Republic of Croatia: Selected Issues
Author:
International Monetary Fund. European Dept.
Search for other papers by International Monetary Fund. European Dept. in
Current site
Google Scholar Close
Publication Date:
29 Jul 2024
eISBN:
9798400285400
Language:
English
Keywords:
NFC sector; interest rate shock; balance sheet vulnerability; Orbis data; asset quality; Corporate sector; Financial statements; Sensitivity analysis; Employment; Global; Eastern Europe
Download PDF (1.6 MB)

Croatian firms have significantly improved their balance sheets since the prolonged recession after the Global Financial Crisis, helped by deleveraging and narrowing country risk premium as Croatia advanced its euro adoption agenda. Micro-level simulations confirm the resilience of the corporate sector against adverse shocks to profitability and financing costs. The well-capitalized banking sector overall is also found to have buffers to absorb negative spillovers from the corporate sector.

Corporate Sector Balance Sheet Vulnerabilities in Croatia1

Croatian firms have significantly improved their balance sheets since the prolonged recession after the Global Financial Crisis, helped by deleveraging and narrowing country risk premium as Croatia advanced its euro adoption agenda. Micro-level simulations confirm the resilience of the corporate sector against adverse shocks to profitability and financing costs. The well-capitalized banking sector overall is also found to have buffers to absorb negative spillovers from the corporate sector.

A. Introduction

1. Croatia’s corporate sector suffered from the prolonged recession after the Global Financial Crisis (GFC) and has not recovered as strongly as the corporate sector in its peers (Figure 1). Both its value-added and number of employees declined and remained below their respective 2008 levels until the strong post-pandemic rebound. This is in contrast to its peers whose corporate sectors have significantly expanded over the past decade, measured by either value-added or employment. The modest growth performance of Croatia’s corporate sector is accompanied by subdued investment activities, with the investment to value-added ratio dropping and stabilizing at the decade low of 20 percent, 7 percentage points below the average of other Central, Eastern and South-Eastern European (CESEE) member states of the European Union (EU).

Figure 1.
Figure 1.
Figure 1.

Aggregate Trends in the Corporate Sector

Citation: IMF Staff Country Reports 2024, 247; 10.5089/9798400285400.002.A001

Sources: Eurostat, Structural Business Statistics; World Economic Outlook; and IMF staff calculation.Note: EU-CESEE includes Bulgaria, Croatia, Czech Republic, Hungary, Poland, Romania, Slovak Republic, and Slovenia.

2. At the same time, non-financial corporations (NFCs) have managed to deleverage and improved their creditworthiness (Figure 2). The gross debt of NFCs plateaued at almost 100 percent of GDP during the recession but declined subsequently; it stood at 77 percent of GDP in 2022, slightly below the EU median. The past decade also witnessed improved asset quality of corporate loans—the ratio of non-performing loans (NPLs) to total corporate loans declined to single digits in all sectors—and a gradual shift in NFCs’ financing pattern. As in other European economies, NFCs in Croatia barely rely on debt securities, but rather on loans and equity. The relative importance of equity financing has been on the rise in recent years and with more active participation of non-bank financial institutions. During the same period, NFCs have reduced the share of loans provided by both banks and nonbanks in their total debt, and instead increasingly resorted to each other for credits. Intercompany loans make up more than half of the overall NFC debt in 2022, compared to around 35 percent in 2013.

Figure 2.
Figure 2.
Figure 2.

Liabilities of the Non-Financial Corporate Sector

Citation: IMF Staff Country Reports 2024, 247; 10.5089/9798400285400.002.A001

3. This paper delves into firm-level data to gauge balance sheet vulnerabilities of the NFC sector and implications for banks and financial stability. It serves as a background analysis to the systemic risk assessment presented in the staff report and follows established approaches to stress-test the corporate sector. We briefly discuss the micro data—the Orbis database of the Bureau van Dijk—and how well they represent Croatia’s NFC sector in Section B,2 followed by a presentation of trends in leverage, profitability, liquidity coverage, and default risks of NFCs. Section C stress-tests the NFCs with shocks to their profitability and borrowing costs, guided by methodologies outlined in the 2013 October Global Financial Stability Report (GFSR) and Klein (2016). We defer technical issues regarding Orbis data cleaning and the handling of missing observations to the data appendix. While the focus of this paper is exclusively on NFCs’ balance sheet vulnerability, a related analysis based on the same data (Annex V of the Staff Report) examines firm-level productivity and its obstacles.

B. Evolution of NFC Balance Sheet Strength

Orbis Data Representativeness

4. The Orbis database offers a reasonably good picture of Croatia’s NFC sector in terms of value-added but tends to underrepresent the role of micro and small firms. Table 1 presents a comparison of the overall coverage and distributions with respect to major private sector industries and firm size between the Orbis data and the Eurostat’s Structural Business Statistics (SBS) for 2021.3 Orbis data miss a large number of micro firms, but the gap becomes much smaller if we instead measure Orbis firms’ representativeness by value-added or employment. For selected industries, the Orbis sample in most cases covers at least 70 percent of the value-added and employment. Nonetheless, micro firms in Orbis are under-sampled and account for a smaller share of total value-added and an even smaller share of employment than in SBS. As a robust check for this under-sampling, we re-weigh firms to align their relative employment shares by firm-size groups to those implied by SBS when calculating the overall corporate defaults in the sensitivity analysis.

Table 1.

Croatia: Data Comparison: Orbis vs. Structural Business Statistics, 2021

article image
Sources: Orbis; Eurostat, Structural Business Statistics; World Economic Outlook; and IMF staff calculation. Note: Enterprises with 1–9 employees are classified as micro, 10–49 employees as small, 50–249 employees as medium-sized, and over 250 employees as large. ICT stands for information and communications technology.

Improved NFC Balance Sheets

5. Croatia’s NFCs underwent significant deleveraging since the GFC. Debt-to-equity ratios declined across firm sizes (Figure 3) and sectors (Appendix) since 2008, reflecting both decreasing indebtedness relative to assets and a gradual buildup of equity. Small and medium-sized enterprises (SMEs) especially cut their nominal debt aggressively and the median firm within each size group held 30–40 percent less debt in 2021 compared with 2008. Debt of large firms, on the other hand, experienced a temporary rise around mid-2010s and ended up at roughly the same level in 2021 as in 2008.

6. Coupled with declining borrowing costs, the lower leverage brought broad-based improvements in NFCs’ financial strength. The effective interest rates facing firms fell to about half of their 2008 levels in 2021. Lower borrowing costs helped to boost firms’ interest coverage ratios (ICRs),4 which rose steadily since 2008 except for the temporary disruption caused by the COVID-19 pandemic. Corporate profitability also recovered after a temporary dip around 2008–09. In line with the improved financial strength of the corporate sector, the estimated risk of default declined over the past decade, as indicated by better Altman Z-Scores (see the note below Figure 3). Small firms outperformed other groups in terms of lower estimated default risks given their more favorable track record of realized earnings before interest and taxes (EBIT) and higher working capital buffers, though they had less equity and lower sales to assets compared with large firms.

Figure 3.
Figure 3.
Figure 3.

Historical Performance by Firm Size, Group Median, 2008–21

Citation: IMF Staff Country Reports 2024, 247; 10.5089/9798400285400.002.A001

Sources: Orbis; and IMF staff calculation.Note: Firm size is defined by the number of employees, with micro firms having 1 to 9 employees, small firms having 10 to 49 employees, the medium-sized firms having 50 to 249 employees, and large firms having at least 250 employees. The Altman Z-Score is calculated using the updated coefficients estimated with the ratio of the book value of equity to the book value of total liabilities. See Altman and others (2014). The median Z-Score calculated with the original coefficients is in general larger (indicating lower default probability) but displays similar trends across sectors and over time.

7. The positive development in the corporate balance sheet strength also manifested in the shrinking share of firms with low ICR (Figure 4). An ICR below 1 signifies that a firm does not generate enough profits to service its debt and thus is technically in financial distress. Firms with ICRs below 2 have some buffers after debt service obligations but the buffers could be relatively easily exhausted by adverse shocks, so these firms are of elevated vulnerability. In Croatia, though the shares of the distressed or vulnerable firms stood at comparable levels in 2021 and 2008, they accounted for much smaller shares in overall debt and employment in 2021. The improvement was more pronounced for employment. Employment at risk, i.e., the share of employment in firms with ICR less than 2, was lowered to a single digit in 2021 from the peak of 30 percent for all groups except micro firms. Similarly, a significant improvement was observed for debt held by large firms, but debt at risk remained elevated for other firms, especially medium-sized ones.

8. In sum, the NFC sector weathered well the COVID-19 pandemic and risks seemed contained as of 2021. Though debt at risks remained elevated in SMEs, the ongoing deleveraging contributed to reduced total exposures and helped limit aggregate risks. To scrutinize the resilience of the NFC sector and assess potential spillovers to banks, we will turn to simulated results to examine how firms would react to shocks to their profitability and borrowing costs in the next section.

Figure 4.
Figure 4.

Evolution of Low-ICR Firms

Citation: IMF Staff Country Reports 2024, 247; 10.5089/9798400285400.002.A001

Sources: Orbis; and IMF staff calculation.Note: Firm size is defined by the number of employees, with micro firms having 1 to 9 employees, small firms having 10 to 49 employees, the medium-sized firms having 50 to 249 employees, and large firms having at least 250 employees. Empty bars denote the distance to the historical maximum over 2008–21; the maximum may refer to different years for different indicators.

C. Sensitivity Analysis to Gauge NFC Resilience

9. This section investigates the resilience of the NFC sector to major shocks and potential spillovers to banks. We focus on the 2021 subsample of Orbis firms (Table 2). These have recovered and resumed their pre-pandemic trends for most financial variables, as seen in the previous section. The shocks under considerations are of two types: a negative shock to firms’ gross value-added, and a positive shock to firms’ borrowing costs.

Table 2.

Croatia: Summary Statistics of Orbis Firms, 2021

article image
Sources: Orbis; and IMF staff calculation. Note: Enterprises with 1–9 employees are classified as micro, 10–49 employees as small, 50–249 employees as medium-sized, and over 250 employees as large.

10. The calibration of shocks follows the macro-financial scenarios of the 2023 EU-wide banking sector stress tests by the European Banking Authority (EBA). The EBA scenarios use 2022 as the starting point and consider a baseline scenario (in line with the December 2022 projection round by the European Central Bank and national central banks) and an adverse scenario over a three-year horizon (2023–25). The shock to sector-specific gross value-added (GVA) is taken as the cumulative difference at end-2025 between the adverse and the baseline scenarios, which ranges from -8.6 percent in accommodation and food and beverage services (I) to -15.7 percent in electricity, gas, steam, and air conditioning supply (D).5 The shock to firms’ effective interest rates is assumed to be 305 basis points, corresponding to the difference of average long-term interest rates in the adverse and the baseline scenarios.

11. We consider separately shocks to GVA, interest rates, and a combined shock. Under the assumption that labor adjusts gradually and thus firms face rigid costs of employees over the horizon examined by the sensitivity analysis, a negative shock to GVA is fully born by reduced earnings (EBITDA and also EBIT assuming that depreciation and amortization stay the same).6 As detailed information on firms’ debt maturity structure is not available in Orbis, we assume that the interest rate shock would affect uniformly all the firms and raise their cost of financing for 50 percent of their outstanding debt. This will cover the rollover needs of short-term debt— aggregate data suggest that 23 percent of NFC debt is short-term as of 2023Q3—and additional financing needs from maturing long-term debt. We also consider the possibility that higher interest rates increase firms’ interest income from the financial assets they hold, which would mitigate firms’ interest burden. The results will be presented as robustness checks.

12. Both shocks to GVA and interest rates raise corporate sector vulnerability with a combined shock increasing the share of firms with ICR<2 and debt held by them by around 15 percentage points (Table 3). Negative GVA shocks reveal that about 10 percent of firms in each size group have just enough earnings to meet debt service obligation and could be in distress if such shocks materialize. However, these firms, especially smaller ones, are relatively lightly indebted, resulting in a smaller increase in debt in distressed firms than their count. Positive interest rate shocks, in contrast, affect a small group of highly indebted firms. The share of distressed firms would rise by about 1 percentage point, but the rise of debt is much more pronounced at 8 percentage points for all firms and as high as 11 percentage points for large ones. As expected, the combined adverse shocks to GVA and interest rates will lead to significant increases in both firms and debt in distress, putting both back to where they were in the early 2010s.

Table 3.

Croatia: Corporate Sector Sensitivity Analysis

(Share of respective firm groups, percentage point difference from 2021)

article image
Sources: Orbis; and IMF staff calculation. Note: The interest coverage ratio (ICR) is calculated using EBITDA. The firm sample is restricted to firms with non-missing debt. Enterprises with 1–9 employees are classified as micro, 10–49 employees as small, 50–249 employees as medium-sized, and over 250 employees as large.

13. The deteriorating financial strengths of NFCs could lead to potential losses for banks due to their exposures to the NFC sector. It is beyond this paper’s purpose to conduct a bottom-up assessment of corporate defaults using historical default instances. We instead rely on corporate default patterns estimated by the October 2013 GFSR as a first attempt.7 The GSFR calculations use historical default rates for five euro area countries and cover, in a few cases, countries that have been heavily affected by the GFC and the subsequent European debt crisis, which may not be a good representation of the corporate default pattern in Croatia. As a result, simulated results presented below should be interpreted with caution and serve only as illustrative examples to gauge how aggregate corporate defaults could respond to shocks, and in turn how banks’ asset quality and capital could be affected.

14. The simulated results suggest that shocks to NFCs could significantly raise banks’ non-performing loans (Figure 5). Using the weighted average LGD of 41.54 percent, both the GVA shock and the interest rate shock would push the share of new corporate defaults in banks’ corporate lending up by 2.6 and 2.1 percentage points, respectively, while the combined shock could lead to an increase of 3.4 percentage points. With a more conservative LGD, the standalone shocks are expected to further increase corporate defaults by about 1 percentage point, and a combined shock could almost double the share of corporate defaults, from 5.1 percent of banks’ corporate portfolio to 10.1 percent.

15. Alternative assumptions have moderate impacts on the simulated aggregate defaults. We reassign sample weights to firms to align the labor share within each size group in the Orbis sample with that of SBS, thus giving the under-sampled micro and small firms higher weights in the aggregation (¶5 in the Appendix). This will increase the share of corporate defaults by 0.2 percentage points against the GVA shock but lower it by 0.1 percentage points against the interest rate shock, as SMEs tend to be less indebted and less vulnerable to changes to the cost of financing. Raising banks’ exposure to SMEs from 36 percent to 75 percent would lower the share of defaults by about 0.3 percentage points almost uniformly across all shock scenarios. Lastly, when interest rates rise, firms could also receive higher interest income from the financial assets they hold. We assume that half of the interest rate increase (around 1.5 percent) passes to the corporate deposit rate. We proxy firms’ financial asset by the cash and deposits reported in Orbis, or if not available, the difference between the reported current assets and inventories and trade receivables. The additional interest income from the interest rate shock would lower the share of corporate defaults by 0.2–0.3 percentage points.

Figure 5.
Figure 5.

Simulated New Corporate Defaults

Citation: IMF Staff Country Reports 2024, 247; 10.5089/9798400285400.002.A001

Sources: Orbis; and IMF staff calculation.

16. The well-capitalized banking sector appears to have enough buffers to absorb losses from corporate defaults. The deteriorated financial position of NFCs after adverse shocks influences banks’ capital adequacy ratio (CAR) via two channels. First, loan losses from new corporate defaults could lower banks’ capital, though they may be offset to some extent by existing provisions. Second, the performing part of the corporate portfolio becomes riskier, thus increasing risk-weighted assets, the denominator of the capital adequacy ratio. Both channels play a notable role in the simulation (Figure 6). The GVA shock induces a more severe impact through the loan-loss channel while the interest rate shock is more influential via the risk-weight channel. With the weighted average LGD of 41.54 percent, the GVA shock would reduce CAR by 0.8 percentage points, the interest rate shock by 1.1 percentage points, and the combined shock by 1.3 percentage points. The reduction in CAR with a more conservative LGD is expected to be more pronounced, from 1½ percentage points for separate shocks to 2.1 percentage points for the combined shock. However, given that the banking system’s tier-1 CAR stood at 23 percent at end-2023, there seems to be sufficient capital buffers overall to absorb losses in NFCs examined in the sensitivity analysis.8

Figure 6.
Figure 6.

Impact on Banks’ Tier 1 Capital

(Percentage Points)

Citation: IMF Staff Country Reports 2024, 247; 10.5089/9798400285400.002.A001

Sources: Orbis; and IMF staff calculation.

D. Conclusions

17. The NFCs in Croatia have significantly improved their balance sheets over the past one and half decades. Relying on macro-financial scenarios calibrated by the EBA, firm-level simulations in this paper suggest that the corporate sector overall is able to weather adverse shocks to GVA and interest rates, which would result in manageable increases in corporate defaults. Distress of the corporate sector will spill over to banks through credit exposures, but the well-capitalized banking sector appears to have enough buffers to absorb losses. It is worth noting, however, that the simulations presented in the paper are based on a static exercise, which abstracts from stochastic shocks and uncertainty that characterize the real-world operating environment for firms and thus their dynamic decisions. The exercise also narrowly focuses on the corporate sector and makes a simplified assumption that adverse shocks stemming from a broad adverse macro-financial scenario leave the other parts of banks’ portfolio unchanged. In reality, when pressures emerge from banks’ corporate portfolios due to negative macroeconomic shocks, they may also emerge from household portfolios or other segments of bank balance sheets. Thus, the sensitivity analysis only presents a partial picture.

18. Further work is needed to understand NFCs’ relatively subdued recovery from the GFC in Croatia. The declining borrowing costs and improving financial strength of NFCs in Croatia point to the need to examine both financing and non-financing related obstacles. Compared with its peers, NFCs in Croatia recovered more gradually after the GFC, making limited contribution to employment, and having low investment rates (Figure 1). A related issue is why NFCs rely increasingly less on bank financing and whether hurdles come from supply or demand sides. Annex V of the Staff Report presents initial investigations from a productivity angle.

Appendix I. Data and Stylized Facts by Sector

Data

1. In addition to macroeconomic indicators on GDP growth and prices from the World Economic Outlook, the paper makes use of the following aggregate data sources:

  • Structural Business Statistics (SBS) by Eurostat, which provides aggregate data on the number, value-added, employment, and investment of the non-financial corporations (NFCs) by industries (NACE Rev. 2) and firm size (defined by the number of employees).

  • Sectoral accounts also compiled by Eurostat as part of the national account statistics, which provide information on financial flows and stocks across domestic sectors (including the NFCs) and the rest of the world. This underlies the financing structure of NFCs by major instruments and counterparts presented in the introduction section.

  • The National Bank of Croatia, which provides the amount and quality of loans to NFCs by major industries.

2. Firm-level data are from the Orbis Database of the Bureau van Dijk. We first perform basic data cleaning by excluding observations with negative number of employees or employee costs, sales, fixed assets, or a few other financial variables (such as shareholder funds, interest, long-term debt, etc.). Following Kalemli-Ozcan and others (2015), we include firms regardless of their filing types (consolidated or unconsolidated financial accounts); for firms that have a mixed filing patterns over years, we choose the filing type that occurs more frequently. Lastly, we restrict our analysis to 2008–21 as the numbers of reporting firms in Orbis are relatively scanty prior to 2008, and very few firms have already reported their 2022 results so far.

3. Among firm-level variables of interest, the value-added (“Added_value”) and interest paid (“Interest_paid”) are missing for most firms. Hence, we calculate them as follows:

Value-added = EBITDA + Costs_of_employees

Interest paid = EBIT – P_L_before_tax if the difference is positive. For the very few firms where EBIT and P_L_before_tax are missing, we proxy the interest paid with the financial expense (“Financial_expenses”).

4. Appendix Figure 1 illustrates how thus calculated value-added and interest paid compare with the firm-reported ones when the latter are non-missing. The calculated value-added closely aligns with the firm-reported value-added, and we also show that it seems to be a good industry-size representation of the aggregate firm distribution revealed by the Structural Business Statistics (Table 1). However, the calculated interest paid tends to exceed the reported one, which could introduce a downward bias to the interest coverage ratio, an important indicator we use to gauge NFCs’ financial strength and vulnerability.

Appendix Figure 1.
Appendix Figure 1.

Orbis versus Own-Calculated Data

Citation: IMF Staff Country Reports 2024, 247; 10.5089/9798400285400.002.A001

Sources: Orbis; and IMF staff calculation.

5. With the calculated variables, we end up with a sample of almost 380,000 firm-year observations for which information on the number of employees, value-added, interest payment, and debt is available (Appendix Table 1). The coverage of medium-sized (with 50 to 249 employees) and large (over 250 employees) firms has been stable over time, while the coverage of the still under-represented micro (less than 10 employees) and small (10 to 49 employees) firms is improving. It is worth noting that the table reflects the common sample for all analyses; the sample size could be different depending on the specific analysis.

Appendix Table 1.

Croatia: Orbis Sample Number of Observations

article image
Sources: Orbis; and IMF staff calculation, Note: Enterprises with 1–9 employees are classified as micro, 1 0–49 employees as small, 50–249 employees as medium-sized, and over 250 employees as large.

6. As micro and small firms appear to be under-sampled, we assign new sampling weights to individual firms when we aggregate firm-level results into the overall impact. To do this, we partition firms by sector and by firm size measured by the number of employees. We then assign probability weights to each sector-size group so that (i) the weight is proportional to the ratio of group-specific total number of employees in SBS to that in Orbis; and (ii) the weighted sum of employees in Orbis is equal to the total number of employees in the SBS after rescaling. The overall new corporate defaults and their impact on bank capital using the weighted Orbis sample are discussed in the sensitivity analysis section as a robust check.

Approximating Corporate Defaults

7. We follow the methodology proposed in the October 2013 GFSR and used by Klein (2016) to quantify these losses. We need probability of default (PD), loss given default (LGD), and banks’ exposure to make the calculation, which are calibrated as follows.1

  • Probability of default is determined using the simulated ICR calculated using EBIT. The ICR for each firm is placed into buckets that roughly correspond to Moody’s rating scale, and each bucket is mapped to the upper bound of the cumulative default rates based on GFSR calculations using historical data for 1970–2012 (Appendix Table 2). 2 To illustrate, a firm with ICR=2 would be mapped to the “Ba” rating and hence its PD is assigned to 4.1 percent over a one-year horizon and 9.6 percent over a two-year horizon. We focus mainly on the two-year default rates since the macro-financial shocks used in the sensitivity analysis are not for the immediate future.

  • Loss given default is assumed to be uniform across NFCs in the simulation. We consider two cases in the simulation. The first, 41.54 percent, is the weighted average of loss given default of corporate portfolio in the EBA sample as of 2023Q3. As the EBA sample covers only banks that are subsidiaries of large international banks, we also consider a more conservative LGD of 60 percent.

  • Banks’ exposure to the NFC sector is inferred from the aggregate flow of accounts data (Figure 2). As of 2023Q3, 36 percent of NFCs debt is held by banks, and we assume that this share applies to all NFCs. Given that SMEs are less likely to access alternative financing, we present as a robustness check a scenario that increases banks’ holding of SMEs debt to 75 percent while the holding of large firms’ debt remains unchanged.

Appendix Table 2.

Croatia: Mapping to Probabilities of Default 1/

article image
Source: IMF, Global Financial Stability Report (GFSR), October 2015.

The table has beer adapted from Table 1.10 on page 60 of the October 2013 GFSR.

Based on 1970–2012.

Simulating Impact on Bank Capital Adequacy Ratio

8. The impact on banks’ capital adequacy ratio is simulated using an excel template developed by Gross and Población (2017), Module F. The reduced-form approach considers separately risk exposures following the Standard Approach and the Internal Ratings Based (IRB) Approach, the share of the latter is calibrated as 17 percent according to data from the Croatian National Bank as of December 2023. To capture the fact that performing firms also incur greater difficulties in servicing their debt after hit by adverse shocks, we adjust the average risk weight for exposures following the Standard Approach by mapping firms into implied credit ratings according to Appendix Table 2 and applying the corresponding Basel risk weights;3 the post-shock IRB portfolio is adjusted using Basel risk weights formula.

Croatia’s NFC Performance by Sector

9. Given their importance and growth potential in Croatia’s private economy, we focus our discussions on six sectors, manufacturing (C), construction (F), wholesale and retail trade (G), transport (H), accommodation and food and beverage services (I), and information and communications technology (K). All sectors are reasonably well represented by the Orbis data in terms of value-added and number of employees, except for moderate under-sampling of employment in accommodation and food (Table 1).

10. In line with the aggregate trends, all sectors deleveraged since the GFC. Nominal debt reduction was most dramatic in construction and accommodation/food, where the median firms shed half of debt as of 2021 relative to 2008. Even for transport which had the most modest debt reduction, the debt in the median firm in 2021 was 30 percent below its 2008 level. All sectors enjoyed declining borrowing costs over the past decade and in general improved their debt servicing capacity and creditworthiness. Construction, especially, underwent a downturn after the GFC but exhibited steady recovery since the mid-2010s. As expected, telecommunication outperformed other sectors on average in both profitability and strength, while the labor-intensive accommodation and food sector had thin buffers and was more susceptible to shocks.

Appendix Figure 2.
Appendix Figure 2.

Historical Performance by Sector, Group Median, 2008–21

Citation: IMF Staff Country Reports 2024, 247; 10.5089/9798400285400.002.A001

Sources: Orbis; and IMF staff calculation.Note: The Altman Z-Score is calculated using the updated coefficients estimated with the ratio of the book value of equity to the book value of total liabilities. See Altman and others (2014). The median Z-Score calculated with the original coefficients is in general larger (indicating lower default probability) but displays similar trends across sectors and over time. ICT stands for information and communications technology.

11. There were also significant reductions of distressed or vulnerable firms. Trade and telecommunication compared favorably to other sectors in terms of both debt at risk and employment at risk in 2021, though telecommunication used to have very high debt concentrated in low-ICR firms, likely reflecting initial setup costs and decreasing-return-to-scale nature of the sector. Construction, and accommodation and food also achieved fast reduction of debt or employment at risk, but both stood elevated at around 20 percent in 2021. The accommodation and food sector, especially, came under severe stress in 2020 when the pandemic hit but recovered swiftly in 2021. Transport had a high share of debt (close to 30 percent) concentrated in distressed firms in 2021; however, this was mainly due to disruptions of the pandemic and the fact that the sector had a few marginal firms whose ICRs could easily fall below 1. In fact, the average debt in distressed transport firms during 2015–19 was 10 percent, while the average debt in vulnerable firms was more than doubled at 24 percent.

Appendix Figure 3.
Appendix Figure 3.

Distressed/Vulnerable Firms by Sector

Citation: IMF Staff Country Reports 2024, 247; 10.5089/9798400285400.002.A001

Sources: Orbis; and IMF staff calculation.Note: Empty bars denote the distance to the historical maximum over 2008–21; the maximum may refer to different years for different indicators. ICT stands for information and communications technology.

References

  • Altman, Edward I., Malgorzata Iwanicz-Drozdowska, Erkki K. Laitinen, and Arto Suvas. 2014. Distressed Firm and Bankruptcy Prediction in an International Context: A Review and Empirical Analysis of Altman’s Z-Score Model. Working Paper.

    • Search Google Scholar
    • Export Citation

  • De Vette, Nander, Stephan Fahr, Pablo Serrano Ascandoni, and Peter Welz. May 2023. Corporate vulnerabilities and the risks of lower growth and higher rates. European Central Bank. Financial Stability Review, Box 1.

    • Search Google Scholar
    • Export Citation

  • European Systemic Risk Board. 2023. Macro-financial scenario for the 2023 EU-wide banking sector stress test. https://www.eba.europa.eu/risk-and-data-analysis/risk-analysis/eu-wide-stress-testing

  • Chow, Julian T. S. 2015. Stress Testing Corporate Balance Sheets in Emerging Economies. IMF Working Paper WP/15/216.

  • Gal, Peter N. 2013. Measuring Total Factor Productivity at the Firm Level using OECD-ORBIS. OECD Economics Department Working Papers No. 1049. https://dx.doi.org/10.1787/5k46dsb25ls6-en

    • Search Google Scholar
    • Export Citation

  • Gross, Marco, and Javier Población. 2017. “Assessing the Efficacy of Borrower-Based Macroprudential Policy Using an Integrated Micro-Macro Model for European Households,” Economic Modelling, 61:510–28.

    • Search Google Scholar
    • Export Citation

  • International Monetary Fund. October 2013. Making the Transition to Stability. Global Financial Stability Report, Chapter 1.

  • Kalemli-Ozcan, Sebnem, Bent Sorensen, Carolina Villegas-Sanchez, Vadym Volosovych, and Sevcan Yesiltas. 2015 (revised 2022). How to Construct Nationally Representative Firm Level Data from the Orbis Global Database: New Facts and Aggregate Implications. NBER Working Paper 21558.

    • Search Google Scholar
    • Export Citation

  • Klein, Neil. 2016. Corporate Sector Vulnerabilities in Ireland. IMF Working Paper WP/16/211.

1

Prepared by Wei Shi. The author is grateful to Xuege Zhang for compiling the Orbis dataset for Croatia, and to the Corporate Analytical Group of EUR at the IMF for their comments. The paper has also benefited from insightful comments and discussions provided by the Croatian authorities.

2

The latest Orbis data refer to 2021 as very few firms have reported their 2022 results so far.

3

The Orbis data were downloaded on January 19, 2024. Following the convention of official statistics, we define the firm size by the number of employees, with micro firms having 1 to 9 employees, small firms having 10 to 49 employees, the medium-sized firms having 50 to 249 employees, and large firms having at least 250 employees. Industries are defined according to NACE Rev. 2. The Orbis data have been retrieved in U.S. dollars and converted to euros using the average annual nominal exchange rate U.S. dollar per euro from the World Economic Outlook database.

4

The ICR is calculated using earnings before interest, taxes, depreciation, and amortization (EBITDA), more properly termed the EBITDA-interest coverage ratio in the literature.

5

The EBA 2023 stress test scenarios divide the manufacturing sector into two subsectors with low and high energy intensity, separately. We calibrate the gross value-added shock to manufacturing using the low-energy-intensity difference between the adverse and the baseline (-13.0 percent), instead of the high-energy-intensity one (-19.1 percent). Given Croatia’s service-oriented economy, the more negative shock to manufacturing only results in a modest increase in the simulated new corporate defaults (by 0.1–0.2 percentage points).

6

This could overstate the actual impact on corporate earnings because the variable costs of production are likely to be lower when corporates face negative shocks to their GVAs.

7

See Appendix for detailed description of how corporate default rates are calculated in the simulation.

8

The sensitivity analysis is performed using the aggregate banking sector data. It cannot be ruled out that individual banks could see their capital fall below prudential requirements if they start with thin capital buffers and/or are highly exposed to the NFC sector.

1

The calibrated PD, LGD, and banks’ exposure, coupled with the 2021 Orbis firm distribution, imply a before-shock expected default of 5.4 percent of the NFC debt if LGD=41.54 percent, or 7.9 percent if LGD=60 percent. Both are between the actual non-performing loan (NPL) ratio in 2021 (9.9 percent) and its current level. We normalize the level of the baseline corporate default to be the latest available actual NPL ratio (5.1 percent as of Dec-2023) in Figure 6 and add to it the difference between the simulated after-shock and before-shock corporate defaults as shares of the corporate portfolio, the latter based on Orbis data.

2

The calculations are based on historical default rates for five euro-area countries, i.e., France, Germany, Italy, Portugal, and Spain. They serve as an approximation to default probabilities of Croatian corporates in our analysis, with the caveat that Croatian corporates could behave differently.

3

See CRE – Calculation of Risk Weighted Assets for credit risk, BIS standardized approach: CRE20.43.

Other IMF Content

Other Publishers

  • Collapse
  • Expand
Cover IMF Staff Country Reports
Republic of Croatia: Selected Issues
Author:
International Monetary Fund. European Dept.
Volume/Issue:
Volume 2024: Issue 247
Publisher:
International Monetary Fund
ISBN:
9798400285400
ISSN:
1934-7685
Pages:
52
DOI:
https://doi.org/10.5089/9798400285400.002
Figure 1.

Aggregate Trends in the Corporate Sector
Figure 1.

Aggregate Trends in the Corporate Sector
Figure 2.

Liabilities of the Non-Financial Corporate Sector
Figure 2.

Liabilities of the Non-Financial Corporate Sector
Figure 3.

Historical Performance by Firm Size, Group Median, 2008–21
Figure 3.

Historical Performance by Firm Size, Group Median, 2008–21
Figure 4.

Evolution of Low-ICR Firms
Figure 5.

Simulated New Corporate Defaults
Figure 6.

Impact on Banks’ Tier 1 Capital

(Percentage Points)
Appendix Figure 1.

Orbis versus Own-Calculated Data
Appendix Figure 2.

Historical Performance by Sector, Group Median, 2008–21
Appendix Figure 3.

Distressed/Vulnerable Firms by Sector