## Box 1. Performance Indicators

## Source details

**Canonical URL:** [Box 1. Performance Indicators](https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08136.pdf)

## Other formats

- [Markdown version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08136.pdf.md)
- [Structured JSON version](/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08136.pdf.json)

---

### Overview and purpose
- Compares Croatia’s spending and performance in health care and education with comparator countries (EU-15, EU-10, Cyprus, Malta, and OECD) using Data Envelopment Analysis (DEA).
- Defines relative efficiency as the distance of a country’s observed input-output combination from an efficiency frontier.
- Uses performance indicators divided into desired outcomes and intermediate outputs to evaluate efficiency and identify reform options.

### Performance indicators used
- Health care
  - Intermediate outputs: density of physicians, pharmacists, and healthcare workers; number of hospital beds; number of immunization vaccines.
  - Outcome variables: infant-, child-, and maternal mortality rates; standardized death rate from all causes per 1,000 people (WHO definition); incidences of tuberculosis; healthy average life expectancy (HALE, WHO).
- Education
  - Intermediate outputs: primary pupil-teacher ratios, enrollment rates, rates of progression to secondary education, graduation (completion) rates.
  - Outcome indicator: average score on PISA 2006 in mathematics (secondary education).

### Health care — structure, outcomes, and resources
- Financing and contributions
  - Health care in Croatia is mainly financed (around 90 percent) by the Croatian Health Insurance Institute (HZZO).
  - Payroll contributions are set at 15 percent of the gross wage; enterprises pay another ½ percent of wages for work-related injury insurance.
- Key health outcome and expenditure figures for Croatia (selected)
  - Total Expenditure on Health (in percent of GDP): 7.9
  - Public Expenditure on Health (in percent of GDP): 6.6
  - Healthy Life Expectancy (years): 66.6
  - Standardized Death Rates (per 100,000 people): 886.9
  - Infant Mortality Rate (per 1,000 live births): 6.0
  - Child Mortality Rate (per 1,000 live births): 7.0
  - Maternal Mortality Rate (per 100,000 live births): 10.0
  - Incidence of Tuberculosis (per 100,000 people): 40.6
- Selected health resource indicators for Croatia
  - Hospital Beds (per 1,000): 5.6
  - Physicians (per 1,000): 2.4
  - Health Worker Density Index (per 1,000): 7.7
  - Pharmacists (per 100,000): 55.8
  - Bed Occupancy Rate, Acute Care Hospitals (percent): 88.1
  - In-patient Care Admissions (per 100): 16.6
  - Average Length of Stay (all hospitals) (days): 10.3
  - Immunization, Measles (percent of children ages 12-23 months): 96.0
- Private funding and per-capita spending
  - Croatia spends 831 PPP dollars per capita on health, of which 143, or 16 percent, comes from private sources.
- Sustainability and inefficiency concerns
  - Stock of health sector arrears at end-2006: 1.1 percent of GDP.
  - About 84 percent of health care spending comes from public sources.
  - Co-payments share in total health spending: less than 1 percent.
  - Coverage features: 600,000 people have supplementary insurance from HZZO; around 1,900 drugs on the A list are fully paid by HZZO and 300 on the B list are partially paid.
  - Government introduced a flat administrative fee of 10 kuna per person (cap 30 kuna per month) in 2005; the government decided to abolish this fee in 2008.
- Demand and supply drivers of high spending
  - Old-age dependency ratio (population aged 65 and older relative to population aged 17–64) was 26 percent in 2006 and is projected to increase to 48 percent in 2051.
  - Only around 35 percent of the population pays contribution.
  - Average Length of Stay (ALOS) in Croatia in 2005: 10.3 days.
    - ALOS comparisons: EU-10 average 8.6 days; EU-15 average 8.4 days.
  - Primary-care referral behavior: 70 percent of patients at primary health care level are referred to hospitals (experts believe this could be reduced to 30 percent).
  - Provider structure: of 66 hospitals, only 3 were privately owned in 2006; only 3 percent of medical doctors work in the private sector.
- Projections and risks
  - Pre-accession Economic Program envisages an increase of 4 percentage points of GDP in public health spending from 2005 to 2050.
  - Population aging and new medical technology costs could increase spending further.

### Education — spending and outcomes (selected)
- Structure and reforms
  - Education mainly financed and operated by the public sector.
  - Government reforms: Strategic Development Framework 2006–13 and Education Sector Development Plan (ESDP) 2005–10.
- Spending and outcome highlights
  - Total spending on education (2005): around 5.6 percent of GDP.
  - Public education spending (2005): about 4.8 percent of GDP.
  - Croatia’s private education spending (around ¾ percent of GDP) is higher than EU-10 and EU-15 averages.
  - Despite spending in line with comparators, educational output and outcome levels are lower than EU-10 and EU-15 peers.

### Education — key findings and outcomes
- Enrollment and completion
  - School enrollment and completion rates are lower than comparator countries.
  - Tertiary gross enrollment was about 46 percent in 2006, compared to about 53 percent in the EU-10.
  - Only one third of tertiary students reportedly complete their programs, with an average completion rate of 6.7 years in four-year programs.
- PISA and staffing
  - Croatia ranked 36th out of 57 countries in the 2006 PISA mathematics test; only Bulgaria and Romania in the EU-10 scored worse than Croatia.
  - Student-teacher ratios in primary and secondary schools have been falling and are lower than comparator countries.
  - About 65 percent of schools have double shifts, and 8 percent have triple shifts (10 percent and 2 percent of students attend these schools, respectively).
  - Teachers with a full position are required to teach 15–21 hours per week (compared with 21–24 hours per week in OECD countries).
- Composition of spending
  - Wages and salaries constitute about 90 percent of recurrent spending in primary education (compared with about 82 percent in the EU-15 and 73 percent in the EU-10).
  - In primary and secondary education, Croatia spends 22 percent on investments (compared with about 7 percent in the EU-15 and 8 percent in the EU-10).
  - Recent increases in education spending have gone mainly to overheads and to a growing pre-school subsector.
- Distributional issues
  - Decision making and financing are fragmented: local governments establish schools while teachers are hired and financed by the central government.
  - Public subsidies largely benefit higher-income households: scholarships and rewards amount going to students from the top-income quintile is almost 10 times higher than the amount going to students from the bottom quintile.
  - Education is free for about 48 percent of students; evidence suggests paid students complete with better grades and earlier.

### Education — comparative statistics (selected)
- Croatia
  - Public Expenditure on Education (percent of GDP): 4.5
  - Primary Completion Rate (percent of relevant age group): 87.3
  - Progression to Secondary School (percent): 91.4
  - Pupil-Teacher Ratio, Primary: 99.9
  - School Enrollment, Tertiary (percent; gross): 17.7
  - Primary School Enrollment, net (percent): 85.0
  - Primary Completion Rate (?) 38.7 (as listed)
  - Average PISA Mathematics Test Scores: 467
- Comparator averages
  - EU-8 average: Public Expenditure on Education 5.3; Average PISA Mathematics Test Scores 497
  - EU-10 average: Public Expenditure on Education 5.0; Average PISA Mathematics Test Scores 481
  - EU-15 average: Public Expenditure on Education 5.6; Average PISA Mathematics Test Scores 498
  - OECD average: Public Expenditure on Education 5.5; Average PISA Mathematics Test Scores 504
- Note: Data are for the latest year available except for primary completion rates (averages over 2003–04) and public education spending and progression to secondary (averages over 2001–03).

### Efficiency assessment — DEA results and system efficiency
- Education DEA results
  - Croatia ranks in the third quartile for primary education and secondary education (and in terms of PISA test scores).
  - Croatia ranks in the last quartile for tertiary education.
  - Identified inefficiencies:
    - Tertiary: low enrollment and graduation rates.
    - Secondary: low enrollment rates and relatively low PISA mathematics scores.
    - Primary: low enrollment, low completion rates, and high overhead costs due to excess number of schoolteachers.
  - System efficiency (secondary enrollment → PISA) was worse than the EU-10 average and significantly worse than the OECD average.
  - Table 9 ratio of percentile rank of efficiency scores in education to OECD average:
    - Croatia: System Efficiency 1.9; Overall Efficiency 1.3
    - EU-8 average: 1.6 and 0.5
    - EU-10 average: 1.8 and 0.5
    - EU-15 average: 1.1 and 1.2
- Health care DEA results
  - Croatia ranks in the 63rd percentile among 37 countries for public spending efficiency.
  - Ranks in the 48th percentile for total spending on health.
  - Individual outcome indicator quartiles:
    - Last quartile for standardized death rates (SDR) and incidence of tuberculosis.
    - Third quartile for HALE, child mortality rate, and infant mortality rate.
    - Second quartile for maternal mortality rates.
  - System efficiency (intermediate outputs to outcomes) is relatively low and worse than EU-15 when using ALOS-to-outcome combinations.
  - Figure 2 findings: Croatia’s efficiency scores for HALE, child mortality rate, infant mortality rate, and incidence of tuberculosis are among the lowest in the sample.

### Drivers, correlations, and caveats
- Key adverse correlations with health efficiency include:
  1. Exogenous and lifestyle factors such as alcohol consumption.
  2. Spending on collective care and administration.
  3. Spending on pharmaceuticals.
  4. Doctors’ wages.
  5. Number of doctor consultations, in-care admissions, and outpatient contacts.
  6. Length of stays in hospitals (weakly).
- Out-of-pocket payment is strongly associated with increased relative efficiency in the sample.
- Caveats:
  - Correlation analysis does not provide causality.
  - Small sample size precludes second-stage regression for causal inference.
  - DEA estimates are sensitive to sample selection, measurement error, and outliers.
  - Finite-sample downward bias in output-oriented DEA scores can be corrected by bootstrapping.

### Reform recommendations and estimated fiscal impacts — Health care
- Increase out-of-pocket spending and co-payments
  - Raising private co-financing to 7 percent of total health spending could generate budgetary savings of 0.5 percent of GDP.
  - Means-testing could grant limited exemptions to protect low-income groups.
- Phase out public supplementary insurance provided by the HZZO to reduce demand and stimulate private insurance.
- Restrain pharmaceutical demand
  - Increase share paid by consumers; reduce number of medicines on the A-list; determine subsidized drugs via periodic competitive tenders; strengthen incentives to prescribe/use generics.
- Accelerate performance-based payments (DRG)
  - Authorities intend to introduce Diagnosis Related Groups (DRG) payment method in all hospitals treating acute diseases in late 2008 (pilot efforts weakened by opt-out options).
  - Expected to reduce length of stays and generate significant medium-term budgetary savings.
- Restructure care toward outpatient services and strengthen GP incentives to treat patients rather than refer.
- Rationalize hospital network via a master plan; consider privatization within that plan.
- Increase private sector role in financing sick leave and reduce replacement rates (about 6 percent of labor force was on sick leave in 2005).

### Reform recommendations and potential savings — Education
- Rationalize teaching force
  - Increasing student-teacher ratios to OECD levels could allow reduction in teaching staff by around 11 percent at primary level and around 17 percent at secondary level.
  - Student population projection: authorities project number of students aged 7–29 to decline by another 358,000 or about 25 percent from 2005 to 2030.
- Rationalize the school network
  - Increase spending on transportation; use multi-grade teaching in small schools; eliminate triple shifts (ongoing); elimination of double shifts needs careful planning.
- Increase teaching hours toward international norms (current: 15–21 hours/week).
- Move toward performance- and per-capita based budgeting using outputs/outcomes (graduation/drop-out rates, student-teacher ratios, international test scores).
- Reduce institutional and funding rigidities by gradually raising local governments’ control and responsibility in delivering educational services.
- Greater cost recovery in pre-school education and university tuition; introduce means-testing for free textbooks, transportation, and dormitories.
- Shift resources to better meet labor market demands to improve human capital utilization.

### Concluding remarks
- Significant inefficiencies exist in government spending on health care and education in Croatia:
  - Health sector inefficiencies are mainly related to high spending rather than weak outcomes.
  - Education sector inefficiencies relate to both poor outcomes and increasing overhead costs.
- There is scope to improve health and education indicators while containing public spending.
- Targeted transfers may be needed to protect vulnerable groups when implementing efficiency-enhancing reforms.
- Menu of reform measures (selected):
  - Health Care: Increase co-payments while minimizing exemptions; reduce subsidization of pharmaceuticals; accelerate DRG implementation; restrict basic benefits covered by HZZO; shift resources to outpatient care; increase private sector role; strengthen GP incentives; rationalize hospitals.
  - Education: Rationalize teaching and non-teaching workforce and wage bill; consider greater cost recovery in tertiary education and means-testing scholarships; increase teaching hours to international norms; target free textbooks, transportation, and dormitories to the vulnerable; rationalize school network and expand multi-grade teaching; move toward per-student or performance-based budgeting; shift resources to better meet labor market demands.

### Appendix — DEA overview and limitations (selected)
- DEA is a non-parametric method estimating production possibility sets to evaluate efficiency of inputs producing outcomes.
- Typical inputs/outputs in public expenditure analysis include public/private expenditure, intermediate outputs (doctors, beds, enrollment), and outcomes (life expectancy, mortality rates, test scores).
- Stylized mechanics
  - Output-based technical efficiency score = ratio FX to EX; output-based scores emphasized because “Croatia will need to improve outcomes without increasing expenditures.”
- Statistical biases and corrections
  - Finite-sample downward bias in output-oriented DEA scores can be addressed by bootstrapping (Simar and Wilson, 2000).
  - Convergence speed = n^-2/(p+q+1); convergence slows as number of inputs (p) and outputs (q) increases.
- Decomposition
  - DEA can decompose efficiency into spending efficiency, system efficiency (intermediate outputs → outcomes), and overall efficiency.

*Source: Box 1. Performance Indicators, _wp08136 - Box 1. Performance Indicators, IMF Working Paper content provided.*

### Box 1. Performance Indicators ..........................................................................................

### Box 1. Performance Indicators

### Overview and purpose
- The analysis compares Croatia’s spending and performance in health care and education with comparator countries (EU-15, EU-10, Cyprus, Malta, and OECD) to assess relative efficiency using Data Envelopment Analysis (DEA).
- Relative efficiency is defined as the distance of a country’s observed input-output combination from an efficiency frontier (see Appendix for DEA details).
- The paper uses performance indicators divided into desired outcomes and intermediate outputs to evaluate efficiency and identify reform options.

### Performance indicators used
- Health care
  - Intermediate output indicators: density of physicians, pharmacists, and healthcare workers; number of hospital beds; number of immunization vaccines.
  - Key outcome variables: infant-, child-, and maternal mortality rates; standardized death rate from all causes per 1,000 people (WHO definition); incidences of tuberculosis; healthy average life expectancy (HALE, WHO).
- Education
  - Intermediate output indicators: primary pupil-teacher ratios, enrollment rates, rates of progression to secondary education, graduation (completion) rates.
  - Key outcome indicator: average score on PISA 2006 in mathematics (secondary education).

### Health care — structure, outcomes, and resources
- Financing and contributions
  - Health care in Croatia is mainly financed (around 90 percent) by the Croatian Health Insurance Institute (HZZO).
  - Payroll contributions are set at 15 percent of the gross wage; enterprises pay another ½ percent of wages for work-related injury insurance.
- Key health outcome and expenditure figures for Croatia (selected, as reported)
  - Total Expenditure on Health (in percent of GDP): 7.9
  - Public Expenditure on Health (in percent of GDP): 6.6
  - Healthy Life Expectancy (years): 66.6
  - Standardized Death Rates (per 100,000 people): 886.9
  - Infant Mortality Rate (per 1,000 live births): 6.0
  - Child Mortality Rate (per 1,000 live births): 7.0
  - Maternal Mortality Rate (per 100,000 live births): 10.0
  - Incidence of Tuberculosis (per 100,000 people): 40.6
- Selected health resource indicators for Croatia
  - Hospital Beds (per 1,000): 5.6
  - Physicians (per 1,000): 2.4
  - Health Worker Density Index (per 1,000): 7.7
  - Pharmacists (per 100,000): 55.8
  - Bed Occupancy Rate, Acute Care Hospitals (percent): 88.1
  - In-patient Care Admissions (per 100): 16.6
  - Average Length of Stay (all hospitals) (days): 10.3
  - Immunization, Measles (percent of children ages 12-23 months): 96.0
- Private funding and per-capita spending
  - Croatia spends 831 PPP dollars per capita on health, of which 143, or 16 percent, comes from private sources.
- Sustainability and inefficiency concerns
  - Stock of health sector arrears at end-2006: 1.1 percent of GDP.
  - Croatia spends about 8 percent of its GDP on health care (Table 1 value reported as 7.9 percent).
  - About 84 percent of health care spending comes from public sources.
  - Co-payments share in total health spending: less than 1 percent (compared with 7–33 percent in Western European countries).
  - Coverage features: 600,000 people have supplementary insurance from HZZO that pays co-payments for services; around 1,900 drugs on the A list are fully paid by HZZO and 300 on the B list are partially paid.
  - Government introduced a flat administrative fee of 10 kuna per person (cap 30 kuna per month) in 2005; the government decided to abolish this fee in 2008.
- Demand and supply drivers of high spending
  - Old-age dependency ratio (population aged 65 and older relative to population aged 17–64) was 26 percent in 2006 and is projected to increase to 48 percent in 2051.
  - Low co-payments plus widespread exemptions from contributions have boosted demand for health services; only around 35 percent of the population pays contribution.
  - Payment incentives: capacity- and input-based hospital payments encourage hospitals to keep beds full and extend length of stay.
    - Average Length of Stay (ALOS) in Croatia in 2005: 10.3 days.
    - ALOS comparisons: EU-10 average 8.6 days; EU-15 average 8.4 days.
  - Primary-care payment design (capitation) encourages primary-care physicians to sign up many patients and refer to specialists; 70 percent of patients at primary health care level are referred to hospitals (experts believe this could be reduced to 30 percent).
  - Limited competition among providers: of 66 hospitals, only 3 were privately owned in 2006; only 3 percent of medical doctors work in the private sector.
- Projections and risks
  - Without reforms, health care expenditures will increase significantly.
  - The Pre-accession Economic Program envisages an increase of 4 percentage points of GDP in public health spending from 2005 to 2050.
  - Population aging and new medical technology costs could increase spending further.

### Education — spending and outcomes (selected)
- Structure
  - Croatia’s education system is mainly financed and operated by the public sector.
  - Government reforms underway since 2005: Strategic Development Framework 2006–13 and Education Sector Development Plan (ESDP) 2005–10.
- Spending and outcome highlights
  - Total spending on education (2005): around 5.6 percent of GDP.
  - Public education spending (2005): about 4.8 percent of GDP.
  - Croatia’s private education spending (around ¾ percent of GDP) is higher than the EU-10 and EU-15 averages (values reported for averages but not restated here).
  - Despite spending in line with comparators, educational output and outcome levels are lower than EU-10 and EU-15 peers.

### Identified reform areas to increase efficiency of government spending on health and education
- Contain demand for services by introducing or increasing existing fees for users of health and education services.
- Reform finance mechanisms for spending (including provider payment systems and pharmaceutical payment arrangements).
- Introduce greater competition in the provision of services (expand private provision/ownership and participation).
- Improve administration of spending on these services (budgeting, control of arrears, incentives for managers).
- Better target health care and education subsidies to those most in need.

*Source: Box 1. Performance Indicators, _wp08136 - Box 1. Performance Indicators, IMF Working Paper content provided.*

### 0.4 percent of GDP),

### _wp08136 - 0.4 percent of GDP),

### Education — key findings and outcomes
- Croatia’s school enrollment and completion rates are lower than those in comparator countries.
- Tertiary education:
  - Gross enrollment was about 46 percent in 2006, compared to about 53 percent in the EU-10.
  - Only one third of the students at the tertiary level reportedly complete their programs, with an average completion rate of 6.7 years in four-year programs.
- In the 2006 PISA standardized test in mathematics, out of 57 countries, Croatia ranked 36th; only Bulgaria and Romania in the EU-10 scored worse than Croatia.
- Student-teacher ratios in primary and secondary schools have been falling and are lower than those in comparator countries, reflecting:
  - Number of students fell at all levels except tertiary education from 1990 to 2005.
  - Number of full-time teachers increased at all levels except primary education (stable).
- School infrastructure and teaching hours:
  - About 65 percent of schools have double shifts, and 8 percent of schools have triple shifts (although only 10 percent and 2 percent of students, respectively, attend these schools).
  - Teachers with a full position are required to teach 15–21 hours per week, compared with 21–24 hours per week in OECD countries.
- Composition of education spending:
  - Wages and salaries constitute about 90 percent of recurrent spending in primary education (compared with about 82 percent in the EU-15 and 73 percent in the EU-10).
  - In primary and secondary education, Croatia spends 22 percent on investments (compared with about 7 percent in the EU-15 and 8 percent in the EU-10), leaving a smaller share for non-wage recurrent expenditures.
  - Share of investments in tertiary education in Croatia is smaller than in peer countries.
  - Recent increases in education spending have gone mainly to overheads and to a growing pre-school subsector.
- Decision making and financing are fragmented: local governments decide on establishing schools while teachers are hired and financed by the central government, contributing to excess spending.
- Public subsidies largely benefit higher-income households:
  - Scholarships and rewards: amount going to students from the top-income quintile is almost 10 times higher than the amount going to students from the bottom quintile.
  - Students from top-income quintile also benefit from free books, dormitories, and transportation.

### Education — comparative statistics (selected figures from Table 3)
- Croatia:
  - Public Expenditure on Education (percent of GDP): 4.5
  - Primary Completion Rate (percent of relevant age group): 87.3
  - Progression to Secondary School (percent): 91.4
  - Pupil-Teacher Ratio, Primary: 99.9
  - School Enrollment, Tertiary (percent; gross): 17.7
  - Primary School Enrollment, net (percent): 85.0
  - Primary Completion Rate (?) 38.7 (as listed)
  - Average PISA Mathematics Test Scores: 467
- EU-8 average: Public Expenditure on Education 5.3; Average PISA Mathematics Test Scores 497
- EU-10 average: Public Expenditure on Education 5.0; Average PISA Mathematics Test Scores 481
- EU-15 average: Public Expenditure on Education 5.6; Average PISA Mathematics Test Scores 498
- OECD average: Public Expenditure on Education 5.5; Average PISA Mathematics Test Scores 504
- Note: Data are for the latest year available except for primary completion rates (averages over 2003–04) and data on public education spending and progression to secondary (averages over 2001–03).

### Education — efficiency assessment
- DEA results:
  - Croatia ranks in the third quartile for primary education and secondary education (and in terms of PISA test scores).
  - Croatia ranks in the last quartile for tertiary education.
  - Inefficiencies at tertiary level: low enrollment and graduation rates.
  - Inefficiencies at secondary level: low enrollment rates and relatively low PISA mathematics scores.
  - Inefficiencies at primary level: low enrollment, low completion rates, and high overhead costs due to excess number of schoolteachers.
- System efficiency:
  - Croatia’s system efficiency from secondary enrollment to PISA scores was worse than the EU-10 average and significantly worse than the OECD average.
  - Table 9 (ratio of percentile rank of efficiency scores in education to OECD average):
    - Croatia: System Efficiency 1.9; Overall Efficiency 1.3
    - EU-8 average: 1.6 and 0.5
    - EU-10 average: 1.8 and 0.5
    - EU-15 average: 1.1 and 1.2
- Correlations (Table 10) — positive relationships with higher overall efficiency include:
  - Share of current expenditure in total education
  - Classroom size
  - Parent’s education
  - School quality and autonomy indicators (student admissions prerequisites; student discipline; principal responsibility for hiring)
- Implication: More spending, especially capital spending, will not automatically improve education outcomes; costs of an excess number of teachers will rise as wages grow.

### Education — reform recommendations and potential savings
- Rationalize teaching force:
  - Increasing student-teacher ratios to OECD levels could allow reduction in teaching staff by around 11 percent at primary level and around 17 percent at secondary level.
  - Authorities project number of students aged 7–29 to decline by another 358,000 or about 25 percent from 2005 to 2030 — implying potential savings if staffing and spending are reduced accordingly.
- Rationalize the school network:
  - Increase spending on transportation, use multi-grade teaching in small schools.
  - Eliminate triple shifts (ongoing); elimination of double shifts needs careful planning.
- Increase teaching hours toward international norms (teachers currently 15–21 hours/week).
- Move toward performance- and per-capita based budgeting:
  - Use number of students and outputs/outcomes (graduation/drop-out rates, student-teacher ratios, international test scores) in budgeting.
- Reduce institutional and funding rigidities:
  - Gradually raise local governments’ control and responsibility in delivering educational services, in line with capacity.
- Greater cost recovery in pre-school education and university tuition:
  - Education is free for about 48 per cent of students; evidence suggests paid students complete with better grades and earlier.
  - Introduce means-testing for free textbooks, transportation, and dormitories to better target vulnerable groups.
- Shift resources to better meet demands in the labor market to improve human capital utilization.

### Health care — key findings and outcomes
- DEA results indicate significant inefficiencies in Croatia’s public health spending, with room to rationalize spending without sacrificing, and potentially improving, health outcomes.
- Efficiency rankings and outcomes:
  - Croatia ranks in the 63rd percentile among 37 countries for public spending efficiency.
  - Ranks in the 48th percentile for total spending on health (reflecting low private health expenditures).
  - Individual outcome indicators:
    - Last quartile for standardized death rates (SDR) and incidence of tuberculosis.
    - Third quartile for HALE, child mortality rate, and infant mortality rate.
    - Second quartile for maternal mortality rates.
- System efficiency (intermediate outputs to outcomes) is relatively low in Croatia and worse than EU-15 countries when using ALOS-to-outcome combinations.
- Figure 2 findings:
  - Croatia’s efficiency scores for HALE, child mortality rate, infant mortality rate, and incidence of tuberculosis are among the lowest in the sample.

### Health care — drivers and correlations (Table 6)
- Key adverse correlations with efficiency include:
  1. Exogenous and lifestyle factors such as alcohol consumption.
  2. Spending on collective care and administration.
  3. Spending on pharmaceuticals.
  4. Doctors’ wages.
  5. Number of doctor consultations, in-care admissions, and outpatient contacts.
  6. Length of stays in hospitals (weakly).
- Out-of-pocket payment is strongly associated with increased relative efficiency in the sample.
- Interpretation: Inefficiencies in health spending in Croatia are partly related to high pharmaceutical spending, long hospital stays, low levels of out-of-pocket spending, and low private participation.

### Health care — reform recommendations and estimated fiscal impacts
- Increase out-of-pocket spending and co-payments:
  - If private co-financing were raised to 7 percent of total health spending (one of the lowest co-payments-to-total-health-spending ratios of Western European countries), through increases in co-payment rates and/or eliminating exemptions, this could generate budgetary savings of 0.5 percent of GDP.
  - Increasing the share of the private sector in financing sick leave and reducing replacement rates would also curb demand and public spending.
  - Note: Co-payments could curtail access for lower-income families; means-testing could grant limited exemptions (e.g., pensioners are exempt, but some of them may not need to be subsidized).
- Phase out public supplementary insurance provided by the HZZO to reduce demand and stimulate private insurance; equity impact likely limited because essential services are covered by basic insurance.
- Restrain pharmaceutical demand:
  - Increase share paid by consumers and expose producers to competition.
  - Reduce number of medicines on the A-list; determine specific drugs to subsidize by periodic competitive tenders.
  - Strengthen incentives to prescribe/use generic substitutes.
- Accelerate performance-based payments (DRG):
  - Move from input- or capacity-based payments to performance-based payments.
  - Authorities intend to introduce Diagnosis Related Groups (DRG) payment method in all hospitals treating acute diseases in late 2008 (pilot efforts weakened by opt-out options).
  - Expected effects: reduce length of stays and generate significant medium-term budgetary savings.
- Restructure care toward outpatient services:
  - Reforms to payment system to incentivize general practitioners to treat patients rather than refer; increase co-payments for inpatient care.
- Administrative measures:
  - Require general practitioners to explain referrals to reduce specialist referrals.
- Rationalize hospital network:
  - Develop master plan assessing population needs by service and location; identify facilities to close, reorient (e.g., long-term care, private practice), upgrade infrastructure/equipment in remaining facilities.
  - Consider privatization program for hospitals within master plan.
- Improve management and competition:
  - Give hospitals more independence, impose hard budget constraints, bring professional management expertise, and expose hospitals to competition.
- Preventive measures:
  - Step up disease prevention efforts (beyond immunizations).
  - Lifestyle issues: Croatia’s share of overweight people is among highest in Europe; share of obese people in Croatia is almost double the average of the EU-15.
  - Smoking-related death incidents and alcohol consumption are high, contributing to circulatory and heart disease mortality.

### Health care — potential systemic measures and caveats
- Increase private sector role in financing sick leave and reduce replacement rates to curb demand (about 6 percent of labor force was on sick leave in 2005; sick leave used to deal with excess employment).
- Restricting basic benefit package provided by HZZO would enhance demand containment and stimulate private participation.
- Caution: Correlation analysis does not provide causality; small sample size precludes second-stage regression for causal inference.
- Sequencing and political economy of reforms are beyond the scope of the chapter.

### Concluding remarks
- The paper finds significant inefficiencies in government spending on health care and education in Croatia:
  - Health sector inefficiencies are mainly related to high spending rather than weak outcomes.
  - Education sector inefficiencies relate to both poor outcomes and increasing overhead costs.
- There is scope to improve health and education indicators while containing public spending.
- Targeted transfers may be needed to protect vulnerable groups when implementing efficiency-enhancing reforms.
- Menu of reform measures summarized (Table 11) — selected items:
  - Health Care:
    - Increase co-payments while minimizing exemptions.
    - Further reduce subsidization of pharmaceuticals.
    - Accelerate DRG implementation.
    - Restrict basic benefits covered by HZZO.
    - Shift resources to outpatient care.
    - Increase private sector role; strengthen GP incentives; rationalize hospitals.
  - Education Sector:
    - Rationalize teaching and non-teaching workforce and wage bill.
    - Consider greater cost recovery in tertiary education and means-testing scholarships.
    - Increase teaching hours to international norms.
    - Target free textbooks, transportation, and dormitories to the vulnerable.
    - Rationalize school network and expand multi-grade teaching.
    - Move toward per-student or performance-based budgeting.
    - Shift resources to better meet labor market demands.

*Italicized source: IMF staff paper (text and tables drawn from the supplied PDF content).*

### References

### References

### Key cited works
- Afonso, A., L. Schuknecht, and V. Tanzi, 2006, “Public Sector Efficiency: Evidence for New EU Member States and Emerging Markets,” European Central Bank Working Paper Series No. 581.  
- Afonso, A. and M. St. Aubyn, 2004, “Non-Parametric Approaches to Education and Health: Expenditure Efficiency in OECD Countries.”  
- Charnes, A., W. Cooper, and E. Rhodes, 1978, “Measuring Efficiency of Decision-Making Units,” European Journal of Operational Research, Vol. 3, pp. 429-44.  
- Coelli, Tim, Mathieu Lefebvre, and Pierre Pestieau (2007), “Measurement of Social Protection Performance in the European Union.”  
- Cucilić, Judita, Michael Faulend, and Vedran Šošic (2004), “Fiscal Aspects of Accession: Can We Enter the European Union With a Budgetary Deficit?” in Croatian Accession to the European Union ed. by Katarina Ott, Chapter 3 in Vol. 2, pp. 49-77.  
- Davies, M., M. Verhoeven and V. Gunnarsson, 2006, “Wage Bill Inflexibility and Performance Budgeting in Low-Income Countries.”  
- Ederer, Peer, Philipp Schuler, and Stephan Willms, 2007, “The European Human Capital Index: The Challenge of Central and Eastern Europe.” Lisbon Council Policy Brief.  
- Farrell, M., 1957, “The Measurement of Productive Efficiency,” Journal of the Royal Statistical Society, Series A, Vol. 120, No.3. pp. 253-90.  
- Funck, Bernard (2003), Expenditure Policies Toward EU Accession, World Bank Technical Paper no. 533.  
- Gupta, S., and M. Verhoeven, 2001, “The Efficiency of Government Expenditure: Experiences from Africa” Journal of Policy Modeling, No. 23, pp. 433-67.  
- Herrera, S., and G. Pang, 2005, “Efficiency of Public Spending in Developing Countries: an Efficiency Frontier Approach,” World Bank Policy Research Working Paper 3645.  
- Mattina, Todd, and Victoria Gunnarsson, 2007, “Budget Rigidity and Expenditure Efficiency in Slovenia,” IMF Working Paper 07/131.  
- Mihaljek, Dubravko (2007), “Health Care Policy and Reform in Croatia: How To See the Forest for the Trees,” in Croatian Accession to the European Union ed. by Katarina Ott, Chapter 11 in Vol. 4, pp. 277-320.  
- Mossialos, Elias, Anna Dixon, Josep Figueras, and Joe Kutzin (2002), Funding Health Care: Options for Europe, European Observatory on Health care Systems Series.  
- Simar, L., and P. Wilson, 2007, “Estimation and Inference in Two-stage, Semi-parametric Models of Production Processes,” Journal of Econometrics, 136, pp. 31-64.  
- Verhoeven, Marijn, Victoria Gunnarsson, and Sergio Lugaresi, 2007 “The Health Sector in the Slovak Republic: Efficiency and Reform,” IMF Working Paper 07/226.  
- Verhoeven, Marijn, Victoria Gunnarsson, and Stéphane Carcillo, 2007 “Education and Health in G7 Countries: Achieving Better Outcomes with Less Spending,” IMF Working Paper 07/263.  
- World Bank, 2007, “Croatia: Restructuring Public Finance to Sustain Growth and Improve Public Services – A Public Finance Review,” World Bank Report No. 37321-HR.  
- Zhu, J., 2003, Quantitative Models for Performance Evaluation and Benchmarking (New York: Springer Science+Business Media Inc.).  

### Appendix: Data Envelopment Analysis (DEA) — overview and application
- DEA is described as a non-parametric method for estimating production possibility sets to evaluate efficiency in the use of inputs in producing outcomes for a sample of production units.  
- Typical application in public expenditure analysis:
  - Outcomes (examples): life expectancy, infant mortality rates, youth literacy rates, test scores, number of roads, telephone lines.  
  - Inputs (examples): public and private expenditure on health, education, infrastructure; intermediate outputs and resources such as number of doctors, hospital beds, enrollment rates, student-teacher ratio.  
  - Production units: often countries or sub-national regions.  

### Stylized example and efficiency scoring
- Figure A.1 (stylized single-input single-output example) mechanics:
  - Efficient frontier connects countries A to D; countries E and G are interior (inefficient).  
  - Convexity allows an inefficient country (E) to be compared to a hypothetical frontier position (Z) via linear combination of efficient units (A and B).  
  - Input-based technical efficiency score = ratio YZ to YE (bounded between zero and one).  
  - Output-based technical efficiency score = ratio FX to EX (reflects output improvements for given inputs).  
- This chapter focuses on output-based efficiency scores because “Croatia will need to improve outcomes without increasing expenditures.”

### Caveats and limitations of DEA
- Key caveats:
  - Results are highly sensitive to sample selection and measurement error; outliers exert large effects on efficiency scores and frontier shape. Proper sample selection is crucial.  
  - Difficult-to-quantify spending attributes (e.g., quality of spending) are not easily incorporated.  
  - Outcome indicators used may not be targeted by policy makers.  
  - Large differences in private health or education spending across countries can bias public spending efficiency scores.  
  - Factors beyond policy makers’ control (example: high incidence of AIDS) can reduce measured efficiency of health spending in some countries.  

### Statistical biases and corrections
- Simple DEA estimation produces biased estimates of efficiency scores due to one-sided error: the best-practice frontier can move outward with added efficient observations but not inward, creating a finite-sample downward bias in output-oriented efficiency scores (i.e., efficiency is overestimated unless corrected).  
- Where appropriate, corrections for finite-sample bias are applied through bootstrapping, as suggested by Simar and Wilson (2000).  
- Convergence speed of estimated efficiency scores to unbiased true values when expanding sample:
  - Convergence speed = n^-2/(p+q+1), where p is number of inputs and q is number of production items.  
  - Example: in the 1 input / 1 product case, convergence speed = n^-2/3.  
  - This is faster than standard parametric regression convergence of n^-1/2, implying reasonable DEA estimates may require fewer observations.  
  - However, convergence speed declines exponentially as the number of inputs and production items increases; at two inputs and two production items, convergence speed is markedly slower than for parametric regression.  
  - Implication: adding inputs and production items significantly increases the sample size needed to draw conclusions on efficiency.

### Decomposition and disaggregation of DEA results
- DEA results can be disaggregated to identify stages where inefficiencies arise:
  - Spending efficiency: overall measure from spending to outcomes.  
  - System efficiency: measure from intermediate outputs to outcomes (e.g., intermediate outputs = hospital beds, immunizations, physicians, health care workers, pharmacists per capita).  
  - Procedure illustrated for health care:
    - Step 1: Assess cost efficiency using health care spending and intermediate output indicators.  
    - Step 2: Calculate efficiency scores using intermediate output index as input and associated outcomes (infant, child, and maternal mortality rates; HALE; standardized death rates; incidence of tuberculosis).  
    - Step 3: Average system efficiency rankings, express as ratio of average OECD ranking, and compare with similar ratios for spending efficiency.  

### Components illustrated in Figure A.2 (Health expenditures → Resources → Outcomes)
- Health Expenditure:
  - Public health expenditure  
  - Private health expenditure  
- Real Health Resources (examples):
  - Hospital beds  
  - Physicians/health workers  
  - Immunizations  
  - Doctors’ consultations  
  - In-patient admissions  
  - Lengths of stay  
  - Bed occupancy rate  
- Health Outcomes:
  - Health adjusted life expectancy  
  - Standardized death rate  
  - Infant mortality rate  
  - Child mortality rate  
  - Maternal mortality rate  
  - Incidence of tuberculosis  
- Efficiency concepts shown:
  - Cost effectiveness  
  - System efficiency  
  - Overall efficiency  

*Source: _wp08136 - References (PDF).*

---


_Source: https://www.imf.org/-/media/websites/imf/imported-full-text-pdf/external/pubs/ft/wp/2008/_wp08136.pdf_