in Accounting Research (1984), pp. 59–82 (24 pages) ↑

7. Stressed or in distress? How best to measure corporate vulnerability, Bank Underground blog, https://bankunderground.co.uk/2024/02/21/stressed-or-in-distress-how-best-to-measure-corporate-vulnerability/ ↑

8. In this paper, fixed costs were approximated by labour costs and depreciated fixed tangible assets. In SUERF Policy Brief, No 256 (2020), Allinger and Huljak empirically confirm that these costs are indeed fixed based on a sample of Croatian companies. ↑

9. The degree of total leverage is the product of the degree of financial leverage and the degree of operating leverage, where financial leverage shows the ratio of earnings before and after interest, and operating leverage shows the ratio of earnings before and after fixed costs. Consequently, the degree of total leverage shows the sensitivity of a company to changes at operational level, i.e., the elasticity of earnings with regard to changes in revenue resulting from operating and financial leverage.

   DTL=DFLxDOL, DFL=EBIT/EBT, DOL=CM/EBIT (1),

   where: DTL is the degree of total leverage, DFL is the degree of financial leverage, DOL is the degree of operating leverage, EBIT is earnings before interest and tax, EBT is earnings before tax and CM is the contribution margin (revenue – variable costs). ↑

10. Decision of the European Systemic Risk Board (ESRB) on the assessment of materiality of third countries for the Union’s banking system in relation to the recognition and setting of countercyclical buffer rates (ESRB/2015/3) ↑

11. Consequently, as of 1 January 2025, borrower-based measures will be actively applied in 26 out of 30 EEA countries. ↑