## Inflation Dynamics in Croatia

## Source details

**Canonical URL:** [Inflation Dynamics in Croatia](https://www.imf.org/-/media/files/publications/cr/2025/english/1hrvea2025002-source-pdf.pdf)

## Other formats

- [Markdown version](/-/media/files/publications/cr/2025/english/1hrvea2025002-source-pdf.pdf.md)
- [Structured JSON version](/-/media/files/publications/cr/2025/english/1hrvea2025002-source-pdf.pdf.json)

---

### Overview and recent inflation performance
- After a period of sharp deceleration, inflation in Croatia has inched up since late 2024 to about 4–4 ½ percent (y/y) lately, among the highest in the euro area.
- Peak and subsequent path:
  - Peak HICP inflation: 13 percent at the end of 2022.
  - Average HICP inflation in 2024: 4 percent.
  - September 2025: 4.6 percent.
  - October 2025 flash estimate: 4 percent (noted as due to lower commodity prices).
- Core and sectoral dynamics:
  - Core inflation exceeded headline inflation in Croatia up to the recent months.
  - Services inflation has been persistently high, above 7 percent (y/y), contributing to stickiness.

### Fiscal stance and labor cost dynamics
- Fiscal balances and spending levels:
  - General government primary balance: surplus of 1.4 percent of GDP in 2022; deficit of 0.9 percent of GDP in 2024.
  - Public expenditure: around 48 percent of GDP on average in 2024 (45 percent of GDP in 2022); close to 50 percent by the end of 2024.
  - After three years of expansionary fiscal policy, fiscal balance shifted from a surplus of 0.1 percent of GDP in 2022 to a deficit projected at 2.9 percent of GDP in 2025.
- Public wage developments:
  - Public wage growth definition: y/y growth in compensation of employees (seasonally adjusted).
  - Public wage growth peaked at over 30 percent y/y in 2024Q2.
  - Annual public wage growth in 2024: 26 percent.
  - The average growth for 2024 in real average monthly gross earnings: 19 percent.
  - Noted that wage pressures appear to have tapered in 2025, with public wage growth normalizing, but past large increases could have persistent inflationary effects.

### Credit dynamics and macroprudential response
- Private sector credit growth:
  - Household credit growth: persistently over 10 percent since 2024.
  - Credit to Non-Financial Corporations (NFCs): growing slower than household credit.
  - Most household credit growth driven by general purpose cash loans (used for consumption and housing related expenses), which do not require collateral and can be requested via mobile apps.
- Macroprudential measures:
  - Croatian National Bank (CNB) introduced explicit borrower-based measures (BBMs) limits, effective in July 2025, to contain systemic risks and reduce inflationary pressures associated with household consumption.
  - CNB announced an increase of the countercyclical capital buffers (CCyB) to 2 percent effective in January 2027 to strengthen banking system resilience and expand releasable buffers.

### Methodology: Bayesian VAR setup
- Sample and model:
  - Period: 2000Q1 to 2025Q2.
  - Model: Bayesian vector autoregressive (VAR), based on Nguyen et al. (2023).
  - Lag length: 4 lags.
  - Priors and estimation: Bayesian estimation with Normal-Wishard distribution and medium-to-loose priors.
  - Endogenous variables (baseline): Croatian and euro area real GDP growth (log difference of real GDP), Croatian and euro area inflation, and Croatian primary fiscal balance in percentage of GDP.
  - Pre-determined factors (non-stochastic elements) included.
- Identification and shocks:
  - Fiscal shock: expansionary fiscal policy that decreases the primary balance but increases GDP growth and inflation; identified via contemporary sign restrictions.
  - Extended model: distinguishes expenditure shocks (increase expenditure, GDP, and inflation), public wage growth shocks, and tax revenue shocks (tax cuts defined as negative tax revenue shocks that increase GDP while inflation is left unrestricted).
  - Euro area aggregate demand and supply identified via block exogeneity (they affect Croatian variables but not vice versa in contemporaneous terms).
  - Exogenous shock includes a COVID-19 dummy.
- Validation:
  - Model demonstrates a strong fit with results statistically significant at the 95 percent confidence level and robust across diagnostic checks.
  - Baseline model diagnostics (λ₁ = 0.7):
    - R2 = 0.566
    - Adj- R2 = 0.446
    - SSR = 25.77
    - RMSE = 0.5128
    - DIC = 1399.77
  - Extended specification with public wages growth:
    - R2 = 0.597
    - Adj- R2 = 0.458
    - SSR = 23.92
    - RMSE = 0.494
    - DIC = 1999.70
  - Sensitivity to prior tightness λ₁ tested: λ₁ = 0.7 (baseline), λ₁ = 0.5, λ₁ = 0.1; impulse responses and historical decompositions remain qualitatively and quantitatively similar.

### Quantitative contributions of fiscal and other shocks to growth and inflation
- 2021–22 period:
  - Fiscal policy supported growth during 2021–22 and acted counter-cyclically.
  - Fiscal policy shocks accounted for 62 percent of the total contribution of endogenous shocks to GDP growth in 2022.
  - Fiscal policy in 2022 accounted for only 18 percent of the total contribution of endogenous shocks to inflation in 2022; inflationary pressures in 2022 were mostly external (exogenous and euro area shocks dominated inflation).
- 2023–2025H1 period:
  - Fiscal policy became pro-cyclical in 2023 and contributed significantly to headline inflation in 2024 and 2025H1.
  - Average impact magnitude since 2024: 0.6 ppts.
  - Since 2024, the negative primary balance shock accounted for more than 40 percent of the impact attributed to endogenous shocks on headline HICP inflation.
  - Considering only shocks that Croatia can influence (i.e., excluding euro area or exogenous shocks), fiscal policy contributed to about 60 percent of the total headline HICP inflation.
- Overall contributions and uncertainty:
  - Fiscal shocks account for around one-quarter of the headline HICP, when excluding only these pre-determined factors.
  - The forecast error variance decomposition confirms fiscal shocks can explain about 18 percent of the total inflation uncertainty forecasted in the longer run.
  - The contribution of fiscal shocks to Croatia’s national CPI is in line with that to the HICP.

### Quantitative impulse-response and cumulative impacts
- Primary balance shock:
  - A decrease in the primary balance to GDP ratio by 1 standard deviation (around 1 ppt) is estimated to increase inflation by 0.2 ppts in the first quarter after the shock.
  - The impact is persistent over time, cumulatively accounting for around ½ ppts one year from the shock and almost 1 ppt in the longer run (3 years).
- Public wages and expenditures:
  - Cumulatively, 0.8 ppts of the headline HICP inflation in 2024 was attributable to the increase in public wages.
  - Cumulatively, 0.7 ppts in the first half of 2025 was attributable to the increase in public wages.
  - The impact of total public expenditure is found to be around ½ ppts (and reached 0.8 ppts in 2025).
- Counterfactuals:
  - Without the increase in public wages in 2024:
    - q/q headline inflation would have moved from 1.3 percent to 1.1 percent;
    - y/y headline inflation would have moved from 4.0 percent to 3.1 percent.
  - In 2025Q1–Q2 the headline HICP would have been below 4 percent (y/y) absent the public wage increases.

### Role of fiscal components (taxes, expenditures, public wages)
- Tax measures in 2022:
  - Reductions of tax rates to mitigate the “cost of living” crisis in 2022 played an important role in explaining the headline inflation in 2022 (especially combined with wage growth).
- Recent years:
  - The impact of tax revenue shocks has been much smaller in the most recent years.
  - An increase in expenditures or in public wages growth played a primary role in explaining headline HICP inflation since 2024, making up nearly half of the total impact from endogenous shocks.
  - The impulse responses to both tax revenues and expenditures shocks are highly comparable.
  - The increase in fiscal expenditure and public wage growth is found to have contributed to around 20 percent of the headline inflation since 2024.

### Timing, persistence, and core inflation
- Timing and lags:
  - Regardless of how fiscal shocks are measured, their impact on inflation took time to materialize; the pick-up in inflation starting at the end of 2024Q3 shows the full effects likely materialized with some lags.
  - In 2024Q4 and 2025Q1, about one quarter of the q/q inflation was due to fiscal shocks, including all shocks and pre-determined factors (i.e., more than half of total endogenous shocks).
- Persistence:
  - The impact of fiscal shocks on core inflation persists longer than in the case of headline inflation, lasting up to 2 years from the time of the shock.
  - The stronger impact of fiscal shocks on core inflation in 2022: 1.6 ppts for core compared to 1.1 ppts for headline inflation.
  - The structural explanatory part of core inflation is bigger than that of the headline inflation (2.5 percent versus 2.3 percent in the baseline), confirming structural persistence in core inflation components such as services.
- Peak contribution:
  - The largest contribution of fiscal shocks is estimated to be in 2024Q4 and 2025Q1, when they accounted for 1/3 of the total endogenous shocks, or over a half of domestic shocks.

### Policy-relevant conclusions and recommendations
- Key diagnostic conclusion:
  - Fiscal policy has played a notable role in Croatia’s recent inflation dynamics: fiscal loosening in 2024, particularly through public wage increases, accounted for more than 40 percent of the total contribution of endogenous shocks to headline HICP and about 60 percent of the total shocks that Croatia can influence.
  - Fiscal shocks, including wage hikes, became more visible with a lag, coinciding with a sharp price increase in late 2024.
- Policy recommendations:
  - Fiscal policy should be mindful of its inflationary impact.
  - With monetary policy set at the euro area level and macroprudential tools now actively deployed to contain the growth of households’ credit—thereby helping to reduce demand-driven inflationary pressures—fiscal policy, while pursuing its own objectives, should be tightened considerably to ensure a coherent overall policy mix to limit imbalances in the economy and safeguard macroeconomic stability.
  - Use of macroprudential measures (BBMs from July 2025) and increased CCyB to 2 percent (effective January 2027) to contain demand-driven inflationary pressures and strengthen banking resilience.

### Data, robustness checks, and model diagnostics
- Data and sample:
  - Data source: Haver based on CNB and ECB data, covering 2000Q1 to 2025Q2.
  - All variables, besides the balance, are seasonally adjusted. Real GDP and prices are taken in quarterly log difference.
  - A COVID-19 dummy equal to 1 for 2020Q1 to 2021Q2 is included.
- Robustness:
  - Sensitivity to prior tightness λ₁ tested: λ₁ = 0.7 (baseline), λ₁ = 0.5, λ₁ = 0.1.
  - Cyclically adjusted primary balance robustness check (subtracting 0.5 times the output gap) yields similar results.
  - Specification that explicitly controls for an exogenous oil shock shows exogenous shocks become more important for 2022 while impact of fiscal shocks in the most recent quarters remains very similar.
  - Results are robust to inclusion of a euro area monetary policy shock identified as in Deskar-Škrbic et al. (2020).
- Model fit and statistical significance:
  - Model results statistically significant at the 95 percent confidence level and robust across diagnostic checks.

*Source: IMF staff analysis, "Inflation in Croatia: The Role of Fiscal Policy" (chapter content provided).*

### 1. Inflation Dynamics in Croatia  ____________________________________________________ 4

### Inflation Dynamics in Croatia

### Overview and recent inflation performance
- After a period of sharp deceleration, inflation in Croatia has inched up since late 2024 to about 4–4 ½ percent (y/y) lately, among the highest in the euro area.
- Peak and subsequent path:
  - Peak HICP inflation: 13 percent at the end of 2022.
  - Average HICP inflation in 2024: 4 percent.
  - September 2025: 4.6 percent.
  - October 2025 flash estimate: 4 percent (noted as due to lower commodity prices).
- Core and sectoral dynamics:
  - Core inflation exceeded headline inflation in Croatia up to the recent months.
  - Services inflation has been persistently high, above 7 percent (y/y), contributing to stickiness.

### Fiscal stance and labor cost dynamics
- Fiscal balances and spending levels:
  - General government primary balance: surplus of 1.4 percent of GDP in 2022; deficit of 0.9 percent of GDP in 2024.
  - Public expenditure: around 48 percent of GDP on average in 2024 (45 percent of GDP in 2022); close to 50 percent by the end of 2024.
- Public wage developments:
  - Public wage growth definition: y/y growth in compensation of employees (seasonally adjusted).
  - Public wage growth peaked at over 30 percent y/y in 2024Q2.
  - Annual public wage growth in 2024: 26 percent.
  - The average growth for 2024 in real average monthly gross earnings: 19 percent.
  - Noted that wage pressures appear to have tapered in 2025, with public wage growth normalizing, but past large increases could have persistent inflationary effects.

### Credit dynamics and macroprudential response
- Private sector credit growth:
  - Household credit growth: persistently over 10 percent since 2024.
  - Credit to Non-Financial Corporations (NFCs): growing slower than household credit.
  - Most household credit growth driven by general purpose cash loans (used for consumption and housing related expenses), which do not require collateral and can be requested via mobile apps.
- Macroprudential measures:
  - Croatian National Bank (CNB) introduced explicit borrower-based measures (BBMs) limits, effective in July 2025, to contain systemic risks and reduce inflationary pressures associated with household consumption.
  - CNB announced an increase of the countercyclical capital buffers (CCyB) to 2 percent effective in January 2027 to strengthen banking system resilience and expand releasable buffers.

### Methodology: Bayesian VAR setup
- Sample and model:
  - Period: 2000Q1 to 2025Q2.
  - Model: Bayesian vector autoregressive (VAR), based on Nguyen et al. (2023).
  - Lag length: 4 lags.
  - Priors and estimation: Bayesian estimation with Normal-Wishard distribution and medium-to-loose priors.
  - Endogenous variables (baseline): Croatian and euro area real GDP growth (log difference of real GDP), Croatian and euro area inflation, and Croatian primary fiscal balance in percentage of GDP.
  - Pre-determined factors (non-stochastic elements) included.
- Identification and shocks:
  - Fiscal shock: expansionary fiscal policy that decreases the primary balance but increases GDP growth and inflation; identified via contemporary sign restrictions.
  - Extended model: distinguishes expenditure shocks (increase expenditure, GDP, and inflation), public wage growth shocks, and tax revenue shocks (tax cuts defined as negative tax revenue shocks that increase GDP while inflation is left unrestricted).
  - Euro area aggregate demand and supply identified via block exogeneity (they affect Croatian variables but not vice versa in contemporaneous terms).
  - Exogenous shock includes a COVID-19 dummy.
- Validation:
  - Model demonstrates a strong fit with results statistically significant at the 95 percent confidence level and robust across diagnostic checks (see Appendix I for details).

### Quantitative contributions of fiscal and other shocks to growth and inflation
- 2021–22 period:
  - Fiscal policy supported growth during 2021–22 and acted counter-cyclically.
  - Fiscal policy shocks accounted for 62 percent of the total contribution of endogenous shocks to GDP growth in 2022.
  - Fiscal policy in 2022 accounted for only 18 percent of the total contribution of endogenous shocks to inflation in 2022; inflationary pressures in 2022 were mostly external (exogenous and euro area shocks dominated inflation).
- 2023–2025H1 period:
  - Fiscal policy became pro-cyclical in 2023 and contributed significantly to headline inflation in 2024 and 2025H1.
  - Average impact magnitude since 2024: 0.6 ppts.
  - Since 2024, the negative primary balance shock accounted for more than 40 percent of the impact attributed to endogenous shocks on headline HICP inflation.
  - Considering only shocks that Croatia can influence (i.e., excluding euro area or exogenous shocks), fiscal policy contributed to about 60 percent of the total headline HICP inflation.
- Other model notes:
  - Pre-determined factors (initial conditions not explained by shocks) consistently explain about 2.3– (text truncated in source).

*Source: IMF staff analysis, "Inflation in Croatia: The Role of Fiscal Policy" (chapter content provided).*

### 2.4 ppts of the headline HICP, depending on the sample period used; this represents the structural

### 1hrvea2025002-source-pdf - 2.4 ppts of the headline HICP, depending on the sample period used; this represents the structural

### Key findings on inflation drivers
- Fiscal shocks account for around one-quarter of the headline HICP, when excluding only these pre-determined factors.
- The forecast error variance decomposition confirms fiscal shocks can explain about 18 percent of the total inflation uncertainty forecasted in the longer run.
- The contribution of fiscal shocks to Croatia’s national CPI is in line with that to the HICP, as the two series largely overlap.
- Fiscal shocks have been unusually large in recent periods, hence their notable historical contributions.

### Quantitative impulse-response and cumulative impacts
- A decrease in the primary balance to GDP ratio by 1 standard deviation (around 1 ppt) is estimated to increase inflation by 0.2 ppts in the first quarter after the shock.
- The impact is persistent over time, cumulatively accounting for around ½ ppts one year from the shock and almost 1 ppt in the longer run (3 years).
- Cumulatively, 0.8 ppts of the headline HICP inflation in 2024 was attributable to the increase in public wages.
- Cumulatively, 0.7 ppts in the first half of 2025 was attributable to the increase in public wages.
- The impact of total public expenditure is found to be around ½ ppts (and reached 0.8 ppts in 2025).

### Role of fiscal components (taxes, expenditures, public wages)
- Reductions of tax rates to mitigate the “cost of living” crisis in 2022 played an important role in explaining the headline inflation in 2022 (especially combined with wage growth).
- The impact of tax revenue shocks has been much smaller in the most recent years.
- An increase in expenditures or in public wages growth played a primary role in explaining headline HICP inflation since 2024, making up nearly half of the total impact from endogenous shocks.
- The impulse responses to both tax revenues and expenditures shocks are highly comparable.
- The increase in fiscal expenditure and public wage growth is found to have contributed to around 20 percent of the headline inflation since 2024.
- Without the increase in public wages in 2024:
  - q/q headline inflation would have moved from 1.3 percent to 1.1 percent;
  - y/y headline inflation would have moved from 4.0 percent to 3.1 percent.
- In 2025Q1–Q2 the headline HICP would have been below 4 percent (y/y) absent the public wage increases.

### Timing, persistence, and core inflation
- Regardless of how fiscal shocks are measured, their impact on inflation took time to materialize; the pick-up in inflation starting at the end of 2024Q3 shows the full effects likely materialized with some lags.
- In 2024Q4 and 2025Q1, about one quarter of the q/q inflation was due to fiscal shocks, including all shocks and pre-determined factors (i.e., more than half of total endogenous shocks).
- The impact of fiscal shocks on core inflation persists longer than in the case of headline inflation, lasting up to 2 years from the time of the shock.
- The stronger impact of fiscal shocks on core inflation in 2022: 1.6 ppts for core compared to 1.1 ppts for headline inflation.
- The structural explanatory part of core inflation is bigger than that of the headline inflation (2.5 percent versus 2.3 percent in the baseline), confirming structural persistence in core inflation components such as services.
- The largest contribution of fiscal shocks is estimated to be in 2024Q4 and 2025Q1, when they accounted for 1/3 of the total endogenous shocks, or over a half of domestic shocks.

### Policy-relevant conclusions and recommendations
- Fiscal policy has played a notable role in Croatia’s recent inflation dynamics: fiscal loosening in 2024, particularly through public wage increases, accounted for more than 40 percent of the total contribution of endogenous shocks to headline HICP and about 60 percent of the total shocks that Croatia can influence.
- Fiscal shocks, including wage hikes, became more visible with a lag, coinciding with a sharp price increase in late 2024.
- Fiscal policy should be mindful of its inflationary impact.
- With monetary policy set at the euro area level and macroprudential tools now actively deployed to contain the growth of households’ credit—thereby helping to reduce demand-driven inflationary pressures—fiscal policy, while pursuing its own objectives, should be tightened considerably to ensure a coherent overall policy mix to limit imbalances in the economy and safeguard macroeconomic stability.

### Econometric methodology and identification
- The analysis uses a Bayesian VAR model based on Nguyen et al. (2023), identifying fiscal shock by sign restrictions in the baseline (expansionary fiscal policy that decreases primary balance but increases GDP growth or inflation).
- Baseline endogenous variables: Croatian and euro area real GDP growth (log difference of real GDP) and inflation, and the primary fiscal balance in percentage of GDP.
- Lag length q = 4.
- Sign restrictions are imposed only on contemporaneous (first period) responses; the euro area aggregate demand and supply are identified separately via block exogeneity.
- Extended model replaces the primary balance with (tax) revenues to GDP and expenditures to GDP (or public wage growth replacing expenditures).
- Identification tables:
  - Baseline: Primary balance (−), Output growth (+), Inflation (+) among others (see Appendix I. Table 1).
  - Extended: Expenditure/GDP and Revenue/GDP identified with respective AD/AS/EA sign restrictions (see Appendix I. Table 2).

### Data, robustness checks, and model diagnostics
- Data source: Haver based on CNB and ECB data, covering 2000Q1 to 2025Q2.
- All variables, besides the balance, are seasonally adjusted. Real GDP and prices are taken in quarterly log difference.
- A COVID-19 dummy equal to 1 for 2020Q1 to 2021Q2 is included.
- Sensitivity to prior tightness λ₁ tested: λ₁ = 0.7 (baseline), λ₁ = 0.5, λ₁ = 0.1. Impulse responses and historical decompositions remain qualitatively and quantitatively similar.
- Model fit and diagnostics (baseline with primary balance, λ₁ = 0.7):
  - R2 = 0.566
  - Adj- R2 = 0.446
  - SSR = 25.77
  - RMSE = 0.5128
  - DIC = 1399.77
- Extended specifications and alternative priors reported; e.g., With public wages growth: R2 = 0.597, Adj- R2 = 0.458, SSR = 23.92, RMSE = 0.494, DIC = 1999.70.
- Model results statistically significant at the 95 percent confidence level and robust across diagnostic checks.
- A cyclically adjusted primary balance robustness check (subtracting 0.5 times the output gap) yields similar results.

*REPUBLIC OF CROATIA — INTERNATIONAL MONETARY FUND*

### Appendix II. Figure 2. Historical Decomposition of Headline HICP and FEVD

### Appendix II. Figure 2. Historical Decomposition of Headline HICP and FEVD

### Baseline: Real GDP Growth
- Decomposition shown: quarter-on-quarter real GDP growth and cumulative annualized decomposition during the post-COVID-19 period.
- Impulse response of GDP growth to an expansionary fiscal shock:
  - The response is large initially but not persistent over time.
  - Implies a positive fiscal multiplier; the multiplier is greater than zero and possibly close to one in the very short run.
- Historical contribution:
  - In 2022, fiscal shocks were a sizable positive contributor to GDP growth.

### Comparison Across Different Fiscal Shocks to Headline HICP
- Impulse responses over a 12-quarter horizon to a 1 percentage point fiscal shock are presented for:
  - Primary balance, public expenditure, and public wages.
- Contributions to HICP inflation from 2022 to 2025H1 are decomposed by these fiscal components.
- Model variants include:
  - Specification with public wages growth (historical decomposition of year-on-year HICP inflation for each quarter of 2024 and 2025).
  - IRFs to a tax revenue shock (e.g., a tax rate reduction) or an increase in expenditure.

### Baseline: Core HICP
- Historical decomposition of core HICP from 2022 to 2025H1 attributes movements to:
  - Pre-determined factors, exogenous shocks, fiscal shocks, and domestic or euro area non-fiscal shocks.
- Impulse response:
  - Core HICP response to a one percentage point fiscal shock over a 12-quarter horizon is shown with 95 percent confidence.

### Differences Between CPI and HICP
- Key conceptual difference:
  - HICP includes the total consumption of institutional households and non-residents in the economic territory; this is not included in the national consumer price index (CPI).
- Robustness:
  - BVAR analysis results using CPI instead of HICP are robust because the two series overlap for most of the period considered.
  - Main empirical differences are seen in services inflation.

### Baseline: Impact of an Exogenous Oil Shock
- Specification that explicitly controls for an exogenous oil shock (as occurred in 2022):
  - Both domestic and euro area aggregate shocks lose prominence.
  - Exogenous shocks—including COVID-19 and oil shocks—become significantly more important for 2022.
  - Impact of fiscal shocks in the most recent quarters remains very similar to the baseline.

### Baseline: with Euro Area Monetary Policy Shock
- Robustness check includes a euro area monetary policy shock identified as in Deskar-Škrbic et al. (2020).
- Results are robust to inclusion of this shock.
- Identification table (variables/shocks) shows signs and zeros for relationships between fiscal shock and AD, AS, EA, MP for variables including primary balance, output growth, inflation, EA output growth, EA inflation, EA interest rates.

### Extended Model with Total Revenues
- Extended model replaces tax revenues with total revenues.
- Impact on results:
  - The impact is slightly reduced in 2024–25 but the overall narrative holds.

### Comparison Across Models: Fiscal Shocks Impact on Headline HICP Inflation
- Across main setups and further checks, the contribution of fiscal shocks (using different definitions) is very robust, especially since 2024.

### Improving Investment in Human Capital: Efficiency of Croatia's Public Spending on Education and Healthcare
- Paper focus and purpose:
  - Assesses efficiency of Croatia’s public spending on education and healthcare, identifying potential for fiscal savings and improved outcomes.
  - Offers structural reform options to increase efficiency while recognizing need for societal consensus.
- Key recommendations (menu of options):
  - Education: consolidate schools, increase mandatory instructional hours, revise curricula to enhance skills acquisition, promote general track education over vocational education to support tertiary enrollment.
  - Healthcare: develop telemedicine and primary care outside hospitals, review central role of hospitals and reorganize to address population aging, increase prevention to improve well-being and reduce fiscal cost of aging.

### Context, Fiscal Position, and Human Capital Metrics
- Fiscal context:
  - After three years of expansionary fiscal policy, fiscal balance shifted from a surplus of 0.1 percent of GDP in 2022 to a deficit projected at 2.9 percent of GDP in 2025.
  - A fiscal consolidation over the medium term is needed; increasing public spending efficiency would facilitate consolidation.
- Human capital metrics and comparisons:
  - World Bank estimate: a Croatian child born in 2020 will enter adulthood (age 18) about 71.0 percent as productive as a peer who receives complete education and proper healthcare.
  - Improvement over a decade: 1.6 percentage points (ppts).
  - Croatia’s human capital index by gender: 0.68 for boys and 0.74 for girls.
  - Potential illustrative gains: if education components were 10 percent higher, the human capital index would be 84.6 instead of 71.0 percent; if health components were 10 percent higher, the index would be 74.2.
- Public spending on education:
  - General government spending on education: 5.3 percent of GDP.
  - Spending on pre-primary to primary education in 2023: 2.6 percent of GDP.
  - Private financing of education is similar to NMS average and slightly smaller than EU average.
- Demographics and spending projections:
  - Student population projected decline: 38 percent between 2022 and 2070 (Ageing Report 2024).
  - Long-term spending implication: education spending in 2070 would be about 0.5 percent of GDP lower than in 2024 (compared to 0.4 percent of GDP lower for the EU and 0.1 percent of GDP for NMS).
  - Near-term projection: public spending on education would decline by only 0.1 percent of GDP between 2024 and 2030.

### Education Outcomes and Distributional Concerns
- Fourth graders and 15-year-old students:
  - Reading (fourth graders): among highest in EU PIRLS assessment.
  - Mathematics (fourth graders and 15-year-olds): among the lowest (TIMSS and PISA).
  - Science: intermediate position but lower than comparators.
- Distribution of performance (15-year-olds, 2022):
  - Share with mathematics score below OECD baseline: 40 percent.
  - Vocational track mathematics low performers: 44.0 percent; general track: 7.8 percent.
  - Share of high performers in mathematics: 5.2 percent (fifth lowest in EU).
  - Share of low performers in reading and science: 31 percent.

*Source: Appendix II. Figure 2. Historical Decomposition of Headline HICP and FEVD (1hrvea2025002-source-pdf)*

### 19.      That relatively high education spending translates to educational outcomes that are, at

### 1hrvea2025002-source-pdf - 19.      That relatively high education spending translates to educational outcomes that are, at

### Efficiency of education spending: findings
- Potential efficiency gains are large: Both Hungary and Poland achieve a higher educational outcome than Croatia at a lower cost. The EU achieves the same outcome as Croatia but at a lower fiscal cost, while Austria achieves a higher educational outcome with similar spending as Croatia. Slovenia’s educational system is less efficient than Croatia’s as a much larger spending translates to an educational outcome that is only modestly higher.
- Using comparators, potential fiscal saving ranges from 0.5 to 1.0 percent of GDP; increases in PISA scores range from 18 to 43 percent.
- Differences in spending can reflect structural differences (e.g., demographic, distribution of students across level of education); therefore estimated efficiency gains and fiscal savings are “potential” and reported as a range rather than a point estimate.
- Alternative measurements confirm that potential efficiency gains are large.

### Efficiency-increasing and cost-reducing reforms (policy directions)
- Structural reforms can raise educational outcomes while reducing public spending, freeing resources for other education needs or fiscal consolidation.
- Reorganization measures suggested:
  - School and class consolidation to allow better allocation of teachers and to address projected strong decline in student population.
  - Consolidation could provide savings because small schools are costly to run and result in duplication of administrative tasks, jobs, and equipment; spending per student and school size are negatively correlated at both primary and secondary levels.
  - Consolidation may require provision of free transportation in remote areas, building on free transportation provided to children attending kindergartens.
- Move to a “whole day school” system:
  - At the beginning of the decade, 40 percent of primary schools (60 percent of students) operated in double-shift work.
  - A revamping of the primary system launched in 2023 will be completed by 2028; eliminating double-shifts provides an opportunity for better use of primary school infrastructure and teachers, potentially lowering fiscal cost.

### Cost drivers and labor market / wage bill facts
- In 2023, compensation of employees of the education system accounted for 3.3 percent of GDP.
  - This was 0.4 percent of GDP higher than EU, 0.3 percent of GDP higher than in Poland, and 1.1 percent of GDP higher than in Hungary.
  - It was 0.2 percent lower than in Slovenia.
- Compensation of employees of the education system was the same as a share of GDP in 2023 as in 2013, but represents a growing share of total government wage bill: an increase of +1.1 ppt in the share of education in the general government’s wage bill.
- The weight of the education wage bill as a share of GDP is likely to have increased since 2023 following the large public sector wage increase implemented in 2024.
- Both wage levels and number of teachers contribute to the high wage bill:
  - Statutory wages in public institutions were high by peer standards at every point of a teacher career in 2022.
  - The gap with peers likely increased after 2022 due to strong public sector wage growth.
- Employment dynamics and student population:
  - Decline in student population over 2015–22 was particularly large in Croatia: -4.8 percent compared to -1.9 percent for the EU, -3.9 percent in Hungary, and -2.4 percent in Poland.
  - Despite the decline in student population, employment in education increased in Croatia (and increased more than in the rest of the public sector), while employment in education declined in the EU and Poland and remained stable in Hungary.
  - In the four most recent years (the school year 2018–19 to 2022–23), the number of teachers at upper secondary level increased by 2.9 percent while the number of students declined by 0.8 percent.
- Student-to-teacher ratios:
  - The decline in student population combined with increased employment led to a drop in the student-to-teacher ratio at all levels in Croatia; the ratio is now lower in Croatia than in any comparator and the EU.

### Structural features and inefficiencies
- Small schools:
  - A quarter of schools have 3 or fewer students per grade.
  - The median primary school has 8 students per grade compared to an OECD average of 27.
  - Many small schools are an important driver of high employment and low student-per-teacher ratios.
  - Small schools contribute to inefficient allocation of teachers, part-time work, and teacher shortages.
- Teacher shortages and subject gaps:
  - The percentage of students in schools whose principal reports that the school's capacity to provide instruction is hindered to some or a large extent by staffing issues increased from 18 percent in 2018 to 46 percent in 2022.
  - Shortages are particularly severe in physics and mathematics, affecting acquisition of basic skills and mathematics performance.
  - This may require a review of tertiary enrollment quotas to meet demand for STEM teachers.
- Part-time work:
  - At secondary education the proportion of teachers working part-time is much higher than in peers.
- Regional disparities:
  - Annex III shows a large regional dimension to the small school issue.
  - OECD (2024b) reports the largest regional difference (across the 4 NUTS2 regions) in enrollment of 15–19-year-olds among EU countries: it ranges 16½ percent in Northern Croatia to 100 percent in Zagreb.

### Outcome-increasing reforms (policy directions)
- Increase instructional time:
  - Croatia has the lowest number of instructional hours in the EU in both primary and secondary education.
  - Examples of instructional hours in 2023:
    - Primary education: 473 hours per year compared to EU25 average 738 hours, OECD average 805, Hungary 679, Poland 558, Slovenia 682.
    - Lower secondary: 663 hours compared to EU25 876, OECD 916, Hungary 803, Poland 754, Slovenia 766.
    - Upper secondary: 568 hours compared to EU25 799, OECD 842, Hungary 741, Poland 656, Slovenia 710.
  - The whole-day school reform aims to increase instruction time for grades 1 to 8 and, if combined with better resource use and revised curricula and teaching methods, could increase educational outcomes.
  - International evidence suggests such reform could also boost women’s labor force participation and help reduce labor shortages.
- Curriculum revision to improve numeracy and labor-market alignment:
  - Test scores point to the need to improve teaching of mathematics.
  - For job prospects, higher numeracy skills correlate with higher labor force participation, lower unemployment, and higher earnings.
  - VET curriculum revision is particularly important: VET students demonstrate lower basic skills. In 2022 PISA, 92 percent of Croatian students of the general track education achieved level 2 or above in mathematics compared to 56 percent for vocational track students.
  - A revision of the curriculum is underway: working groups to draft a proposal for a national early childhood and pre-school curriculum; VET curriculum changes to be more skill-based are being implemented in all secondary vocational schools.
  - Frequent curriculum review and strong collaboration with firms are necessary to identify labor market needs and ensure smoother transition to less classroom-based VET.
  - Employer involvement in VET curricula and funding in Croatia is limited; nearly half of VET graduates work in a field outside their specialization and there are larger imbalances between job qualifications and requirements for medium skilled occupations than in peers.
- Change teaching methods and teacher training:
  - New curricula and a shift to more skill-based VET require changes in teaching methods, teacher education, and in-service training and pedagogical support.
  - Only 45 percent of teachers in upper-secondary schools used ICT for teaching in 2018 and 38 percent feel adequately prepared to teach using digital aids.
  - Promoting group work and team teaching may improve creative thinking and problem-solving skills.
  - PISA 2022 results: VET students score lower than general track students on “collaborative problem solving” (VET score 448 compared to national average 473).
- Targeted tertiary enrollment and teacher supply:
  - Review tertiary enrollment quotas to better meet demand for STEM teachers.

### Empirical relationships and supporting evidence
- Lower student-to-teacher ratio in Croatia is not associated with comparatively higher educational outcomes.
  - Hattie (2005) meta-analysis: reducing class-size has only a “tiny” positive effect; smaller classes often use the same teaching methods as larger classes and thus do not optimize opportunities from fewer students.
- Empirical statistics cited throughout the analysis are drawn from Eurostat, OECD, Croatian authorities, and IMF staff calculations.

*Source: https://www.imf.org/-/media/files/publications/cr/2025/english/1hrvea2025002-source-pdf.pdf*

### 33.      Postponing the start of the VET may help increase the basic skills of younger workers.

### 33.      Postponing the start of the VET may help increase the basic skills of younger workers.

### VET timing and promoting the general track
- VET in Croatia starts earlier than in peers (14–15 years).
- Delaying the start of VET would:
  - Complement the whole school day reform and the revision of the curriculum to give more time to students to better acquire basic skills and non-routine and cognitive and social skills such as creative thinking and collaborative problem-solving.
  - Diminish gaps in educational performance between VET students and general track students.
- Croatia has the third highest enrollment rate of upper secondary students in VET among EU countries (Dalvit and others, 2023).
  - VET enrollment rate increased from 68.2 percent in 2018–19 to 69.7 percent in 2021–22 (Croatian Bureau of Statistics, 2024).
- Emphasis on VET contributes to higher spending per student in upper secondary education than in peers but lower overall educational outcomes because VET students demonstrate lower PISA performance than general track students.
- Promoting the general track is argued to:
  - Increase overall educational outcomes.
  - Foster enrollment in tertiary education, given large returns to higher education (higher education associated with an increase in earnings of 9.8 percent in 2022, World Bank, 2024).
  - Be potentially more effective than attempting to reform VET, because difficulty to join and succeed in tertiary education after VET is substantial (failure rate to the “matura” exam was 3.9 percent for general track students versus 37.8 percent for VET students in 2019–20, World Bank, 2023).

### Tertiary education, skills mismatch, and labor market implications
- Tertiary promotion would reduce skills mismatch, labor shortages, and boost innovation and competitiveness.
- Skills mismatch is high in Croatia (Figure 20); contributes to low labor force participation despite labor shortages and reflects insufficient graduates from tertiary education.
- Employment outcomes:
  - Employment rate of people with tertiary education is much higher than for the rest of the population but lower than comparators and the NMS average (Figure 18).
- Field composition matters: a relatively high share of tertiary students enroll in fields with limited job prospects (World Bank, 2023; EC, 2025).
- Quality and relevance of tertiary education need to be ensured:
  - Past reforms linking funding to quality should be built upon.
  - Measures should ensure quality and adequate curricula; graduates reportedly lack practical knowledge, application skills, and adaptability in some areas (Pažur Aničić and others, 2023), with the largest deficit in biomedical and health sciences.
- Over-qualification and wage impacts:
  - A recent survey: overqualification at 15 percent, below OECD level of 23 percent.
  - Being overqualified leads to a wage on average 7.3 percent lower than peers in well-matched jobs, compared to 12.5 percent average in OECD and 11.4 percent for EU countries that participated in the survey (OECD, 2024e).
- Capacity and institutional consolidation:
  - Enrollment in tertiary education is below capacity; NRRP indicates at least 12 faculties, polytechnics and/or institutes should be merged or integrated (Kršul, 2025).

### Early childhood education and childcare infrastructure
- Development of childcare infrastructure (supported by Cohesion policy and NRRP) targets rural and less-developed areas to increase early childhood education enrollment (EC, 2025).
- Expected benefits if accompanied by adequate, high-quality staffing:
  - Increased educational achievement and long-term productivity.
  - Increased women’s labor force participation (Hallaert and others, 2023).
- Staffing considerations:
  - Adequate staffing includes addressing teacher shortages and ensuring proper teacher training; Croatia offered primary school teachers the possibility to become early childhood educators to help staffing.
- Inequality in access:
  - Enrollment among children aged 0–2: 11 percent for bottom income tertile versus 29 percent for top tertile — a larger gap than EU and OECD averages (OECD, 2024b).
  - Children aged 0–2 whose mother has tertiary education benefit more from early childhood education than children whose mother does not: 35.1 percent versus 20.5 percent — a "14½ ppts gap" noted as about 4 ppts larger than the EU25 average and twice as large as the gap in Hungary and Slovenia (OECD, 2023a).
- Affordability and support:
  - Childcare cost paid by parents accounts for a lower share of household income than EU average and declined over time for most family types (Figure 22).
  - Government provides subsidies to socially disadvantaged families and minorities (including exemptions for Roma parents) and transport to facilities in remote areas (OECD, 2024b).

### Healthcare: public spending and outcomes
- Public spending on healthcare:
  - In 2023, public spending on healthcare reached 7.8 percent of GDP (Table 7).
  - This is 0.5 percentage point of GDP more than the EU27 average and almost 2 ppts higher than the NMS average.
  - In 2022 capital spending was 0.5 percent of GDP; total public spending on healthcare in 2022 also reached 7.8 percent of GDP.
- Composition and trends:
  - Healthcare accounted for a higher share of general government expenditure in 2023 than in the EU or any comparators; this share has increased steadily over the last decade and was larger than in the EU and comparators as early as 2017 (Figure 23).
- Public vs private financing:
  - Public spending accounts for 85 percent of current healthcare spending in Croatia — about 10 ppts more than EU or NMS averages.
  - As a result, total current healthcare expenditure as a share of GDP is smaller in Croatia than the NMS average, while public spending is larger (Figure 24).
- Health system features:
  - The system is centralized; most providers are publicly owned though private providers have grown, especially in primary care, dental services, and specialized clinics.
  - CHIF is the single purchaser of services under mandatory health insurance and offers voluntary complementary insurance.
  - Mandatory health insurance covers all residents, including foreign nationals residing longer than three months; more than 99 percent of the population benefits from health insurance.
  - Amendments to the 2023 Law on Mandatory Health Insurance led to 13,184 people being deregistered from health insurance in 2024, about 0.3 percent of insured persons; the Ombudswoman submitted a constitutional review request in June 2023.
- Generosity and out-of-pocket costs:
  - The system covers most health services and medical goods and a large share of costs, resulting in the lowest out-of-pocket cost in the EU (Figures 25 and 26).
  - Exemptions from co-payments exist for vulnerable groups; some complementary insurance memberships for people with disabilities are financed from the state budget.
- Spending per capita:
  - Total (public and private) current health expenditure per capita (PPP) increased rapidly and outpaced comparators and EU27 between 2019 and 2022, but remained in 2022 one of the lowest in the EU (Figure 27).

*Source: Republic of Croatia — IMF staff report excerpts (content unit).*

### 43.      Against most metrics, health outcomes are lower than in the EU:

### 43.      Against most metrics, health outcomes are lower than in the EU:

### Health outcomes and healthspan
- Death rates (except for maternal death) are all higher than EU averages.
- Life expectancy:
  - Life expectancy at birth is comparatively low.
  - Life expectancy at 65 is even lower (Figure 30).
- Healthspan:
  - The number of years a person lives in good health, estimated by Healthy Life Years (HLY) and Health-Adjusted Life Expectancy (HALE), are comparatively low.
  - Both HLY and HALE declined over past decades (Figure 31).
- Notes on indicators:
  - HLY measures the number of remaining years a person of specific age is expected to live without any severe or moderate health problems (disability dimension).
  - HALE measures the number of years of life lived in full health; it is “based on a much more complex approach that requires a large volume of data to describe the health status of the population, taking into account the overall burden of diseases and injuries, as well as some valuation (weighting) of the severity of disability related to all these diseases and injuries” (OECD-EC, 2024).
  - Both HALE and HLY are used in this paper to assess efficiency.

### Access, unmet needs, and self-perception of health
- Despite poorer outcomes, self-reported unmet medical and dental needs are low.
- Near universal healthcare insurance coverage and its generosity contribute to unmet medical needs due to healthcare organization that are markedly lower than in peers (Figure 32).
- Self-perception of health status has improved in recent years and is now close to the EU27 average, though more unequal (Figure 33).
- Geographic and income inequalities likely underlie persistent disparities in access (remote and poorer areas affected).

### Potential efficiency gains
- A higher level of spending than in most peers is not reflected in better outcomes, indicating lower spending efficiency.
- Benchmarking (Figure 34 and Appendix V) shows:
  - If healthcare spending were as efficient as the best EU performers, Croatia could either achieve higher health outcomes for the same level of spending or achieve the same outcomes at significantly lower cost.
  - Potential efficiency gains are large; reforms can significantly affect both outcomes and fiscal cost (Table 8).
- Table 8 — Croatia: Potential Efficiency Gains in Healthcare
  - If Croatia's healthcare system was as efficient as:
    - Fiscal saving (ppts of GDP)
      - Hungary: HALE (2019) 2.7; HLY (2019/2022) 3.1 / 3.5
      - Poland: HALE (2019) 2.4; HLY (2019/2022) 2.8 / 2.7
      - Slovenia: HALE (2019) 0.8; HLY (2019/2022) 0.9 / 0.8
      - EU: HALE (2019) 0.4; HLY (2019/2022) 1.1 / 0.5
    - Increase in healthspan (years)
      - Hungary: HALE (2019) 38; HLY (2019/2022) 42 / 49
      - Poland: HALE (2019) 33; HLY (2019/2022) 36 / 32
      - Slovenia: HALE (2019) 8; HLY (2019/2022) 9 / 7
      - EU: HALE (2019) 4; HLY (2019/2022) 10 / 4

### Efficiency-increasing reform areas (high level)
- Four types of structural reforms with potential to increase spending efficiency:
  - (a) Reallocating resources to reduce regional disparities.
  - (b) Reviewing the role of hospitals, including preparing them for population aging.
  - (c) Increasing the focus on prevention.
  - (d) Improving the use and distribution of pharmaceuticals.
- Reforms should address potential distributional impacts in design and implementation.

### Reallocating resources to reduce regional disparities
- Geographic distribution of health workers varies considerably; urban concentration of specialized services; primary care practitioners lacking in rural areas and on islands (Džakula and others, 2024).
- Physicians per 100,000 inhabitants in 2023:
  - Sjeverna Hrvatska: 245
  - Zagreb region: 667
  - Difference: 442 per 1000,000 inhabitants
  - Country average: 401 per 1000,000 inhabitants
- Hospital bed access and geographic differences:
  - Number of hospital beds by 100,000 inhabitants was 25 percent higher in the Zagreb region than the national average and 52 percent higher than in Sjeverna Hrvatska (2023).
  - While 83 percent of the EU population lived within 15 minutes driving time of a hospital, in about one third of the NUTS3 Croatian regions the share was below 50 percent and in almost two third of the NUTS3 regions the share was below 65 percent (Eurostat, 2024).
  - In Jadranska Hrvatska less than 8 percent of the population lives within a 10-minute drive of the nearest health center compared to the EU average of about 29 percent (EC, 2025).
- Pharmacy access is unequal: “pharmacies are mainly located in cities and towns, while the pharmacy network in rural and underdeveloped areas remains poorly developed” (Džakula and others, 2024).
- Regional access implications:
  - 54 percent of unmet medical needs due to healthcare organization are because service is “too far to travel.” This is the largest share in the EU.
  - Czech Republic: 20 percent (second largest share in the EU); EU level: 4 percent.
  - For dental needs, the share is 20 percent in Croatia, compared with 8 percent in the Czech Republic and 3 percent at the EU level.
  - Limited access in some regions affects mortality rates; mortality from cardiovascular disease and cancer among the highest in the EU, reflecting limited prevention focus and limited accessibility (EC, 2025).
- Solutions suggested:
  - Better geographical allocation of resources to improve outcomes and spending efficiency.
  - Facilitate patient transport to hospitals (including helicopter emergency medical services) complemented by developing primary care outside hospitals in rural and remote areas.
  - Provide incentives for health professionals to practice in underserved areas.
  - Develop telemedicine to increase access in remote areas; Croatia’s decline in in-person consultations during COVID-19 was almost fully offset by teleconsultations, and teleconsultations remained relatively high in 2022 (OECD-EC, 2024).
  - Promote telemedicine while ensuring availability of healthcare professionals for remote consultations and addressing digital infrastructure/skills bottlenecks.

### Reviewing the role of hospitals and preparing for aging
- Hospitals are central to the Croatian healthcare system:
  - On the eve of the COVID-19 pandemic, hospitals accounted for a larger share of total healthcare spending than in any other EU country.
  - Their share has increased more than in any other EU member except Hungary.
  - In 2023, more than half of the total current healthcare spending was dedicated to hospitals (Figure 35).
- Outpatient care delivered by hospitals:
  - In Croatia, inpatient care accounts for only 40 percent of hospital spending; outpatient care and day care activities (same-day elective surgeries, dialysis) are often performed in hospitals (OECD-EC, 2024).
  - Reasons: hospital system not substantially modified to match medical advances enabling outpatient/day care provision; lack of trust in primary care physicians leading to more referrals to secondary care (Džakula and others, 2024).
  - Developing outpatient care outside hospitals could be more cost-effective and reduce regional imbalances and unmet needs.
- Data gaps and difficulty assessing efficiency:
  - Data limitations hinder evaluation of primary care and hospital efficiency.
  - Key data like up-to-date standardized 30-day hospital mortality rates for conditions (acute myocardial infarction, stroke) are not available.
  - Data on avoidable hospital admissions exist for asthma, chronic obstructive pulmonary disease, and diabetes, but not for hypertension or congestive heart failure.
  - Average length of stay: below NMS and EU averages for acute care, above these averages for childbirth without complications (Figure 36).
  - Some statistics, e.g., relatively high rate of death within a year of discharge for ischemic stroke patients, suggest scope for more and better integrated care (OECD-EC, 2024).
- Hospital capacity and occupancy:
  - Density of hospital beds: 565 beds per 100,000 inhabitants in 2023.
  - This is 11 percent higher than the EU level and 40 percent more than in Slovenia; still below Hungary and Poland.
  - Occupancy rate of inpatient curative care beds is 4 ppts lower than the EU22 average, suggesting scope to reduce curative beds (Figure 37).
- Preparing for aging:
  - 2024 Ageing Report (EC, 2024a) baseline projects health spending would increase 0.7 percent of GDP between 2025 and 2070 (0.8 percent if long-term care cost is added).
  - Aging will increase demand for geriatric beds and require reorganization of medical processes, implying fiscal cost and need for increased cooperation and coordination across services.
  - Elderly patients have higher emergency admission risk and longer hospital stays; complex comorbidities require efficient coordination with rehabilitation institutions and nursing homes.
  - Investment in long-term care facilities is needed to reduce aging pressure on hospitals (Hallaert, 2025b).
- Geographical distribution of hospitals:
  - Hospital network has not been substantially modified in recent decades to match migration patterns, changing demographics, or medical advances, potentially resulting in inadequate provision and higher costs (Džakula and others, 2024).

### Policy implications and recommended focus areas
- Reallocate resources geographically to reduce disparities in physician density, hospital bed distribution, and pharmacy access.
- Strengthen primary care capacity outside hospitals and provide incentives for practice in underserved areas.
- Expand and promote telemedicine in remote areas while ensuring professional availability and addressing digital barriers.
- Review hospital roles: shift appropriate outpatient and day care from hospitals to community/outpatient settings to improve cost-effectiveness and access.
- Improve data collection (e.g., standardized 30-day hospital mortality rates, broader avoidable admission metrics) to evaluate and monitor hospital and primary care efficiency.
- Reorganize hospitals and invest in long-term care and coordination mechanisms to address demographic aging and associated fiscal pressures.
- Consider efficiency gains identified through benchmarking (Table 8) to either increase healthspan or achieve fiscal savings equivalent to up to 3.5 ppts of GDP under best-practice comparators.

*Source: IMF staff summary of chapter section “Against most metrics, health outcomes are lower than in the EU.”*

### 56.      Shortages of healthcare professionals are limited. Though regional imbalances and

### Shortages of healthcare professionals are limited. Though regional imbalances and staffing issues in some hospitals in regions where doctors are scarce may give the impression of physician shortages, the overall number of doctors per 100,000 inhabitants is, actually, close to the EU average and higher than the NMS average and in the three comparators.

### Healthcare workforce: current status and distribution
- Doctors:
  - Overall number of doctors per 100,000 inhabitants is close to the EU average and higher than the NMS average and in the three comparators.
- Nurses:
  - Number of practicing nurses is lower than the EU average and in Slovenia but higher than the NMS average.
  - Practicing nurses increased from 6.8 per 100,000 inhabitants in 2018 to 7.7 per 100,000 inhabitants in 2023.
- Mobility and foreign-trained nurses:
  - Nurses from Croatia are among the 10th largest foreign-trained group in Switzerland in 2023 (OECD-EC, 2024).
  - In 2023, foreign-born nurses accounted for 1.2 percent of the nursing workforce in Croatia compared to the EU19 average of 19 percent and 4.0 percent in Slovenia.
  - In 2023, 51.3 percent of the nursing workforce in Ireland was foreign-trained (OECD-EC, 2024).
- Remuneration:
  - Hospital nurses’ remuneration was 1.1 times the national average wage in 2022.
  - Comparative figures: 1.2 in the EU, 1.3 in Hungary, and 1.4 in Slovenia.
  - Large increases in public sector wages (including in the healthcare sector) took place in 2024 and may have changed the situation.

### Training and retention policy priorities
- Retention vs. education:
  - Increasing the number of students in nursing education is effective generally (OECD-EC, 2024), but for Croatia—already among the highest rate of nursing graduates in the EU—the priority is retaining nurses.
- Policy levers to retain or supplement nursing workforce:
  - Increase wages for hospital nurses (current 1.1 times national average wage in 2022).
  - Build on immigration policy to increase foreign-born nurses (Croatia: 1.2 percent in 2023).
  - Example: Ireland’s active recruitment from countries with a well-established diaspora (e.g., India and the Philippines).
- Reforming training:
  - Surveys point to inadequacy in the qualification of health associates professionals (Annex IV).
  - Pažur Aničić et al. (2023) highlights large deficits among biomedical and health sciences graduates in:
    - Practical knowledge and its application
    - Adaptability
    - Entrepreneurial spirit
    - Leadership
  - Training should prepare graduates for the health needs of an aging population.
  - Hallaert (2025b): training policies need to be implemented rapidly despite short-term costs to meet rising geriatric demand.

### Increasing the focus on prevention: findings and priorities
- Avoidable deaths and expenditure:
  - Death rates from treatable and preventable diseases are both higher than EU levels.
  - Death rates from treatable diseases compare well to countries with comparable health spending; preventable disease death rates are worse.
- Major risk factors and Croatia’s performance:
  - Smoking:
    - Croatia has a high share of population aged 15 and over that smokes and smokers smoke more intensively than in any European peers.
    - Tobacco consumption is more widespread among adolescents than in most EU countries.
    - Lung cancer accounts for 5.6 percent of deaths in Croatia (1 ppt more than in the EU or NMS) and accounts for 22 percent of deaths from cancer in Croatia (2.2 ppts more than in the EU27 and 1.4 ppts more than in the NMS).
    - Colorectal cancers account for 3.9 percent of deaths (more than 1 ppt higher than in the EU or NMS) and 15.3 percent of deaths from cancer (4 ppts more than in the EU). Ninety percent of colorectal cancers are thought to be preventable.
  - Diet:
    - Reported consumption of fruits, vegetables, and whole grains is below the EU27 level.
  - Obesity:
    - Croatia has the largest share of population reporting being obese in the EU: 24 percent.
    - Croatia has the second largest share reporting being obese and overweight.
  - Alcohol:
    - Overall alcohol consumption has declined significantly in the past decade and is now lower than the EU average and in any of the three comparators.
    - Episodes of heavy drinking are less frequent in Croatia than in the EU27.
    - Adolescent consumption of alcohol is more frequent than in most EU peers.
  - Physical activity:
    - Share of 11-year-olds and 15-year-olds meeting WHO recommended daily physical activity is among the highest in the EU for both girls and boys.
    - Share of population performing non-work-related physical activities is below EU levels.
    - 56 percent of adults never perform physical activity outside working time.
- Prevention spending and allocation:
  - Until the COVID19 pandemic, Croatia dedicated a slightly larger share of its current healthcare spending to prevention than the EU27 as a whole.
  - Preventive colorectal cancer screening is among the lowest in the EU despite higher share of death from colorectal cancer.
  - Fighting overweight and obesity is likely cost-effective given higher average total healthcare expenses for people with overweight and obesity and higher absenteeism-related costs.
  - Example cost estimates elsewhere: in Belgium, additional annual healthcare cost of overweight and obesity estimated at 0.6 percent of GDP; cost of absenteeism estimated at 0.2 percent of GDP (Gorasso and others, 2022).
- Advocacy and regulatory reforms:
  - Campaigns targeting younger teenagers on smoking, excessive alcohol consumption, and proper diets could have significant impact.
  - Strengthen anti-smoking policies, which remain comparatively lenient.
  - Strengthen alcohol control policies (Džakula and others, 2024; EC, 2023b and 2025).
  - Such measures would entail limited costs and improve health status, increase health span, promote healthy aging, and reduce fiscal healthcare spending including the expected fiscal cost of aging.
  - The Ageing Report (EC, 2024a) estimates that the healthcare dimension of the fiscal healthcare cost of aging would be reduced from 0.7 percent of GDP between 2025 and 2070 to 0.4 percent of GDP under a “healthy aging” scenario.
- Environmental toxins:
  - Prevalence of premature deaths attributable to air pollution remains one of the highest in the EU despite a decline.
  - Figure 48 measures premature deaths attributable to annual PM2.5 concentrations above 5μg/m3.

### Improving the use and distribution of pharmaceuticals
- Spending:
  - In 2023, Croatia dedicated about 16½ percent of healthcare spending to pharmaceuticals.
  - This is 1.7 ppts more than the average for the 18 EU countries for which data is available.
  - Pharmaceuticals represented one-third of private out-of-pocket payments in 2021; dental care represented 28 percent and outpatient medical care 18 percent.
- Distribution and retail competition:
  - Counties are responsible for organization and management of pharmacies; pharmaceuticals are provided by pharmacies contracted by the CHIF (Džakula and others, 2024).
  - Limited competition in retail sale of medicine may result in higher prices of non-prescription drugs.
- Generics and procurement:
  - Share of generics is comparatively high but international comparisons suggest scope to increase it.
  - Ministry of Health aims to:
    - Increase the use of generic drugs.
    - Reduce waste through joint public procurements with other EU Member States.
    - Better control consumption of expensive medicines (Džakula and others, 2024).
  - Complementary measures:
    - Avoid over- or inadequate prescriptions of some drugs (World Bank, 2021).
    - Improve national, evidence-based treatment guidelines and pharmacotherapy guidelines (World Bank, 2021).
  - These measures could address hospitals’ arrears (World Bank, 2021).

### Conclusions and reform implications
- Potential gains:
  - The potential to enhance the efficiency of public spending on education and healthcare is large.
  - Increased efficiency could increase outcomes, boost human capital formation, support productivity, and foster higher potential growth necessary for faster income convergence with EU partners.
  - Efficiency gains could free fiscal resources to finance other needs or support planned fiscal consolidation.
- Reform horizon and considerations:
  - Increasing efficiency requires structural reforms that take time to develop, implement, and yield results.
  - Reforms can have important distributional impacts that should be carefully considered in design and implementation.
- Education-health nexus:
  - Main purpose of educational reform should be to increase educational outcomes and reduce skills mismatch to support labor market needs and technological adaptation.
  - Reforms in education would eventually increase employment, creation and diffusion of innovation, and productivity.
  - Consolidation of small schools has the potential to reduce staff shortages and fiscal spending without reducing educational outcomes.

*International Monetary Fund, Republic of Croatia chapter (excerpts).*

### 70.      Reform in the healthcare sector should aim at improving health outcomes and access

### 70.      Reform in the healthcare sector should aim at improving health outcomes and access to healthcare for the population in remote and rural areas.

### Key findings and diagnosis
- Croatia dedicates a larger share of its resources to healthcare than peers, but outcomes are lower.
- Remote and rural areas are facing inadequate access to medical services, creating geographical inequality in access to healthcare services.
- The central role of hospitals in the healthcare system—notably in the provision of outpatient care—contributes to higher costs and may not align with the needs of an aging population.
- Inefficient use and distribution of pharmaceuticals contributes to higher healthcare cost for both public finances and the population.

### Policy recommendations and efficiency-increasing reforms
- Reduce geographical inequality in access to healthcare services by:
  - Developing telemedicine.
  - Expanding primary care outside hospitals.
- Review the central role of hospitals in the healthcare system, with particular attention to the provision of outpatient care, to:
  - Help reduce costs.
  - Free resources to prepare for the reorganization of hospitals needed to respond to the needs of an aging population.
- Strengthen focus on prevention to:
  - Significantly improve population well-being and health outcomes.
  - Eventually reduce spending and mitigate the fiscal cost of aging by fostering healthy aging.
- Improve the use and distribution of pharmaceuticals to reduce healthcare costs for public finances and the population.

### Expected benefits and channels
- A stronger emphasis on prevention and community-based care would:
  - Improve health outcomes and reduce inequality in health status.
  - Foster healthy aging, yielding longer healthspan and a healthier population.
  - Boost workforce productivity and the supply of labor by reducing disease burden and the need to care for disabled elderly relatives.
- Telemedicine and primary care outside hospitals would improve access for remote and rural populations and reduce reliance on hospital outpatient services.

*Source: Republic of Croatia — IMF staff analysis (page 70).*

---


_Source: https://www.imf.org/-/media/files/publications/cr/2025/english/1hrvea2025002-source-pdf.pdf_