Report No. 52536-PL Poland Public Expenditure Review (In Two Volumes) Volume II: Background Papers March 8, 2010 Poverty Reduction and Economic Management Unit Europe and Central Asia Region Document of the World Bank TABLE OF CONTENTS PENSIONS ............ ,............ ...,...,..... ,......,..,,..,............,.............,.....,..,.....,........... ..... . 1 ........................... 11. RECENT DEVELOPMENTS IN POLAND ON PENSION REFORM ........................ 3 111. SUSTAINABILITY AND BENEFIT LEVELS FISCAL Iv. RECOMMENDATIONS ........................................................................................................................................ 7 AGRICULTURAL PENSIONS .................. ........, .,...,...............,...,... ..,....................11 ............ ........ INTRODUCTION ........................................................................................................................................................ 1 1 I. BACKGROUND ........ ......... 11. KEYRECOMMENDATIONS REFORM FOR DIR 111. ESTIMATTNG AGRICULTURAL INCOME ......... Iv. SHORT- AND MEDIUM-TERM EFFECTS UNIFICATION ... FISCAL OF ..................................... 18 v. ADDRESSING ECONOMIC SOCIAL IMPLI AND APPENDIX ................. ...................................................................................................................................... 21 1 EDUCATION , ......,,.,.................................................. .............,...........,,...,.,..............,.....,...,......,.............. 23 INTRODUCTION .................................................. .................................. 23 I. POLAND'S RECENT HISTORY REFORM OF 11. OVERVIEW AND OUTCOMES OF THE EDUCATION SYSTEM ............................................................................. 25 111. FINANCTNG EDUCATION SYSTEM ............. THE ...................... ...................... 29 1v. SUMMARY AND POLICY OPTIONS FOR THE UNFINISHED AGENDA.. ANNEX 1: DIVISION OF RESPONSIBILITIES IN THE POLISH EDUCATIO SECONDARY ........................................................ LEVEL .......................................................................... 51 ANNEX SCHOOLS AND STUDENTS IN THE POLISH EDU 2: YSTEM, 2 0 0 7 / 0 8 A . ... 52 ANNEX CALCULATING SAVINGS FROM MINIMUM 3: INDICATIVE CLASS SIZE STANDARDS IN POLAND ..................52 HEALTH .................................................................................................................................................... 55 OVERVIEW ............................................................................................................................................................... 55 INTRODUCTION .......... .......................................... .........56 1. HEALTH STATUS 11. FINANCIAL PROT ............................................................. 59 111. EXPENDITURES HEALTH, ON 2005-2009 ......... IV. RECENTHEALTHSECTORREFOR V. MANAGING HEALTH CARE EXPENDITURES IMPROVE EFFICIENCY CONTROL COSTS ........................ TO AND 65 VI. QUALITY~ACCREDITATION ............. .................................. VII. HEALTH EXPENDITURES, OF S E USE ................................................ VI11 ACCOMPLISHMENTS CHALLENGES ......................................................................................................... AND 86 IX. POLICY FRAMEWORK GOING FORWARD ........ ...................89 STATISTICAL ANNEX............................................................................................................................................... 91 SOCIAL ASSISTANCE CASH TRANSFERS...... ............................................................................... 103 INTRODUCTION AND STRUCTURE OF THE CHAPTER ............................................................................................... 103 I. TRANSFERSPOLAND ................................... THE PARAMETERS AND DELIVERY SOCIAL ASSISTANCE OF IN 104 11. SPENDING SOCIAL ON r~ ASSISTANCE POLAND ACROSS REGION AND THE ................................ ,108 111. COVERAGE, TARGETING PERFORMANCE OF SOCIAL ASSISTANCE: AND IMPORTANCE N HOUSEHOLD .................................................................................. ........ Iv. DECENTRALIZATIONSOCIAL ASSISTANCE AND ACROSS ....... POLAND'S REGIONS V. ................... CONCLUSIONS .... ............................ 125 LABOR MARKET PROGRAMS ......................................................................................................... 131 INTRODUCT~ON.................................................. ..................................................................................... 131 I. RECENTLABOR MARKET DEVELOPMENTS ........................................................................................ 132 11. EVOLUTION LABOR OF MARKET EXPENDITURES ......................................................................................... 135 111. LABORMARKET EXPENDITURES N POLAND COMPARATIVE IN PERSPECTIVE ............................................. 140 Iv. LABOR MARKET PROGRAMS AND THEIR EFFICIENCY .................................................................................. 142 V. LABORMARKET POLICIES IN RESPONSE TO TH ...................................................... 155 VI. CONCLUSIONS AND POLICY IMPLICATIONS ....... ..... MEDIUM-TERM BUDGETING AND PERFORMANCE BUDGETING ....................................... 165 INTRODUCTION ......................................................................... I. MEDIUM- ETING ........................................................................................... 11. PERFORMANCE BUDGETING ...................... .................................................. 111. FUTURE DIRECTIONS MEDIUM-TERM FOR RMANCE BUDGETING ........................... I79 IV. ACTION PLAN ................................................................................................................................. TABLES Table 1 : Main Parameters of the Polish Pension System ................................................................................. .....2 Table 2: Eligibility Conditions for and Benefits of Disability Pensions in Poland under the First-Pillar Eamings- Related Scheme ................................................................. ....................................... ......................................... 6 Table 3: Main Characteristics of ZUS and KRUS, 2008 ............................................................................................. 13 Table 4: KRUS - Selected Financial Indicators ..... ....................................................................................... 13 Table 5 : Farms by Area Groups of Agricultural La ............ ........................................................... 17 Table 6: Shares of Collected PIT and CIT Staying in the Budgets of Territorial Self-Governments........ Table 7: Structure of Tertiary Students in Poland and OECD Countries by Field, 2006 .......................... Table 8: Public Spending on Education in Poland in Comparison with Emerging Economies (2007 or latest available) ...... ................................................................... ......30 Table 9: Funding Level of the Polish Education System in 2007 (As a percent of GDP and in PPP USD) ................ 1 3 Table I O : Total Spending on Education by Tier in 2004-2007 ........................ ............... Table 1 1 : Poland Per-Student Expenditures (billions of PLN, constant 2007 prices) .................. Table 12: Local Expenditures (without preschools) as a Percentage of the Education Subvention, Considering Wealth of Municipalities, 2006 ...................................... ................................. ........................ 37 Table 13: Structure of Local Expenditures on Primary and Secondary Education ......................................... Table 14: Fiscal Equalization Effects in Polish Municipalities (2006)a ......................................................... Table 15: Ratio of Annual Teachers' Salary to GDP and Salaries in PPP Converted USD in Poland and OECD Countries ........................ .......................................................................................................................... Table 16: Fiscal Equalization Effects in Polish Municipalities, 2006 (Data in PLN) ............................. Table 17: Health Status Indicators, Poland and Comparator Countries (2007 or latest available year) ...................... 58 Table 18: Hospitals and Medical Personnel, Poland and Comparator Countries (2007 or latest available year) ........66 Table 19: Monthly Per Capita Spending on Health (in 2005 Zloty) ................................................................ * 75 Table 20: Distribution of Health Spending (in 2005 Zloty) ............................................................................ 76 Table 21: Mean Monthly Per Capita Expenditures on Health in 2007 (in 2007 Zloty) ........... ............................ .76 Table 22: Distribution of Responses for the Subsample Reporting Income versus the Entire Sample ( Table 23: Central and Local Government Responsibility for Delivering Social Assistance ................. Table 24: Eligibility for Poland?s Main Social Assistance Programs ..................... Table 25: Total Spending on Social Protection in Poland and Central and Easte of GDP) ..................................................................................................................................................................... 108 Table 26: Spending on Social Protection Programs in Poland and Central and Eastern European EU States by Function (Percentage of GDP in 2007) ...... ....... 109 Table 27: Trends in Spending on Social Protection Programs in Poland, 2002 - 2007 (Percentage of GDP) .........110 Table 28: Trends in Spending on Social Assistance in the Central and Eastern European New Member States, 2002 - 2007, (Percentage of GDP)....................... ................................ Table 29: Trends in Spending on Means n 2000 - 2006, (P Social Protection) ......................................................................................................................................... Table 30: Trends in Spending on Means Tested Family and Child Benefits 2000- Spending on Family and Child Benefits) .................................................................................................................. 11 1 Table 3 1: Social Protection Coverage, ( percent of Households in Each Group who Report Receiving Benefits) ... 112 Table 32: Social Protection Targeting, ( percent of Total Transfers Reported Going to Households in Each Group) ................................................................................................................................................................................... 1 14 Table 33: Distribution of Households by Number of Dependent Children and Consumption Decile ..,.. 1 15 Table 34: Social Protection Benefits as a Share of Household Income in 2007 ........................................................ 116 Table 35: The Coverage Gap Across Poland?s Voyvods (Percentage of the population eligible for social assistance but not receiving benefits) ................... ................... ,123 Table 36: Targeting Efficiency of Family Allowances and Minimum Income Assistance Across Voyvods, (Coady- Grosh-Hoddinott Ratio*) ........................................................................................................................................... 124 Table 37: Unemployment Incidence and Profile by Socio-demographic Characteristics, 2008 (4th quarter)...........136 FIGURES Figure 1 : Pension Contribution Rates in New EU Countries ........................................................................................ 2 Figure 2: Public Pension Scheme fiscal Balance as Percent of GDP ..... Figure 3: Projected Benefit Levels as Compared to Average Wage .......... Figure 4: Number of Insured Persons and Beneficiaries in 1991-2008....................................................................... 12 Figure 5: Phasing-in and topping-up of Direct Payments, (in percent ofEU15 payment level) Figure 6: Change in Enrollment by Education Tier, 1990-2007 ............ Figure 7: Disciplines of Study among Tertiary Students in Poland, 2007 ................................................................... 28 Figure 8: Local Government Expenditures on Preschools (1999-2008) .............. Figure 9: Expenditure on Preprimary Educational Institutions as a Percentage of GDP, Combined Public and Private Sources of Funds (2005)a ............................................................................................. .....33 Figure 10: Expenditures on Primary and Secondary Educational Institutions as a Percentage of GDP, by Source of .............................................................................................................. 35 constant 2007 prices) ................................... 36 Figure 13: Total Spending Compared to Wage Bill on Primary and Secondary Education, 2002-07 (billions of PLN, constant 2007 prices) .................................................................. Figure 14: Total Enrollment and School Number in Poland (primary and secondary education) Figure 15: Dynamics of Enrollment and Teacher Employment, 2000-07 .................................................................. 41 Figure 16: Per Capita Spending on Education in Polish Households as a Percent ...................................... Figure 17: Expenditure on Tertiary Education as a Percentage of GDP, by Source of Fund (2005)a ......................... 44 Figure 18: Higher-education Funding in Poland (2004-07), (PLN billion, constant 2007 prices) .............................. 45 Figure 19: Share of 19-26-years-olds Not Attending any School, by Quintile of Household Monthly Consumption Per Capita ..................... .......................................................................................................................................... 46 Figure 20: Life Expecta at Birth in Poland and Neighboring Countries .. Figure 21: Self Reported Inability to Use Health Services Due to Financial Difficulties .... Figure 22: Health Expenditures ( percent of GDP, 2007) ................ Figure 23: Public Expenditures on Health ( percent, by financing a Figure 24: Health Expenditures ( percent of GDP) ,..,,.. Figure 25: Composition of Health Expenditures (percent of Total) ............................................................................ 63 Figure 26: Number of Hospitals and Le Figure 27: Hospitals and Hospital Beds Figure 28: By Income Quintile -- Use of NHF-Financed Services During the Past 6 Months, Survey of the Public Opinion Research Center 2007 ..... 78 Figure 29: By Income Quintile -- Use of Private Services Not Financed by NHF During the Past 6 Months, Survey of the Public Opinion Research Center 2007 .............................................................................................................. 79 Figure 30: By Residence -- Use of NHF-Financed Services During the Past 6 Months, Survey of the Public Opinion Research Center 2007 ................ ............................................... ........................................................................ 79 Figure 31: By Age -- Use of NHF-Financed Services During Past 6 Months Months, Survey of the Public Opinion Research Center 2007 .............................................. ........................................................ 80 Figure 32: By Quintile- Use of NHF Services During the L 06 for Outpatient and Last 12 Months for Inpatient Care, Household Budget Survey ............................................................................................................. 8 1 Figure 33: By Quintile -- Use of Private Services Not Financed by NHF during the Last 3 Months of 2006 for Outpatient Care, Household Budget Survey ................................................................................................................ 82 Figure 34: Expenditures on Pharma Figure 35: Use of Non NHF Providers Figure 36: Reasons for Using Non NHF Providers ... ........................ .................. Figure 37: Household Satisfaction wi Figure 38: Expenditures Connected to Hospital Stays in 2006 ......................... ............................. Figure 39: Household Perception o f t Figure 40: A Steep Increase in Unem Figure 4 1 : The Increase in Unemplo Figure 42: Lay-Offs Led to a Surge in Unemployment Benefit Claims .................................................................... 133 Figure 43: Job Losers, Who are Eligible to Unemployment Benefit, Replaced New Labor Market Entrants, Who are Not Eligible, as a Dominant Group Among the Unemployed .,.. Figure 44: Most of the Unemployed are Jobless for Less than 6 Figure 45: Labor Market Expenditures are Countercyclical and Figure 46: Expenditure Per Unemployed Increased with the Fall in the Unemployment Rate ,. Figure 47: The Increase in Labor Market Expenditure was Driven by the Growing Spending Programs, Whose Share in Total Expenditure Markedly Increased .......................................................................... 139 Figure 48: Poland Spends on Labor Market Programs a Higher Percentage of GDP than Most Middle-Income OECD Countries but Less than High-Income Countries ........................................................................................... 140 Figure 49: Expenditures on Active Programs are Modest in Poland ......................................................................... 141 Figure 50: Unemployment Benefit is Received Mainly by the Poor .............................................. Figure 5 1 : Labor Fund Spends Most of its Budget on Business Support Program, but the Major Enrolled in Training and Apprenticeships Programs................................................................................................. 148 Figure 52: Officially Reported Job Placement Rates are High, but They are a Weak Measure of Programs? Effects ................... ................ .............. ............ 149 Figure 53: Substantial Variation in Cost-Effectiveness Across Programs ................................................................ 150 Figure 54: Program Costs Vary Across regi0ns.R .............. ............................................................................ 150 Figure 5 5 : Cost of Most Programs Increased Over Last Year ...................................... ............................ 151 Figure 56: Most Active Labor Market rams in Poland Improve the Employm s of the Unemployed ............................................................... ....... .......... .................................................... 153 Figure 57: Targeting of Active Labor Market Programs by Education and UrbadRural Location .......................... 154 BOXES Box 1 : The Single Area Payment Scheme (SAPS) Provides for Direct Payments Per Hectare of Agricultural Land. 16 Box 2: Estimating Agricultural Income........................................................... .................................................... 18 Box 3: The Evolution of Hospital Management and Financing in Poland .................................................................. 70 Box 4: Labor Fund deficit Under Different Unemployment Scenarios .................. ...... Box 5 : Unemployment Benefit System ................................ .................................................... Box 6: Labor Market Programs: Their Potential and Limitations in Addressing Unemployment during Economic Crisis ......... ................................................................................................. ........ Box 7 : Active Labor Market Programs in Poland: Key Features .............................................................. Box 8: Special Programs to Mitigate the Effects of the Economic Crisis on Employment ....................................... 156 ACKNOWLEDGMENTS This report was prepared by a team from the World Bank for the Government of Poland from April 2009 to December 2009. Findings of the report were discussed during two workshops at the Ministry of Finance on July 8, 2009 and November 13, 2009. The report team benefited greatly from collaborative work with the Government of Poland during three missions (April, June and November 2009). The team is very grateful for the comments, feedback, and guidance received from the Government during the preparation of the report. Government counterparts for this reports included officials from the Ministry of Finance, Chancellery of the Prime Minister, Ministry of Labor and Social Policy, Ministry of Health, Ministry of National Education, Ministry of Science and Higher Education, Ministry of Interior and Administration, Social Insurance Fund and National Health Fund. The World Bank team included Leszek Kasek, Ewa Korczyc, and Kaspar Richter (public financial management and synthesis report); Ufuk Guven, Anita Schwarz, and Sergiy Biletsky (pensions); Alberto Rodriguez and Mikolaj Herbst (education); Mukesh Chawla and Charles Griffin (health); Truman G. Packard, Anna Ruzik and Katarzyna Pietka (social assistance); Jan Rutkowski (labor market programs); and Marc Robinson (medium-term budgeting and performance budgeting). Monika Bazyl, Maciej Bukowski, Malgorzata Guzowska, Adam Kozierkiewicz, and Piotr Lewandowski provided important contributions. Luca Barbone (Sector Director), Bernard Funck (Sector Manager), Thomas Blatt Laursen (Country Manager) and Swati Ghosh (Lead Economist) provided excellent overall guidance. Fabienne Ilzkovitz, Theresa Jones, Maureen Lewis, Kathy Lindert, Ahmadou Moustapha Ndiaye, Pia Schneider, Jos Verbeek, Hermann Von Gersdorff, and Penny Williams provided very useful comments. Malgorzata Michnowska and Mismake D. Galatis provided excellent administrative support. Jus PENSIONS Introduction: 1. The Polish pension system underwent major reforms in 1999. Similar to the pension systems of the transition economies, the Polish pension system was strained by the country's transition to a market economy characterized by massive restructuring and a significant drop in formal sector employment. Revenues declined as formal sector employment fell; expenditure rose as a result of low retirement ages, early retirement provisions in many sectors, and relatively generous benefits, creating pension system fiscal deficits. The deficits were projected to worsen over time under the projected aging of the Polish population. Following a long debate, in 1999 the Polish government introduced a new multi-pillar pension system. 2. The traditional pay-as-you-go system inherited from the socialist times was replaced with a multi-pillar pension system that included a notional defined contribution (NDC) scheme and a mandatory fully funded defined contribution scheme. These reforms significantly improved the fiscal balance of the pension system. Following the 1999 reform, Poland continued reforms to address the remaining issues. These reforms placed Poland in a better situation than the rest of the neighboring countries that also introduced major reforms. While most of the neighboring countries felt the pressure to introduce changes to certain pension system parameters with the impact of the financial crisis, Poland managed to keep the existing parameters intact. However, as the Polish authorities are well aware, there are still items on the reform agenda which remain to be completed. 3. This chapter analyzes the structure of pension sector, evaluates the long term fiscal impact of reforms, discusses recent reforms and future challenges, and is organized as follows. Section 2 presents the characteristics of the pension system in Poland and provides international comparisons. The impact of reforms on fiscal balance and on benefit levels is discussed in Section 3 followed in Section 4 by a description and analysis of recent pension sector reforms in Poland. Section 5 presents conclusions. I. Main Characteristics of the Polish Pension System and Comparison with Other Countries 4. The Polish mandatory pension scheme consists of two pillars. The first pillar is a Notional Defined Contribution (NDC) scheme financed by mandatory contributions. Under the NDC scheme, benefits are computed using a defined-contribution formula with tracking of each individual's contributions in an individual account to which a government-determined interest rate accrues, but the scheme's underlying financing remains pay-as-you-go (contributions from the current workers are used to finance the benefits of the current retirees). Individual account balances and investment earnings are purely notional, meaning that they are an administrative record of contributions and credited interest without any underlying fundse2When an individual retires, the amount of money in his account ' Prepared by Anita Schwarz and Ufuk Guven. This analysis covers the main pension scheme only and does not cover the pension scheme for the agricultural sector (KRUS) and separate pension schemes with different rules that exist for certain occupations in Poland. 'Rather than changing the parameters of the existing defined-benefit pension scheme, Poland closed the scheme to younger workers (those born after 1948) and enrolled them in a new first-pillar scheme based on NDCs. Their determines the amount of his pension. The second pillar is a privately managed fully funded defined- contribution scheme. The contributions in this case are actually invested and actual investment earnings accrue on these benefits. Benefits from both schemes are payable in the form of annuities, the values of which are computed on the basis of contributions; credited investment earnings (which, in the case of the NDC scheme, are notional and based on the rate of growth in covered wages) and life expectancy at retirement. Table 1: Main Parameters of the Polish Pension System Parameter First Pillar (NDC) Second Pillar (FDC) Retirement Age 65 for men, 60 for women, 65 for men, 60 for women, Pension from conversion notional Pension from conversion capital Benefit Rate capital accumulation into annuities. accumulation into annuities 75 percent of growth rate of covered Interest rate provided by fund Interest Rate on contributions wage bill chosen by the individual Mixed price-wage formula, with 20 Indexation Post Retirement percent share of wages and 80 Under discussion percent inflation Contribution Rate 12.22 percent (pure NDC 7.3 percent (Pension only) participants pay 19-52) Source World Bank Staff 5. The total pension contribution rate compares favorably with the neighboring new EU countries though not as favorably with the high income OECD countries. Total old-age contributions are evenly split between employers and employees at the rate of 19.52 percent levied on wages up to 2.5 times the national average Fi ure 1: Pension Contribution Rates in New EU Countries wage; 12-22 percent of ,E, wages is credited to the 35 NDC pillar and 7.3 30 percentage points go to the 25 funded second pillar. 2o ' 15 There is an additional 6 percent contribution for disability that leads to a , total contribution rate of High Bulgaria Slovenia Poland Latvia Lithania Czech R Slovak R Romania Hungary Income 25.52 percent. Among the - OECD new EU countries, Source: World Bank Staff contribution rates for Poland are on the low side. (Figure 1) After Bulgaria and Slovenia that have lower contribution rates, Poland has the lowest contribution rate among the new EU member states. Poland's 25.52 percent contribution rate is higher than the average contribution rate of about 20 percent in the high income OECD countries. 6. Retirement ages for women are relatively low potentially making the retirement benefits low for those women who choose to retire at the set retirement age. The retirement age for men is set at 65, comparable to the standard retirement age of most OECD countries. The retirement age for women is 60. The low retirement age for women make the length of period of collection much longer leading to a more erosion of the pension income over the longer period of benefit collection. Also the lower retirement age for women would potentially lead to lower pension capital (and therefore a lower capital pensionable rights accrued under the old system until 1998 were converted into initial capital, which was credited to their individual accounts. 2 accumulation for pension calculation) at the time of retirement as women will likely retire with a shorter contribution period at the lower retirement age, divided by a longer life expectancy, resulting in a double reduction of benefits for women.3 Most OECD countries, by comparison, have e ualized retirement ages between men and women, and some have increased the retirement age beyond 65. 7 7. The notional interest rate provided in the Polish pension system is linked to some component of the growth of the wage bill and is less generous than the notional interest rate that is provided in the other notional account systems. Linking the notional interest rate to the growth of the wage bill contributes to fiscal sustainability in a declining population as in the case of Poland since the notional interest rate takes into account both the decline in the number of contributors as well as the growth in wage rates. However, the Polish notional interest rate does not even fully reflect the wage bill growth as the notional interest is set at 75 percent of the growth of the covered wage bill. This is less generous than the notional interest in other notional account systems. For example, Sweden links notional interest rate to the growth of average wages; Latvia links it to the growth of the full wage bill. 8. The indexation rule for pensions consists of a combination of inflation and wage growth is not fully in line, but not far from, international best practice. Currently, the post retirement benefits in Poland are indexed to 20 percent of nominal average wage growth and 80 percent of inflation. This indexation rule provides not only for maintaining the real value of pensions (that the pure inflation indexation would do) but also allows for reflecting some component of the wage growth in pension levels and therefore allowing the pensioners to benefit from some component of the wage growth.' 11. Recent Developments in Poland on Pension Reform 9. The Polish Government has been busy completing and complementing issues which were not addressed in the original reforms. The 1999 reform, while comprehensive, did not fully address issues regarding early retirement, disability pensions, and specification of payout mechanisms for the second pillar. Poland, like many of the other transition countries, made heavy use of early retirement and disability pensions as a mechanism for easing the required change in the labor market, but has been tightening the use of these systems since then. Similarly, given that the earliest retirees from the funded system would not appear until 2009, there was no need at the outset to have legislated how pensions would eventually be paid. 3 Although unisex mortality tables are expected to be used in the calculation of the pension annuity which normally compensate for the longer lives of women, unisex life expectancy at age 60, the women's retirement age, will still be higher than at age 65, the men's retirement age. Retirement Ages in the OECD Countries .Most OECD countries have a standard retirement age of 65 for men. In Iceland, Norway and the United States, the retirement age is either 67 or it is being increased to this age. Denmark, Germany and UK are in the process of legislating increases. Most OECD countries have equalized or in the process of equalizing retirement ages for men and women at age 65. These countries include Australia, Belgium, Canada, Denmark, Finland, Germany, Greece, Italy, Japan, Luxembourg, Netherlands, New Zealand, Portugal, Spain, Sweden, and UK. In Iceland, Norway and, United States, retirement ages are equalized at 67 while in Ireland the retirement age is equalized at 66. Most OECD countries that have an earnings related first pillar pension scheme link the adjustment of pensions to prices. These countries include Belgium, Canada, France, Japan, Korea, Spain, Turkey, United Kingdom and United States. Some countries such as Finland, Hungary, Slovak Republic, and Switzerland have adopted indexation to a mix of price and wage inflation, as pioneered by Switzerland. In some countries such as Germany, Netherlands, Luxembourg, and Norway, the adjustment is based on increases in wages. 3 Reducing the Number of Groups Eligible for Early Retirement 10. Occupational categories eligible for early retirement were revisited and substantial tightening of early retirement provisions was introduced to reduce the fiscal burden of early retirement on the pension system. Poland had numerous occupational categories that were eligible for early retirement. The rules for defining special/hazardous occupations were identified in the existing legislation. Based on this definition there were about 1.2 million individuals who were part of special occupational schemes and therefore eligible for early retirement. This situation decreased the effective retirement age significantly to about 57 from the standard retirement age of 60/65 for women/men.6 The Polish government decided to re-visit the occupational categories eligible for early retirement and a new set of criteria was developed for eligibility for early retirement in 2008. It is expected that these reforms reduce the number of people eligible for early retirement pensions from about 1.2 million to around 300 thousand. 1 1. Bridge pensions for those employed in special difficult conditions, effective from 2009, replaced generous early retirement schemes. Based on the new law some occupational groups that work under hazardous conditions remained eligible for early retirement. The new law introduced bridge pensions to be paid to eligible workers (those who remained eligible under the new legislation) Employers of workers in these special groups are required to pay 1.5 percent of contributions to the Fund for Bridge Benefits which is used to finance bridge pensions. The financing gap between revenues of the Fund and the spending on bridge pensions is to be financed from the central budget. Bridge pensions are calculated similar to the old age benefits, meaning that the capital accumulation is divided by the life expectancy at the time of early retirement (which is usually 60)7. Establishing the Pay-out Provisions from the Second Pillar 12. Given the enormous changes undertaken in the pension reform, some of the critical elements of the funded system were not addressed in the initial reform. This included specifying the pay-out provisions for the second pillar benefits. Based on the rules of entry, women are expected to retire from the funded system in 2009 and there was a need to specify the procedures on benefits to be provided to the funded system beneficiaries retiring between 2009-201 3. 13. Legislation identifying the principles on how the funded system would pay-out benefits for those retiring between 2009-2013 was issued in 2008. The new legislation specifies two types of benefits (i) temporary pension and, (ii) lifetime pension (annuity). Temporary pensions are basically for women as men will start collecting funded scheme benefits in 2014. The temporary pension is calculated like the notional account pensions from the first pillar are calculated. The minimum accumulated funds required for eligibility for a temporary pension is twenty times the nursing benefit (about zloty 163/month, total minimum: about zloty 35,000). If the amount of capital accumulation is lower than this specified amount, the capital accumulation is transferred to the ZUS and treated as a single contribution; the calculated benefit cannot be lower than the minimum pension. The indexation of the temporary 6 The Law on Social Insurance and Disability already mentioned the restriction of early retirement benefits for those born after December 3 1, 2008. The deadline for early retirement was postponed several times and the final deadline was December 31, 2008. Those who were eligible (those categories eliminated for early retirement with the new law) before this date could retire before the standard retirement age. 7 If the beneficiary starts working in a job characterized as hazardous conditions, their benefit is stopped as the idea behind providing bridge pensions for certain groups is to protect their health status and going back to a job of hazardous conditions is inconsistent with this purpose. If the beneficiary goes back to a job outside these special occupational groups, then the general principles in the pension legislation apply. 4 pension is the same as with the regular old-age benefit. The legislation refers to the life time pension leaving the details to be identified with subsequent legislation, 14. The Government has made substantial efforts to secure passage of the annuity law which regulates the lifetime pensions, but has faced political opposition. An annuity law that defines the payout provisions for the lifetime pension from the funded pillar was presented and adopted by the Parliament in December 2008, but then vetoed by the President. While legislation would be desirable now, it is not urgently needed, and it will not be essential until cohorts with a significant balance of accumulated savings in their private retirement accounts begin to retire in 2014. At that time, those who have received temporary pensions above the minimum level can be converted to lifetime pensions. 15. The Annuity Law also included specified provisions on the investment rules for retirement savings in the funded pillar. Initially, Poland opted for a conservative approach in the portfolio options in the investment of second pillar funds. The investment rules were designed such that each pension fund would offer only one portfolio. This approach was chosen mainly to ensure that the existing capacity to supervise was sufficient to ensure safety of the portfolio in the initial years of implementation of the second pillar. Now that the supervision capacity has been well established, moving to multiple portfolio options would allow individuals to optimize portfolio allocation over their life cycle, avoiding potential disruptions from events like the financial crisis which could have had adverse effects on Poles who were just retiring. Aligning the Disability Benefits with Old-Age Benefits 16. While the benefit provision for old-age pensions was completely changed in the 1999 Reform, the old formula remained for the disabled. Rather than integrating the disability formula with the old age formula in the time of reform, Poland kept the old formula for disability with financing coming from a 13 percent contribution (Table 2). As noted earlier, the disability pension system provided a safety net during the labor market upheavals in the early days of the transition. As the labor market recovered, disability eligibility conditions have been tightened, allowing a reduction in the contribution rate required to finance the benefits. The contribution rate for financing disability was reduced from 13 percent to 6 percent in two phases, reaching 6 percent in January 2008. Tightened eligibility conditions reduced the total number of disability pensioners from 2.6 million in 2000 to 1.47 million in 2008. 17. Similar to the Annuity Law, substantial effort was made to secure passage of legislation to align the disability benefit rules with that of old-age benefits. The Polish Government recognizes the importance of making the disabilityisurvivor benefit consistent with the reformed first pillar old-age benefit. In some cases, individuals eligible for disability benefits could receive higher benefits than under the old age system. To avoid providing new incentives for disability certification which undo the progress made to date, the Government sought to realign the disability benefit structure with the old age benefit structure. A draft law was presented by the Government and adopted by the Parliament in 2008, but again was vetoed by the President. The law proposed to link disability pensions to old-age pensions, with a view to making the system consistent across all of its component programs. 5 Table 2: Eligibility Conditions for and Benefits of Disability Pensions in Poland under the First- Pillar Earnings-Related Scheme Required years of service Contribution rate Eligibility Benefit rate Partial pension for eligibility at different ages Flat component (24 percent of reference Below 20: 1 year wage) plus 1.3 6 percent (4.5 percent for each year 20-22: 2 years 22-25: 3 years percentemployer, Or of contributions and 75 percent of total 1.5 percent lity pension 25-30: 4 years Above 30: 5 years employee) topup the total to 25 years Source European Commission 2007. 111. Fiscal Sustainability and Benefit Levels 18. The analysis shown below is based on the Bank's PROST model using data provided by the Polish authorities. However, there were a number of variables missing, and several others where the Bank team was unable to double check with the Polish authorities on the accuracy of the data. As a result and as with any projections, the results particularly those in the distant future should not be regarded as point-accurate, but rather as a reflection of the overall trends expected in the future. 19. Currently, the first pillar pension scheme in Poland is running a deficit but the system will achieve a fiscal surplus in the long term. Currently expenditures are over 9 percent of GDP while the revenues are at 7.9 percent of GDP leading to a deficit of 1.3 percent of GDPe8Over time, projected revenues will go down and level off at around 6.0 percent of GDP in the long term. Expenditures are projected to fall more substantially to reach 4.5 percent of GDP in 2075. As a result the pension system is estimated to generate a surplus of 1.5 percent of GDP in the long term. (Figure 2) 20. However, the future expected benefits are quite low. Given the expected aging of the Polish population, a pension system which runs a long run surplus is unusual. But the cost Source: World Bank Staff Calculations based on Prost model of achieving the surplus is borne The 2009 deficit is underestimated for mainly one reason. Due to data limitations, the model is using historical data up to 2007. Hence, the spike in retirements in 2008 triggered by the phasing out of early retirement options is not reflected. , which means that expenditures in 2009 and later will be higher than what is projected. 6 by the relatively low levels of benefits to be paid in the future. Figure 3 shows the benefits as a percentage of gross average wage and as a percentage of average wage, net of social security contributions. The benefits include both the benefits paid by the NDC system and the benefits expected to be paid from the funded system, assuming a real interest rate of 4 percent over the lifetime of the worker. As shown in Figure 3, benefits fall from the current level of approximately two-thirds of average wage, net of social security contributions, to less than one-third, which is quite low by international standards. With regards to composition of benefits, while the contribution rate is split with 63 percent going to the NDC and 37 percent of the total contribution to the funded, the NDC only pays 46 percent of the total benefit in 2075 compared to more than half paid from the substantially lower contribution to the funded system. The low average future benefit rates are also heavily influenced by the number of women among the elderly. Women, who are longer-lived than men, are expected to make up almost two-thirds of the elderly in 2075 and have lower pensions because of their lower wages, but also because the earlier retirement ages give - - them less time to accumulate -3: P r o j e c t e d l s as Cr- on pension capital and a longer , retirement period over which I the lower capital accumulation ~- Gross average replacement rate - - - Net average replacement rate ' ;1 I must be spread. The average ~ 700% - - - - - - - - - -- - - _-_ . I _ I - - - __ - - - - , I replacement rate for male pensioners in 2075 is expected to be closer to 33 percent of 500% - - _ . . " _ - gross average wage or 44 1 percent of average wage, net I of social contributions, which i the internationally acceptable 20.0% 9 range. 10 0% 00% ' 2 1. Unreformed 'mm I 1 ~ N ~ m ~ P ~ O m w ~ N ~ ~ b ~ O disability benefits threaten ~ ~ ~ ~ ~ ~ ~ 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 N N N N N N N N N N N N N N N N N N N N N N N m m m b to swamp the system. , Currently more than two- . thirds of pension benefits go Source World Bank Staff Calculations based on Prost model to old age beneficiaries while only 15 percent go to disability beneficiaries. With the sharp reduction in old age benefits, an unreformed disability system with its much higher level of benefits will result in a roughly equal share of spending allocated between old age benefits and disability benefits. As people see higher benefits being paid by the disability system, they will have a huge incentive to qualify for a disability pension rather than an old age pension, which would threaten the gains that the government has achieved thus far in limiting the disability pension growth and in containing the costs of the pension system. IV. Recommendations 22. While the Polish pension reforms have done a superb job at addressing the issue of fiscal sustainability, the projected long-term benefits are low. A number of measures can be phased in over time to improve the level of benefits, but this needs to be done in a manner that does not undermine the These projections assume that men continue to contribute for a full 40 years and women for a full 35 years. Should contribution periods decline substantially as they have in many of the other ECA countries, benefits would be even lower, while expenditures of the pension system would be higher. 7 hard-won gains in fiscal sustainability. Appropriate measures include (i) raising the retirement age, particularly that of women, which is below the typical age in other countries, (ii) reforming the disability system and reallocating some of the contribution currently allocated to disability to the funded system, (iii) reallocating some of the notional contribution in the long run to the funded system, (iv) maintaining and improving the efficiency of the second pillar, and (v) integration of special schemes. Raising the retirement age, particularly for women, will substantially raise benefits. As noted earlier, women who make up a large proportion of the elderly will face reduced benefits in the future because the earlier retirement age that they face limits the amount of pension capital they can accumulate and then spreads this limited capital over a larger number of retirement years. Benefits for men are double the benefits for women. Part of this is due to higher men's wages, but average male wages are only 22 percent higher than average female wages, while male pensions are 100 percent higher. The greater differential in pensions is almost entirely attributable to the lower retirement age. Reforming the disability system, as has already been legislated by the Polish Parliament, can open the fiscal space to provide better old age benefits. Reforming the disability program would reduce expenditures on disability pensions and create the fiscal space to reallocate some of the contributions currently financing disability to the old age system. As noted in the text, the low rate of return to the NDC contributions in the long run suggests that if such a reallocation were to take place, increasing the contribution to the second pillar would have a more sizable impact on the level of future benefits than increasing the contribution to the NDC. And as the NDC system begins to generate surpluses, some of the existing NDC contribution could also be reallocated to the funded system to improve the benefit level further. Reforms that aim to improve the efficiency of the second pillar will contribute to better pension outcomes. Given the increased importance of the second pillar in financing future pensions, the funded pillar of the pension system will need to become more efficient. The Polish second pillar has outperformed the other second pillar systems in the region, generating an 8.9 percent rate of return from its inception to end-2007. But as the economy matures and economic growth slows down, there are reforms which can be undertaken to maintain its good performance and to improve further. Payout options need to be designed so that they are flexible enough to withstand market fluctuations while protecting workers. The use of multiple portfolios with a mandatory shift toward a fixed income portfolio in the years just prior to retirement is being pursued as a way of protecting workers against sudden drops in the valuation of financial assets just prior to retirement. Measures to lower costs to affiliates will help increase the balances available to participating workers when they retire. Most countries in Central Eastern Europe that introduced funded individual retirement savings pillars - including Slovakia, Hungary and Estonia - have allowed multiple portfolio options for workers' mandatory savings. Multiple portfolio options provide more instruments with which affiliates can manage financial risk as they approach retirement." 10 The portfolios of Polish pension funds are not well diversified. A preponderance of government bonds is observed in the pension portfolios of most countries that have recently introduced mandatory funded pension systems (the second pillar), although in Poland the funds have relatively larger holdings of domestic equity and investments in privately-issued fixed income instruments. The lack of diversification of pension fund portfolios reflects in part some investment restrictions. The crisis and its impact on capital markets reinforces the need to combine greater choice in funded private pension pillars with better default options for portfolio allocation. The multi-fund or "lifestyle model" was pioneered in Chile at the beginning of the decade to give affiliated workers more options to manage financial risk as they approached retirement. Proponents of greater investment choices did not expect low 8 Introduction of legislation to further reduce management fees could also be considered. Poland's pension funds operate with relatively low administrative costs and fees. This outcome is due to centralized revenue collection arrangements and regulation that caps management fees. Total fees amount to 160 basis points on assets, which is reasonable given the short period the funded pillar has been operating and the small volume of assets under management. However, to increase the accumulated savings balances workers can rely on at retirement, and to safeguard the funded pillar from political threat, the Government is correctly considering further measures to reduce fees." Integration of special schemes within the pension system. Poland still maintains a number of separate occupational schemes, such as those for farmers, the military, and police. While integrating these occupations into a national scheme is difficult in most countries because of occupational limitations, labor market flexibility is sacrificed by maintaining many separate schemes. Integration issues are beyond the scope of this note, but will need to be considered by future Polish governments. financial literacy to be a barrier. Affiliates were slow to exercise their choices, those who did selected portfolios irrationally, and once they chose a portfolio rarely changed their allocations. Lifecycle funds provide a better risk allocation along the working life of individuals. They alter the default "do nothing" portfolio allocation progressively away from variable return securities to fixed return options as individuals get older. (Rudolph, H. and R. Rocha, "Competition and Performance in the Polish Second Pillar'' mimeo). The Bank is advising the authorities on the importance of setting age-related default options as more investment choices are offered. II Fees on contributions are capped at 7 percent for all new contracts issued after 2004. This cap will be progressively reduced to 3.5 percent in the year 2014. All but one fund manager currently charges the maximum contribution fee. Fees on assets are capped according to the volume of assets under management (Rudolph, H. and R. Rocha, "Competition and Performance in the Polish Second Pillar" mimeo). The Government is drafting legislation that would accelerate the pace of fee reductions. 9 AGRICULTURAL PENSIONS Introduction 23. Poland reformed its pension system in 1998, but in contrast to the intention of its architects some occupational groups were not included into the general system, mainly for political reasons. The new system consists of a pay-as-you-go financed, publicly managed notional defined contribution (NDC) scheme, a mandatory second pillar serviced by privately managed, defined contribution pension funds and voluntary pension savings. At the outset of the pension reform process, all Polish citizens - including farmers, miners, the uniformed and armed forces- were supposed to be covered by the new arrangements. Also, the concept envisaged the unification of the Agricultural Social Insurance Fund - KRUS (KRUS - Polish acronym from Kusu Rolniczego Ubezpieczenia Spolecznego) and the Social Insurance Institution ZUS (ZUS - Polish acronym from Zaklad Ubezpieczen Spolecznych). Later, however, the plans were scaled down and the actual reform did not touch KRUS. This was done for political reasons - reforming ZUS was hard in itself without having to deal with the traditionally powerful agricultural lobby - but also because the necessary time and resources were lacking to prepare those changes in the fields of social assistance, income taxation, and social security administration - all deemed crucial conditions for the unification. 24. As of end 2008, when so called bridge pensions were introduced, around 2.14 million people were still covered by special retirement systems. They included both other than general occupational system, led by ZUS, or special privileges within it (miners). The separate systems exist for: J farmers, separate contributory system operated by KRUS [ 1,456 thousand people], J miners, enjoy a special retirement status within the general system, based on the law passed in July 2005 [323 thousand people], J the uniformed forces (policemen, firemen, intelligence agents), a budgetary system administered by the Ministry of Internal Affairs and Administration [ 182 thousand people], J the armed forces (soldiers), a budgetary system administered by the Ministry of Defense [ 157 thousand people], J the Prison Service, a budgetary system administered by the Ministry of Justice [26 thousand people], J judges and prosecutors, a budgetary system administered by the Ministry of Justice, J Veterans, Military Disabled, who received pension rights based on legal regulations or international agreements, and whose benefits are financed from state budget, and paid by ZUS, KRUS, or budgetary system administrators. 25. There has been broad agreement within successive Polish governments that KRUS needs to be reformed in order to improve its financial sustainability. There are competing views as to whether the reform should leave KRUS as a separate entity functioning along different - be it improved - rules or whether perhaps KRUS rules should be harmonized with those of ZUS and thus a universal pension system be created. A related but separate issue is whether there should be a single entity administering public pensions or whether KRUS and ZUS should remain separate agencies. In our view, the ultimate goal must be that of harmonizing ZUS and KRUS regulations and that the most efficient way of doing so is by unifying the governance and management structures of the two entities. Thus, any measure rendering KRUS financially more sustainable - expansion of the contribution base or raising the rate of farmer's social security contribution, contraction of coverage or lowering of the average benefit levels received by KRUS pensioners - will be a much needed positive development even if not implemented within the framework of a structural pension reform. I. Background 26. Poland's pension system is divided into two social security regimes: one run by KRUS for farmers and their dependants, and the other by the ZUS for all other workers and the self-employed. KRUS offers pay-as-you-go financed, defined benefit pensions to nearly 1.45 million people (as of 2008) and collects contributions from and on behalf of almost 1.57 million farmers and their family members. The parameters of the ZUS and KRUS pension systems differ significantly with KRUS offering access to public health care, old-age and disability pensions, and accident and maternity benefits at a much lower premium than ZUS''. As a result, KRUS revenues only cover a small part of its expenditures, with the shortfall being financed by budget transfers. In 2008, budget subsidies financed 88.5 percent of expenditures. Many people owning land or related to farmers have the defacto option to choose between ZUS and KRUS. Figure 4: Number of Insured Persons and Beneficiaries in 1991-2008 2.10 .- 6 5 2.00 1.90 1.80 1.70 160 1.50 1.40 --C benticiaries 4- insured Source: KRUS (www.krus.gov.pl) I' Monthly contribution to KRUS in 2008 was about 7.5 times lower than the minimum monthly contribution paid by ZUS members. 12 Table 3: Main Characteristics of ZUS and KRUS, 2008 - KRUS zus Contribution rate (old-age and disability 10 percent 25.52 percent pensions) Contribution base Basic (minimum pension) Actual wage Average old-age pension PLN/in percent of PLN896 / 30 percent PLN 1370,81 / 46 percent average wage Minimum pension PLN/in percent of average PLN 675 zl / 23 percent PLN675 / 23 percent wage Retirement age 65/60 (5 years less with 30 65/60 years of contributions) Full career contribution history 25 years 25120 years Beneficiaries 1.45 million 7.4 million Contributors 1.57 million 14.3million Expenditures as percent of GDP (old-age and 1.3 percent 9.6 percent disability) Budget subsidy as percent of Expenditures 88.5 percent 24.5 percent Source: KRUS (www.krus.gov.pl), ZUS, Central Statistical Office, World Bank Staff calculations Table 4: KRUS - Selected Financial Indicators Subsidies as Average Revenues Contributions Subsidies Expenditure percent contribution expenditure per month -- -- 1995 6536.4 364.3 6172.1 6483.2 95.2 percent 21.28 1996 8104.9 469.6 7635.3 80 19.4 95.2 percent 28.2 1 1997 10069.3 583.7 9485.6 10055.5 94.3 percent 34.30 1998 11352.7 614 10738.7 11310.6 94.9 percent 36.16 1999 139 12.4 688.8 13223.6 13737.6 96.3 percent 40.19 2000 14086.7 736.3 13350.4 14308.0 93.3 percent 42.25 2001 15835.9 844.5 14991.4 15793.O 94.9 percent 46.85 2002 16406.0 932.5 15473.5 16402.3 94.3 percent 49.83 2003 16101.3 1000.5 15100.8 16226.7 93.1 percent 52.46 2004 16306.9 1108.7 15 130.0 16338.6 92.6 percent 59.99 2005 15886.7 1077.715 14710.0 16035.5 9 1.7 percent 56.77 2006 16204.9 12 12.022 14932.7 16039.5 93.1 percent 62.53 2007 15942.1 1 189.485 14680.7 15460.3 95.0 percent 62.02 2008 16185.3 1224.658 14867.9 16793.8 88.5 percent 64.82 Source; KRUS (www.krus.gov.pl), ZUS, GUS, Annual Budget Reports. 27. While KRUS is an occupational insurance scheme, it meets broader goals of social and agricultural policy. For instance, by reducing wage taxes, the significantly lower premium increases farming households' disposable income and therefore functions as an income subsidy in the economically less developed rural areas of Poland. KRUS is, therefore, seen by both the Government and agricultural lobby groups as a means of addressing rural poverty and unemployment. It may also be seen as an input 13 subsidy intended to improve the competitiveness of Polish agriculture. However, it can also be argued that the budget subsidy enjoyed by KRUS members acts as a dually-edged sword and delays necessary adjustments in terms of productivity and farming vs. non-farming employment as the primary means of income generation. The parallel systems also undermine ZUS by offering the opportunity of on opt-out and taxing prohibitively the movement from informal to formal employment. 28. KRUS plays an important social role by providing old-age, disability and survivor pensions to retired farmers and their dependants but also has a perverse redistributive impact. Without KRUS benefits, poverty would be higher both among the rural elderly and the younger, disabled people. At the same time, by subsidizing farmers across the income distribution by the same amount, it has a perverse redistributive impact within its own affiliates as well as between KRUS members and other taxpayers. KRUS is financially unsustainable and requires substantial budget subsidies for the foreseeable future. This is due to numerous factors, including the legacy of farmers brought into the system when it was created, to the prospect of progressive decline in the number of contributors as agriculture employment contracts in line with the experience in the rest of the EU, and the low level of contributions from farmers. 11. Key Recommendations for Reform Direction 29, Operating a separate social insurance scheme for farmers is not sufficiently justified. Operating a separate, mandatory, occupational pension scheme with different rules than the general scheme is justifiable if: (a) the characteristics of the target population which are relevant from a pension point of view are significantly different than that of the rest of the insured population; (b) the specificities of the target population - including the ability to observe their income and monitor compliance- cannot be accommodated within the general scheme; and (c) the special scheme is financially sustainable and its existence does not impose significant (administrative) efficiency losses on the social insurance system. KRUS and its members do not meet the above criteria to an extent that would justify operating a separate social insurance scheme. 30. In order to reduce the fiscal burden created by agricultural pensions, two approaches can be taken: reducing eligibility and increasing contribution revenues within (a) a reformed KRUS or (b) by merging it into the general social security scheme (ZUS). Since most KRUS retirees already receive a pension equal to the basic (minimum) benefit, a significant reduction of the average replacement rate is not a viable option. The solution b) would reduce KRUS membership to zero and raise farmers? social security contributions to the same level as those paid by ZUS members. The main difference between the two solutions does not lie in the nature of the changes (increasing farmers? contribution burden per unit of pension promise or benefit) but the extent to which they are applied and the institutional arrangements of the reform. 31. The purpose of the unification is to treat every citizen the same way regardless of their sector of employment but it is advisable that a minimum contribution be paid by everyone except the unemployed. Unemployment needs to be defined slightly differently in agriculture than elsewhere, approximating unemployment through status and income. Hence, from a contribution point of view, an unemployed person would be someone who is registered as such: or (a) holds no business license or commercial tax identifiers; and (b) has no income in excess of an amount equal to an unemployment benefit based on the statutory minimum wage, net of the minimum contribution. Defining unemployment in such a manner would make it a relatively good proxy for generating a very low income. 14 32. The current functions of ZUS and KRUS are not comparable and therefore KRUS' efficiency in what they do is not an argument against administrative unification. KRUS has a reputation for being an efficient organization. We must bear in mind that for social security agencies it is rarely the payment of established pension benefits that causes difficulties - it is maintaining contributor and contribution databases and collecting contributions on the basis of actual wages which is the most difficult and costly operation. Traditional paygo, DB schemes do not concern themselves with liabilities accrued before a beneficiary's retirement - at the same time ZUS, as a result of the NDC reform, transcribes past contributions and their implied yield quarterly. 33. It is crucial that contribution revenues from farmers a r e gradually but significantly increased, although it is obvious that a unified social security agency would continue requiring budget subsidies. The parallel operation of two publicly managed social insurance schemes increases operating expenses, creates an opportunity for tax arbitrage and weakens the incentives for employment in' the formal sector. 34. A unique window of opportunity, opened from 2004, to increase farmers' social insurance contribution and re-assess their income tax status is presented by the EU Common Agricultural Policy (CAP) payment per hectare. EU payments to farms have been Figure 5: Phasing-in and topping-up of Direct Payments, decoupled from production, and are (in percent Of EU15payment level) I now paid on a hectare basis, currently averaging about EURl60 per hectare in the EU 15 countries. Poland's farmers loo 80 are entitled to an annually increasing 60 fraction of that amount, equivalent to about 50 percent in 2008, and rising to 40 100 percent in 2013. Beyond this 20 Simplified Area Payments Scheme (SAPS) allocation, there are other EU 0 subsidy schemes which may be 1 2 3 4 5 6 7 8 9 10 accessed along with limited "top-ups" year of E U membership from national budgets, as their Source: European Commission information, World Bank Staff legislatures prefer. But the SAPS payment is a good concrete instance of a windfall cash income gain for Polish rural landholders. 35. Full harmonization of ZUS and KRUS regulations will take time, partly because the new contribution rules can only be applied to farmers gradually, partly because of the incompatible information technology and also because pensions assessed by KRUS will remain in payment for decades to come. At the time of designing the reform process and its early stages of implementation ZUS would be responsible for managing two separate schemes while maintaining the operational infrastructure inherited from KRUS. As the reform proceeds, reorganizations affecting both the inherited KRUS structure and that of ZUS can gradually take place. 15 Box 1: The Single Area Payment Scheme (SAPS) Provides for Direct Payments Per Hectare of Agricultural Land. The simplified SAPS option was offered to NMS because most of them did not operate CAP-conforming direct payments during the "reference period" and because of the high administrative requirements for implementing direct payments on a Single Payment Scheme basis, operating in the EU15. Income support payments across the EU differ between old and new EU member states, as farmers in the EUlO do not immediately receive the full EU15 level of support payments. Eligibility was strongly debated during negotiations as these payments were initially intended to compensate EU-15 farmers for past reductions in price support. It was eventually agreed that direct payments would apply to NMS and that they would be gradually phased-in over the first decade of membership, with initial payment levels amounting to 25 percent of those granted in the EU15. However, the NMS were allowed to speed-up this process through "topping-up'' the EU-hnded SAPS'3 by an additional 30 percent of the EU-I 5 payment level. 111. Estimating Agricultural Income 36. The extent to which unification can be expected to deliver short-term fiscal improvements depends solely on the speed of phasing in higher contributions for farmers. Their benefits are already low in comparison to ZUS pensions; agricultural cash incomes potentially serving as the contribution base are also low. As contribution revenues gradually increase and benefits are maintained at their current levels, the fiscal position of KRUS - or of a unified social security agency - is expected to improve. 37. Given that the number of small farms has grown very slightly but the number of insured people was growing over the past decade, it means that KRUS eligibility did encourage people without a formal sector job to become insured as family members. KRUS membership may have also discouraged the sale of small plots and turning farmers into ZUS affiliated employees. Many farmers are already affiliated to ZUS if their non-agricultural income exceeds a legally set threshold. Unification may speed up the merger of small and medium-sized farms on the verge of commercial viability into large holdings. (This process will also change some formerly independent land-owners into employees.) It is not expected that, after the initial impetus given to mergers, unification in itself will have a lasting effect on land size. 38. In the future, a natural extension would be to subject farming income and farmers to the same personal income tax provisions as others taking into account problems with estimation and the volatility of farming incomes. This raises a question whether farmers are indeed a special category of workers/owners or are self-employed or employees whose income is generated from agriculture instead of industry or services. Subjecting farmers to the same registration, licensing, taxation and reporting requirements as all other private entrepreneurs would make the administration of both social security and other taxes more uniform and possibly more efficient. Nonetheless, the assessment and collection of contributions will need to take the seasonality and volatility of arming incomes into consideration by applying longer assessment periods and less frequent (quarterly, semi-annual) contribution payment schedules. 16 Table 5: Farms by Area Groups of Agricultural Land With agriculture land area in ha - YEARS Total 50 and 0-1 1-2 2-3 3-5 5-10 10-15 15-20 20-30 30-50 mOre In absolute numbers 2000 2,859,196 973,492 448,419 269,080 344,868 447,991 185,842 84,014 62,401 27,252 15,837 2004 2,844,168 987,939 484,194 264,338 329,841 403,312 177,337 79,954 62,615 32,491 22,147 2007 2,579,178 771,134 422,610 273,779 340,485 400,152 166,595 77,597 65,351 37,372 24,104 In percent 2000 100.0 34.0 15.7 9.4 12.1 15.7 6.5 2.9 2.2 1 .o 0.5 2004 100.0 34 7 17 0 93 11.7 14.2 6.2 28 2.2 1 1 08 2007 100.0 29.9 16.4 10.6 13.2 15.5 6.5 3.0 2.5 1.5 0.9 Source: Central Statistical Office, World Bank Staff. 39. Agricultural production indeed has its peculiarities. It is subject to seasonality within the year as well as to natural and business cycles. It is also affected by unpredictable weather and veterinary problems, making agricultural income's volatility less predictable. Agriculture is also specific to the characteristics of the land, which limits the room for shifting farming cultures in response to natural or commercial problems - the flexibility of agricultural production is much less than in industrial production or services. Farming is also special in that the same output can often be considered a final product ready for consumption or as an input for further processing. Farmers can choose if they wish to market their produce to end users or sell it for further processing. This complicates tracing income, too. Farmers also produce for own consumption which is difficult to measure - not so much because of the actual human consumption but because own consumption includes using own-produced input for other lines of production, such as animal feed used in milk production. Given seasonality and price uncertainties, farmers need to stockpile on both own-produced and externally purchased inputs to maintain the continuity of their own productions. As a result, accounting for farmers' income is more complicated than for other workers or for self-employed working in other sectors. If farmers are to be treated as self-employed, clear rules must be established for separating personal income from the income of the enterprise, the valuation and amortization of inventories. Special rules must be established to filter out income volatility through either applying different (longer) assessment periods for the taxable income base and/or operating an income tax credit system with carryover of accumulated gains and losses. 40. Most of the above issues are already addressed by the tax code, however. Self declaration of actual realized contribution-liable income can apply to farmers as well although control over the veracity of these declarations will require additional efforts and expenses. Similarly to private entrepreneurs and small businesses, most farmers will probably report income at the minimum and better compliance will depend as much on administrative efforts as on the growth of per capita income. If the two systems are unified, the two contribution minima (one which is explicit but lower for farmers and one which is higher for all others) may be merged or be kept separate. 17 Box 2: Estimating Agricultural Income Estimation of farm incomes, by governments or other external agents, for purposes of determining payments by farms, is everywhere a major problem. First, the farmer himself may not know exactly how much he has produced, of wheat, milk, cherries, etc., unless he has used modem equipment to harvest and store the output. Selling, to middlemen or cooperatives, provides a point of quantification (and indeed, of pricing), but this clarity is diminished to the extent that the farmer and his family and friends have consumed a part of the output, which should not be underestimated in Poland. Next, even these estimates are only of gross output. While costs of purchased inputs should be relatively straightforward, estimates get much more complicated (on both the input and the output sides) if the farm itself produces some of the feed for the animals, if the farm pays workers (who may also be relatives), and if depreciation of buildings and equipment is significant, which is often the case. All of this holds true if the farmer himself is a willing participant in the effort. In fact, of course, if the aim is to assess capability to pay for anything, from social insurance to income tax, this willingness cannot be assumed. What can be assumed is moral hazard, and the pleading of poverty which any enumerator will encounter on even the largest and most prosperous farms in the world. Without the active assistance of farmers themselves, external agencies are reduced to very expensive field or aerial photo-based surveys of each farm's land use: crops produced, animals held, milked, sold, or slaughtered; and then most likely use of regional averages for yields in terms of kilograms per hectare or liters per cow or kilograms per hog sold, as well as the prices received for each of these things. Fields of the same crop in the same village can differ by up to 100 percent in terms of yield per hectare, similarly in terms of milk yield per cow, depending on inputs, breeds, and skill. One approach to inducing some incentive for farmers to participate actively in this process, could be to prepare estimated incomes for each farm as outlined above, submit these estimates to each farmer, and then permit him or her an appeals process in which the farmer must bring documented proof of input purchases above those assumed, output sales at prices below those assumed, and the like. This will still be an expensive process bound to result in some injustice. IV. Short- and Medium-Term Fiscal Effects of Unification 41. Fiscal effects of the unification will depend on the relative dynamics of expected expenditures and revenues. Expenditures, in the short run, are determined by t h e pensions already in service and pension entitlements accrued t o cohorts close t o retirement. Expenditures attributable t o retirees are determined by benefit indexation rules and the demographic indicators of t h e beneficiary population. A unification will have n o effect o n either of these since indexation rules are already t h e s a m e (both schemes observe the s a m e minimum pension, too) and longevity is not influenced by administrative arrangements. Expenditures attributable to cohorts within ten years of retirement, as observed earlier, will depend on whether contributions paid between t h e date of unification and retirement increase pension benefits beyond t h e minimum pension and, if so, by h o w much. That, in turn, will depend o n t h e formulae used to convert past contributions into notional account balances. If we assume that t h e interest rate applied t o ZUS account balances result in a n actuarially correctly priced annuity, then t h e overall fiscal effect will depend on how much of t h e additional contributions made after unification are actually increasing account balances beyond the necessary m i n i m u m and h o w much is used t o m a k e u p for the shortfall which, under t h e current system, would b e financed b y budget transfer. 18 42. On the revenue side, unification with ZUS should avoid not only cross-subsidization but also lack of clarity as to the destination of the needed budget subsidies during the transition period. Unification with ZUS will not reduce benefits and, therefore, the incentives to contribute would not diminish. Indeed, if farmers start contributing on the basis of their actual income but also have to observe a minimum contribution above today's fixed amount, a revenue increase is very likely, even if we allow for some farmers stopping contributing. V. Addressing Economic and Social Implications of the Reform 43, The reform, whether involving parametric adjustments or unification, will increase farmers' average contribution burden and limit access to KRUS as an occupational pension scheme, needs to address welfare consequences, political, and economic repercussions. It is important to: (a) increase contributions gradually; and (b) to ensure that the most vulnerable groups - including retirees and contributors close to retirement - are protected against the negative welfare impact through special "grandfathering" arrangements. It is also important that the impact of a KRUS reform on labor and land mobility, tax incidence and poverty be surveyed in order to better design mitigating measures within the pension system, in other social transfer schemes, in tax policy and administration as well as in agricultural policy. The subsequent analytical work need to focus on the reform's impact on the lower end of the income and land distribution. Preparations for the reform will also need to include quantitative pension models forecasting the long-term fiscal and distributional effects of the reform. A poverty and social impact analysis focusing on the impact of higher contributions is necessary, with a special focus on farmers with less than 5 hectares of land, living in areas with higher than average rural unemployment indicators and with reported income below the statutory minimum wage. Furthermore, if unification is considered, the translation of past KRUS contributions into ZUS account balances will need to be calibrated after carefully modeling typical individuals' past and expected contribution records and the resulting pension benefits. 44. The purpose of harmonizing pension rules is to increase both farmers' contributions and the resulting pension benefits - but in a way which increases revenues more than expenditures. The ultimate goal is to bring every farmer into a uniform social security system. Those who cannot or cannot be made to contribute would also be treated uniformly - in case they have the minimum qualifying contribution history they would receive the minimum pension established within ZUS. If they fail to accrue sufficient pension rights, they would have to rely on discretionary social transfers.- such as social assistance - in old-age. 45. An alternative solution is to keep farmers' social insurance separate - but, instead of increasing their contributions, reduce the benefits received to reflect the low contributions paid. This solution would lower farmers' average pension significantly, although the reduction would be gradual as pension rights already accrued would be honored according to the old rules and benefits would be calculated as the average of the pre- and post-reform entitlements. Farmers would be allowed to shift to ZUS in case they wish to accrue higher entitlements. This alternative is seemingly simple to introduce and is politically more appealing as it keeps KRUS in existence and allows farmers to continue contributing at the current, low level. There are various problems with these options, however. First, benefits significantly lower than today's KRUS minimum pensions are probably politically untenable and would lead to extensive take up of social assistance or to pressure to increase benefits. Second, the option to shift between KRUS and ZUS is difficult to accommodate by either system since both are pay-as-you- go financed. Third, as farmers age, the incentive to shift to ZUS becomes smaller. At the same time, younger people, when facing the choice between ZUS and KRUS may be inclined to opt for KRUS in their majority to maximize their consumption. Consequently, this option would not increase contribution 19 revenues, would lead to a gradual reduction of benefit which would, over time, lead to pressures to increase income support to KRUS members. 46. However, pensions already in payment and benefits to be paid to people close to retirement cannot become lower than without unification. If contribution histories accrued in KRUS are turned into notional defined contribution account balances, KRUS affiliates' balances could only finance an annuity lower than the minimum pension according to ZUS regulations. Assuming that the minimum pension is indeed the lowest socially and politically acceptable benefit, farmers would need to receive that, too, regardless of their past contributions. 47. In terms of contribution-liable income, the two schemes are different by design. If farmers become subject to the same contributory regime as ZUS members and are required to pay a uniform percentage of their income (but not less than the minimum contribution), the contribution-liable income of farmers would need to be legally defined. To do this, one can either positively list the items that count into contributory income or list the exceptions. The former approach has the drawback that failure to continuously update the list with the changes in the tax code leads to revenue losses. It is therefore advisable that the contribution-liable income be determined through defining the exemptions. 48. The higher contribution rates will have a pronounced effect on farmers' income. If the minimum contribution-liable income equals the statutory minimum wage, the contributions payable by farmers at that income level will grow 3 to 6 fold (using 2004 data), depending on whether farmers are required to pay only for pension insurance or all other social insurance benefits. For farmers with incomes at the economy-wide average wage, the increase will be even higher, between 8.5 and 17 fold. Although the rate of increase will grow with income, it is the low income farmers who will find it difficult to pay the higher premium. The reformed system should, therefore, exempt farmers (similarly to the unemployed in other sectors) with incomes lower than the statutory minimum wage from contributions. Farmers may also be given further easements in the form of exempting the first tier of agricultural income from the contribution-liable income base, although this provision should only be applied as a transitory arrangement. The negative welfare effects of the contribution increase can also be countered by the gradual approximation of current ZUS rates and by existing social transfers. 49. The proposed reform would gradually harmonize the regulatory framework of social insurance for farmers with the rules applying to other occupational groups and unify KRUS and ZUS into a single social security agency. Pension already in payment would remain unchanged, KRUS affiliates within 10-15 years of the statutory retirement age would either have the same rules apply to them as before or have their contribution histories translated into notional defined contribution account balances in a manner which leaves their expected old-age benefit unchanged. Younger workers would have their past contributions credited in a way which provides incentives to continuing compliance and, in the case of low earners, seeking better-paid formal sector jobs. Since the increase of the farmers' contribution burden can only happen gradually and younger KRUS members have a better opportunity of finding other jobs, the formula interpreting past contributions into account balances needs to be nonlinear in age. 20 rl N U e! 0 a 0 x a N N Introduction 50. This report analyzes the public expenditures of the education sector in Poland, with special interest in uncovering trends overtime, establishing international comparisons to benchmark Poland's fiscal efforts in education, and linking those expenditures to the outcomes of the education system. The report then presents a set of concrete policy options to improve the effectiveness and the efficiency of the education expenditures in Poland. 51. Perhaps the most important finding is that , while the expenditures in education are somewhat lower compared to those of OECD countries, the outcomes in coverage are impressive (with the exception of preschool education, where coverage is comparably low) and the quality outcomes are at the average of OECD countries, and have been steadily improving, as measured by the PISA exams. This suggests that the Polish education system is efficient and continues to improve. 52. Still, the report highlights areas where improvement in the quality of expenditures can be achieved. For example, while Poland's student population in primary and secondary education continues to fall, the number of teachers has not followed the same trend as the demographic shift in the student population. There are inflexibilities in the management and deployment of teachers that have made this adjustment difficult. Incentives for local governments, which manage basic education, to quickly adjust the teaching force to demographic changes by consolidating schools and classrooms would liberate significant resources to increase the availability of financial resources to be used in non-recurrent, quality- enhancing investments in the education sector. Some specific budget management suggestions as well as the need for developing indicative standards for service delivery as a tool to determine the overall cost for delivering the educational service are presented among the policy options that this paper outlines. 53. Another area where improvement of expenditures is possible is in higher education. Indeed, while the growing role of private institutions and financing in Poland's tertiary education has been an important factor in expanding coverage, adjustments in the public financing scheme for higher education may lead to a more equitable and fiscally sustainable system. The public financing strategy should be revised to change the allocation of subsidies from one that only considers merit to one that considers merit and means-testing to increase the equity of the subsidy allocation. Evidence suggests that higher- education subsidies (free education for "day students") are resulting in a regressive system that disproportionately benefits the wealthy. By establishing a mechanism through which the subsidies are allocated on the basis of need, the public system may simultaneously address issues of equity and quality. To facilitate this process and efficiently allocate the public subsidies, this paper outlines the policy option of expanding and improving the student loans/scholarship scheme. 54. The policy options outlined in the final paragraphs of the report are intended to spur discussion and analysis among policy-makers in Poland as they look into different possibilities to enhance the quality of expenditures and improve the outcomes in the education system. l4 Prepared by Alberto Rodriguez, Ph.D. (The World Bank) and Mikolaj Herbst, Ph.D. (University of Warsaw). I. Poland's Recent History of Reform The Polish education system has undergone profound changes during the last 20 years. Prior to 1990, the structure of the education system corresponded to the needs of a centrally planned economy, with a large number of specialized vocational schools subordinated to different sectoral ministries. Primary education, general secondary schools, and the universities were managed by the Ministry of Education. In 1990, a gradual process of education decentralization began, in line with the practice in other EU/OECD countries seeking to strengthen democratic participation and local decision-making, Initially, municipal governments (gminus) became responsible for maintaining and managing preschools. In 199 1, municipalities were also allowed to manage, on a voluntary basis, primary schools. However, in 1996, the management of the primary school network-both human resources and physical buildings- became a compulsory task of local governments at the municipal level. Another major step in the process of decentralization was taken in 1999, when management of secondary schools and most non-school educational tasks (for example, youth hospices, boarding-schools, reformatories), became the responsibility of a newly created county (powiat) administration (EU NTS 4 level). Annex 1 summarizes the current division of responsibilities (with respect to pre-primary, primary, and secondary education) between the central and local administrations. 55. Thus, decentralization of education in the 1990s brought significant reforms to the mechanisms through which preschool, primary, and secondary educational financing was handled. As the management of schools became the responsibility of local governments, they had to be endowed with revenue and other resources adequate to the challenge. Thefirst of those sources was a lump-sum grant (called the "education subvention") provided by the central government and distributed among local governments through a complex algorithm with a heavy emphasis on student enrollment. The second source is the right to keep in the local budget a substantial share of personal income tax (PIT) and corporate income tax (CIT) funds collected by the state from local citizens. Currently, 5 1 percent of PIT and 22 percent of CIT are the revenues of territorial self-government, with differentiated shares going to the budgets of municipalities, counties and regions (see Table 6) Table 6: Shares of Collected PIT and CIT Staying in the Budgets of Territorial Self- Governments Territorial self-governmenttier Share of PIT Share of CIT Municipal 39.34 percent 6.71 percent County 10.25 percent 1.4 percent "_ Regional - _-._. 1.6 percent 14.0 percent Source: Law on the revenues of territorial self-government 56. The third source is an "equalization subvention," also provided by the central government and distributed among local governments based on the local tax base levels. These three main sources of revenue (supplemented by targeted grants) are intended to cover all services delivered by the public sector at the local government level, of which educational services represents, on average, 34 percent of expenditures (as of 2006). 57. The "education subvention" is a two-stage funding procedure for financing education. First, the funds are transferred from the Ministry of Finance to 2,487 municipal governments and (starting in 1999), to over 300 counties and 16 regions. Second, local authorities are responsible for direct school financing, including the subvention and additional resources coming from their own local revenues. Local governments have total autonomy in their decisions on how to spend this money. In fact, they are even allowed to finance non-educational expenditures with the education subvention, since it enters local 24 budgets as an "unmarked" grant, and joins the "equalization subvention" and the local revenues with no specific earmarking. Adopting such a radically decentralized solution was widely seen as increasing the financial potential and autonomy of local governments under the then newly adopted decentralization, and critical to breaking old management structures relying on central administration. 58. In 1999, the Government of Poland also launched a critical reform with the intention of improving secondary education and realigning the supply of skills with the needs of the emerging knowledge economy. This reform focused on the integration of the content of general secondary and vocational education and on delaying the tracking into vocational training to expand students' exposure to general secondary curriculums. The purpose was to prepare young people to have flexible skills to enable them to be effective learners of new skills and knowledge. Since then, other efforts to improve the education system delivery have been implemented in the form of discrete and incremental policy improvements. This Public Expenditure Review (PER) discusses the reform efforts and their impact on the efficiency of use of the educational financial resources. 11. Overview and Outcomes of the Education System 59. The Polish education system (all tiers) includes about 36,000 schools and almost 8 million students. The primary and secondary tiers remain predominantly public, while a significant share of preschool and higher education is private (see Annex 3). Between 1990 and 2008, enrolment in primary and secondary education fell by 24 percent as a result of the demographic shift experienced in the region. However, during that period, Poland experienced a boom in enrollment in tertiary education. The number of students increased from 400,000 to almost 2 million, with an enormous growth in the number of private tertiary education institutions. Figure 6 plots the changes in enrollment starting in 1990. Figure 6: Change in Enrollment by Education Tier, 1990-2007 -total primarytsecondary -higher 9000 8000 7000 3000 Source: Based on Central Statistical Office data. 60. Currently, enrollment rates in education are high, with the exception of preprimarv education, where enrollment rates are below the level of comparable countries. During the 2007/08 school year, more than 95 percent of students of the relevant ages were enrolled in elementary and 25 secondary school^.'^ The net enrollment rates for different tiers vary from 96.8 percent for primary education, to 95.3 percent for the lower secondary tier, to 89.3 percent for the upper secondary tier. Of the student who complete lower secondary schooling, 94 percent continue their education at the upper secondary tier. Of citizens aged 20-24, 88.1 percent have satisfactorily completed at least upper secondary education (including a basic vocational path). This compares quite well with the 77.6 percent average among European Union (EU)-25 countries.I6 61. In preschool education, Poland has a low participation rate among children aged 3-5 (52 percent). More specifically, only 53 percent of 4 year olds" attend preschool, which compares poorly to 100 percent in Belgium, France, and Italy, and 90 percent in the Czech Republic, Hungary, and the U.K. 62. In hipher education, net participation rates have increased impressively and are estimated at 49.7 percent of 19-24 year-olds (while the gross rate was equal to 51.1 percent) in 2007/08,'8one of the highest in Europe. With about 500 tertiary students per each 10,000 habitants, Poland ranks sixth in the world regarding the accessibility of higher education. An enrollment boom in higher education over the last 20 years was a result of the expansion of public institutions (in which the enrollment increased from 400,000 to 1.2 million over the period 1990-2007) but also due to fast development of private sector in tertiary schooling. In 1992 there were only 18 private higher education institutions in Poland with 16,000 students enrolled. Today years, students in private institutions account for one third of total enrolment in tertiary education (now reaching 2 million) and the number of private institutions has increased to 324. 63. Adult participation in education and training, which measures the use and availability of lifelonp learning opportunities, is low. Eurostat data show that 4.7 percent of the population aged 25-64 participated in any form of education in 2006. This is about half the EU average and five times less than in Finland, Sweden, and the U.K., the leading countries in the development of lifelong learning. 64. The quality of basic and secondary education, as measured by the Program for International Student Assessment (PISA) results, is below the Organization for Economic Co- operation and Development (OECD) country average but is steadily improving, especially in reading outcomes. Between the 2000, 2003, and 2006 tests, Poland showed significant improvement, particularly in reading literacy. Indeed, in 2000, Poland's performance in reading was significantly below the OECD average. In 2006, however, its performance was significantly above the OECD average, and was one of the countries with the strongest improvement. The PISA 2006 outcomes revealed that 16.2 percent of Polish 15-year-old students were at or lower than level 1 (minimum reading literacy competency), whereas the OECD average was 20.3 percent. Conversely, 34.7 percent of Polish 15-year- old students obtained high achievement (levels 4 and 5 on the PISA scale), above the OECD average of 29.6 percent of 15-year-old students.'' Overall, in the 2006 reading test, Poland placed in the top 10 of OECD counties. One of the major reasons behind recent improvements in reading performance is the establishment of a lower secondary tier in the Polish education system (in 1999), which delayed professional tracking by one year. The creation of the lower secondary education tier also contributed to enhanced equity of the system by improving the school environment (physical conditions), particularly in rural areas. Is Net enrollment rates reported by the Central Statistical Office Regional Data Bank. 16 Eurydice database, Eurostat, Labor Force Survey. 17 Minist@of National Education, 2008 data. IB Central Statistical Office (GUS), "Szkoly wytsze i ich finanse w 2007 r.", GUS 2008. 19 PowerPoint presentation by the Ministry of National Education in Poland, based on the OECD Report on Results of PISA 2006. 26 65. PISA results in math and science are less strong. Poland ranks between 16thand 21" within 30 OECD countries in math and science, respectively. Of special note is the low performance of Polish students in the area of "identifying scientific issues"--the average of 483 points is below the OECD reference value (500). In general, it seems that Polish students perform better in acquiring formal knowledge than in applying it to solve problems.20Poland's scores in mathematics are also a matter of concern. Although the average of 495 points in 2006 is similar to the achievement of France, Norway, and the U.K., and well above the results of Italy, Spain, and the United States, it is still approximately 30 percent lower than in the leading European countries, such as Finland, the Netherlands, and Switzerland. 66. Nonetheless, the outcomes of the reforms in secondary education of the last 20 years have been encouraging with regard to their impact on realigning the skills of graduates to the needs of the labor market. First, there has been a shift in the profile of the type of secondary education pursued by students. In September 1990, the share of students receiving vocational training after leaving primary school was 78 percent, while only 12 percent were choosing general secondary school (lyceum). Half of those attending vocational schools were receiving only basic training, without a full secondary education certificate (rnatura). Currently, approximately 50 percent of secondary students choose the general education path. Specifically, 44 percent of the 16-1 8 age cohort attends general secondary school, 32.6 percent attend secondary vocational school, and 13 percent attend basic vocational school. Second, these reforms have been important in reshaping the skills of secondary education graduates and, coupled with a dynamic economic environment, probably contributed to a steady decrease in youth unemployment since 2004.2'These shifts point toward a slow but steady realignment of the education system to better respond to the needs of the labor market in the global economy. 67. Still, modernization of secondary curriculums needs to continue to make the skills taught more relevant to the demands of the labor market. As an example, according to Eurostat data, the average secondary school student in Poland is exposed to one foreign language, and about 10 percent of students are exposed to two. This is similar to the situation observed in Austria and the Czech Republic, but significantly behind France (where 50 percent of students are exposed to a second foreign language), Sweden (where 70 percent are exposed), and Finland and the Netherlands, where students leave secondary school having being taught two foreign languages. 68. The relevance of tertiary education is widely perceived as low by employers. Because of the educational boom, public universities experienced an enormous increase in demand for their services, but the developments on the supply side-quality of teaching and research-have not been dynamic enough. Indeed, the disciplines of study among tertiary students may not correspond to the challenges of the modem economy, because a strong emphasis is placed on social sciences, law, and management, contrasting with a weak pipeline of students in science, mathematics, and information technologies (Figure 7). 2oFollowing Paciorek, J.WiSniewski, P.Szyrmer, ,,Co wie i potrafi 15-letni uczeri - sprawdzila OECD w badaniach PISA", CASE 2007. ''Youth unemployment for the 15-24 cohort has diminished from 37 percent in 2005, to 30 percent in 2006, to 21.7 percent in 2007. 27 Figure 7: Disciplines of Study among Tertiary Students in Poland, 2007 arts&humanities social sciences, law, management @ science, mathematics, informatics r: medicine,social care engineering, architecture, construction agriculture,veterinary servicesand other Source: Based on Central Statistical Office data. 69. There are substantive differences between Poland and the EU-27countries when it comes to the disciplines chosen by students. The comparison presented in Table 7 highlights important issues. When compared to the student structure of the EU-27, Polish universities educate more students in the social sciences (48.3 percent of all students, compared to 37.1 percent in OECD countries). In turn, the shares of those studying computer sciences, mathematics, engineering, manufacturing, life sciences, physical sciences, and agriculture of the total number of students in Poland are significantly lower than in OECD countries. These differences between the structure of the Polish and OECD tertiary education market have been deepening in recent years as a result of the fast development of the private sector in Polish higher education. Private universities have played a crucial role in boosting tertiary enrollment, but most of them choose to operate in disciplines that do not require heavy investments in infrastructure and equipment but, at the same time, disciplines that may be less relevant to the needs of the labor market*2. Among students attending private universities, 57 percent study social sciences and only 1 percent study engineering or architecture. 22 This seeins to suggest that the Polish job market is increasing13 flooded with management students and has high demand and low supply for technical and science students. It is, therefore, interesting that private education students continue to pay for a less useful education in such numbers Perhaps this points out to two separate issues: (1) the labor market can still absorb management students - at least for now- and ( 2 ) a secondary education that does not prepare students for rigorous, technical higher education. Further analytic work on labor market demands can clarify this issue. 28 Table 7: Structure of Tertiary Students in Poland and OECD Countries by Field, 2006 Poland OECD Average Health and re Life Sciences, Physical Sciences, and Agriculture 5.1 6.9 Mathematics and Computer Science 4.8 5.2 Humanities, Arts, and Education 25.2 24.9 Social Sciences, Business, Law, and Services 48.3 37.1 Engineering, Manufacturing, and Construction 8.6 11.9 Not Known or Unspecified n 0.6 Source, OECD Education at a Glance 2008 70. The contribution of Polish higher-education institutions to global scientific knowledge is limited. Despite the country's academic potential, Polish researchers rarely publish their articles in internationally recognized journals. The "Report on the Intellectual Capital of Poland" shows that in 2003, Poland had 177 such publications per 1 million habitants, which accounted for only 50 percent of the indicator value for Greece, 30 percent of the Austrian score, and 15 percent of the result for Sweden, the leader among the researched countries.23 Moreover, the links between higher education institutions and the business sector remain very weak. Polish industry's contribution to funding research and development (R&D) activities is only 30 percent of total R&D expenditures, compared to an average of 62 percent in OECD counties.24 A survey conducted in 2006/07 by the World Economic Forum also shows that the cooperation of higher-education institutions with the business sector is underdeveloped. According to the synthetic indicator value, Poland ranked 56`h, achieving 46 percent of the maximum possible score, while leading countries-Finland, Germany, and Sweden-achieved close to 80 percent of the maximum score. 111. Financing the Education System 71. Overall, Poland is making important financial efforts to finance its education system. Table 9 summarizes the 2007 expenditures as a percentage of GDP per educational level, identifying public expenditures and calculating the per-student public expenditure. The table also compares the per-student public expenditures in Poland with those in OECD countries. As the table shows, Poland currently spends 5.25 percent of GDP on education; this is slightly higher than the 5.0 percent OECD country average. When compared with OECD countries, the per-student allocation (at all levels) is lower than average. However, Poland's per capita income level is the lowest of all 29 OECD member states, and less than half of the OECD median per capita income level in 2007. As spending on education tends to increase with per capita income, this suggests that comparisons with the OECD average have to be treated with caution. For example, Poland's public expenditure of 4.6 percent of GDP is higher than public expenditure on education in the emerging Asian economies of Korea, Malaysia, Thailand and Singapore. 23 "Report on the Intellectual Capital of Poland," Strategic Advisors Team, Prime Minister's Office, 2008. 24 OECD, Main Science and Technology Indicators, 2006. 29 72. Several noticeable patterns emerge from Table 9: Aside from tertiary education, the expenditure on education in Poland is overwhelmingly public. In 2004, the reported expenditure as a percentage of GDP was 6.1 percent, but it fell to 5.25 percent in 2007, mostly as a result of a significant increase in GDP over the period. Per-student expenditures in preschool are surprisingly high (representing 20 percent more than the per-student expenditure in general upper secondary education), revealing inefficiencies in service delivery at that level. Overall, per-student expenditures at all levels of education are significantly lower in Poland than in OECD countries. However, the outcomes in coverage (discussed in paras. 5-9 above) and in student learning (discussed in paras. 10 and 1 1 above) are generally comparable with those of the OECD average suggesting that the Polish education system is efficient: achieving OECD average outcomes with lower-than-OECD expenditures. Table 8: Public Spending on Education in Poland in Comparison with Emerging Economies (2007 or latest available) Public Education GNI Per Capita (current Spending (% of GDP) US$, Atlas method) Poland 4.6 9870 Malaysia 4.6 6420 Korea 4.4 21210 Thailand 3.9 2660 Chile 3.4 8160 Si ngapo re 3.4 31890 Source World Bank's Edstat and Word Development Indicator databases. 73. Variations in the trend of expenditures over time have been quite limited, as shown in Table 10 below. Still, when analyzing the trend of public expenditures between 2004 and 2007, a slight shift from primary and secondary towards higher education can be observed. 30 Table 9: Funding Level of the Polish Education System in 2007 (As a percent of GDP and in PPP IJWl 4,888 Whole System 5.25 4.6 0.65 8,031,681 3,494 7,527$ PPP = Purchasing power parity. USD = U.S. dollars. I Including all spending on education institutions and related activities. 2 Estimated based on unit cost in public preschools. 3 2005 data from OECD Education at a Glance 2008. 4 Only public expenditures. 5 Excluding preschools. Source: OECD Education at a Glance 2008, and calculations based on "Szkoiy wyzsze i ich finanse w 2007," Polish Central Statistical Ofrice (GUS), 2008. Ministiy of Finance data on local government budgets 2007, GUS Regional Data Bank 2007. Table 10: Total Spending on Education by Tier in 2004-2007 (billions of PLN, constant 2007 prices) 2004 2005 2006 2007 preschools (public) 3.96 3.75 4.00 4.25 as percent of Total 9.1 percent 8.6 percent 9.0 percent 9.6 percent primary (public) 13.16 12.86 13.04 12.59 as percent of Total 30.3 percent 29.4 percent 29.3 percent 28.4 percent secondary (public) 14.39 14.41 14.60 14.09 as percent ofTota1 33.2 percent 3 3 .O percent 32.8 percent 3 1.8 percent teritary (public) 9.59 10.29 10.40 10.84 as percent of Total 22.1 percent 23.5 percent 23.3 percent 24.4 percent teritiary (private) 2.28 2.40 2.54 2.60 as percent of Total 5.3 percent 5.5 percent 5.7 percent 5.8 percent ' Total 43.39 43.71 44.59 44.37 Source: Based on Central Statistical Office and Ministry of Finance data. 74. The characteristics and challenges of educational expenditure, however, are quite different for each level of the educational system. Thus, we will now disaggregate the characteristics and trends of expenditures in education according to each level in the system and identify the most critical challenges for each level. 31 Preschool Education Financing: Expenditures 75. Preschool education became the first educational task assumed by local governments in the process of decentralization. Starting in 1991, municipal authorities became responsible for the provision of early education, the maintenance of related infrastructure, and the employment and payment of teachers. Because municipalities (gminus) do not receive central resources for this purpose, the preschools are intended to be financed from their general local revenues. The initial period of municipal management over preschools was associated with a significant drop in accessibility. The number of preschools fell by almost 20 percent between 1991 and 1993 and the "capacity" (places offered) fell by 14 percent. However, since 1995, the accessibility of preschool gradually improved, and so did the financial resources dedicated by municipalities to this educational level (see Figure 8). Figure 8: Local Government Expenditures on Preschools (1999-2008) (in billions of PLN, constant 2007 prices) Source: Based on Ministry of Finance data 76. Expenditures on preschool education mimic the OECD average, but coverage is low. Municipalities spend over 4 billion PLN yearly on preschool education. Once the estimated private contributions are added, Poland invests about 0.45 percent of GDP in this level of education, which is identical to the OECD average (see Figure 9). However, as discussed in para. 6, the coverage is comparatively low, with a per-student allocation that represents 15 percent more than what is spent for each student in primary education and about 74 percent of what is spent for each student in tertiary education. This mismatch in financial inputs and their outcomes appears to be the result of a highly regulated subsector with inflexible service delivery standards. The result is an expensive program that municipalities can barely afford, thereby hampering expansion. 32 Figure 9: Expenditure on Preprimary Educational Institutions as a Percentage of GDP, Combined Public and Private Sources of Funds (2005)a 0.9 I 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 a. 2007 data for Poland Source: OECD Education at a Glance 2008. Preschool Education Financing: Equity 77. Access to preschool education is inequitable, because it is mostly limited to students in urban areas. Perhaps as a result of the rigid service guidelines, coupled with budgetary constraints among small, rural municipalities, only 18 percent of enrolled children are in rural areas. In fact, the net enrolment rate among children in rural communities is only 19 percent, compared to 62 percent among urban children. This difference in educational opportunity, so early in life, creates far different outlooks for the future of children in rural areas, who tend to come from poorer backgrounds, and the future of children in urban areas. Preschool Education Financing: Service Provision 78. Provision of preschool education has been mostly public, but there are signs that the trend may be changing. Almost 90 percent of pupils aged 3-5 attend public kindergartens, while 10 percent go to the non-public preschools established by private persons or nongovernmental organizations (NGOs). Until 2007, only one form of public (publicly funded) preprimary education was allowed by law. The operation of these officially approved preschools has been highly regulated, with strict norms imposed on teacher employment and wages-in accordance with the Teacher's Charte$'-and on curriculums, infrastructure, and organization. The non-public preschools are, by law, supported through municipal grants equivalent (in per-pupil terms) to at least 75 percent of local per-pupil spending on public pre- primary education. In 2008, as a result of pressure from NGOs and education analysts, the government introduced important modifications to the law, allowing the creation (by local governments, private 2s Teacher's Charter is the legal act, established by the Parliament, defining the rights and duties (including statutory working time and wages) of teachers in public preschools and schools. Some of these regulations apply also the employees of non-public education institutions. 33 providers, and NGOs) of "alternative preschool forms," with more flexible service delivery requirements. The new law mandates that local governments subsidize these alternative preschools at 40 to 50 percent (depending on legal form) of the per-pupil expenditures on "standard" preschools in a given municipality. Primary and Secondary Education Financing: Expenditures 79. Primary and secondary education financing together represent 72 percent of total public education financing. As a percentage of GDP, public expenditures on primary and secondary education are 3.3 percent, nearing the OECD country average of 3.5 percent (see Figure 10). In 2007, per-student expenditure for primary education was around 6,000 PLN, and for lower secondary education it was about 5,500 PLN. The spending on upper secondary education was also close to 5,500 PLN, with a .. significant difference between Table 11: Poland Per-Student Expenditures (billions of PLN, general schools (5,100 PLN) and constant 2007 prices) vocational schools (5,800 PLN). 1998 2005 2007 Compared to other OECD countries, the Per-student exPenditure on Per Capita GDP 2 1,246 27,181 30,634 prima& and secondiry education Per-student expenditure(PLN) 4,030 5,345 5,800 and is still low (see Table 11). As a percent ofper capita GDP 18,97 19,66 18,93 However, the trend has been one of percent percent percent increased funding: the Source, Based on Central Statistical Office and Ministry of Finance data. yearly per-student expenditure on primary and lower secondary education increased from 2,948 PLN in 1998, to 5,066 PLN in 2005, representing an approximately 34 percent real increase (the latter calculated based on 1998 constant prices),26 to about 5,800 PLN in 2007, reflecting (a) important efforts on the part of municipalities to invest in education, and (b) the slightly increasing education resource envelope distribution among a smaller number of students. It is noticeable, however, that this increase in expenditures has not been a result of an increase in fiscal effort towards the financing of education. As Table 1 1 indicates, as a percentage of GDP the per-student expenditures have been quite steady overtime. 80. As discussed already, the main financing mechanism for education is a grant provided to each local government called the "education subvention." Until 2003, the total amount for the education subvention was defined as a percentage of the national budget. Starting in 2004, the law was modified and it was determined that the yearly education subvention envelope would be at least the same nominal amount as the previous year. The complex yearly discussions to set the subvention envelope have increased the total amount of funding every year, as shown in Figure 1 1. 26 M.Jakubowski and LTopinska, "Impact of Decentralization on Public Service Delivery and Equity," Centre for Social and Economic Research, Warsaw, 2006. 34 Figure 10: Expenditures on Primary and Secondary Educational Institutions as a , Percentage of GDP, by Source of Funds (2005) a. 2007 data for Poland. Orange bar is public expenditures; green bar is private expenditures. Source: OECD, Education at a Glance 2008, Figure 11: Total Local Government Spending on Education Excluding Preschools (1999- 2008) (In billions of PLN, constant 2007 prices) 4s - - - - ~ t l ~ l < ~ t ~ ~ l lW l l l r m SUh 40 Told rl)n#lhlg 011 rdrlt moll %'l si! 25 to IS 10 5 II IW'3 2000 1001 )UUI 2003 LUUI 2005 2006 LOO? 2008 Source; Based on Ministry of Finance data. 81. The distribution of the education subvention is made through an algorithm, which is adjusted regularly. When distributing the subvention among local governments, the Ministry of Education is obliged to use an algorithm that takes into account enrolment plus other specific features of the school networks. The algorithm is the subject of constant political bargaining and trades, and some argue that it hardly reflects any policy goals or real cost differences among localities or schools. Yet, it is widely accepted and has played an important role in a more transparent distribution of resources. 35 Currently, it contains 4 1 different parameter^.^' Indeed, between 2000 and 2008, the algorithm underwent significant changes. The government embarked on a gradual departure from the quasi-voucher approach, under which all funding was distributed mostly on a per-student basis, toward a mixed approach, in which funding criteria include both students (their number and characteristics) and teachers (only their average qualifications; number is not taken into account) as the main cost drivers in the education system. Most changes to the algorithm over the years have been the result of adjustments to the dynamic political and financial conditions. 82. Because the education subvention envelope is largely defined by law, not policy, it does not operate as an effective tool to adjust educational expenditures to demographic changes. As a result, despite being a capitation-based grant, the subvention does not adjust expenditures to demographic changes and, thus, does not serve as a cost-containment mechanism. 83. The education subvention grant is the greatest contributor to educational expenditures. In 2007, Poland spent over 39 billion PLN on primary and secondary schools, and on some non-school educational tasks (for example, youth hospices, boarding-schools, reformatories). The subvention, about 28 billion PLN in 2007, represented about 72 Figure 12: Flow of Public Spending on Preschool, Primary, and percent of this public Secondary Education, 2007 Central Government expenditure on primary and secondary education, and about 65 percent of the total local expenditures on education (which includes preschool). Indeed, the education subvention is not Local authorities (municipal, county, and regional level) expected to cover all local governments' education expenditures. In fact, local I 1 38 2 bn I PLN governments are currently 4 bn PLN spending approximately 11.2 billion PLN of their own revenues, about 4 Preschools Primary. secondary schools, and non-school tasks billion PLN for preschool Source: World Bank Staff and 7.2 billion PLN to complement the subvention for primary and secondary education. Figure 12 explains the flow of the resources for preschool, primary, and secondary education. 84. The proportion of the contribution to primary and secondary education financing out of own revenues varies little among local governments. As can be concluded from Figure 12, local governments contribute about 35 percent of educational expenditures (primary and secondary, excluding preschool), and this contribution has been increasing over time. In 2006, only 27 municipalities (about 1 percent of all municipalities) managed to cover all primary and secondary education expenditures without exceeding the amount provided by the subvention. In contrast, for 46 percent of municipalities, the education subvention was too low to cover even the wage expenditures in education (including the preschool tier). While it might be expected that the wealthier local governments (because of higher fiscal 27 The criteria that carry the greatest weight in the distribution of the education subvention are: (a) location in the rural area or small city (population) under 5,000 (weight = 1-38),(b) number of disabled students (weight depending on the type and degree of disability), (c) number of students in vocational schools (weight = 1.20), and (d) average teacher qualifications (multiplier mechanism). 36 capacity) would contribute significantly more to education expenditures to supplement the education subvention, the data reveal that the local fiscal effort is, in general, quite similar between wealthy and poor municipalities and that, on average, the subvention just about covers the wage bill (Table 12). This seems to be the effect of strong fiscal equalization mechanisms included in both the "education subvention" (indirectly, through rural coefficients) and the "equalizing subvention". The effects of these subventions on equity will be addressed later in this paper. Table 12: Local Expenditures (without preschools) as a Percentage of the Education Subvention, Considering Wealth of Municipalities, 2006 Quartile - Own Average Own Revenues Education Spending as Education Wage Bill as Revenues Per Capita Per Capita Percent of Education Percent of Education (In PLN) Subvention ( percent) Subvention ( percent) 1 391 143 101 2 564 142 100 3 764 143 100 4 1499 158 107 Total 804 146 102 Source: Based on Central Statistical Office and Ministry of Finance data. 85. The structure of education expenditures in primary and secondary education has slightly shifted over time. As Table 13 indicates, the recurrent expenditures constitute approximately 94 percent of the total cost, while the investment takes 6 percent, and the slight changes over time are erratic with no specific trend. The wages (including both teacher and non-teacher remuneration) represent about 70 percent of educational expenditures in both primary and secondary education and-just as with total expenditures-the wage bill has been increasing despite the fall in student enrollment (Figure 13). Moreover, the wage-bill effect has no relationship to the type of municipality. That is, rural and urban, and rich and poor municipalities are similarly affected by the impact of the wage bill (Table 14). This suggests that the increases in total expenditures are a result of the proportional increase in the wage bill and in non-wage expenditures. Why is the wage bill increasing despite expectations to the contrary? The following section addresses this question. Table 13: Structure of Local Expenditures on Primary and Secondary Education 200 1 2002 2003 2004 2005 2006 2007 (percent) (percent) (percent) (percent) (percent) (percent) (percent) Share of Recurrent Expenditures in Total 92.2 93 .O 94.9 94.7 94.5 93.6 94.1 Local Spending on Education Share of Wage Expenditures in Total 71.1 68.3 72.2 71.6 69.9 65.3 69.7 Local Spending on Education Share of Investment Expenditures in Total 7.8 7.0 5.1 5.3 5.5 6.4 5.9 Local Spending on Education Source: Based on Ministry of Finance data. 37 Figure 13: Total Spending Compared to Wage Bill on Primary and Secondary Education, 2002-07 (billions of PLN, constant 2007 prices) -Tibtrcl /)I 11)irtt \ -1rital ~t~ 1 391 987 1,362 1,966 2 564 1,144 1,425 1,998 3 764 1,295 1,460 1,992 4 1,499 1,986 2,025 2,5 15 Average 804 1,353 1,568 2,118 * "others" includes targeted grants, which play an important role in poor municipalities, and a small subvention (on average, I O PLN per capita) based on a horizontal mechanism. The richest local governments pay the 15 percent tax on their tax revenues above 150 percent of the national average, and 30 percent tax on the revenues above 300 percent. Then the subvention is distributed according to "social needs," mostly based on housing benefits needs. Source; Based on Central Statistical Office and Ministry of Finance data. Primary and Secondary Education Financing: Summary 93. The above-mentioned financing strategy has resulted in an equitable and generally transparent distribution of resources; however, the strategy does not promote efficiency in expenditures. As a result, there has been no decline in the education wage bill expenditures despite a sharp drop in student enrollment. The subvention system (education plus equalization subventions) are generally effective in providing financial support to the local governments that need it in order to deliver the education servic,e. This explains why the fiscal effort among municipalities-rich and poor- to complement the subvention does not depart significantly from the average local revenue financing (35 percent of total education expenditures). Moreover, the capitation nature of the algorithm of the education subvention helps in providing some transparency to the distribution of resources, despite the complexity of the algorithm itself. However, the lack of incentives for efficiency should be a matter of concern. At the center of this issue are: (a) the presence by Law of a floor for the total envelope of the subvention, which effectively limits the subvention's ability to adjust to demographics, and (b) the difficulties adjusting the teaching force to the falling number of students. The effect of this is that the per-student allocation of the subvention (and therefore the expenditure in primary and secondary education) is increasing. This would be a welcome development for Poland, a country with comparably low per-student expenditures. Unfortunately, the lion's share of these resources seems to be financing inefficiencies nested in the poor deployment and management of the teaching force. Indeed, with a falling number of students, the wage bill continues to increase and its participation over the total expenditures continues to be relatively constant. Tertiary Education Financing: Expenditures 94. Public expenditure on tertiary education in Poland represents 0.9 percent of GDP, lower than the 1.1 percent average of OECD countries. Once private investment is factored in, Poland spends about 1.4 percent of GDP (2007), which is close to the 2005 OECD country average (see Figure 17). Clearly, private provision and financing of tertiary education in Poland has been a critical element in explaining the rapid enrollment growth and the impressive enrollment patterns of the country. Tertiary education public spending represents about 19 percent of total public spending on education, comparable 43 to the 21 percent average for OECD countries. Moreover, the ratio of tertiary to primary per-student public spending is 1.55, lower than the 1.9 average for EU-15 countries, suggesting that expenditure within subsectors is progressive, because it does not overspend on tertiary education at the expense of lower levels of education. This indicator, incidentally, would also suggest some level of under-spending in tertiary education. Indeed, in 2007, the per-student expenditure on tertiary institutions in Poland was 9,422 PLN (equivalent to PPP EURO 4,932) which represented about 59 percent of the average 2005 public expenditure in EU-I 5 countries. Figure 17: Expenditure on Tertiary Education as a Percentage of GDP, by Source of Fund (2005)a 3.0 d Public 1 Private a. 2007 data for Poland. Note: Orange bars are public expenditures: green bars are private expenditures. Source; OECD. Education at a Glance 2008. 95. An algorithm is also used to distribute public resources to public higher-education institutions. Public tertiary institutions receive funds from the Ministry of Science and Higher Education through an algorithm that takes into account variables such as the number of regular students (evening and weekend students excluded), the structure of teaching staff (professors, assistant professors, others), the research activity (number of projects in the previous year), the number of Ph.D. and professor titles, and the intensity of international exchange. The total envelope of resources to be distributed under this algorithm was 7.5 billion PLN, representing about 80 percent of the resources that the Ministry of Science and Higher Education spends on higher-education activities. Overall, it has been observed over the last few years that the allocation to a given institution from one year to another varies little. 96. Total public expenditures by higher-education institutions in Poland exceeded 1024 billion PLN (2007), including spending generated by revenues from both budget revenues and some student fees. Figure 1 8 summarizes the private/public expenditures in higher education. Private financing is high and growing. With regard to public expenditures, aside from the resources distributed through the algorithm, additional public spending from the budget is a result of (a) direct assistance to students for educational materials (1.5 billion PLN), which is also distributed to universities through a per-capita algorithm and which is allocated by a committee in each institution (private and public) and paid in cash to students; (b) subsidies for interest rates on student loans (0.15 billion PLN); and (c) other expenditures including support for specialized universities and administrative costs. Public funds go almost exclusively to public institutions (more than 99 percent). However, approximately 20 percent of the budgets of public 44 institutions comes from student fees paid by weekend and evening students. In contrast, private tertiary schools are financed almost exclusively by student fees. Figure 18: Higher-education Funding in Poland (2004-07), (PLN billion, constant 2007 Drices) 16 . 12 10 m Ptivateexpencliturrr 011 8 tertiary Irh o o II ed iic a ti o 11 n P u h h expenditiit es on 6 1~ tertiat y ~ c 1 ~ 0 0education 4 2 . 0 LOO4 2005 LO06 2007 Note: The numbers in the figure refer to the revenues of tertiary schools. They do not include the expenditures in the higher-education sector that do not enter Higher Education Institution's budgets. Source: Central Statistical Office. Tertiary Education Financing: Equity 97. Higher education in Poland presents an equity challenge. With the expansion of higher education and the growth in private service delivery, the competition for the "day spaces" that are free of charge in public universities became fierce. In 2003, for example, Warsaw University could take only 1 of 10 applicants, with even worse ratios in fields such as economics, management, biotechnology, and law. Students prepare assiduously for this selection process by taking private lessons. Indeed, a pilot survey at Warsaw University demonstrated that 63 percent attended paid courses or took private lessons to prepare for admission. The result is not surprising: students from more affluent, urban families are overrepresented in the group that benefits from the full subsidy. Unfortunately, few data exist about this, but a 1999 study based on a survey of second-year students at Warsaw University revealed that students whose parents had higher education had better access to the "day spaces" than those whose parents did not. On average, 64 percent of survey participants paid tuition fees, but when disaggregated, if a student's father had higher education, this percentage fell to 48 percent, and if a student's father had only primary education, the percentage was 78 p e r ~ e n t . ~ ' enrollment rates in higher education by household The consumption quintile confirm this inequitable pattern, since the rate was 63 percent for the highest quintile in 2007 compared to 37 percent for the lowest quintile (Figure 19). Moreover, while the enrollment rate within the highest quintile has increased significantly since the 2003 Household Survey (from 53 percent), the value for low-income families remained at the same level, which indicates that the inequality of access to higher education rises over time. 35 Information obtained from "Tertiary Education in Poland," EIB and the World Bank, July 2004. 45 Figure 19: Share of 19-26-years-olds Not Attending any School, by Quintile of Household Monthly Consumption Per Capita 70% 60% 50% 40% 30% 20% 10% 0% 1 2 3 4 5 Source: 2007 Household Survey 98. Poland created the Student Loan and Credit Fund, in 1998, to help students pay tuition and living expenses. This fund is operated by the Bank of the National Economy (BGK) and loans are granted to students through commercial banks. The fund covers the costs of interest payments during studies and the administrative costs on behalf of the students. Loans are available for low-income students36(regardless of type of institution attended), provided they are under age 25 when starting their studies. Students can borrow 400 PLN or 600 PLN per month during the 10 months of the academic year, for a maximum of six years for undergraduate students and 10 years for graduate students. Repayment begins two years after completing studies and lasts twice as long as the duration of the loan disbursements. After studies are completed, the interest rates are paid by students at about 50 percent of the rediscount rate of the national Bank of Poland (between 5.25 percent and 6.25 percent in 2008). The debt is discounted by 20 percent for students graduating among the top 5 percent of their class, the debt is discounted on a case-by-case basis for students with difficult living conditions (that is, accident, illness), and the debt is fully eliminated for students who die or suffer permanent disability. 99. However, the impact of the loan program is limited, since few students actually use it. Indeed, during the 2007108 academic year, only 16,423 new loans were approved, bringing the total number of active loans to 209,383 (representing 10.5 percent of coverage among the active student population). Of the new loans, 76 percent went to full-time students and 81 percent went to students attending public universities. Officials believe the loan program does not have wider coverage because the involved commercial banks require physical collateral to approve the loans, thus making this option less viable for poorer students. Limited access to loans may be considered to be one of the reasons for low enrollment rates in higher education among individuals aged 19-26 from low-income families. Household Survey data show that only 9 percent of private-day-school students come from households within the first quintile of per capita consumption, while 40 percent come from households from the last quintile. 36 For the 2008109 school year, this was defined as students from families with less the 2,500 PLN per capita of income. 46 Tertiary Education Financing: Summary 100. The growing role of private institutions and financing in Poland's tertiary education has been an important factor in expanding coverage without overburdening public expenditures, but also raises new issues of quality and relevance. Nevertheless, within the public financing strategy there is a need to revise the allocation of subsidies from one that only considers merit to one that considers merit and means-testing to increase the equity of the subsidy allocation. This implies some politically difficult decisions, such as establishing tuition charges for all students in public tertiary education institutions that can be supported by a stronger and more effective student loan program, and perhaps more important, that can also be condoned (or partially condoned) through consideration of merit and financial need of the student. Countries in Western Europe (i.e. Netherlands), in Asia (i.e. China) and more recently in Latin America (Le. Colombia, Chile) have successfully made this shift in higher education financing. IV. Summary and Policy Options for the Unfinished Agenda 101. Enrollment in preschool education is lower than expected for the amount of resources spent. Preschool education in Poland has low coverage (only 53 percent of 4 year olds attend) despite representing an investment (0.45 percent of GDP) equal to OECD countries, and using 75 percent of the comparable PPP per-student expenditures. This mismatch in outcomes is the result of a highly regulated subsector with inflexible service delivery expectations. The result is an expensive program that can barely be afforded by municipalities, thus hampering expansion. However, it is widely recognized that expanding preschool education will have important effects on equity and will increase the school readiness of students, potentially improving student learning throughout their educational experience. Investing in the early years of education (pre-primary) has been widely studied as one of the investments with highest returns (according to the High/Scope Perry study in the United States, USD 17 in benefits over the lifetime of a student for every USD 1 spent in quality pre-primary programs). Policy options to achieve these objectives include: 0 Diversify service delivery for 3-4 year olds: While a first step was taken in 2008 to increase the opportunity for service delivery through social and community organizations, the efforts to diversify service delivery and create a flexible but effective framework for monitoring service delivery should be deepened. This implies focusing on diversifying service delivery methods (non-public institutions are in the best position to explore these), with lower per-capita costs, and creative staffing and management arrangements. The government should continue to gradually support the expansion of preschool (for 3- and 4- year-old children) through increased flexibility for non-public institutions to flourish. This can be a costly endeavor, and will likely need to be implemented overtime. However, for the purpose of quantifying the effort, we estimate that if all 4 year olds that are out of school in 2009 were to be enrolled in a pre-primary education program during 20 10, the cost for Poland would represent an additional PLN 1 billion, estimating per-capita savings of 25 percent through the use of alternative delivery mechanisms (Le., non-government organizations). Enhance participation gradually for 5 year olds: The government should strengthen the schooling of 5 year olds by linking their classrooms to primary schools and by developing age-appropriate curriculums to better respond to the needs of these children. By linking to 47 primary schools, new shared service delivery mechanisms can be developed that will allow wider accessibility, even maintaining the existing teaching force in primary schools. This is especially crucial in rural areas, where participation is low and primary schools small. 102. Efficiency and quality of instruction of primary and lower secondary education can be further improved by addressing challenges in teacher management. While a reduction of 25 percent of primary schools since 1996 reflects important adjustments in the school network to respond to falling demand, these reductions have not been as significant among the teaching force, whose wage bill represents 70 percent of educational expenditures. Moreover, as discussed, teacher salary structures present negative incentives to quality of teaching. Therefore, it is important to adjust both the size of the teaching force and the gradient of teacher salaries. Policy options to adjust the teaching force to the size of the demand include: Set indicative standards on class size to be applied mainly by urban local governments, and determine the per-student allocation of the subvention based on those standard average class sizes. The low student:teacher ratios in Poland (1 1 : 1 in primary and 12.3:1 in secondary) suggest that there may be ample space for this policy adjustment. The savings here can be substantial: in a simulation with expenditures for 2006, an increase from 23 students to 25 students per class in urban districts would have resulted in savings of about 391 million PLN that year. This calculation is explained in Annex 4. 0 Establish a central fund for severance packages to be used as an incentive for local authorities to rationalize their teaching force. For example, a fund generated with 0.5 percent of reserves from the education subvention would cover the severance packages of about 11,000 teachers every year, generating yearly savings of about 280 million PLN in the wage bill, which represents cumulative savings of 4.2 billion PLN over five yearse3' 0 Remove the established floor for the subvention envelope, and instead determine total funding based on the development of indicative standards for service delivery (class size - as indicated above, teacher ratios, teaching time, wages, non-recurrent educational inputs, facilities, etc.). These indicative standards should be revised at least every five years in a transparent and technical manner. It may also be advisable to separate the budget lines of the education subvention (which is meant for administering and maintaining the operations of the system) from the budget lines for investments in quality improvement. While the distribution of both can be done under the same per-capita guidelines of the subvention, separating the budget lines will avoid an undesirable trade-off between quality and efficiency. Overall, these indicative standards would be used to calculate the total funding envelope, which then is distributed through the education subvention's funding formula. It must be emphasized, however, that developing these standards as a tool to identify the total funding envelope is not intended to interfere with the freedom and obligations of local governments for the quality of education, which are clear and should be maintained. 0 Further expand the weekly hours of work of teachers to rationalize the teaching workforce. The Ministry of Finance estimates that with an increase of six hours per week (that is, from 18 37 The cost of a severance package is six months' base salary of the teacher. The severance is paid only to teachers that have reached one of the two upper tiers of the professional hierarchy (expert teachers or nominated teachers). These two tiers account for 80 percent of teachers. Given that the average base wage for expert and nominated teachers is 1,991 PLN (1,037 PPP euro), the average severance pay is estimated at 1 1,946 PLN (6,222 PPP euro). 48 hourdweek to 24 hourdweek of classroom work) nearly 70,000 teaching posts could be rationalized, equivalent to about 14 percent of the workforce (about 1.6 billion PLN per year). A policy option to adjust the gradient of teachers' salaries is: 0 Increase entry salaries but reduce the tenure-related salary increases throughout a teacher's career over time to attract an energetic and well-prepared teaching force that may leave the profession earlier and be replaced by a new cohort. This rapid rotation will convert the teaching force into a younger cadre, open to innovation and circulating more fluidly between schools and industry throughout their careers. This is a budget-neutral option. A transition toward this policy option was initiated in 2009, by providing higher salary increases for junior teachers (34 percent) than for expert teachers (1 0 percent). 103. Skills attained in secondary education should continue to shift toward further alignment with those demanded by the labor market. Building on early successes in secondary education reform, the government should continue to strengthen and expand the teaching of general secondary education skills among all students. Policy options for achieving further alignment with labor market demands include: 0 Further delaying tracking into specialized training to lothgrade. Aside from the educational merits of this option, some fiscal savings can be expected because in calculating the education subvention, the delivery costs for general education are lower than for diversified training. 0 Continue reform of the upper secondary education curriculum to better integrate vocational training with general education. This implies a further consolidation of tracking options within secondary schools to focus more on. the development of a family of broad skills instead of on specific occupation training. 0 Establish stronger and more stable national examinations, including revamping the matura test, and including mandatory math testing at the exit of secondary education. Starting in 2002, Poland developed a National Assessment system that yearly tests all students in the last grade of elementary school and the last grade of lower secondary school. These two tests are complemented by the matura, which tests all aspiring candidates for entrance into higher education. Both the National Assessment tests and the matura have been important in introducing a culture of information and accountability into the system, but clearly, they need to be further strengthened from a technical perspective to ensure their validity and reliability and, ultimately, their long-term sustainability. 104. Adjustments in the financing scheme for higher education may lead to a more equitable and fiscally sustainable system. Evidence suggests that higher-education subsidies (free education for "day students") are resulting in a regressive system that disproportionately benefits the wealthy. By establishing a mechanism through which the subsidies are allocated on the basis of need, the public system may simultaneously address issues of equity and quality. Policy options for establishing a more equitable and fiscally sustainable higher-education system include: Expand student loan opportunities. The current student loan program is small, in part due to some of the loan conditions (real collateral, for example). International experience suggests that revising the loan scheme and making it more flexible, and affordable (which includes some 49 embedded subsidies), will expand its reach with little risk to repayment given that the students' human capital and associated future income are perhaps the strongest guarantee of repayment. As a key element of higher-education reform, revise the tuition policies of public universities so that those who can pay will pay independently of the type of program (day, evening, or weekend) attended. While this may be a politically difficult transition, it is possible to enact the following intermediate policies: o Mandate that students who prolong their studies over and above a standard period of five years will have to pay for the additional study period. o Mandate that students will benefit from only one fully subsidized degree, and those pursuing additional degrees will have to pay for their studies. 105. The relevance of higher-education programs can be improved and the quality-assurance mechanisms must .be strengthened in light of the rapid expansion of private service delivery. The government and private sector providers can complement and support each other given their shared goals of expanding access to quality and highly relevant higher-education training for students. It is important, therefore, that the higher-education reform currently under discussion establishes mechanisms to address quality and relevance for both public and private higher-education institutions. Policy options to address quality and relevance include: 0 Establish a quality-assurance framework. The government should consider strengthening the existing accreditation system by linking it to international accreditation agencies and strengthening the consequences for a poor evaluation of a program or institution. This is a critical instrument in providing a quality improvement road map for public institutions, but also for private institutions. The accreditation system does not need to be controlled directly by the government, it may also be managed by an independent organization, respecting the autonomy of higher education institutions. 0 Adjust the mechanisms through which financing flows to higher-education institutions, including research institutes. The algorithm for the distribution of public resources should be strengthened to provide stronger incentives for outcomes on technological inventions and patents. Moreover, all resources for R&D should be assigned on a competitive basis. 0 Centralize a national one-point access to the web-based labor market observatories created at regional levels. Several regional governments have created web-based labor market observatories to assist in student decision-making and to broadly inform both providers and users of higher education. These are important because they are able to address information failures that result from a rapidly changing labor market. Creating a one-point access that leads to the regional observatories will accelerate their use and foster stronger labor market fluidity. Accelerate the development of a lifelong learning framework, strategy, and plan. Lifelong learning is a critical instrument to ensure that skills of those employees that are already in the job market are constantly renewed and updated to the rapidly changing demands of the knowledge economy. However, Poland appears to be lagging behind in tackling this challenge as a result of institutional coordination deficiencies and unclear roles and responsibilities among different Government agencies. 50 a 9 vocational I I I - = N o t applicable , a. Without schools for students with special educational needs Source; Central Statistical Office. Annex 3: Calculating Savings from Minimum Indicative Class Size Standards in Poland 106. The average class size in urban primary schools in Poland is 23 (2006 data). At the same time, regression analysis shows that in urban municipalities the elasticity of per-student spending to the average class size is close to 1, This means that, statistically, a 10 percent increase in the average class size is associated with an identical drop in per-student expenditures. Estimation of Per-student Municipal Spending on Primary Schools, 2006 Variable All Municipalities 65 Large Cities Rural Municipalities Intercept 6.71 (37.2) 11.3 (6.29) 6.37 (17.1) In OW, 0.09 (16.9) 0.15 (4.50) 0.08 (12.2) In SO, 0.36 (20.4) -0.10 (-0.53) 0.39 (9.22) 0.01 (4.25) 0.12 (2.77) 0.01 (2.22) In DYP4 In YO, -0.58 (-39.6) -0.98 (-6.19) -0.54 (-3 1.2) N 2,471 65 1582 -- R2A I--"-._-- 0.58 ~~ 0.65 0,42 -- Endogenous variable: log (recurrent spending on primary schools per student in 2006). Parameters significant at 5 percent level are bolded. LnWO is the natural logarithm of the average class size. 52 107. Given that Polish cities (urban municipalities) spent around 4.5 billion PLN on primary schools in 2006, if we assume that the indicative standard of the class size was set at 25 (compared to the actual average class size of 23), and assuming that local governments follow this standard, a decrease in total spending by 391 million PLN in 2006 could be expected. 108. The analysis becomes more complex in the case of rural areas, in which the elasticity of spending to the class size (around 0.54) is lower than in cities, revealing some inefficiency in the distribution of the education subvention. If the indicative standard for primary school class size was set at 20 (with 17 being the average in 2006), and if rural municipalities adjust their network to this standard, the estimated drop in total spending would be of 475 million PLN. 53 HEALTH38 Overview 109. The health policy agenda of the current government was passed by Parliament in late 2008. The agenda includes definition of a basic care package to be financed by NHF (finally signed into law in June 2009), introduction of private insurance to finance services not covered by that package, the transformation of hospitals organized as independent public institutions (SPZOZ) into joint stock commercial code companies (NZOZ), creation of a single Patient's Ombudsman, a new legal framework for accreditation, and a new legal framework for health professionals. Implementation of this agenda has been slowed by Presidential vetoes of most of it. 1 10. The Polish health system is widely, characterized as achieving poor health outcomes, suffering from an overload of hospitals and hospital beds, low public spending, inequitable and impoverishing because of high out-of-pocket spending, corrupt because of side payments for care, providing unsatisfactory services, and-consequently-characterized by low patient satisfaction. The conventional wisdom may still be accurate to an extent, but this review of the health system suggests that the Polish authorities have had considerable success in addressing every one of these problems. The criticisms to some degree describe a system that no longer exists. Through persistent tinkering and efforts to fix these problems, the health care system and the financing of it have been transformed. The new challenges resemble to some degree the old ones, but it is time to recognize that some things have been fixed, and future challenges are more pressing than the old battles. Let us take each of these items in turn: 0 Poor health outcomes: Poland has made progress in improving life expectancy and catching up to its western neighbors. Today, well over 80 percent of the remaining adult mortality deficit with the original European Union States is due to cardiovascular disease. A concerted effort in this area would be a sensible public health priority going forward. 0 Suffering from an overload of hospitals and hospital beds: Poland is now among the lowest of the EU States in hospital beds per 100,000 population. There are still too many beds because the EU has too many beds, and possibly there are too many hospital employees, but the more important issues now are more detailed: the distribution of beds; incentives for and use of an appropriate mix of inpatient, outpatient, and day patient services; and making sure surviving hospitals are economically efficient and proficiently using modern medical technology. The government is making strides to rationalize hospital incentives, to improve payment methods for them, and to solve their debt and governance problems. Poland has now gotten hospital expenditures down to about 33 percent of spending, similar to OECD countries. Local government and university owners of hospitals have done less well to development networks of services to improve patient care, and there are still major problems, but this is excellent progress. '' Mukesh Chawla and Charles Griffin wrote this chapter with contributions by Adam Kozierkiewicz on details of the corporatization program, DRGs, and other aspects of health policy. Monika Bazyl analyzed the household budget survey and opinion poll data used in the household expenditure/equity section. Low public spending: Public spending per capita in constant 2008 prices rose 35 percent from PLN1,293 in 2005 to PLN1,745 in 2008. By 2008, after a surge in spending, health care exceeded 7 percent of GDP for the first time. The share of overall spending for health care services through NHF has steadily increased. Inequitable and impoverishing because of high out-of-pocket spending: Public spending has fairly consistently accounted for 70 percent of all spending in recent years, with about 30 percent coming from out-of-pocket payments. France has driven out-of-pocket spending to about 12 percent of the total, and the Netherlands, to about 18 percent, but at much higher levels of per capita income and fiscal effort for health. With virtually 100 percent coverage of the population by NHF and very low out-of-pocket expenditures for hospital care, Poland has achieved a high level of financial protection for catastrophic costs. Between 2005 and 2007, our analysis of the Household Budget Survey suggests that out-of-pocket payments fell even as incomes were rising quickly. There are problem areas: households consider pharmaceuticals to be excessively costly and out of reach; the elderly in particular merit attention because of their high drug expenditures. Corrupt because of side payments for care: Although corruption is intrinsically hard to measure, available evidence suggests that the incidence of gifts and under-the-table payments has dropped steadily over the past 20 years. A 2006 HBS survey found almost negligible reporting of payments in NHF inpatient facilities for care; from 10 to 20 percent of each income group gave a gift of some sort to their doctors. There has been a steady drop since 2000 in people reporting financial difficulties in using health services. 0 Providing unsatisfactory services, and-consequently-characterized by low patient satisfaction: But the 2006 HBS data suggest a more nuanced evaluation by the public. HBS data show high levels of satisfaction (responding "good or very good") with access to general practitioners and hospitals, overall medical services, and the system for purchasing medicines. Problem areas for patients are dentists and specialists, which is reinforced by higher reported use of non-NHF providers in these areas even by the poor mainly because of perceived higher quality or accessibility and courtesy. 1 1 1, We find that the challenges for health policy are narrower and more subtle than they were even five years ago. One of the key missing ingredients is a stronger analytic and evaluative capacity to use new sources of data and policy instruments to improve incentives in the system, monitor the results, and make further adjustments to improve results. Introduction 112. Since the introduction of universal health insurance in 1999, the Polish health care system has undergone several reforms in the production, delivery and management of the health system directed at protecting and improving the health status of Polish citizens. On the delivery side, the Semashko-style system was replaced with a health system staffed by family doctors and general practitioners providing outpatient care and a vast loosely-defined network of hospitals. On the financing side, the budgetary funded system was replaced by a social insurance system which provided universal health insurance to all citizens of the country through 16 regional health insurance funds, reorganized in 2003 into a single centralized insurance agency. However, substantial scope for further reforms remains, especially aimed at managing cost pressures, encouraging more efficient use of resources, improving equity of access, and consolidating the financial strength of the sector to speed the improvement in health outcomes to converge with the pre-2004 European Union states. 56 113. To further improve health outcomes, Poland faces several challenges: (i) achieving higher efficiency in resource allocation and management; (ii) improving service delivery; and (iii) combining these reforms with better assessment and analytical tools to guide spending decisions. The health sector is well placed to protect the vulnerable during the economic downturn because of full coverage of the population. However, Poland has a persistent problem of high out-of-pocket spending. Although progress has been made in addressing this issue, challenges remain, particularly for specific groups and services. 114. The efficiency agenda includes both debt management and provider payment reform. With important exceptions, hospitals have consistently accumulated public debts (failure to pay social security and other taxes) and civil debts (borrowings and receivables owed to suppliers), quite apart from debt for investments. The current government is aiming to solve this problem permanently through an initiative to corporatize hospitals under commercial law, but this will be done initially using a voluntary approach. Close monitoring will be required for hospitals that are unlikely to be corporatized but for which debt accumulation has been a persistent problem, as well as for hospitals affiliated with universities. In addition, Poland has implemented Diagnosis Related Groups (DRGs) as a cost accounting and payment mechanism in all hospitals in 2009. This represents a major step forward in improving the reimbursement system. Lastly, the corporatization of hospitals allows the MoH and NHF to take steps to improve the accreditation and regulatory framework under which hospitals and clinics operate. The MoH has drafted the enabling regulations and will take steps to develop this framework, which will create a tool that can be used in conjunction with DRG-based reimbursements to raise the cost-effectiveness and quality of health services provided. Other possible reform areas include the introduction of a multiple competitive private insurance system and the expansion of long-term care services, though these are medium-term challenges. 1 15. This chapter analyzes the structure of health sector expenditures and identifies key reform issues and options, and is organized as follows. Section 1 presents a short account of health status of the people of Poland. Details regarding health expenditures are discussed in Section 2-3 followed in Section 4 by a description and analysis of recent health reforms in the health sector in Poland. Sections 5 to 7 highlight issues related to efficiency and equity. Section 8 concludes. A statistical annex is placed at the end of the chapter. I. Health Status 116. The health status of the people of Poland compares favorably with the health status of neighboring EU countries though not as favorably with the health status of pre-2004 EU countries (see Table 17). Standardized death rates (ratio of total deaths to total population) - 842.17 per 100,000, all causes - are lower than post-2004 EU average, but are over 40 percent higher than the EU average. Life expectancy in Poland of 75.4 years is also higher than the post-2004 average, but lower by 5 years when compared to the pre-2004 EU average. Across almost all indicators, Poland does better than the post-2004 EU average, but continues to lag behind pre-2004 averages. 57 Table 17: Health Status Indicators, Poland and Comparator Countries (2007 or latest available year) Life Reduction Infant SDR all SDR, SDR, SDR, Tuberculosis AIDS expectancy of life deaths causes, chronic diabetes, disease incidence per incidence at birth, in expectancy per all ages, liver all ages, of blood 100,000 Per years through 1000 per disease per & 100,000 death live 100,000 and 100,000 blood- before 65 births cirrhosis, I forming years all ages organs, Per all ages 100,000 Per 100,000 EU preZ004 80.27 4.17 3.96 580.73 11.03 13.35 1.53 9.15 1.6 EU 79.09 4.73 4.6 648.19 14.1 13.27 1.41 15.49 I .37 Slovenia 78.53 4.93 2.78 668.71 26.63 9.1 1.18 10.5 0.2489 Czech Republic 77.1 5 3.14 767.48 16.82 17.1 0.72 7.65 0.1169 Poland 75.38 6.66 5.98 842.17 15.41 13.12 0.98 20.85 0.299 EU post2004 74.47 6.86 7 36 91 1.06 25.55 13.7 0.9 39.45 0.4945 Slovakia 74.3 6.29 7.2 944.98 24.76 12.26 0.91 12.46 0.0741 Hungary 73.02 7.58 6.23 I01547 44.51 25.55 1.48 15.31 0.2184 Estonia 72.89 8.24 5.44 993.64 21.72 12.2 0.62 33.99 2.38 Latvia 71.2 9.34 8.76 1094.84 21.55 11.45 1II 53.91 2.67 Lithuania 71 9.91 5.87 1095.45 43.32 6.94 0.77 66.21 0.7955 Source: WHO, Health for All Database, http://data.euro.who.int/hfadb/; accessed April 04, 15:22, EST (USA). 117. The higher levels of standardized death rates for all causes draw attention to preventable deaths due to treatable conditions in Poland, if its health system performed at levels of pre-2004 EU countries. If EU 15 age-, gender-, and cause-specific mortality rates are applied to Poland, we can derive a mortality profile for Poland as if it were one of the EU15. Looking at this "ideal" profile relative to its actual profile reveals that 84 percent of the excess deaths are due to cardiovascular disease, suggesting a clear priority for the public health authorities. 1 18. Figure 20 summarizes the success and challenge of health status in Poland. Although health outcomes reflect all elements of life, the health system certainly is one determinant. As explained above, Poland has done well in increasing life expectancy and other measures of health. It is among the best performers of the new member states of the EU. The historical perspective in Figure 20 shows a clear departure from decades of stasis to a rapid increase in life expectancy at birth beginning in 1991, The departure from the record of Poland's immediate neighbor to the east, Ukraine, is extraordinary. However, the challenge is that Poland today has only caught up the EU of 1985. Its improvement however, is not fast enough to converge with the pre-2004 members of the EU. The problem primarily is related to excess mortality, as discussed above, among adults. The challenge is to plot a strategy to increase the rate of improvement in life expectancy so that Poland can close the 5 year gap with its richer neighbors to the west. 58 Figure 20: Life Expectancy at Birth in Poland and Neighboring Countries i ~ Source; WHO Health for All Database 11. Financial Protection and Coverage 119. A major accomplishment of the Polish system has been maintenance of near-100 percent coverage of the population by health insurance as the health system has evolved since Soviet times?' With only minor exceptions, participation in the National Health Fund is mandatory. All children under 18 are automatically covered; pregnant women, the disabled, and the elderly are also automatically eligible. For the working population, participation and payment of "premiums" are workplace-based. However, the level of contribution is based on the personal income tax base, so the financing system departs from the social insurance model in two ways -all personal income - not just wage income - is ~~ taxed for health, and there is no employer c o n t r i b u t i ~ n .The only participation of the employer is to collect the tax from the employee, for which the employer receives compensation of 0.1 percent. 120. Originally the contribution was set at 7.5 percent of the base. However, the rate was allowed to rise incrementally to 9 percent in 2007. 7.75 percent of the contribution is deductible from the personal income tax, but the remaining 1.25 percent is an additional contribution for health insurance. The tax base is the same as the personal income tax, which is aggregate income from all sources after being reduced by allowances and deductions in the tax law. In 2009, there were two tax rates, 18 percent on income between PLN557 and 85,528, and 32 percent on higher amounts. Only the personal income tax finances health care, except for subsidies financed on behalf of farmers from general revenues. Other taxes, such as corporate taxes or the value added tax, are not part of the revenue stream of NHF. 121. The major issue on the revenue side is the differential treatment of farmers, as it is for pensions. The premium for farmers is paid through the Farmers' Social Insurance Fund (KRUS) and is based on the price of rye. The value of the calculated premium is several times lower than for a non- 39 The 2006 special household budget survey on health care (described later in this chapter) found that 98.6 percent of residents were covered by NHF. About 3 percent of the population intended to purchase private insurance to top up NHF coverage, financed either by their employer or themselves. The survey suggests that ineligibility for NHF, while infinitesimal, is higher within the two poorest quintiles, 2.1 percent in the poorest and 1.8 percent in the next poorest. Maybe this is a statistical fluke, but it would make sense to try to understand who is not covered among the poorest quintiles and how it could be fixed if appropriate. 40 Economists typically assume that the employer's share of a tax is shifted to the employee; in Poland it is explicit that the employee pays. 59 farmer. In 2007, the State budget transferred PLN1.8 billion to NHF to supplement farmers' contributions, a subsidy of 60 percent of the total cost for them. In contrast, the self employed pay a premium based on their taxable personal income, but the base cannot be lower than 75 percent of the average enterprise salary. In some cases, it is beneficial for an employee to be redefined as a self- employed contractor to lower the tax. 122. Although distortions in any system are inevitable, Poland is one or two steps from a health insurance system financed from general revenues and could reduce distortions by moving further in that direction. Already 60 percent of farmer contributions come from the general revenue base, and embedding the "insurance premium" in the PIT represents a fundamental and sensible shift away from payroll or labor taxes. 123. All participants are eligible for the same benefits. Cost-sharing is built into the system, particularly for nursing homes, spas, dental care, drugs, and medical appliances. The most important expenditure category is pharmaceuticals, which accounts for almost two-thirds of personal spending. If a person is admitted for inpatient care, all drugs are free. For outpatient care, "basic" medicines require a PLN2.5 co-pay; "supplemental" medicines require a 50 to 70 percent co-pay. Access to care is rationed through waiting lists. At the end of the chapter, we present current evidence on use patterns and expenditures across consumption quintiles. 124. In short, Poland achieved universal health insurance coverage in the late 1990s and has continued to develop this system. In an economic recession, this can serve as an automatic stabilizer because the State provides subsidies to pay the premiums of the unemployed without unemployment insurance income, for social assistance recipients, and for others. However, this equitable feature of the system is somewhat distorted by contribution rates that favor particular employment groups and the unpredictable effects of waiting lists and copayments on access. Overall, however, there seems to have been success over time in reducing financial barriers to use of services. As Figure 21 shows, there has been a steady and substantial reduction since 2000 in the percentage of respondents citing financial problems as a reason for not seeking needed services (except for diagnostic tests). For costly care in hospitals, it is near zero, a major accomplishment in any health financing systems4'In 2000, almost 40 percent of the population reported financial difficulties in accessing medications and dental services. While the frequencies for these two elements had dropped to well below 30 percent by 2007, they are still high and would be two areas to consider for action, particularly for specific groups, such as dentistry costs for children and medication costs for the elderly or chronically ill. 4' Czapinski J., Panek.T. (2000 - 2007), Diagnoza spoieczna 2007 [Social Diagnosis 20071. Warszawa: VizjaPress&IT. [Reports are produced for the relevant years. See www.diagnoza.com.] 60 Figure 21: Self Reported Inability to Use Health Services Due to Financial Difficulties hospital sanatorium rehabilitation B E a 0 2007 0 2005 I . i H 2003 physicians services 1 0 2000 i dental prostetics dental services I r J i purchase of medications i ._____ 0 10 20 30 40 50 Percent of Households Source: Staff estimates based on Czapinski J. and Panek data (see footnote 40). 111. Expenditures on Health, 2005-2009 125. There are three principal sources of health financing in Poland: (i) income-based health insurance premiums; (ii) government revenues; and (iii) out-of-pocket payments. As discussed above, the first two categories collectively constitute public financing of health, and are mostly directed to the National Health Insurance Fund (NHF) that acts like the "single payer" purchasing agent on behalf of the insurees. Expenditures incurred over and above the budgetary allocations and Figure 22: Health Expenditures ( percent of GDP, 21 insurance premiums remain uncovered EU (2009 classification) and are deemed financed by "debts" in the Slovenia public domain.42 Private expenditures on Hungary health care comprise direct out-of-pocket Slovak Republic payments on health care. Czech Republic Latvia 126. Poland spent 6.2 percent of Poland GDP on health in 2007, of which Lithuania approximately two-thirds came from Estonia the public domain and the balance one- third from private out-of-pocket 4 6 percent of GDP 8 10 sources. Compared with other post-2004 .Public Expenditure on Health HTotal Health Expenditure EU countries, Poland spends a moderate amount of GDP on health (Figure 22). Source: WHO Health for All Database 42 Debt figures as announced at end of each quarter and each year represent net end-of-period debts. For the purposes of this calculation, the flow of debt for a specific period is obtained as the difference between the current and previous end-of-period figure. 61 127. Total expenditures on health care in Poland (in real 2008 prices) increased from PLN68,812 million per year in 2005 to PLN93,077 million in 2008, distributed approximately in the ratio of 70:30 between public and private spending. State Figure 23: Public Expenditures on Health ( percent, by financing budgetary expenditures on agent) health (through the Ministry of Health) increased in absolute terms from PLN3,438 million to PLN4,3 12 million during this period, but dropped from 6.97 percent to 6.39 percent of total public spending on health. Expenditures on health channeled through NHF increased from PLN36,250 2005 2006 2007 2008 2009 million (73.5 percent of Debt total) to PLN55,182 million 8) Local Governments zloty (83 percent of total) during this period (all 2008 Source Staff estimates based on data from MoF, MoH. and NHF prices). Total public spending, including the MOH, NHF, local governments, and net additions to debt, was about PLN66,487 in 2008 and is projected to be flat in 2009 at 66,585 billion zloty (equivalent to PLN65,734 million in 2008 prices), with reductions in real terms from the State and Local Government units and a slight increase by NHF. Financing through additions to debt fell during this period from PLN7,288 million in 2005 to PLN2,364 in 2008 (2008 prices). As a share of total public expenditures on health, debt financing fell from 14.8 percent in 2005 to 3.6 percent in 2009 (Figure 23). 128. While total spending on health increased in absolute terms during each year of the 2005- 2009 period, it remained stable at about 6.5 percent of GDP in 2005 and 2006, falling slightly to 6.4 percent in 2007. The year 2008 saw a huge Figure 24: Health Expenditures ( percent of GDP) increase in public spending on health, with State expenditures increasing by 74 percent and NHF spending W increasing by 3 1 percent, 2 2 1 01 though local government spending on health fell 2005 2006 2007 2008 2009 by 25 percent. The net +State Budget +Local Governments result of these changes National Health Fund Total Public Expenditure was that total spending Private Out-of-Pocket -Total Expenditure on Health on health in 2008 Source: Staff estimates based on data from MoF, MoH, and NHF increased even faster than GDP, resulting in total health spending as a percentage of GDP rising to 7.35 percent of GDP, an increase of 15 percent in one year ( 129. Figure 24). 62 130. Per capita expenditures on health increased from PLN1,803 in 2005 to PLN2,043 in 2007 and then jumped to PLN2,442 in 2008 (all 2008 prices), for a gain of 35 percent between 2005 and 2008. Per capita public expenditures during this period increased from PLN1,293 to PLN1,745 in 2008 (at 2008 prices), with private out-of-pocket expenditures accounting for the balance. 13 1. The composition of health expenditures has remained fairly stable during the last few years. Inpatient expenditures, which account for the largest share of health spending, fell from 36.5 percent in 2005 to 33.1 in 2008. This share is close to the OECD average, which signals a transformation from the Soviet era. Even today in Russia and Ukraine, hospitals account Figure 25: Composition of Health Expenditures (percent of for over 65 percent of spending, Total) almost double the proportion in Poland. The share of outpatient expenditure increased from 14.4 percent in 2005 to 16.1 percent in 2008 (Figure 25). "Other" includes long-term care, emergency medical services (EMS), health care policy programs, and other unclassified expenditures. It has risen as a proportion from 16.6 to 20.2 percent due to steady increases in 6 2007 2008 2009 Outpatient Drugs rn Inpatient long term care (from 1.1 to 1.5 percent) and adoption of a Source: Staff estimates based on data from MoF, MoH, and NHF separate line item for EMS in the State budget (shifting it out of the NHF), which caused an increase from 0.1 to 2.1 percent of the total. 132. Expenditures on pharmaceutical products constitute a significant share of total health expenditures, 32.6 percent in 2005 and 30.8 percent in 2008. Drugs constitute the single largest cost group in the entire healthcare system, and drug costs have been one of the most dynamically growing elements in overall costs of healthcare services. The Polish market is dominated by imported drugs, which account for 63 percent of the market share, many of which are original drugs. The total value of imported drugs has grown annually at over 12 percent during 2005-2008, while the value of domestically produced drugs has grown at around 10 percent per annum. Both the retail price of pharmaceuticals as well as quantity of drugs consumed (especially expensive drugs) have grown during this period. Yet, a large number of modern innovative drugs are not included in the NHF list of partially reimbursed drugs, which puts Poland behind most EU countries in terms of public financing of drugs, The high share of private expenditures in financing drug costs (almost 60 percent overall and 35 percent for subsidized drugs) leads to equity, financial protection, and health concerns. Therefore spending on pharmaceuticals by households would be a good target for thorough analysis of the impacts and possible additional public financing in the future. 133. From 1989 to 2005, Poland struggled to pay the bills, caused by a combination of low fiscal effort for health and an expensive and unaffordable endowment of inappropriate infrastructure (which is discussed below). 2005 seems to have been an inflection point when, through a combination of higher incomes and a higher NHF tax rate, resources flowed into the system in volumes sufficient for hospitals to incur less debt for operations, while NHF started to experience surpluses. Health care expenditures around the world are highly income elastic, rising faster than incomes. Inevitably this will 63 + happen in Poland, and increasingly the challenge will be how to manage cost escalation pressures. The next section reviews efforts to gain control over costs. IV. Recent Health Sector Reforms in Poland 134. The most significant and far-reaching health sector reforms were initiated in January 1999, when Poland introduced systemic changes in the health sector by thoroughly overhauling health financing, management and organization. Social health insurance was introduced and the institution of the Sickness Fund (Kasa Chorych) was created. All citizens belonged to one of 16 Sickness Funds organized on a regional basis (plus one Sickness Fund for military services), to which mandatory health insurance contributions were paid. The Funds contracted with a wide variety of individual and institutional providers of preventive, ambulatory, specialist and in-patient care to supply health services for the insured. Some highly specialized procedures under tertiary care were financed directly from the State budget. An equalization fund - financed by a proportion of each Fund?s revenues - adjusted for disparities in income and expenditure across regions according to a formula risk-adjusted for age. Individuals, organizations and other legal entities paid health insurance premiums, which were collected by the ZUS or KRUS, and then transferred to the sickness funds. 135. The introduction of social health insurance brought about significant changes in the organization and flow of funds in the health system, and it changed the incentive system for service providers. Prior to the reforms, almost all physicians were paid on a salary basis, and hospitals were given an annual budgetary allocation. With the reforms, primary care providers came to be paid on the basis of a per-enrollee capitation fee, while outpatient specialist care came to be paid on a fee per service basis. Most hospitals came to be reimbursed on the basis of a fixed amount per admission per department, irrespective of the type of stay or severity of illness. 136. Along with the health financing reforms, the MoH launched a series of programs aimed at restructuring the provision of health services. In a bid to contain healthcare costs -which were already facing pressure from rising incomes, an aging population, new medical technologies, and increased patient expectations - MoH initiated programs to bring about a rationalization of health care service delivery and provision, including reductions in the number of health care personnel and organizational changes in inpatient, primary and emergency care. Implemented together with regional and local governments, the restructuring program saw many initial successes, especially a reduction of health personnel and hospital beds, along with efficiency-enhancing measures in hospitals. 137. These reforms were not always received well by the public or by many of the providers. Following a huge drop in the number of services contracted, long waiting lists emerged for specialist, dental and emergency services. Health care was available free at the point of delivery before 1999, but the reforms led to uncertainty related to procedures governing choice, access, utilization and benefit package. The new funds and the hospitals found it difficult to manage their budgets, and the system as a whole became increasingly indebted. The poor financial situation of some health facilities only added to the general confusion and the impression that the health reforms had failed to arrest the growing debts. In 2003, the Sickness Funds were abolished and replaced with a single National Health Fund, which effectively reversed the 1999 decentralization. 138. The consolidation of Sickness Funds into one centralized National Health Insurance Fund brought about greater transparency in the flow of funds, as well as greater uniformity in the scope of services and entitlements. However, a number of issues have remained unresolved. First, there is a wide variation in the availability of hospitals and hospital beds across the country, with excess capacity in some regions and shortages in some regions. Second, despite a great deal of progress in hospital autonomy, financial accountability among hospital managers remains low, and indebtedness remains a serious problem. Third, the health system assures the supply and finance of a generous range of services, 64 but in reality is unable to provide all these services. And fourth, the reforms thus far have not focused in any meaningful way on improving the clinical quality of care, which varies considerably across regions. In addition to highlighting the shortcomings in the allocation of resources and in the flow of financing within the system, these and related issues also provide the basis for moving away from reforms which aim merely at stabilizing resources towards those that realign incentives for improving the efficiency and effectiveness of the health system. Further, the reforms in the past have been criticized for poor performance on the equity front, and it has been argued that the level and scope of informal payments that patients make for health services has remained high. Recent data reviewed later suggests that recent increases in public.spending may have made headway in improving equity. 139. In sum, there a r e three major issues with respect to improving health services in Poland, which relate to improving efficiency and controlling costs in the health care system, enhancing equity and access to health care as well as patient satisfaction, and taking steps to improve the analytical basis for financing decisions, the organization of care, and quality of services to improve health outcomes. These are discussed below. V. Managing Health Care Expenditures to Improve Efficiency and Control costs 140. In 1999 the core reform was implemented: separation of financing from the provision of services with the creation of the purchasing function that has now migrated to the NHF and the change in form of providers to independent institutions that could contract with the NHF. For all practical purposes, primary care was privatized in this process - shifting to capitation-financed, gate- keeping primary care physicians - and institutional care was shifted to independent organizations owned primarily by local governments (although purely private hospitals and other institutions have developed as well). Providers were expected to become self-sustaining based on revenue connected to providing care to patients, without additional subsidies from governmental units. Local authorities retained responsibility for investments in the infrastructure they owned. 141. Recent health sector reforms in Poland have included a number of organizational and management measures aimed a t improving the functioning and efficiency of this system. While some of these reforms have produced results in desired directions, others have not. Efforts to restructure and downsize health facilities have been generally successful; the numbers of hospital beds and staff have been significantly reduced. Similarly, hospital autonomy has proceeded rapidly, with almost all hospitals enjoying a large degree of autonomy with respect to personnel and financing decisions. Development of networks to reduce redundant investments and to simplify use of the system for patients has not been emphasized. Quality assurance and competitive contracting to improve services have not been approached systematically. 65 i. Hospitals: Measures to Reduce Expenditure Growth and Increase Efficiency 142. Hospital restructuring was first undertaken over a three-year period starting in 1999, and the process was repeated in 2003. a Figure 26: Number of Hospitals and Length of Stay result, the number of hospital beds in Poland has fallen by almost 9 percent since the reforms began in earnest in 1999. At the same time, the average length of hospital stay has fallen by 45 percent, from 9.3 days --c- Hospitals per 100,000 per admission to 6.4 -Average length of stay, all hospitals days per admission ( Hospital beds per 100,000 143. Figure 26). Source Staff estimates based on data from MoH 144. Table 18 presents some indicators on hospitals and medical personnel for Poland and comparator countries. Poland has 5 16 hospital beds per 100,000 population, which is lower than most countries in the region and lower than the EU average. Likewise, it has the lowest number of physicians and nurses, and a low average length of hospital stay. Table 18: Hospitals and Medical Personnel, Poland and Comparator Countries (2007 or latest available year) Hospitals Hospital Acute Psychiatric Nursing Physicians Nurses Average Average Bed Per beds care hospital and per (PP) per length of length of occupancy 100,000 per hospital beds per elderly 100,000 100,000 stay, all stay, rate in 100,000 beds 100,000 home hospitals acute percent, Per beds care acute care 100,000 Per hospitals hospitals I Slovenia 144 466 I8 37674 67 79 100,000 23729 76479 6 80 only 5 51 6"9 67- on1 Hungary I76 71255 41343 3043 85 53 27802 903 85 7 92 6 02 69 I5 Poland 208 516 I7 41075 64 78 203 18 468 84 6 40 70 01 E l (post 2004) 258 62549 46484 25323 55726 7 63 7 58 70 65 Slovakia 265 67884 601 58 82 44 31325 63160 8 70 7 80 68 30 EU 297 570 18 32238 74564 9 00 6 51 76 28 EU (pre 2004) 3 10 554.20 37532 337.87 804.96 9 35 6 46 Latvia 4 13 757 13 52326 137 03 30491 54795 9 44 Estonia 425 55699 38020 5620 54991 32853 65521 7 81 5.92 70 90 Lithuania 489 81395 50933 102 29 437 26 40671 73480 9 90 6 60 75 50 Source WHO, Health for All Database, http //data euro who intkfadbl, accessed April 04, 15 22, EST (USA), Statistical Bulletin of Ministry of Health. Poland, 2008 66 145. At first blush this may suggest a limited scope for further rationalization. However, another comparator, the United States, decreased acute care hospital beds by 13 percent from 1996 to 2006 to reach 246 per oo,ooo. In the Figure 27: Hospitals and Hospital Beds part of the U S . , most regions had well below 200 beds per 100,000. At the same time, hospital-based full time T 550 500 t Number of Hospitals equivalent employment increased in the U.S. by 6.2 percent. The two trends reflect a 20-year shift of inpatient care to E 450 4oo outpatient settings, including hospital- based clinics and outpatient surgery units. Clearly there is considerable 5 k3Oo' scope in Poland for further reductions in inpatient capacity based on current r p p pX o =o , l $ 3 2 3 { 2 i 3 g z { p r r % G % m : = p . ' yr p s " 9 " , $ medical practice, albeit in another part of g%gj';m_E 3 p oNmg%!ps= G g n u u the world facing different price signals. ; ~ B d i s , 2 2 2 However, data on the regional w distribution of hospitals and hospital beds poland suggest that despite Source Staff estimates based on data from MoH hospital closures and bed rationalization, it still suffers from the legacy of many public hospital beds located in the wrong places. The problem is especially severe in Slaskie, Lodzkie, Lubelshkie and Podlaskie voivodships - which have public hospital beds well above the national average (Figure 27). The problem of debt growth has been particularly visible in regions with an excessive concentration of public hospitals. According to MoH data, 80 percent of matured (or past due) debts originated from 15 percent of establishments, while half of the matured debts came from only 5.5 percent of establishments. Regions that generate the most debt - Slaskie, Lodzkie, Dolnoslaskie - are also the regions that have a large number of hospital beds or hospitals or both. On the other hand, Opole region is the least in debt, and it has the smallest healthcare infrastructure in Poland. 146. The Government's plan to corporatize hospitals will give the regional and local government hospitals greater autonomy and will create conditions that will enable market-like incentives to generate pressures to improve performance. In essence, corporatization will change the degrees of decision rights, residual claim, and market exposure. It will create indirect accountability arrangements that give managers more day-to-day freedom. Obligations to adhere to the commercial code would require that money-losing services and other social functions be made explicit, and funded from other sources to ensure continued delivery. Corporatization is a further extension of autonomy in that provisions for managerial autonomy are stronger in a corporatized organization than under autonomization, giving managers virtually complete control over all inputs and issues related to service delivery. A corporate unit would have to be legally established as an independent entity, making the transfer of control more durable than under autonomization. Thus, a corporatized hospital would have a hard budget constraint or financial bottom line, which would make it fully accountable in theory for its financial performance. 147. Technical and allocative efficiency can be expected to increase in corporatized hospitals, especially if an incentive structure is associated with these reforms that will promote managers and employees to make efficient decisions. Further, the greater freedom associated with these reforms provides enhanced opportunities for managers to choose their optimal mix of inputs. In effect, corporatization will create even stronger incentives than exist today for managers to employ personnel optimally, improve staff performance, increase availability of drugs and services, and improve maintenance of facilities and equipment. 67 148. Corporatization will be done initially using a voluntary approach because a law making it compulsory was vetoed in 2008 (see Box 3 ). The voluntary window was opened by a resolution of the Cabinet of Ministers taken in February 2009, with implementing arrangements approved in May 2009. In the past 8 years, 72 hospitals have been corporatized as a result of local government decisions. Ninety percent of these hospitals are current in their accounts. In contrast, the roughly 630 hospitals operating as government organizations have run up about PLN 10.5 billion of debt on roughly PLN25 billion in annual revenue (a 42 percent debt-to-revenue ratio, about 25 percent in public debt and 75 percent in debt to private firms and banks). The Government expects about 210 hospitals to join the corporatization program. 149. The public hospitals will be liquidated, with the MOF financing the public debt and the local authorities taking on most of the private debt. When a hospital emerges from liquidation as a new entity under the commercial code, it will be debt free, will be operating under an approved 5-year restructuring plan, and will be monitored twice yearly for financial and technical adherence to the plan. The MoH and NHF estimate that virtually all hospitals can operate on a cash basis with the revenues they earn, once current debt service requirements are removed. The corporatized hospitals are expected to have surpluses once they emerge from liquidation, which would allow them to invest in themselves as well. The program will be monitored through yearly reports to the Council of Ministers. 150. At the same time, it is important to be aware that several issues may limit the gains from corporatization, and public hospitals may not be able to create appropriate private sector-type incentives to improve efficiency and quality of services. These limitations include their near-monopoly situation, constraints on staffing procedures, and the ultimate safety net of a government bail-out. Autonomy and corporatization may also fail to meet their intended objectives if decision-making and management capacities are poor. 151. Ultimately, the corporatization program will succeed if it results in the consistent enforcement of a hard budget constraint on hospital managers so as to provide the necessary incentives for accountability and financial discipline. Healthcare organizations that are not economically viable would have to be allowed to fail or be reorganized. Because there are still excess numbers of hospitals and beds, this result is probably inevitable, so the Government cannot escape developing a medium term strategy for further reducing the number of hospitals and hospital beds. It is entirely possible that local governments and corporatized hospitals will take the lead in doing this if the corporatization program succeeds, but for the regions burdened with extreme excess infrastructure, more assistance may be necessary. Because the corporatization program is voluntary at least at this initial stage, it is likely that some of the best-managed and most viable hospitals and localities will be the first to participate in the program. Thus the least viable institutions may be the last to become eligible, so the Government may not be able to avoid having to deal directly with hospitals and localities that are financially marginal. 152. The difficulty of successfully implementing the corporatization program cannot be over emphasized. The Pomorskie Region, for example, has the lowest bed-to-population ratio in the country (Figure 27). It has developed a strategic plan for restructuring the health system in the region and by taking advantage of the corporatization program to facilitate the reform. Currently, each institution seeks to maximize its independence and self-sufficiency. The result is duplication of investment, overlapping service provision, competition for funds from NHF, few economies of scale, ineffective use of equipment and medical staff, and inadequate quality of care in facilities that do not have the volume of patients necessary to justify the services they offer. The fragmentation and lack of synergy across institutions results in low patient satisfaction and confusion. To address these problems, and focusing only on the 12 institutions owned by the regional authorities, Pomorskie would eliminate smaller local facilities by 68 shifting them to counties, selling them outright, merging them, or liquidating them. It would continue to support only those facilities considered strategic for the region. 153. These "strategic institutions" would be converted to Commercial Code Companies under the corporatization program. However, a single corporation would be created initially to put all the regionally owned hospitals, which currently operate individually, under a single umbrella in an attempt to solve the fragmentation problem described above. But to participate in the program, these 12 hospitals would have to first be liquidated and their debts settled. Their debts at the end of 2008 were approximately PLN423 million, 46 percent of revenues of about PLN918 million (of which 90 percent comes from NHF). Note that this debt-to-revenue ratio is just slightly higher than the average of 42 percent for the country. In addition, the regional authorities estimate a need for PLN600 million in investments over the next 3 years to modernize facilities and comply with environmental laws. 154. To corporatize these hospitals, the region estimates that it would receive a subsidy of PLN180 million from the State, leaving it with an obligation of PLN240-250 million. Regions are legally prevented from incurring debts above 60 percent of revenues. With yearly revenues of about PLNl billion, the hospital debt would likely cause the region to exceed its debt ceiling (ignoring the fact that it would also crowd out borrowing for other needed investments). As regional authorities typically own half the hospitals in the region - the larger ones - this situation would be common. The MoH estimates that 1 1 out of 16 regions may face similar obstacles. It has developed options for managing the problem, which revolve around the disposition of physical assets when the institution is corporatized. The regional would rent or sell these assets to the facility, with the revenues earmarked to service or retire the debt - in effect making the debts entirely or in part once again an obligation of the facility. Whether some or all of the regional government's residual debt could be restructured in a way that the new corporation could service it successfully is the key detail in realizing this plan. 155. A second major hurdle is what to do about 47 university hospitals, which are legally barred from participating in the corporatization program. The university hospitals are autonomous healthcare units, whose "funding bodies" are the medical universities. The universities operate with nearly complete autonomy under the law on higher education. There is no local government behind these hospitals, so they operate on a weaker financial platform. These hospitals are among the most indebted. Their average age is 48 years; nearly 60 percent were built between 1917 and 1969; only 12 percent are operating in buildings constructed after 199043.University hospitals tend to provide complex healthcare services, but they also undertake research activities and education. They have a tendency to charge as much of their operations as possible to the NHF. The MoH anticipates a new law covering university hospitals that would separate the hospital from the university so that it could be corporatized, put a firewall between its accounts and those of the parent university, and creates a supervisory role for the MoH. However, this reform is not yet in place and will depend in part on the experience of the corporatization program as it is implemented. 156. These are major managerial, financial, and political hurdles for the corporatization program to overcome. In December 2008, when the MoH expected to implement the corporatization program universally, it estimated an overall cost of PLN2.7 billion to the State budget. With 100-150 hospitals participating in the voluntary program over the next two years, it expects the State share to be about PLN1.2 billion with outlays beginning in 2010. The mix of cash outlays needed relative to write- offs and guarantees once the program gets underway would affect the total. The hope is that this effort will solve the problem once and for all; that of course depends on willingness to use the instruments afforded by corporatization to enforce financial discipline in the hospital sector, 43 Golinowska S., Sowada Ch., Wozniak M. (2007), "Sources of Inefficiency and Financial Deficits in Poland's Health Care System, WHO Regional Office for Europe, memo 69 Box 3: The Evolution of Hospital Management and Financing in Poland Beginning from 1995, the legal structure of Polish healthcare institutions was changed from "budgetary units" to autonomous public healthcare institutions (SPZOZ). Budgetary units were fully dependent on the state as the owner, as administrative units of government. SPZOZs were more independent and self-governed. Budgetary healthcare units left an accumulated debt of approximately PLN8.5 billion (over EUR 2 billion), by the end of their existence in 1998. This debt constituted approximately half of the annual healthcare expenditure at the time. The change to SPZOZ was intended to generate a more active management of these institutions, fashioned on private companies, as well as to prevent new debts. The directors of these entities were given greater autonomy. In particular, the authorities stopped regulating salaries, dictating staffing levels, and funding through line-item budgets, which became the responsibility of the SPZOZs. Managers were to become more responsible and accountable for the operation of the organization. However, the principle was maintained that SPZOZs could not operate for profit and would not be allowed to become bankrupt. Following a two-year period of relative economic stability, the wave of debt of SPZOZs returned in 2001 and has continued ever since. This occurred despite a large scale program of SPZOZ restructuring carried out by the Government in 2000/01, which led to a decrease in staffing levels at public healthcare institutions by about 100,000 employees (out of approximately 600,000). It is commonly recognized that, in their current form - with over-employment, debt-generation and without appropriate motivators to control costs - SPZOZs threaten to destabilize the finances of regions (voivodships), counties (poviats), and municipalities (gminas). As owners of SPZOZs, they are ultimately responsible for the debts. Moreover, in the event of the debts exceeding 60 percent of the annual budget, a receiver can be appointed, signifying a loss of power by elected officials. Several counties and regions of Poland are currently facing the prospect of administration by receivers due to the debt generated by SPZOZs, and the continuity of the provision of healthcare services is jeopardized by seizure of SPZOZ bank accounts. By the end of 2006, there were approximately 150 hospitals run as non-public entities, including approximately 50 established by commercial companies owned by local governments. These 50 entities were established as a result of "closing down" the public facility (SPZOZ) and creating a non-public Healthcare Unit (NZOZ) established by a joint-stock or limited liability company, the shares of which are mostly or completely owned by the local authority. 44 The other 100 hospitals were privately owned. In September 2008, the Government introduced to Parliament a draft law to convert all SPZOZs into commercial code companies, or NZOZs, within two years. The process was supposed to be obligatory and supported by government. Although eventually passed by Parliament, on 26 November 2008, the President vetoed it. On 18 December 2008, the Parliament voted again but did not gamer the necessary two-thirds majority to overturn the Presidential decision. As a consequence, the Government developed Plan B, a voluntary program to achieve the same goals. It will have the following characteristics: Participating local governments will liquidate the SPZOZ and open it as an NZOZ run by a commercial code company, owned by the local government, As part of the liquidation, the local government will cover the debts of the healthcare units. The State may offer financial support estimated at over PLNl billion to cover public liabilities (for taxes and social security debts), as well as to provide loans for covering part of other debts. Each new NZOZ will prepare a business plan in advance of this transformation that will be appraised by the Bank Gospodarstwa Krajowego (BGK) for the State. The National Health Fund (NHF) will participate in the appraisals, validating the revenue side of the plans. The Ministry of Health will provide a technical appraisal of the health aspects of each institution's plans. The major difference comparing to the previous restructuring programs (in 1995-1998 and 2000-2001) is that the hospitals will be converted into companies operating under the Trade Code. In effect, this is a formal effort to introduce a hard budget to the health care units. It cannot be guaranteed that this change will 44 Note from the text that the total number of corporatizations had risen to 72 by 2009, 3 years later than the data cited in the box. 70 eliminate the need for future rescues, but it is probable that trade code mechanisms will result in quicker action and make it more difficult for owners to ignore problems. It puts the organization at risk of bankruptcy, creating an opportunity to close down institutions that are no longer financially viable. A recent study by The Gdahsk Institute for Market Economics (GIME) found, after reviewing four cases of reforming hospitals, that the three successfully restructuring to reduce the stock of debt and to gain control of the accumulation of new debt shed employees to make revenues cover expenditures. One of the cases, the Powiat Hospital in Kluczbork, illustrates a corporatization experience dating from April 2004. The hospital's accumulated debt was PLN2O million in 2003,45 percent of the powiat's annual revenues. The powiat took on this debt with a plan to retire it over five years. Upon being corporatized, the hospital hired 200 staff to replace the 325 staff previously employed. By July 2008, there were still only 198 staff, of whom 9 were in administration. The SPZOZ hospital had 1 1 surgeons; the new NZOZ hospital had 4, but they produced nearly double the volume of surgeries. The powiat owns 100 percent of the shares of the hospital but does not interfere in its management; the Board includes 3 outside members to reduce the opportunity for political interference. By 2005, the hospital was able to purchase a building from the powiat; by 2007 it was able to finance capital expenditures exceeding PLN1.4 million from its own resources; and by 2008 it paid a dividend to the powiat. The powiat is considering selling shares to an investor willing to invest further in the hospital. Many risks remain, but the progress made by this hospital is tangible. 45 , Source; Bank staff analysis. ii. Provider Payments: Measures to Reduce Expenditure Growth and Increase Efficiency in Hospitals 157. The increase in the cost of hospital care has been a persistent and growing problem in Poland. While a substantial portion of the increase in hospital costs has been attributed to the large number of hospitals and hospital beds, the rising use of new technologies is threatening further increases in costs of hospital care. To address this problem, Poland moved away from a retrospective cost-based reimbursement approach whereby hospitals could recover most of what they spent - and thus had little incentive to control costs or rationalize the use of high-end technology. In 2002, Poland adopted "prospective payment" systems, in which rates were set prior to the period during which they apply and the hospital incurs at least some financial risk. Effectively, this system controlled the amounts hospitals could charge for specific services and paid hospitals an all-inclusive rate per day of hospital care. However, this had its own problems, since it set up incentives for hospitals to increase lengths of stay and admissions, and in some cases the controlled charges also encouraged hospitals to increase the number of services they provided. In 2004 Poland moved to a per-case payment, under which the hospital is paid a specific amount for each patient treated, regardless of the number or types of services provided. This system rewards the hospital for reducing the cost of treating a patient over the entire course of the hospital stay. 158. To adjust for differences in the kinds of patients treated by the hospitals, Poland initiated a system of measuring the hospital's case mix - or the relative frequency of admissions of various types of patients, reflecting different needs for hospital resources. This has led to the development of Diagnosis Related Groups (DRGs) called Jednorodne Grupy Pacjentow (JGP, Homogenous Patients Groups), which were introduced in July 2008, although hospitals during the first 6 months were guaranteed incomes at least at the level that would be provided under the previous system, primarily to prevent problems caused by errors in using the new system. Full implementation for reimbursement is underway in 2009, without the safety net of a separate funding stream based on historical values. 45 Elzbieta Malinowska-Misiqg, Wojciech Misiqg, Marcin Tomalak (2008) The Use of Public Resources In Hospitals: Case Study of Poland, Gdansk Institute for Market Economics, Warszawa, Poland August 29. See http://ibngr,pl/index,php/pl/content/download/ 103311 2 1 67/file/GIME-Final-Report-August-29,pdf. 71 159. In essence, DRGs classify treatments according to their resource costs, by employing a complete and consistent coding system of patient-level information obtained from medical records to establish and cost bundles of appropriate inputs for one or more diagnosis-based treatments. Typically requiring about 10-20 data items per discharge over a few years, the DRG system provides a valuable tool not only for reimbursing hospitals but also for overall planning and resource management. There is no inconsistency in employing DRGs in a system subject to global payment limits. In fact, retaining global budgets is perhaps sensible, at least in the short run, to ensure that hospital expenditures are contained within defined limits. 160. Development and implementation of DRGs in Poland has been controversial and slow. The first year of experience with DRG-based reimbursement for inpatient care has been conducted in such a way that hospital revenues have increased substantially. The average value of hospitalizations rose by 14 percent within 3 months of the introduction of DRGs in July 2008. In a subsample of 5 5 high complexity clinical hospitals, the value of a hospitalization in the second half of 2008 was 14 percent higher than during the first half; by March 2009 it had risen another 12.3 percent over the third quarter of 2008. 161, DRGs help to unify medical and financial information and vastly improve the statistical understanding of what happens inside a hospital. Because reimbursement is fixed for a diagnosis, it encourages hospitals and physicians to monitor and improve treatment protocols for each diagnostic group and to understand the determinants of costs. DRGs give NHF the ability to analyze in detail efficiency measures across hospitals as well as financial and health outcomes by diagnosis. Poland is already beginning to benefit from the improved information both nationally and within hospitals. 162. However, as a reimbursement tool, DRGs are intended to allow the payer to encourage efficiency by paying the average cost across hospitals. As higher cost hospitals lower their costs to survive at the average reimbursement rate, the average cost and the variance in costs across hospitals by diagnostic group should be driven downward. To placate opposition to DRGs, however, NHF made a strategic choice to ensure that hospitals were at least as well off under DRGs in 2008 and 2009 as they were under the old system. The good luck of windfall revenues in 2008 (well above its forecast) accommodated NHF's generosity. 163. In Poland, three types of reimbursement are used that lead to known distortions. Primary care physicians are financed based on capitation fees. They have an incentive to limit their costs and effort by referring patients to higher levels of care, particularly to outpatient specialty care and emergency hospital services that operate around the clock. Outpatient specialist care is financed through fee-for- service payments by NHF via fees for three categories of consultations that include appropriate diagnostic tests, plus some other services that are reimbursed separately. Specialists have an incentive to avoid funding diagnostics that are included in the consultancy fee and to encourage unnecessary follow-up visits. Hospitals have an incentive to admit patients, even with minor problems, if outpatient treatment for the same condition is less remunerative. As a result of these incentives, patients may have the impression of being unwelcome at most service points, as they are referred from place to place or make repeated visits for treatment. Each episode of care is paid, one way or another, by the NHF. Additionally, there are long waiting times (lists) for admissions to specialist clinics as well as to some hospitals. 164. One initial finding related to the brief implementation of DRGs is that 28 percent of inpatient stays are for one or two days, which is consistent with all of the incentives described above. An important element of DRGs is to encourage least cost treatment of any specific diagnosis, so extending them to outpatient care and developing an economic framework for encouraging medically sound decision making between outpatient and inpatient care could further reduce costs while making medical treatment more appropriate to the needs of the patient. NHF plans an outpatient DRG that pays 72 for outputs. The new system is expected to encourage specialist outpatient care and ambulatory/day care in lieu of hospital admissions. It generally does not refer to the practice of the GPs, whose services are likely to continue to be financed through capitation. 165. By itself, the use of DRGs will improve knowledge of economic and medical aspects of hospital care. The power of additional knowledge and information on how the system functions can only be beneficial. However, it seems inevitable that the power of DRGs to improve hospital efficiency as a reimbursement mechanism will eventually be used by the NHF, possibly sooner rather than later as economic conditions reduce its revenue. Along with corporatization to increase accountability of hospital managers, DRG-based reimbursement will give the NHF, MoH, and local authorities a clear picture of which hospitals are producing the best medical and financial results and thus are viable as the slimming down of the hospital sector proceeds. Extending DRGs to outpatient care will reduce distortions in incentives and encourage sound tradeoffs among inpatient, day patient, and outpatient interventions. DRGs are a powerful instrument to improve quality while driving down costs, but only if they are used to do so, which requires considerable analytic capacity that weighs in on policy decisions. 166. The reforms in hospital governance and compensation of hospitals hold promise to improve their incentives and ability to operate in an economically efficient manner. The result should be a further rationalization of their infrastructure and personnel while adapting dynamically to modern medical practice that reduces the need for hospital hotel functions. These changes will require a firewall between hospital decisions and their founders? treasuries as well as an aggressive use of DRGs to encourage integration of care, adoption of money-saving approaches to care, and greater efficiency in service delivery. Yesterday and today the issue has been primarily hospital rationalization; tomorrow the issue will be escalating costs and how to manage them as pressures build from a richer society for more care. VI. Quality/Accreditation 167. Voluntary accreditation of health care units has been practiced in Poland since the middle 1990s, when Rafal Niiankowski, former department director in the Ministry of Health created a Ministerial agency, the Centre for Quality Monitoring, located in Cracow. The Centre has developed an accreditation system based on American Joint Commission standards. The accreditation process is time-consuming and costly for hospitals, with no obvious benefit to them. Improving accreditation is an element of the health system development program announced by the current government. 168. The proposed framework would make the following changes: accreditation would become a formal process governed by law; accreditation would be granted by the Minister of Health instead of the Centre for Quality Monitoring, as it is presently; the Center of Quality Monitoring would play a technical role in the process and would receive additional funding; and the Minister of Health would nominate a 12-person Accreditation Committee that would prepare an opinion for the Minster whether to grant the accreditation certificate for a given unit. 169. Despite initial expectations, the proposed law (blocked, probably temporarily, by a Presidential veto) does not offer any immediate benefits for accredited units. The most obvious reason for an accreditation system would be to qualify institutions to compete for contracts with the NHF. If that were the case, a full set of instruments to improve dynamically the functioning of the hospitals and specialist care system would be in place: a. corporatized hospitals to try to get incentives right for managers and to raise the specter of bankruptcy or reorganization for service units that are not economically viable; 73 b. reimbursement based on inpatient and outpatient DRGs that, if adopted regularly to evolving medical practice and cost data, would create incentives to economize and improve the quality of care; and c. accreditation as part of a competitive NHF contracting system to improve basic standards of care. VII. Health Expenditures, Use of Services, and Equity 170. We have two types of information related to equity outcomes, the distribution of out-of- pocket payments and utilization of services. Out-of-pocket Payments 17 1 . Out-of-pocket payments have emerged as an important source of health care financing and take two forms: formal copayments and informal out-of-pocket payments. Formal copayments were almost non-existent prior to the transition, reflecting the constitutional guarantee of free health care for all citizens, but they now apply to many services, including dental care, pharmaceuticals, and medical prosthetics. They have become significant over time. Informal payments, in cash or in kind, to public health care providers existed even during the socialist times, and continue today. 172. Table 19 shows medical expenditure information from the 2005 and 2007 Household Budget Surveys by quintile. The 2007 data have been deflated to 2005 prices for comparison. Spending is quite skewed. In 2007, the poorest quintile spent about PLNlO per person each month on health, compared to PLN87 for the richest quintile, representing 3.2 and 5.5 percent of monthly income respectively in the two quintiles. If a financial protection system is working, this is the pattern of spending. If it is not working, the poorest quintile tends to devote a higher percentage of its income to health care consumption. The Polish health financing system passes this simple but important test. The question for the authorities is whether they would like to drive any of these rates lower (e.g., for the poorest) or let them drift higher (e.g., for the richer quintiles). 173. In the most recent experience, they have been driven lower across the board. Households spent 12 percent less out of pocket in 2007 relative to 2005, which is astonishing considering that GDP per capita grew 3.6 percent in 2005, 6.3 percent in 2006, and 6.8 percent in 2007. In other words, while incomes increased rapidly, out-of-pocket spending on health - which typically increases faster than income - declined in real terms and, obviously, as a percentage of income. All quintiles lowered their spending in absolute terms on health services and products, with the poorest and richest quintiles lowering their spending the least and the middle quintiles lowering their spending by over 10 percent. The final column in Table 19 shows how overall consumption changed in real terms during this two-year period. It increased by almost 10 percent overall, with the highest consumption group registering an enormous increase. The drop in out-of-pocket spending for health is an extraordinary achievement. Even if the richer households captured most of the benefits, the drop in spending extended as well to the poorest. 74 Table 19: Monthly Per Capita Spending on Health (in 2005 Zloty) Quintiles of Per Capita Consumption 2005 2007 ~~'~~~~ Health Spending Percentage Change in Overall onsumption on onsumption 2o05 2o07 Consumption Poorest 20 Percent 9 98 3 5 percent 9 47 3 2 percent -5 4 percent 2 0 percent Quintile 2 21.42 4 8 percent 19 36 4.2 percent -10 6 percent 3 5 percent Quintile 3 36 55 6 1 percent 29 84 4 8 percent -22 5 percent 3 7 percent Quintile 4 55 87 6 8 percent 46 82 5.4 percent -19.3 percent 5.0 percent Richest 20 Percent 92 14 6 6 percent 87 30 5 5 percent -5 5 percent I8 7 percent ----- Total _ - _ _ ~ - 43- 19 _ _ - _ -L E 6 ?!- !E 38 5 5 4.6 percent -12 0 percent 9 7 percent Source Staff estimates based on HBS 2005 and 2007 174. Looking a t the last row of Table 19, on average, a Polish household in 2007 spent almost PLN39 per month, per person, on health (in 2005 Zloty), an amount equivalent to about 4.6 percent of its total expenditure (down from 6.1 percent in 2005). Overall, total private expenditure on health was PLN 18 billion, equivalent to 1.6 per cent of GDP (down from 1.8 percent of GDP in 2005). 175. Table 20 shows the distribution of household per capita health spending across types of expenditure, in declining order of importance. The last column shows the percentage change from 2005 to 2007. Just three categories - pharmaceuticals, dental services, and medical services - account for over 88 percent of out-of-pocket health expenditures. Within that group, pharmaceuticals continue to dominate out-of-pocket spending, accounting for just over 61 percent of the total in 2007. However, it dropped by almost 20 percent from 2005. Despite substantial percentage increases in spending on dental services, rehabilitative equipment, medical goods, and hospitals, these increases were from low initial amounts that have little weight in overall spending, except for dental services. Drops in pharmaceutical and medical services spending more than offset these increases. 176. In essence, the public financing system between 2005 and 2007 has whittled away a t private spending in the two most important categories, pharmaceuticals and medical services, where there has been persistent criticism about high out-of-pocket costs. With the large increases in outlays by NHF in 2008, this pattern probably continued. The extremely low out-of-pocket spending for hospital services should be noted also because it suggests extraordinary success in protecting the population from catastrophic costs. 177. Often data that include the entire population mask the underlying issues that give rise to complaints about high costs, because many households have zero o r low expenditures on health- related goods and services unless a problem strikes. Table 21 compares households with and without spending, using the 2007 household expenditure data. The first column, which includes zeroes, is the same as the 2007 column in Table 20 but is not deflated to 2005 Zloty. The second column excludes households with no spending, and suddenly categories that had many zeroes, especially dental services, rehabilitative equipment, sanatorium services, unconventional therapy, hospital services and other medical services increase substantially. However, the heavy weight of pharmaceuticals in the mix, with high frequency of nonzero spending, keeps the overall average below PLNSO. If we ignore Sanatorium, Unconventional, and Other as areas discretionary spending, we are left with dental, medical, and hospital services as potential trouble spots for specific patients. 75 Table 20: Distribution of Health Spending (in 2005 Z-) -- Expenditure Category 2005 -_---_I - - 2007 ~ Percentage - Mean Monthly Distribution Mean Monthly Distribution Change Per Capita Per Capita Expenditures on Expenditures on Health (PLN) Health (PLN) Pharmaceuticals 29.55 68.4 23.67 6 1.4 percent -19.9 oercent percent Dental Services 5.02 11.6 t 23.2 per cent percent Medical Services 4.53 10.5 4.26 1 1 .O percent -6.0 percent percent Rehabilitative Equipment 1.79 4.2 percent 2.03 5.3 percent 13.4 percent Sanatorium Services 0.75 1.7 percent 0.74 1.9 percent - 1 .O percent Lab and X-ray Services 0.58 1.3 percent 0.55 1.4 percent -5.8 percent Auxiliary Medical Services 0.36 0.8 percent 0.38 1.O percent 3.7 percent and Unconventional Therapy Medical Goods 0.34 0.8 percent 0.39 1.O percent 17.8 percent Hospital Services 0.20 0.5 percent 0.28 0.7 percent 37.4 percent Rehabilitation Equipment 0.06 0.1 percent 0.05 0.1 percent -14.0 Repair percent Other Medical Services 0.01 0.0 percent 0.01 0.0 percent 21.4 percent Total 43.19 100 percent 38.55 100 percent -10.8 percent Source, Staff estimates based on HBS 2005 and 2007. Table 21: Mean Monthly Per Capita Expenditures on Health in 2007 (in 2007 Zloty) Expenditure Category All Households By Age Including Os Excluding Os Average Average Household Age Household Age Below 60 Above 60 Excluding Os Excluding Os Pharmaceuticals 27 57 34.21 24.11 91.14 Dental Services 6.42 28.59 25.93 70.38 Medical Services 4.42 19.72 16 87 36.48 Rehabilitative Equipment 2.11 62.1 1 52.12 111.14 Sanatorium Services 0 77 140 33 107.29 207 06 Lab and X-ray Services 0.57 9.00 7.97 15.82 Auxiliary Medical Services 0.39 22 24 18 65 39 49 and Unconventional Therapy Medical Goods 0.41 3.93 3.44 8.67 Hospital Services 0.29 54.21 42 13 151.81 Rehabilitation Equipment 0.05 16.45 12.73 30.42 Repair Other Medical Services 0 01 30 44 17.84 82.39 Total 40.02 49.37 38.03 119.87 Source, Staff estimates based on HBS 2005 and 2007 178. The last two columns of the table check for problems for elderly households, and a few alarms definitely go off. If the average age of household members is over 60, the household spends over 3 times more out of pocket for health goods and services relative to households with an average age below 60 - and this is only for households in both categories that incur expenses. Spending on 76 pharmaceuticals and hospital services is nearly four times higher in older households. It would be easy and probably not exceedingly expensive to provide some additional protection to older households, particularly for pharmaceutical purchases. Targeting by age and type of drug might make a considerable difference for them at relatively low cost to the NHF, but of course to do so would require a more detailed understanding of the current situation. 179. In summary, it is astonishing to see out-of-pocket expenditures on health goods and services fall, especially in a country where incomes have increased rapidly. While this is an important accomplishment in an area that has been a source of criticism in Poland, an opportunity may have been missed to target reductions in expenditures (particularly on pharmaceuticals) to the poor or the elderly, which would have required introduction of sliding scales or differential cost sharing arrangements on services, equipment, and specific drugs that affect them more. The reductions in out-of-pocket spending for the upper three quintiles could have been steered to more vulnerable groups, or it could be done in the future. Public Subsidies - Public Opinion Data46 180. This section reports utilization data by income quintile based on the Survey of the Public Opinion Research Center in 2007. The survey has a sample size of 38,866 and is nationally representative. It was conducted in late 2007, so the information it collected on use of health services during the previous six months would refer to the first part of 2007. It includes questions on income, but it has missing values on income for nearly 12,000 of the observations (about a third). The question we face is whether we can use the subsample with reported income to approximate the full sample. With such a large sample, if the loss of observations is roughly random across income quintiles, it may still give us a reasonable indication of how use of publicly financed health services through NHF is distributed. 181, There are several indications that we could use the data in this manner. For the subsample of non-missing income values, the frequencies of use are almost identical to those for the full sample. Table 22 reports these frequencies. Even without formal statistical tests, it is clear that the mean values of the responses are close between the full sample and the subsample. For the subsample we use for the subsequent figures, roughly 28 percent reported seeing a dentist in the past 6 months; 71 percent, a general practitioner; 45 percent, a specialist; 38 percent reported using a laboratory; 15 percent had a hospital stay, and 2 1 percent reported some problems getting the services they needed. 46 Sondat Centrum Badania Opinii Spolecznej ,,Warunki iycia spoieczenstwa polskiego - problemy i strategie" zrealizowany na reprezentatywnej probie losowo adresowej dorosiych mieszkancow Polski, N=38, 866 w okresie wrzesien - listopad 2007 r. Kierownik Zespolu Badawczego: prof. dr hab. Krzysztof Zagorski. [Public Opinion Research Center "The Conditions of Life of Polish Society - Issues and Strategies." Implemented on a representative sample of adult residents in Poland, sample size 38,866 for the period September - November 2007. Professor Krzysztof Zagorski, PhD, Research Team Leader.] 77 Table 22: Distribution of Responses for the Subsample Reporting Income versus the ( 1 1.983 Responses) Responses) ractitioner LsiboratoryiDiagnostic Center 38 2 35 8 t4 8 Any Difficulties Getting Services? 20 8 20 6 _ . ~ .- Source Staff estimates based on Public Opinion Research Center Poll September-Nokember 2007 182. Figure 28 shows the distribution of these visits across income quintiles (each bar represents 20 percent of the income distribution in the sample). For all practical purposes, these services are used in about the same proportions by each income group - and the poorest are heavier users of public- financed dentists and hospitals, To achieve fairly high levels of contact with the health system, particularly primary care services, and for use patterns to be similar across income quintiles, is an achievement. Moreover, although 20 percent reported difficulties accessing the services, the poorest were the least likely to report problems. 183. Figure 29 shows the use of private services not covered by NHF. The richer quintiles are more likely to opt out of NHF-financed services and either pay out of pocket or use additional coverage to finance this care. This behavior is particularly obvious for dental care, specialist services, and laboratory services. It is least likely for the primary care physician and inpatient care. Figure 28: By Income Quintile -- Use of NHF-Financed Services During the Past 6 Months, Survey of the Public Opinion Research Center 2007 80 a Poorest 2 0 Percent i p Quintile 2 Quintile 3 &+ Quintile 4 Richest 2 0 Percent 70 60 g y 50 5 2 UJ 40 B 5 30 5 5 20 El Y a 10 0 Dentist GP Specialist Lab/Diagnostic Hospital Any Difficulties? Center Service Source Staff' estimates based on Public Opinion Research Center Poll September-Novcmber 2007 78 Figure 29: By Income Quintile -- Use of Private Services Not Financed by NHF During the Past 6 Months, Survey of the Public Opinion Research Center 2007 80 6 P o w e r t 2 0 Percent *x Quintile 2 Uulntilcl3 L Pulntile 4 Richest 20 Percent 70 60 8 6 30 : . E 20 B 10 0 Dentist GP Specialist Lab/Diagnostic Hospital Service center Source Staff estimates based on I'ublic Opinion Research Center Poll September-Nohember 2007 184. NHF is mandatory for everyone, so all are paying to support it. However, because it is possible to opt out and pay for private services, which happens increasingly as household income rises, NHF expenditures are likely to be either proportional to income or slightly pro-poor. A pro-poor result is consistent with the previous evidence of a lower share of poor households' income being spent on out-of- pocket payments. 185. Figure 30 shows that use of NHF-financed services tends to be slightly lower in rural areas, but the differences are small. Even so, rural people are less likely to report having trouble using services than their urban counterparts. We do not show the data for use of private services, which is much higher in urban settings. Figure 30: By Residence -- Use of NHF-Financed Services During the Past 6 Months, Survey of the Public Opinion Research Center 2007 80 L Rural .e Urban f0 60 g 50 5 f : 40 6 30 5 5 g 20 10 0 Dentist GP Specialist Lab/Diagnostic Hospital Any Ddflculttes? tenter SeWiCe Source Staff estimates based on Public Opinion Research Center Poll September-November 2007. 79 186. Finally, Figure 31 suggests that those over 60 are heavier users of all services except for dental care. We do not include the private service data, but the elderly are slightly less likely to use private services, especially for dental care and their GP. 187, If we combine the information on use of NHF-financed and private health services, the opinion poll data suggest that poorer quintiles use fewer health services overall than do the richer quintiles, but the latter have moved some of their demand to private providers, As would be expected, the elderly are heavier users of medical care, but there are few significant differences between urban and rural households. The data show a high frequency of contacts with the health system across the board, with about 20 percent of the population reporting difficulties in using NHF-financed care. Household Budget Survey 2006`7 188. This section reports frequencies from the special health utilization and expenditure survey of 2006. For clinic services, it asks use of care during the last 3 months of 2006, except for hospital care it asks for the previous 12 months. It differs from the public opinion survey in that it is based on a subsample of households interviewed for the regular household budget survey, so robust estimates of household consumption can be used to construct quintiles rather than self-reported income as used in the opinion survey above. Figure 31: By Age -- Use of NHF-Financed Services During the Past 6 Months Months, Survey of the Public Opinion Research Center 2007 *) C60 I60+ 80 70 60 50 40 30 20 10 0 Dentist GP Specialist Lab/Diagnostic Hospital AnyDifficultIes? Center Service Source. Stafl'estimates based on Public Opinion Research Center Poll September-Notember 2007 189. Results from both surveys are reported because each has its peculiarities. The opinion poll has the income-reporting problem, but it has the advantage of a straightforward survey instrument and a huge sample. The special Household Budget Survey module has the advantage of the connection to the HBS, but the sample size is much smaller (4728 households out of the regular sample of 14,017), the `'Ochrona Zdrowia w Gospodarstwach Domowych w 2006 R, GIJS, Departament Radan Spoiecznych [Health Care in Households in 2006, CSO, Social Surveys Division]. See h m . ' M R U stat ~ov.pl'zus 5840 2657 E N G HTML,&m. 80 survey oversampled richer households, 48 and the survey instrument is extremely complex. The two surveys provide similar but not fully consistent pictures of utilization. 190. The results are presented below in Figure 32 and Figure 33, Figure 32 is a rough counterpart to Figure 28. The first set of bars show the overall response for whether someone used outpatient services during the last quarter of 2006. If they did, questions were asked about dental care, basic medical services, and specialist services. "Medical Advice Basic Services" roughly equates to General Practitioner care. As before, we see high contact rates with the health care However, this survey suggests a stronger correlation between income (actually, consumption) and use of services, particularly for NHF-financed dental and specialist care. The poorest quintile was almost half as likely to use these services relative to the richest quintile. Inpatient stays are much less skewed across quintiles. 191. The patterns of use of private sector services (Figure 33) are similar to the opinion survey, but with lower implied use of dentists and specialists. For the richest quintile, about 12 percent used a non-NHF dentist compared to 16 percent using an NHF dentist; for specialists, the frequencies are 11 non-NHF and 27 percent NHF - nearly half their dental and a third of their specialist visits were to private providers. These, however, are much lower ratios than shown in the opinion poll data, which suggested a more pervasive use of non-NHF providers especially by the upper quintiles. Figure 32: By Quintile- Use of NHF Services During the Last 3 Months of 2006 for Outpatient and Last 12 Months for Inpatient Care, Household Bud get Survey 70 + Poorest 2 0 Percent IJ6 Quintile 2 Quintile 3 D, Quintile 4 Richest 2 0 Percent 60 50 p 40 z 30 = 0 e 20 5 G a 10 0 Used O u t p a t i e n t Dentist M e d i c a l Advice Basic Specialist Hospital Care Services Service Source: Staff estimates based on special HBS health survey September-December 2006 48 The consumption quintiles are constructed based on the regular household budget survey. so each quintile contains 20 percent of households in that survey. {lowever, for the special healthcare survey, the poorest quintile contributed only 13 percent of the households to the sample; quintile 2, 15 percent; quintile 3, 19 percent; quintile 4, 24 percent; and the richest quintile, 30 percent. Overall results from the survey would be skewed by behavior of the two upper quintiles. Most of the results reported in this chapter are for each quintile, so do not suffer from that problem, but the sample sizes of sick people are relatively smaller for the poorer groups. 49 Keep in mind that frequencies in Figure 32 are for 3 months while frequencies in Figure 28 are for 6 months. To be comparable, we would as a first approximation double the bars in Figure 32. ai Figure 33: By Quintile -- Use of Private Services Not Financed by NHF during the Last 3 Months of 2006 for Outpatient Care, Household Budget Survey - -- - ~~ 16 (I Poorest 2 0 Percent YE QUmtIIe 2 Quinflle 3 e Quintile 4 RICheSt 2 0 Percent 14 Wentlst GP spec lalt6t scrvtca Source Staff estimates based on special IJBS health s u n ey September-December 2006 192. The HBS data suggests a more skewed distribution of utilization that favors the richer quintiles. To be able to monitor better the distributional aspects of health care financing, it is important to invest in better population-based data and to couple this with regular monitoring of NHF's payments data from a distributive standpoint. 193. Household Perceptions 2006 194. The special health utilization and expenditure survey of 2006 also queried households on perceptions of the health care system and financial issues. Figure 34 shows self-reported difficulty of paying for pharmaceuticals by consumption quintile in the vertical bars (left axis), along with average expenditure in each quintile joined by a line (right axis). The picture is self-explanatory: at much lower expenditure levels, 30 percent of Figure 34: Expenditures on Pharmaceuticals the poorest quintile B o u g h t & No Financial P r o b l e m s w- B o u g h t but a Financial B u r d e n 2S claimed they could *W S o m e t i m e s Not A b l e to B u y - E x p e n d i t u r e o n M e d i c i n e s 4 4 2006 ~ not purchase drugs 100% 290 during the previous 90% 270 12 months. Only 80% 15 percent had no 250 70% problem, compared to about 5 5 percent 60% 230 in the highest 50% 210 quintile. This result 40% is to be expected. A 30% p' 190 possible policy 20% response would be 10% 170 to make the entire 150 picture look like the 0% Poorest Quintile 2 Quintile 3 Quintile 4 Richest top quintile's profile when the next survey is Soirrce. Staff estimates based on special HBS health survey September-December 2006 conducted. 82 195. Another issue that comes up is why people opt out of NHF providers to use services in the private sector that are not covered, and for which they must pay either out of pocket or with other insurance that they carry. Figure 35 shows the distribution of use of non-NHF physicians and dentists across quintiles for late 2006. The pattern is consistent with what we have come to expect, Figure 36 illustrates the reasons, Very few households opt out of NHF-financed services because of eligibility problems or because they are unable to get a referral from their GP for what they want to do, For physicians, approximately equal proportions of households opt out based on perceived higher quality of equipment and personnel versus the alternative of waiting time, distance, or courtesy." For dental care, quality of personnel and equipment is a stronger driver. 196. For the rich to Figure 35: Use of Non NHF Providers opt out is not a major 45 concern because they can u Urad No" N H F M a d i c a l Provider S Usad N o n NHF Dentist 40 certainly choose to pay for real or perceived 35 differences in services, 30 However, for the poorer 25 two quintiles to opt out 20 suggests there may be some serious performance 15 issues to address in these 10 areas, as they are making a 5 choice that has momentous financial consequences for POO,CPf Ruintilo 2 Quintlia 3 Quintil@4 Rlchert them, to pay outside the Source Staff estimates bdsed on special HBS health survel September-December 2006 NHF system. Figure 36: Reasons for Using Non NHF Providers c Medical c Dental QUdlltY of Cqulpmentor Personnel Wait, Distance, or Courtesv r Not Eligible o N o G P Referral 056 10% 20% 30% 40% 50% 60% 70% Source Staff estimates based on special WHS health surwq September-December 2006 Each characteristic was a possible separate response in the questionnaire; responses were combined to make the presentation simpler. 83 197. Figure 37 shows households' perception of the performance of the different elements of the system. There is some good news in this figure. There is high satisfaction with access to NHF's general practitioners, treatment of F i e o u s e h o l d Satisfaction with Aspects of Health Services patients in NHF-financed facilities, and the system for Satisfaction with Medical Service purchasing prescribed Abilityto Get Referrals medicines. Each of these elements of the system is Purchase o f Prescribed Medicines rated good or very good by Treatment o f Patients 60 percent or more of the L Good or Very Good Specialist Wait respondents. Overall Average p1 satisfaction with medical Dentist Wait services is lower, but still Specialist Access Bad or Very Bad with only 10 percent calling it bad or very bad. Access Hospital Access to hospitals is also General Practitioner Access considered good. The weak 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% areas are referrals, access, and waiting time to see Source Staff estimates based on special HBS health survey September-December 2006 specialists, The worst category, specialist waits, gets a 62 percent negative rating of bad or very bad, with dentist waits not far behind. There are no questions about quality of specialist and dentist services, but judging from Figure 36, there would be some dissatisfaction with perceived quality of these services as well. 198. Figure 38 Figure 38: Expenditures Connected to Hospital Stays in 2006 addresses expenditures on 60% 400 hospital stays." As we have seen, the system perhaps 350 50% performs best in providing 2 I financial protection at the 300 u Any Payment E I hospital level. People have U w : 40% 250 Medicines and Medical Goods low out-of-pocket expenses, c 0 a l a have good access, and there t-2 5 30% 200 E" Gifts, Flowers, Cash t o Medical are minimal differences in wm > a Staff Gifts, Flowers, Cash t o Doctor m use across consumption 5 $ Y 150 quintiles. Nevertheless, 20% a U Paymentfor Careofpatient ? there are persistent claims of d 100 8 3 Total Expenditure rf a Payment pervasive side payments to 10% Made get care and other problems. 50 Figure 38 shows, across 0% quintiles, the frequency of s Poorest Quintile 2 Quintile 3 Quintile 4 Richest various payments during 2006 (left axis) and the Source Staff estimates based on special HBS health surve) September-December 2006 average expenditures (right Not all responses are included in the figure, only the ones with the highest frequency and weight in expenditures. 84 axis), "Any Payment" shows the proportion of households that had a payment, which is about 40 percent among the poorest and 50 percent and more among the two richest q~intiles.'~ Among all of the costs, medicines and medical goods once again emerge as important, with roughly 20 percent of all households incurring expenses in this category. Other costs may be important for specific households (such as travel and obligatory fees for health resorts, which is not included in Figure 38), but they do not affect many households. 199. Payment for care of the patient, the shortest of the vertical bars, should not be necessary. It is reported by 2 percent or less of households. However, gifts to caregivers are common and - except for the poorest quintile - the proportion providing gifts seems to be equal at 20 percent across groups rather than rising significantly with ability to pay. Clearly the practice exists; whether it is good or bad, expected by the provider or voluntary, cannot be known from this information. The good news is that roughly 60 percent of the poorest 2 quintiles incurred no expenditure for their hospital stays in 2006. For those who did, expenses were on average no more than PLN150, primarily for pharmaceuticals and gifts. Higher income households spent more, but less relative to their incomes.53 Golinowska et al. report that 36 percent of respondents in 200 1 reported making extra payments to health professionals; a similar question in 2006 elicited a positive response of only 8 percent. 200. Finally, Figure 39 shows household perceptions of costs incurred to use care in 2006. At this point it is almost anticlimactic to present the data. It is definitely consistent with all of the evidence provided to this point. Over 80 percent of the population considers the cost of hospital stays in 2006 to have been small or very small. Perceptions of costs of outpatient services - medical diagnosis, doctors, and dentists - indicate that for at least 50 percent of users, the costs were considered small or very small, and Only 2o percent Or less Figure 39: Household Perception of the Amount Paid for Health found the costs high or Servicesin2006 very high. These responses include people Medicines who used and paid for non-NHF providers. As would be expected from Dentists the earlier evidence, the * High or Very High percentages are flipped Doctors when it comes to Average lg medicines, with half saying costs were high or Med'ca'D'agnos's small or Very Small very high and only 17 percent responding "small H a p i t a l Stay or very small ," Pharmaceutical costs are 0% IOk 20% 30% 40% 50% 6 0 1 70% 80% 90% 100% clearly a sore point with Source Staff estimates based on special HBS health surveq September-December 2006 the population. '' Each household could have had expenses in more than one category; "Any Payment" means they had an expense in at least one category. 53 For example, an average hospital expenditure of PLN1.57 is 53 percent of average per capita monthly expenditures in the poorest quintile. PLN341 in the richest quintile is about 19 percent of their average per capita monthly expenditures. In both cases, the amount spent out of pocket is low relative to the actual cost of the services, but it is a larger percentage of total resources for the poorest quintile. 85 VIII. Accomplishments and Challenges 201. Poland has made huge strides in the development, financing, and management of its health system. It is instructive to begin with some of the major accomplishments. Accomplishments i. It has achieved, for many years, universal coverage of the population. For much of the past 20 years, it has been a struggle to make this system function well, but now most of the population has regular `contact with the health system, particularly at the primary care level, and real financial protection is provided for catastrophic care. Overall satisfaction with the system is reasonably good. ii. While out-of-pocket expenditures remain an important feature of health financing, they increase as a percentage of income as income rises, suggesting that to some degree they a r e discretionary. Between 2005 and 2007, a period a rapidly rising incomes, out-of-pocket health expenditures dropped in absolute terms and as a percent of household consumption. iii. Although access to health services is an entitlement to all Polish citizens, NHF spending is likely to be pro-poor because of more frequent opting into privately financed services as incomes increase. iv. Although health outcomes reflect all elements of life, the health system certainly is an important contributor. Poland has done well in increasing life expectancy and other measures of health. It is among the best performers of the new member states of the EU. Figure 20 at the beginning of the chapter shows a simple comparison with its immediate neighbor, Ukraine, and the EU members before 2004. The departure from Ukraine's record is extraordinary, and there is now clear potential for Poland to catch up to its richer EU neighbors. Catching up is in fact probably the core challenge facing Poland - how to speed up the rate of improvement in health outcomes so that it can close the gap with its richer EU counterparts. Health data suggest that success will require progress in reducing excess mortality from cardiovascular disease. v. A rarely remarked accomplishment is that health facilities, whether public o r private, have hard budget constraints. They must pay their bills out of revenues from NHF or from patients. That has created a repeated debt problem, but it makes financial performance transparent and helps to prevent subsidies from creeping into the system untransparently from other sources. Debts are explicit, their causes can be identified and corrected, and debt relief entails subsidies that all can see. Debt accumulation exposes management and structural problems that would not otherwise be obvious. Poland's approach to a hard budget constraint for facilities is a model for other countries. The corporatization program will reinforce this accomplishment. vi. The system itself and patient behavior have disengaged from the inherited Soviet model of hospital- and specialist-based care and shifted to a primary care system with family doctors as gate keepers. Changing this behavior is a major accomplishment and is probably more advanced in Poland than in most of the former Soviet bloc countries, all of which are still struggling with inherited patterns of supplier and patient behavior. Poland's separation of 86 financing from provision and development of primary care as a gatekeeper to specialist and hospital services has probably been the cause of its success. The facts that the population is reasonably well satisfied with the system and with access to general practitioners are markers of the success of this policy. vii. The system is adequately financed. Poland has been criticized for under financing health services in part because of the various caps in place - principally the percentage of the PIT devoted to health, copayments for drugs and some services, and the inability of local authorities to subsidize running costs of institutions they own. 54 The percentage of PIT going to NHF has increased over the years, and spending increased substantially in 2007 and 2008 with the tremendous economic growth, rise in incomes that occurred, and increase in taxable income caused by the reduction in disability tax rate. The result has been a reduction of hospital debt and declining out-of-pocket payments. Moreover, Poland is doing well in terms of outcomes and has been for many years. Any arguments for higher spending at this point would need to be focused on specific goals, such as reducing excess mortality due to cardiovascular illness, reducing the burden of pharmaceutical costs for the elderly, or further reducing out-of-pocket costs for the poor. Even these improvements, however, probably could be financed through adjustments in priorities and policies within the existing expenditure envelope. viii. The shift to a prospective payment system (DRGs) with global budgets has promise to contribute to a further rationalization of hospitals and, potentially, to better integration of outpatient and specialist services and shifting short hospital stays to outpatient or day services. 202. With virtually every one of these accomplishments comes an additional challenge in the never-ending evolution of health problems and policy. Challenges I. What would it take to catch up to the richer EU states in health outcomes by 2020? What strategies would work in Poland at different levels of available resources? 11. How to make the voluntary corporatization program succeed, how to apply it successfully to university hospitals, and how to use the program to encourage modernization and development of the sector. iii. Development of stronger referral networks and integrated care networks to further de- hospitalize the system, increase the role and responsibilities of primary care and specialist outpatient providers, and develop proactive approaches to reduce further adult morbidity and mortality. Programs to improve the quality of care might sensibly focus on the network issue in the short run and how to use payment mechanisms to reinforce good medical practice. Careful development of inpatient and outpatient DRGs - in tandem with a focus on how to improve medical outcomes and medical practices by diagnostic group - has the potential to integrate financing incentives and good medical practice through a partnership among NHF, the MoH, academics, and caregivers. 54Although debt accumulation becomes a subsidy once taken over by government, and municipalities do provide loans to facilities they own. a7 iv Similarly, accreditation of facilities would be far more meaningful if addressed in the context of NHF financing of services. Quality, cost reduction, and incentives for network development could be advanced through competitive contracting among accredited institutions. V. How to address public perceptions of low access, long waits, and quality issues with respect to dental and specialist services. In particular, the NHF is not adequatelyserving the poorest segments of the population when they feel they need to opt out of NI-IF-financed care and pay out of pocket. vi. How, in a constrained economic environment over the next couple of years, to reduce growth in expenditures while building on the positive equity characteristics of the system and spending carefully to improve health outcomes for the population. vii. Better monitoring of out-of-pocket costs and developing surgical, low-cost interventions to solve specific financial problems, such as reducing out-of-pocket costs for the poor and taking specific actions to reduce the burden of pharmaceutical costs for elderly households. viii. As use of private services increases, demands for alternatives to NHF and competition among insurers will escalate. The challenge will be how to maintain universal coverage for a basic level of adequately financed services for all, while increasing competition and choice that allows those who want optional services to pay for them, but does not lead to unnecessarily high administrative costs. It is a difficult set of policy issues. ix. How to build a strong evaluation framework into the system to increase its cost- effectiveness, the health impact of payment and coverage decisions by insurers, and to improve horizontal and vertical equity both on the revenue and expenditures sides of the system. This is probably the key missing element in Poland's institutional framework for financing and delivering health services. X. How to manage long term care as a social issue rather than a health issue, while assuring health care for the disabled, handicapped, chronically ill, and elderly in the most cost efficient yet humane manner. xi. We have only touched on the revenue side of the equation, but a broader tax base for health financing with lower contribution rates, as well as greater horizontal equity in contributions (people of similar incomes contributing similar amounts) would be beneficial to the economy (although it would have no impact on the health sector specifically). Poland's unique method of financing NHF lends itself to further improvement without a dramatic departure from the existing approach. For example, as the economy continues to grow, lowering the contribution rate of employees to NHF and substituting general tax revenues for it would move forward on all three goals - lowering the tax rate, broadening the tax base, and moving the rate paid in the nonfarm sectors incrementally closer to the rate paid by farmers. 88 IX. Policy Framework Going Forward 203. The agenda of the current government was developed during 2007 and 2008. It includes definition of a basic care package to be financed by NHF, introduction of private insurance to finance services not covered, the transformation of SPZOZ hospitals to NZOZ commercial code companies, creation of a single Patient's Ombudsman, a new legal framework for accreditation, and a new legal framework for health professionals. Most of this agenda and how it is proceeding in the face of a Presidential veto has been discussed in this chapter. In July 2009, however, the President signed a law that clarifies what patients are entitled to receive free of charge under NHF, thus paving the way for the potential introduction of fees for services not included in the basic package and further development of the private insurance market. 204. Based on the analysis in this chapter, we conclude that Poland has moved in a sensible direction on many fronts. It still has many challenges, and the internal debate on every aspect of health care and health care financing is intense. The development of private insurance was characterized as a method to inject additional resources into the system to allow the NHF portion to finance the core set of services. Our conclusion is that additional resources are not the priority they once were, and Poland can take its time to carefully design future changes to its health financing system. Economic growth has done wonders in expanding the resource base. Improving health outcomes within the current expenditure framework via efficiency-improving, medical practice-improving, and network-friendly payment policies is a priority. Addressing patient concerns on specialist and dental services can be done with clever policy making that does not turn the whole system upside down. Addressing persistent concerns on drug costs as well as addressing the negative equity impacts of out-of-pocket spending (primarily for drugs) can also be done with fairly marginal and well targeted changes. Developing and implementing these changes would require stronger analytical work than is currently available to the MoH and the NHF. While we find most of what the government is doing to be sensible, we encourage development of a greater capability to use available data to analyze the impact of policies and to institutionalize such analysis in NHF and MoH decision making. 89 90 8 5 3 d ? 3 D, * l u N m o 22 n a 0 'Z .3 I n a 0 .I 5 cL( d I 'S 8 5 k .I & c) a $ u 2 L . L I a a E 0 W c1 a 2 L a z c) a u W a a Y 00 4 0 0 . cl 6 e 2 4 I a s c) k 6 a 0 a 4 0 .I c) e .I d I E ki s a. r( N 3 Y Y 2; F 3 3 J 2 D i 3 2 D 3 3 L E Y m h 0 9 % c m m w Y Y Y G X 3 n C .z? U m C a 3 SOCIAL ASSISTANCE CASH TRANSFERS" Introduction and Structure of the Chapter 205. This chapter of the PER focuses on social assistance: the set of social protection transfers for which there is no explicit contribution required from households and firms for eligibility, other than the taxes they pay that pool into the general revenues of the State. As in other countries in the region and around the world, these "non-contributory" transfers are designed to alleviate and eliminate poverty, as well as to supplement household income to ensure a minimum standard of welfare. Like many countries in Central and Eastern Europe, Poland does not strictly limit eligibility to social assistance transfers based on income, and provides a number of "universal" or non-targeted forms of assistance. However, Poland is one of few countries in the Region that targets family and child benefits to its poorest households. This has been a policy since the mid 199Os, and as family and child benefits are typically the largest category of social assistance transfers, targeting in Poland has led to a relatively progressive distribution of spending on non-contributory social protection. Reforms that further consolidated benefits related to children into the targeted category in 2004 were intended to increase progressive outcomes. Steps taken to decentralize social assistance since are also expected to have had an impact on performance. 206. This chapter of the PER has six sections. Section 2 briefly describes the main features of social assistance transfer programs in Poland. Section 3 presents trends in aggregate spending on social assistance relative to spending on other programs in the Polish social protection system, and compared with spending on similar programs in neighboring countries. Section 4 examines the outcomes of public spending on social assistance, focusing on three specific measures of performance: the extent to which the Government is covering eligible households, targeting spending to these households, and providing an adequate supplement to household income. Section 5 uses these same performance measures to assess the outcomes of social assistance spending at the district level ("voyvods", Poland's sixteen NUTS 2 district administrative units), motivating a discussion of the advantages and disadvantages of decentralizing the delivery of social assistance, and the experience of other countries in the region with decentralizing different functions. It is important, from the outset, to clarify that this discussion can only be indicative and outline general concerns that should be examined further, as the data currently available do not allow for a full assessment of the performance of local self governments ("gminas", Poland's 2,478 Municipality/commune administrative units) in administering social assistance. Section 6 summarizes the findings of this chapter and concludes. 207. In order to place social assistance in context, the spending and performance statistics discussed in this chapter are presented alongside those for "contributory" social insurance, such as pensions and disability, as well as unemployment benefits financed by contributions from employers to the Labor Fund. However, these branches of Poland's social protection system are covered in depth in chapters Pensions and Labor Market of the PER. The analysis in this chapter is supported with statistics drawn from three waves of Poland's Household Budget Survey (from 2003, 2005 and 2007). 55 This chapter of the PER was written by Truman G. Packard (Sr. Economist, ECSHD), World Bank, Anna Ruzik and Katarzyna Pietka, from the Center for Social and Economic Research (CASE) in Warsaw,. Ruzik and Pietka estimated the coverage, targeting and adequacy statistics from Poland's Household Budget Survey on which the analysis and assertions made in the chapter rests. The conclusions and any errors in interpreting estimation results are the sole responsibility of the authors. 208. The main findings of this chapter can be summarized as follows. Relative to other branches of the social protection system, Poland's spending on social assistance transfers is low, even by regional standards. Of the 15.7 percent of GDP Poland spent on social protection in 2007, only 1.4 percentage points was spent on family and child allowances and minimum income assistance, compared to 9 percentage points spent on old age pensions. In a region where long-term unemployment and dependence on social assistance has long been a problem, and where targeting social assistance benefits in order to improve efficiency is still a relatively new and rare approach, a relatively low level of spending is not on its own a cause of concern. 209. However, even among the countries of Central and Eastern Europe that target family and child benefits to poorer households, spending in Poland on these benefit categories is still relatively low. In aggregate the share that Poland spends on targeted social protection programs (as a share of all social protection spending) has risen since 2004 with the consolidation of family and child allowances. Other than Poland, the countries in the region that target family and child benefits to poorer households are the Czech Republic and Slovenia, where governments spent 2.5 and 2.3 percent of GDP on family and child benefits and minimum income assistance, respectively, in 2007, compared to Poland's 1.4 percent of GDP. In 2009, the Government passed important reforms substantially tightening eligibility for early retirement, which might create some fiscal space for increasing spending on targeted social assistance transfers. But these changes will take several years to generate significant fiscal savings. 210. Furthermore, coverage estimates presented in this chapter show as much as 31 percent of the population who met the eligibility criteria for some form of targeted social assistance in 2007 was not receiving benefits. The social insurance branches of the social protection system help make up this short fall: when all social protection benefits are taken into account, the share of the population eligible for social assistance benefits but not receiving any form of social protection falls to 14.8 percent. For an EU member country, this coverage gap is still surprisingly large, and could point to a lack of information among potentially eligible households about entitlements, as well as a need for more proactive outreach to vulnerable groups from local social welfare offices. 21 1, Although there have been measurable improvements since 2003 in targeting when performance statistics are disaggregated to capture differences across voyvods, the impact of the steps taken to decentralize social assistance has been mixed. Large gaps in coverage remain a cause for concern. The evidence presented in this chapter does not point to any clear improvement in performance from giving local governments a greater role in administering social assistance. In several cases the targeting of family allowances and minimum income assistance has actually worsened. Having said this, a more thorough examination of benefit administration at the gmina level is required using data that can capture improvement at the local level, in order to come to a more a definitive conclusion about what is contributing to large gaps in coverage, leakage to non-poor and otherwise ineligible households, and the impact decentralization has had. Nevertheless, the results presented in this chapter motivate a thorough review of social assistance cash transfers, and a careful discussion and caution in determining the optimal role for local governments in the delivery of assistance. I. . The Parameters and Delivery of Social Assistance Transfers in Poland 212. Most social protection programs in Poland - and certainly those accounting for the largest share of spending- are centralized. This is true of social insurance programs, such as old age pensions, survivor and disability protection (as well as health insurance), of unemployment benefits and employment programs (covered elsewhere in the PER), as well as of the largest social assistance program, family allowances and child benefits. 104 2 13. However, examined as a separate category of social protection, social assistance transfers are distinguished in that programs are partially financed by local governments, with the share of local expenditure increasing since 1999 and particularly from 2004. This increase is likely to continue. Table 23 shows the current division of responsibilities for financing and delivering social assistance to households in Poland, - Table 23: Central and Local Government Responsibility for Delivering Social Assistance Central I Local Social Assistance Program Implementation Financing * Responsibility C' Family allowance + other Local: eligibility assessment Central via targeted subsidy to benefits complementing and payment local government fam i 1y allowance C Career's benefit + nursing Local: eligibility assessment Central via targeted subsidy to benefit and payment local government C/L2 Housing Benefit Local: eligibility assessment local (from taxes) and payment C Social pension Central Central C/L3 Minimum income assistance Local: criteria assessment For some benefits - central benefits, cash + in-kind and payment financing via targeted subsidy to local government; for other programs - local (from taxes) 4 Child allowance during - prolonged maternity + for - single + single unemployed Notes: * decision about granting a benefit: - Central (C) if the rules are set centrally and local authorities are responsible only for management of granting and payments of benefits Local (L) - if the rules are set by local authorities and benefits are financed from their local budgets 1) Gminas are allowed to increase benefits complementing family allowances above the legally set minimum level of the benefit; Gminas are allowed to grant an additional one-off benefit in case of child birth 2) Gminas are allowed to increaseidecrease benefits up to a certain level 3) Permanent and temporary minimum income assistance benefits or benefits for special purposes are granted from targeted subsidy from the central to local governments; Gminas are allowed to grant additional benefits for special purposes and increase the level of temporary benefits from their own resources. They are also allowed to raise the eligibility criteria. 4) Those benefits were transformed into benefits complementing family allowances in 2004 214. Family allowances and child benefits are fully financed from the state budget. Until 2004, social assistance coverage for families and children was highly fragmented in a complex array of sometimes overlapping benefits. In 2004, the system was reformed by consolidating the various benefits related to families into single family allowance and child benefits programss6The core family allowance 56 Family and child benefits were reformed in May 2004. The main elements of the reform include (i) linkage of the family support level with the age of the child and not the number of children (as it used to be before), (ii) replacement of various benefits paid to families with a single benefit, supplemented with additional allowances as required; (iii) additions to the family benefit: for child bearing (a lump-sum), taking care for a child during child- care leave, for lone parents and the loss of rights to unemployment benefit, for the beginning of the school-year (lump-sum, paid in September), for education and rehabilitation of a disabled child (until the age of 16 or 24 in the case of moderate and severe disability); (iv) a decrease of the income threshold test to be eligible for family benefits. 105 is supplemented with benefits granted for specific circumstances and situations: such as for disabled children; grants to cover costs at the beginning of the school year; benefits to cover the costs of commuting to school, benefits for prolonged maternity and others. Child allowances have been income- tested based on family income since 199SS7and the income criteria for eligibility were lowered in 2004. Child allowances are paid at three different benefit levels, depending on the age of children in the household, and given monthly up to the age of 18, or 21 if a child is in education, or 24 if a child is in higher education and disabled. Although financed centrally, the program is operated by local government^.^^ Alongside the targeted family allowance and child benefit program and apart from a 1000 PLN benefit for all families who had a child and are entitled to family benefits, there is a universal, one- off cash payment made to families who have just had a child ("becikowe"). This universal benefit is paid regardless the family's means and financed from the state budget. 2 15. Minimum income assistance is also based on income-testing. Financing of minimum income assistance is currently shared between the national budget and local government budgets. The bulk of the financing still comes from the central budget, in the form of open capitation. Since the administrative reform of 1999, administrative tasks that were once performed by central administration have been shifted to local governments. Today benefit delivery is firmly the responsibility of gminas and conducted by gmina social welfare offices. There are two forms of minimum income assistance: permanent benefits and temporary benefits. Eligibility for one form of the other is the decision of local government social workers based on their assessment and in accordance with the law. Table 24 presents the main eligibility parameters of these programs. 216. Per capita net income is used in the means test to determine eligibility for both family allowances (including child benefits) as well as minimum income assistance. In the case for family allowances, income in the previous fiscal year is assessed from income tax declarations, while for minimum income assistance a household's earnings in the previous month are used. The income eligibility threshold is higher for family allowances than for minimum income assistance. However, eligibility for benefits is also determined by factors other than poverty. Family benefits are paid for a longer period to households with a disabled child. Whether minimum income assistance is permanent or temporary depends on whether the recipient is disabled or able to work. Local social assistance workers take additional factors that contribute to a household's vulnerability into account when determining eligibility for minimum income a s s i ~ t a n c e . ~ ~ (Law on Employment Promotion and Labor Market Institutions, April 20th, 2004; Law on Family Benefits of November 2gth,2003; Law on Social Assistance of March 12`h,2004) 57 Before, the reference for the income threshold was an individual income of a beneficiary and only social security contributions were covered. 58 The level of family allowances and all benefits complementing them are adjusted every three years to reflect changes in the cost of living. The last such adjustment took place on the 1st of September 2006, and the next is planned for the second half of 2009. Since September 2006 the level of the basic family allowance depends on the age of the child (and not - as before - on the number of children in a family). The lowest benefit is granted for children aged up to 5 years, then higher for children aged over 5 to 18 and the highest for children in education aged over 18 and below 24. 59 These can include homelessness, long-term unemployment, long-term sickness, being an orphan, home abuse, addictions, single parenthood, problems with adaptation to society after leaving prison, and special crisis situations. 106 Table 24: Eligibility for Poland's Main Social Assistance Programs Age Residence Means testing es Zn Of Income threshold means testing for Max. family solidarity testing ( 1 ) (2) Children aged up Family to 18 or to 24 Family income per capita below PLN 504 monthly, or and child when in higher yes Yes 3 benefits education or PLN 583 monthly if a child disabled is disabled. Income for a one-person household below 477 PLN, Minimum in larger households: 351 income no Yes yes 3 PLN. per person. Local 5 assistance government can increase income thresholds entitled to -- temporary benefits. Notes; 1) Type of means testing = means taken into account: I = income out of labor only: 2 = income out of pension only; 3 = all income; 4 = income and assets. 2) "Maximum level of family solidarity'' indicates that the following persons are taken into account in assessing eligibility: I = person/benejciaty only; 2 = I -+ cohabitating official spouse only; 3 = I + partner cohabitating: 4 = 1 + official spouse non-cohabitating; 5 = 3 + children cohabitating; 6 = 3 -t children non-cohabitating; 7 = 3 -+ parents cohabitating; 8 = 3 + parents non-cohabiting; 9 = 5 + grandparents/grandchildren cohabiting; IO = 5 + grandparents/children non-cohabiting: I I = 5 + all family members in third degree. 217. There are no compulsory "activation" measures for people who are entitled to social assistance in Poland. Existing programmes designed to increase the labour force participation of the unemployed and/or disabled people are voluntary for beneficiaries. In order to better link employment and social assistance, the Government is planning to take a number of steps to support activation of long- term unemployed, and to reduce dependence on social assistance transfers (temporary benefits). The MLSP is eager to improve coordination between public employment services (which are under voyvod governments) and social assistance administered by the social welfare offices (which are run by gminas). Better information sharing and coordination of benefits between employment and social assistance offices could encourage the movement of people out of long-term unemployment and into the labor market, as well as reduce dependence on social assistancea6' 218. An important distinction of Poland's main social assistance transfers should be kept in mind when assessing the performance indicators presented in the sections that follow. Social assistance transfers are legal entitlements for eligible households, and not simply programs rationed by funds allocated in the national budget. The state has a legal obligation to pay these transfers to eligible individuals and eligible households who fail to receive the benefits they are entitled to, can compel payment through the courts. 219. Funding for mandatory social assistance benefits paid by gminas is transferred from the national budget to voyvods. At the start of each budget cycle, voyvods present their estimate of financing needs for benefits based on estimates submitted by gminas. These estimates are revised at the start and the end of every month as benefits are granted and paid. The head of a voyvod administration 6o As a part of this effort, from 2004 the MLSP has been expanding the use of a `social contract' that sets out job- search effort expectations and commitments between a social worker and a beneficiary. As of 2007, 65,800 beneficiaries had signed a social contract that clearly outlines their commitment to job-search. The Ministry would like to see the use of the contracts increased, and has set itself the target of having had 150,000 beneficiaries sign contracts by 20 10. 107 (wojewoda) is responsible for shifting financing between gminas within the voyvod, from those that have overestimated their needs for social assistance to those that underestimated their needs. If at the voyvod level there are not enough resources to cover demands from households found to be eligible, the wojewoda is responsible for shifting money from other, non-obligatory budget line items in order to finance obligatory social assistance transfers. If there is still a short fall, the MLSP is required to provide resources from a I`eserve that is kept for this purpose (rezewa celowa). 220. The Ministry maintains a sophisticated system of statistical monitoring to ensure payment of benefits at the gmina level. Using this system the Ministry monitors and sometimes corrects the estimates submitted by the gminas of what they will need. These adjustments are done based on monitoring each gmina on a quarterly basis. If a gmina is in need of additional funding, it will receive resources from the voyvod. If there isn't sufficient money at the voyvod level then the Ministry can use its reserve. 221. Finally, although not directly related to the social assistance branch of Poland's social protection system, changes were introduced to the personal income tax in January 2008 which are likely to have a large impact on household welfare. The Government increased the annual personal income tax deduction for every child almost 10-fold from 120 PLN to 1145 PLN (1 173.70 PLN in 2008). The change covered all people paying personal income tax and was effective also for 2007 taxes. In 2008 over 4.2 million people used this opportunity of tax deduction and lowered their personal income tax liability by over PLN 6 billion. The tax deduction affected 6358 thousand children. From 2009, for the children born in a given calendar year, the amount of the tax deduction is proportional to the number of months after the child birth. This change will not be reflected in the household data analysis in this chapter, but could have an impact on observable performance outcomes measured in later waves of the Household Budget Survey. 11. Spending on Social Assistance in Poland and Across the Region 222. Social protection policies and programs are examined in three chapters of this PER. An adequate analysis of spending on social protection requires this sort of extensive treatment, given the different objectives and financing models of each branch of the system and the different levels of government involved in their administration. Although this chapter focuses on just one branch of the system - non-contributory social assistance transfers - it is important Table 25: Total Spending on Social Protection in Poland and to put family benefits and minimum Central and Eastern European EU States, (Percentage of income assistance into context. This G D a~ ~ - - 2002 2003 2004 2005 2006 2007 -_11_ ~ section on spending attempts to do just that by presenting Poland's Hungary 15.4 15.9 16 17 17.6 17.3 public investment in protecting 18.3 18.8 17.6 17 16.9 15.7 vulnerable and poor families 17 17 17.4 .3 16.9 15.5 alongside spending on other forms 14.4 13.5 11.3 -1 12.2 13.1 of social protection. Czech 14.6 14.4 13 12.8 1 12,9 Republic Lithuania 10.8 10.3 10.2 9.9 9.9 11.1 223. Table 25 presents Slovakia 14.9 16.3 12.1 13.2 12.4 10.6 comparable aggregated Romania 10.4 9.3 9.6 9.8 9.7 9.9 expenditure data for social protection as a whole, and various Estonia 10 9.9 10.2 9.9 9.6 9.8 , program categories classified Latvia 11.5 10.7 10.4 9.8 9.8 8.4 according to the Europeans ~ o ~ r cEurostat e: 108 Union's System of Integrated Social Protection Statistics (ESSPROS). The social protection benefit categories of particular interest in this chapter are "Family & Child" which capture Poland's family allowances and complementary benefits, and "Social Exclusion" which captures minimum income assistance and other means-tested benefits. 224. Among the eight states that joined the EU in May 2004, and Romania and Bulgaria that joined in January 2007, Poland ranked the second highest for spending on social protection as a percentage of GDP in 2007 (15.7 percent). In the same year (the last for which comparable data are available from EUROSTAT-ESSPROS), Poland spent the second highest amount on old-age pensions (9.0 percent of GDP) and survivor benefits (1.7 percent of GDP). The large share of spending for the elderly reflects the generous disability benefits and widespread access to early retirement pensions. The Government has taken steps to substantially curtail both, but these will take time to translate into fiscal savings. 225. In contrast, when ranked for spending on the programs that are the main interest of this chapter - family benefits and minimum income assistance - Poland is close to the bottom of the list (1.2 percent of GDP on family benefits and 0.2 percent of GDP on minimum income assistance). This last observation in part reflects Poland's consolidation of family benefits in 2004. Among the three countries in the table that target family benefits according to household income (Slovenia, the Czech Republic and Poland), Poland spent the least on this category of benefits in 2007. Table 26: Spending on Social Protection Programs in Poland and Central and Eastern European EU States by Function (Percentage of GDP in 2007) Total Sickness & Old Survivors Family Unemployment Housing Social Other Disability Age & Child Exclusion Benefits Hungary 17.3 3.7 6.8 1.3 2.4 0.6 0.9 1.5 0.2 Poland 15.7 2.3 9 I .7 1.2 1 0.I 0.2 0.1 Slovenia 15.5 2.8 8.2 1.5 1.7 0.5 0 0.6 0.3 Bulgaria 13 1 02 9.8 2.4 01 0.6 Czech 12.9 2.7 6.5 0.6 2.1 0.3 0.1 0.4 0.2 Republic Lithuania 11.1 28 5.1 0.5 1.1 0.4 0.1 0.2 1 SIov akia 10.6 Romania 9.9 Estonia 9.8 1.7 5.6 0.1 1.6 0.4 0 0 0.3 Latvia 8.4 I .5 4.9 0 0.9 0.3 01 0.2 0.4 Source; Eurostat 226. The pattern of spending on social protection programs has remained fairly consistent in recent years. Table 27 shows the break-down of social protection spending across the various branches of the system from 2002 - 2007. With regard to the benefits of specific interest to this chapter - family and child benefits and minimum income ("social exclusion") assistance - Poland has remained fairly constant in its spending (see Table 28) since 2002. The consolidation of family benefits in 2004 cut spending on these programs only slightly from 1.3 percent of GDP in 2004 to 1.2 percent annually in the years since. 109 227. Although Poland has targeted family allowances based on family income since 1995, the shift to means testing social assistance is relatively new Table 27: Trends in Spending on Social Protection Programs in - elsewhere in the region. Table Poland, 2002 - 2007 (Percentage of GDP) 29 shows expenditure on 2002 2003 2004 2005 2006 2007 means-tested programs as a Old Age 11.1 11.4 10.7 10.3 9.9 9.0 share of total spending on social Sickness & 2.8 3.0 2.4 2.4 2.5 2.3 protection across Central and Disability Eastern Europe. Poland and its Survivors 1.7 1.7 1.7 1.7 1.8 1.7 neighbors are ranked from Family & Children 1.3 1.3 1.3 1.2 1.2 1.2 highest to lowest, according to Unemployment 0.9 0.9 1.0 `0.9 1.0 1.0 the amount they spent on means Social Exclusion 0.3 0.2 0.2 0.2 0.2 0.2 tested programs in 2006 - the Housing 0.1 0.1 0.1 0.1 0.1 0.1 last year in this series. The Others 0.1 0.1 0.1 0.2 0.1 0.1 consolidation of family benefits Source: Eurostat may explain the rise in Poland's spending on means tested programs from 4.4 percent of total social protection spending in 2003 to 5.2 in 2004 and higher in 2005. Table 28: Trends in Spending on Social Assistance in the Central and Eastern European New Member States, 2002 - 2007, (Percentage of GDP) - Social assistance 2002 2003 2004 2005 2006 2007 - Bulgaria type Family & children 3.4 1 2.1 1.1 2.9 2.4 Social exclusion Hungary Family & children 2.1 2.3 2.1 2.1 2.4 2.4 Social exclusion 1.3 1.2 1.3 1.6 1.8 1.5 Czech Family & children 1.6 1.7 1.5 1.5 1.5 2. I Republic Social exclusion 0.6 0.6 0.8 0.5 0.5 0.4 Slovenia Family & children 1,9 1.9 1.9 1.9 1.9 1.7 Social exclusion 0.7 0.7 0.8 0.7 0.7 0.6 Estonia Family & children 1.5 1.4 1.6 1.7 1.6 1.6 Social exclusion 0 0 0.2 0 0 0 Poland Family & children 1.3 1.3 1.3 1.2 1.2 1.2 Social exclusion 0.3 0.2 0.2 0.2 0.2 0.2 Lithuania Family & children 0.5 0.5 1.1 1 1 1.1 Social exclusion 0.3 0.2 0.2 0.2 Latvia Family & children 0.9 Social exclusion 0.2 Source: Eurostat 110 Table 29: Trends in Spending on Means Tested Social Protection 2000 - 2006, - - (Percentage of Total Spending on Social Protection) 2000 2001 2002 2003 2004 2005 2006 Slovenia 9.4 9.1 9.1 9.8 9.8 9.7 9.1 Bulgaria 6.6 6.1 Slovakia 12.8 13.0 10.8 7.2 5.1 5.4 5.9 Poland 5.0 4.7 3.7 4.4 5.2 6.4 5.2 Czech 7.8 7.3 7.1 6.7 6.9 5.4 5.2 Republic Romania 8.1 5.4 6.1 6.4 9.4 7.4 5.1 Hungary 6.8 6.2 5.7 5.3 5 .O 5.3 4.6 Lithuania 4.6 4.5 4.8 4.6 3.4 2.1 1.8 Latvia 1.7 1.8 1.8 I .6 1.5 1.5 1.4 Estonia 2.5 2.6 2.4 2.1 1.3 1.o 0.8 Source: ESSPROS 228. As a share of spending on all family and child benefits, Poland spends the most on means tested programs, and - unlike in other countries- this share has grown in recent years. This has lead to a relatively more progressive distribution of spending on social assistance in Poland when compared with the performance of similar programs in neighboring countries, as will be discussed in the next section of this chapter. Table 30: Trends in Spending on Means Tested Family and Child Benefits 2000- 2006 (Percentage of Total Spending on Family and Child Benefits) 2000 2001 2002 2003 2004 2005 2006 -- Poland 63.3 58.9 56.9 56.2 71.9 76.6 71.9 Slovenia 75.1 74.2 73.7 74.4 71.4 70.5 67.8 Bulgaria 57.2 48.8 Czech Republic 57.6 55.3 54.4 51.4 49.7 38.6 35.6 Romania 0.8 0.5 1.2 1.2 15.3 18.1 12.7 Lithuania 18.6 20.0 19.6 19.1 13.1 7.6 6.9 Hungary 11.9 12.7 12.6 10.1 10.8 10.7 3.5 Latvia 4.7 3.8 3.8 2.3 2.0 1.9 2.0 Slovakia 54.9 54.5 25.6 1.o 0.9 1.1 1.7 Source: ESSPROS 111. Performance of Social Assistance: Coverage, Targeting and Importance in Household Budgets 229. In 2007 about 47 percent of the population received some form of social assistance (see Table 31). About 24.2 percent received family allowances and complementary benefits, and 6.7 percent received minimum income assistance cash and in-kind benefits. Not surprisingly, the coverage of social insurance benefits is relatively even across the deciles of the consumption distribution, although a significantly higher share of households in the poorest quintile receive social insurance benefits. Taken as 111 a whole, social assistance covers about 64 percent of the population in the poorest decile. The pattern of coverage of social assistance, when compared with social insurance and labor market programs, is dramatically more progressive, showing a steep decline up the consumption distribution. This is particularly the case of family benefits and even more so for minimum income assistance in cash and in- kind. 230. Coverage of social assistance as a whole fell from 54.6 percent of the population in 2003 to 50.6 percent in 2005 and to 47.6 in 2007. Notably the coverage of family benefits in the population as a whole fell during this period from 32.1 percent in 2003 to 27 percent in 2005 to 24.2 in 2007. The coverage of minimum income assistance, also fell from 8.1 percent of the population in 2003 to 7.4 percent in 2005, and then to 6.7 percent in 2007 reported in Table 3 1. 231. A critical measure of coverage performance is the share of poor or otherwise entitled households that fail to receive benefits. Although the data in Table 3 1 help in determining the degree to which poorer groups are served by the system, they are not a fully reliable measure of this aspect of performance. The equivalent consumption deciles do not correspond to the eligibility criteria currently stipulated in the law. In the case of family allowances, eligibility can extend to households in the fourth consumption decile and sometimes higher. Most of the eligible population according to the legal criteria are concentrated in the 1" consumption decile (62.3 per cent in 2007) and in the 2nddecile (32.3 per cent). However, there are also eligible households among the 3rddecile (5.1 per cent) and even among the 4`h decile (although just 0.1 per cent of the eligible population). Table 31: Social Protection Coverage, ( percent of Households in Each Group who Report Receiving Benefits) Deciles by Household Consumption Total 1 2 3 4 5 6 7 8 9 10 All social protection 72.3 84.8 82.2 79.8 77.5 76.4 72.6 70.6 66.1 61.6 51.0 All social insurance 45.6 44.0 47.5 49.2 49.4 50.6 48,2 47.7 45.5 41.8 32.3 Old Age Retirement Pension 30.5 23.2 27.7 30.5 33.4 34.7 33.7 33.3 33.2 31.1 24.5 All labor market programs 5.3 6.3 6.8 6.0 5.9 6.5 5.9 5.5 4.1 3.8 2.6 Unemployment benefits & 2.7 3.9 4.0 3.0 2.9 3.1 3.0 2.3 2.1 1.8 1.1 assistance Early retirement benefits 2.7 2.5 2.9 3.0 3.1 3.5 2.9 3.4 2.0 2.0 1.5 All social assistance 47.6 63.6 58.9 56.0 51.7 49.2 45.6 43.8 40.7 36.6 30.5 Family allowance (and 24.2 47.3 40.6 36.0 29.9 25.0 21.1 16.9 12.2 8.6 4.0 complementing benefits) Career's benefit & nursing benefit 3.7 5.3 6.2 5.0 4.8 3.6 3.4 2.9 2.5 2.1 1.3 Housing Benefit 5.0 11.3 9.6 8.1 6.2 4.4 4.2 2.6 1.7 1.1 0.6 Minimum income assistance 1.6 2.8 2.6 1.9 1.9 1.7 1.5 1.0 1.0 0.7 0.6 (permanent social pension) Minimum income assistanae 5.1 15.9 9.6 6.5 5.6 3.7 3.5 2.5 2.2 1.4 0.5 (temporary benefits, cash & in kind) - Source: Estimates using HBS 2007 232. Using these eligibility criteria, a coverage statistic can be calculated that identifies the extent to which Poland's social assistance transfers fall short of meeting their coverage objective. In social assistance policy literature this shortfall is sometimes called a "coverage gap" or an "error of exclusion". Readers should note that the eligibility criteria set in the law cannot be exactly replicated, and for this reason, the estimates of the coverage gap are just approximations, although the best that can be achieved with existing household survey data. 112 233. In 2007, 31 percent of the Polish population that was entitled to some form of social assistance (either to family allowances and child benefits, or to minimum income assistance) was not receiving any of these benefits. This is cause for concern as the coverage gap has remained stubbornly stable despite reforms since 2003 (30.1 per cent) and has actually worsened since 2005 (28.2 percent, after a short period of improvement from the level in 2003). 234. Some advances in covering the entitled population and closing this coverage gap since 2003 have been noted in rural areas. From 2003 to 2005 the share of the entitled population without social assistance benefits fell by 3.3 percentage points in rural areas. From 2003 to 2007 the coverage gap in rural areas was narrowed by 1.1 percentage points showing some gains. At the same time the coverage gap in urban areas increased by 3.6 percentage points. 235. When the coverage of the broader set of social protection programs is taken into account, the "gap" in coverage closes significantly. Taking account of social insurance and labor programs, the share of the population eligible for social assistance who are not receiving any form of social protection was 14.8 percent in 2007. Worryingly, this gap has also increased from 13 percent in 2003 and 13.4 in 2005. Does Social Assistance Spending Reach its Target? 236. Income tested social assistance programs in Poland are relatively well targeted when compared to Income tested programs in other countries in Central and Eastern Europe. However, substantial leakages to non-poor groups are still a concern. Poland, along with the Czech Republic and Slovenia are the only countries among those in the region that joined the EU in May 2004, Bulgaria and Romania, that target family benefits based on an income test. Since 2004, all benefits targeted to families with children have been consolidated with family allowances, contributing to the relatively more progressive distribution of Poland's social assistance programs. 237. However leakage of targeted benefits to non-poor and otherwise ineligible groups remains a challenge. The coverage data in Table 31 show nearly 20 per cent of the population receiving social assistance benefits belong to the five highest deciles of the consumption distribution. This is a problem as the official social assistance income eligibility criteria lie somewhere in between the 1'' and the 4`h decile. Around 47 per cent of the population receiving social assistance benefits is ineligible (according to the legal threshold and based on their consumption). For a portion of these households, the social assistance transfers themselves could be pushing them above the eligibility threshold. However, even allowing for this, the size of those numbers clearly suggest a leakage problem. 238. Estimates for 2007 show that gains in targeting efficiency could be made to mop up leakage of social assistance benefits being received by households in the higher consumption deciles (see Table 32). The last two columns of Table 32 present a measure of targeting efficiency that increases with better targetinga6'By this measure, Poland's means tested programs perform relatively well, scoring from 1.7 to 2.1. As a means of comparison with neighboring countries, a regional study on social assistance in 6 ' In Table 8, CGH is the "Coady-Grosh-Hoddinnott" indicator as cited from Ringold, D. and L. Kasek, (2007) "Social Assistance in the New EU Member States: Strengthening Performance and Labor Market Incentives" Working Paper No. 117, Washington DC. The CGH indicator is constructed by dividing the actual outcome by the neutral outcome for the bottom quintile (or decile) of the distribution of households (by consumption). If the poorest 20 percent of households received 30 percent of benefits, the CGH indicator would be (30120) = 1.5, i.e. targeting has led to the poorest quintile receiving 50 percent more than they would have received had benefits been distributed evenly. 113 the new EU member states found social assistance transfers in Lithuania to be the best targeted with a score just over 3.00. At the other extreme are social assistance benefits in Russia with a score of 1 .lo. 239. The distribution of benefits across households from the lowest to the highest consumption deciles has not changed dramatically from 2003 to 2007, although changes from 2005 to 2007 are more easily observed. After 2003 there were minor changes in targeting performance across social assistance programs observed in the 2005 data, but showing mostly positive results: while the targeting efficiency of social assistance as a whole improved slightly, the efficiency of majority of programs improved as well. Comparing 2003 with 2007, the targeting efficiency of programs such as family benefits, permanent minimum income assistance, nursing benefits, housing benefits, scholarships improved slightly using efficiency measures calibrated to the 1 '' quintile; the targeting efficiency worsened only in the case of temporary minimum income assistance and the remaining benefits. This remains the case when the higher income eligibility threshold for means-tested family allowances is used. Targeting efficiency measured at 2ndquintile improved slightly as well from 2003 to 2007. Table 32: Social Protection Targeting, ( percent of Total Transfers Reported Going to Households in Each Group) Deciles by Household Consumption Total 1 2 3 4 5 6 7 8 9 10 CGH CGH dl ql All social protection 100.0 7.0 8.0 9.0 98 10.2 108 11.5 11.6 11.7 10.4 0.7 0.7 All social insurance 100.0 5.6 7.0 8.4 9.6 10.2 11.0 12,l 12.3 12.6 11.3 0.6 0.6 Old Age Retirement 100.0 4.6 6.0 7.7 9.2 10.2 11.1 12.3 12.8 13.6 12.5 0.5 0.5 Pension All labor market programs 100.0 9.0 11.4 10.8 11.3 12.2 11,3 11.9 8.5 7.8 5.7 0.9 1.0 Unemployment benefits 100.0 13.9 13.9 12.0 9.4 10.6 10.9 9.3 8.5 6.5 4.9 1.4 1.4 & assistance Early retirement benefits 100.0 6.6 10.1 10.2 12.3 13.0 11.6 13.2 8.6 8.5 6.1 0.7 0.8 All social assistance 100.0 16.9 14.7 12.7 11.4 9.3 9.1 74 7.1 6.4 50 1.7 1.6 Family allowance (and 100.0 19.4 16.5 15.0 12.6 10.0 8.3 7.0 5.4 4.1 1.8 1.9 1.8 complementing benefits) Career's benefit & 100.0 13.5 15.2 14.3 14.2 9.3 10.1 8.2 7.1 4.9 3.3 1.3 1.4 nursing benefit Housing Benefit 100.0 19.6 19.1 16.5 13.5 9.8 8.6 5.3 3.8 2.2 1.6 2.0 1.9 Minimum income 100.0 17.0 16.1 11.9 12.6 9.6 11 4 7.0 64 4.2 3.7 1.7 1.7 assistance (permanent social pension) Minimum income 100.0 25.2 17.4 11.5 10.2 6.8 7.4 6.3 6.1 7.2 2.0 2.5 2.1 assistance (temporary benefits, cash & in kind) - Source: Estimates using HBS 2007 Notes; CGH d l : CGH indicator constructed for targeting to decile 1 CGH q 1 : CGH indicator constructed for targeting to quintile 1 240. As part of its preparations to face the social costs of the economic slow-down, the Government planned to target the triennial adjustment of family benefits to recipients with three or more children. This measure was proposed to balance the tightening fiscal constraint on spending one the one hand, and the need to better target limited resources to the poorest families on the other hand. Given the close correlation between the number of dependent children and poverty (see Table 33), this targeted adjustment was expected to redistribute between 0.4 and 1 percentage point of spending on family benefits from non-poor to poor households (depending on the size of the adjustment, still to be 114 determined), and to reduce the share of families with 3 or more children who are poor, by between 1.3 to 3.4 percentage points.62 241. The final shape of the government decree on family benefits stipulates for freezing the income thresholds for both family benefits as well as for minimum income assistance. It also introduces increased level of a core family benefit by over 40 percent (around 30 percent in real terms with CPI inflation of 11 percent between the last amendment in September 2006 and expected rate in November 2009). Such a growth was required by the law: it is supposed to be at least 40 percent of the consumption basket of a child in the given age brackets (calculated by the Social Policy Institute). However, the income thresholds have been kept stable since 2004. This means substantial narrowing of the coverage - during that time e.g. the minimum wage has grown by nearly 60 percent (from PLN 824 in 2004 to 1317 to be valid since January 2010; by 46 percent between 2006 and January 2010). Such a policy of freezing thresholds must have fed into substantial fall in the number of children covered by the family allowances: from 5.55 mln in 2004 to 3.26 mln currently. The trade unions expect a further fall of around 0.7 mln children. Less children will be covered with higher allowances. The effective growth of the average level of payments to children will be smaller than 40 percent nominally because the levels of supplementing benefits remained unchanged (including the supplement for family benefit recipients with three or more children). 242. Primary proposal of the Ministry of Labor included an increase in the thresholds (by 23 percent) and a raise in a supplement for family benefit recipients with three or more children by PLN 20 however the budget situation did not allow for that. 243. Freezing the thresholds for the minimum income assistance means that the newly calculated existential minimum for 4-person family is only PLN 0.50 below the relevant threshold. It may happen soon that families living below the existential minimum do not qualify for the minimum social assistance. And even if the coverage gap would formally remain unchanged or even improved the real situation of the poor would worsen because the legal criteria became unrealistic. Table 33: Distribution of Households by Number of Dependent Children and Consumption Decile Households by Consumption Decile _-l__l I - - - - - - - - - _- -- I- -"_--.11--.-~.- - --~-- --- 1 2 3 4 5 6 7 8 9 10 no 6.9 7.9 8.6 9.5 10.1 10.6 10.9 11.3 11.7 12.8 children 1 child 6.8 6.4 8.1 8.9 9.9 9.3 11.7 11.7 12.6 14.5 2 children 9.1 9.8 10.6 10.5 10.2 10.3 9.9 9.8 10.2 9.5 3 children 14.7 15.7 12.9 12.1 9.6 10.0 8.3 7.2 5.7 3.7 4+ 25.3 17.7 14.4 10.2 9.3 6.8 5.8 4.8 3.6 2.1 - children - ~ - -- Source HBS 2007 Note The distribution across the deciles sums to I00 percent, I e 25 3 percent of households with 4 or more dependent children are In the 1st consumption decile 62 Three targeted benefit adjustments have been simulated for households who in the 2007 HBS and who reported 3 or more dependent children: an increase in monthly benefits of 20 zlotys, 40 zlotys and 60 zlotys. The three adjustments redistribute 0.4 percent, 0.7 percent and 1 percent, respectively, of spending on family benefits from non-poor to poor households. 115 The Importance of Social Assistance in Household Incomes 244. The average social assistance cash benefit accounts for about 15 percent of the average net wage and 44 percent of the gross minimum wage (2007). The average family allowance accounts for 6.5 percent of the gross minimum wage (2 percent of the net average wage). However, in about 36 percent of cases the base family allowance is supplemented, which doubles its amount. By way of comparison, the regular unemployment benefit accounts for 20 percent of the average wage and 50 percent of the minimum wage. 245. As a share of average household income social assistance is minor at 3.3 percent in 2007. This varies, of course, across the household decile groups. In 2007 social assistance accounted for 11.9 percent of the income of households in the lowest decile, declines to 3.7 percent of income for households in the 5'h decile, and then to just 0.8 percent of income for the wealthiest households. However, among only those households entitled to some form of social assistance, benefits are more important making up 13.4 percent of income in 2007. In 2003 social assistance accounted for 12.0 percent of the income of eligible households. Table 34: Social Protection Benefits as a Share of Household Income in 2007 Share of total household income Deciles All 1 2 3 4 5 6 I 8 9 10 All incomes 100 100 100 100 100 100 100 100 100 100 100 Any social protection 30 2 322 363 40 5 39 2 315 349 33 7 29 8 24 9 14 6 Any social insurance 25 8 311 310 32 4 32 5 321 302 30.0 27 0 22.8 13.6 Old Age Retirement Pension I9 2 189 199 22 1 23 3 23 7 221 22 9 21 0 I8 4 II 1 Structural pensions in agriculture 02 03 03 02 04 01 02 0.2 03 0.2 0.1 Disability Pension 35 85 69 64 50 43 37 33 30 19 10 Survivor's Pension 2.7 3.2 3.8 36 3.7 38 35 3.4 2.5 2.1 1.o - Maternity Benefit social insurance 01 01 01 00 01 02 01 0.1 01 01 01 Other Social Security benefits 0.1 0.0 0.0 0.1 0.0 0.0 0.0 0.1 0.2 0.2 0.3 Any labor market programs 11 22 22 18 1 .7 17 14 13 08 06 03 0.4 1.1 0.9 0.7 0.5 05 04 0.3 0.3 0.2 0.1 benefits + unempi, assist. Early retirement benefits 08 11 13 12 13 12 IO 10 06 05 02 Any social assistance 3.3 11.9 84 6.3 5.0 37 3.2 2.4 2.0 1.5 0.8 Family allowance / Family 13 54 37 29 22 16 12 09 06 04 01 allowance + other benefits complementing family allowance Career's benefit t nursing benefit 0.2 07 06 05 0.5 03 03 0.2 0.1 0.1 0.0 Housing Benefit 02 I O 08 06 04 03 02 01 01 00 00 Minimum income assistance - 02 0.8 0.6 04 0.4 03 03 0.2 0.1 0.1 0.0 permanent / social pension Minimum income assistance - 05 21 15 09 07 04 04 03 03 03 00 temporary / Social assistance benefits, cash+in kind Assistance from non-commercial 0.0 0.1 0.0 0.0 0.0 0.0 00 0.0 0.0 0.0 0.0 units Other non-insurance / Scholarships 02 03 03 02 03 02 03 0.2 0.3 0.2 0.2 Other 0.5 10 0.8 07 0.6 07 06 0.5 0.5 0.5 0.3 Source: HBS 2007 Note: deciles defined based on expenditures. 246. For all households, the far more significant benefits come from the contributory branch of the social protection system. Social insurance benefits accounted for 25.8 percent of average household 116 income in 2007. Across the decile groups, social insurance benefits ranged in importance from 3 1.1 percent of the income of the poorest households, to 13.6 percent of the income of the wealthiest. Among households entitled to some form of social assistance, social insurance benefits accounted for 22 percent ' of household'income in 2007, down from 25.4 percent in 2003. Therefore, during the period when social assistance became more important to the income of eligible households, social insurance became less important. IV. Decentralization and Social Assistance Across Poland's Regions 247. The Ministry of Labor and Social Policy (MLSP) is continuing a process started with the Administrative Reform of 1999 that is gradually decentralizing the administration of benefits and services for poor and vulnerable households, and since 2004 devolves a portion of financing for these services. Local governments in Poland perform a set of tasks delegated to them with financing by .~~ the central g ~ v e r n m e n t Prior to 2004, local governments were only responsible for delivering social benefits and services, but since then, have taken responsibility for fullyfinancing a growing set of "own tasks" in this sphere, along with a greater share of tax revenue to cover these obligations. 248. MLSP is pushing ahead with its policy of shifting social assistance activities to the set of "own tasks." It is not yet clear whether the Ministry's movement in this direction will be limited to certain programs, or whether the process is eventually intended to shift administration and financing responsibility for all forms of non-contributory social protection to local governments. Given the experience of other countries in the region and around the world that have decentralized social assistance to varying degrees and with very mixed results, this is a matter of concern for the Polish authorities to consider carefully. 249. With the shift in financing responsibility, local governments have a strong incentive to look for savings and to be more efficient. In other parts of the region, this incentive has sometimes led to compromised standards and uneven delivery of entitlements to the poorest households across municipalities of different incomes. In this section we examine - to the extent that available data will permit - how the shift in financing responsibility has changed the coverage and targeting efficiency of spending on social assistance. An important consideration is whether the process of decentralizing not only administration but financing of targeted social assistance programs to sub-national governments, has impaired their ability to respond to the needs of households eligible for targeted social assistance. From the experience in other countries, the devolution of financial responsibility for social assistance could create strong incentives for sub-national governments to cut back or ration benefits, particularly during economic downturns. This is a concern national authorities should keep in mind while monitoring the performance and standards maintained by local authorities in the current economic crisis. The rationale for decentralizing social assistance 250. The main motivation for decentralizing responsibility for services to local governments is the lack of efficiency, flexibility, innovation and responsiveness on the part of central government. Decentralizing responsibility for service delivery can improve outcomes by encouraging competition 63 The reforms implemented in 1999 shifted the management of public employment services and social assistance services to local governments. However, the financing of these programs remained centralized, with funds distributed along specified criteria between regions. This is changing in the case of social assistance financing that is being gradually shifted to local governments. 117 between local authorities and between an enlarged set of private service providers in a given locality. Decentralizing services can also improve outcomes by opening the possibility of local innovation and differentiation of services to meet the particular needs of specific groups, which often differ widely across regions and municipalities. Furthermore, local authorities can bring greater knowledge to service delivery, which is particularly useful in accurately targeting services to the households that need these most, reaching out to socially disconnected groups, and to closely monitor the impact of services. Finally, by placing responsibility for services on the shoulders of locally elected officials, decentralization can lead to improved services by increasing accountability to households and the pressure these households can bring to bear on local government and non-government service providers when performance is unsatisfactory (World Bank, 2004). 25 I . The policy literature distinguishes between three specific forms decentralization of functions can take. The first, and most complete transfer of responsibilities is called devolution, by which central government transfers the authority for decision making, financing and management of services to politically autonomous units of local government. The second, more limited transfer is called de-concentration by which national governments transfer activities and responsibilities of central ministries and agencies to local offices of these agencies, but retain control over financing and formulation of policy. The third and most limited transfer is called delegation, by which local governments are given purely administrative responsibilities for delivering public services, but are subject to supervision by central government agencies. 252. A seemingly obvious point underlying the theoretical arguments in favor of decentralization - although often forgotten in practice - is that the extent to which the benefits of decentralizing services can be had is directly determined by how quickly the capacity of local authorities can be built to manage newly decentralized responsibilities. However, even where responsibility is decentralized to highly capable local governments, when considering social assistance in particular, two fundamental distinctions have to be made. The first fundamental distinction is between financing and administration, and the second is between targeted cash benefits and social welfare services. These distinctions can help determine what social assistance functions can be improved, and to what extent, by shifting responsibility to the local level. These distinctions should be kept in mind when considering arguments along four principal economic and political dimensions. 253. The first dimension is the negative externality of poverty and the corresponding public- good nature of social assistance. Many public goods (i.e. government interventions to increase positive externalities and correct negative externalities) such as national defense and macroeconomic management can only be effectively provided uniformly. The public-good nature of social assistance - that is, its primary function to alleviate poverty and ensure minimum consumptions and social welfare - also requires a strong degree of uniformity. When considering minimum levels of publicly-financed support, it is difficult to justify disparities across localities or discrimination in the level and quality of provision based on place of residence (Bird, 2000). Some authors even speculate that allowing differences in social assistance provision could lead to a ?race to the bottom? as local authorities seek to discourage settlement by the poor and vulnerable by lowering cash benefits and cutting services in their area (Matkovic, 2006). From the standpoint of balancing overall economic efficiency and equity, the challenge of how to use public spending to achieve the highest possible rates of growth while ensuring the gains improve the welfare of all citizens, it is critical that social assistance are in effect uniformly across regions, particularly given the strong market forces that drive geographic concentration of economic production (World Bank, 2009). 254. The second dimension is the economics of insurance, risk management, and savings from scale. If considered ?poverty insurance?, social assistance is most efficiently financed from the largest possible ?risk pool?. Decentralization can fragment the risk pool, if responsibility for financing benefits is 118 devolved and the tax base on which the social assistance system is built is constrained. This would create smaller, less efficient risk pools in which there are not an actuarially sufficient number of "winners" (Le. net contributors to the pool) supporting beneficiaries (i.e. net recipients from the pool) (Barr, 2001). In a related point, there are substantial savings from scale from centralizing provision of social assistance, which could be lost. Smaller municipalities would bear higher marginal costs to deliver the same minimum package of services than larger municipalities. These higher administrative costs could lead to lower level and quality of benefits and services in smaller, poorer communities. 255. Thirdly is the argument for diversifying the set of service providers and capturing the benefits of market forces. As with education and health care, gains in efficiency and quality can be had from greater competition and innovation. The education and health policy literature advocates a "split" between the purchase and provision of publicly financed services (which are typically - but not necessarily - bundled together in, large centralized systems of public service provision), that allows a larger and more diverse set of service providers to compete and for households to choose between them. Similar arguments can apply to social assistance, particularly to the provision of assistance above a socially defined minimum package. These arguments, however, are more valid with respect to social welfare services, than to cash benefits, as there is greater scope for differentiation across service providers and thus a more obvious basis for competition (Barr, 2001). 256. Finally, there are important political considerations that have to be kept in mind. In contrast to services that benefit the majority of households, such as education and health, social assistance in Central and Eastern Europe is increasingly targeted to the most vulnerable groups. This is efficient from a social welfare standpoint. However, not only are these groups often a minority in any given community in terms of their relative and absolute number, they also typically consists of those households who are the least able to exert political pressure, due either to their poverty, social disconnection, or the prejudices of the majority and local elites, Thus, with respect to benefits and services for the poorest and most vulnerable, without the right mechanisms to strengthen accountability and participation, local governments have less incentives to respond to demands for better funded and higher quality services. Indeed, local authorities may have strong incentives to divert funding away from social assistance to services enjoyed by the majority or more powerful groups in their constituencies. 257. In sum, economic and political theory presents several arguments that distinguish social assistance from other services when balancing the merits and risks of decentralizing responsibilities to local government. While these arguments apply to social assistance as a whole, they indicate that the distinctions between financing and administration on the one hand, and between cash benefits and werfare services on the other, should not be ignored when taking decisions about what to decentralize and to what extent - that is, devolution, de-concentration, or simple delegation. 258. To recap, the public good function of social assistance requires a substantial degree of uniformity to meet its objective of providing a minimum level of welfare. While uniform level and quality is important for a minimum package of both cash benefits and services, above this minimum package the need for uniformity can be relaxed, particularly in the provision of services for which there is greater need for innovation to meet locally-specific challenges. Actuarial principles suggest that targeted social assistance should be financed from the largest and most diverse risk pool available, national general revenues. As above, while critical for financing the minimum package of social assistance benefits and services, this is arguably more applicable to the provision of minimum cash benefits. And although there are important gains to be had from competition between a wider set of providers, these gains are most likely to be reaped from innovation and competition in the provision of welfare services, where there is greater scope for differentiation. Finally, the exclusion suffered by the poor, vulnerable and socially disconnected, is a strong challenge to the argument that shifting service delivery to lower levels of government will increase the accountability of service providers to households. The political economy of 119 providing social assistance is, in this way, very different from the provision of education and health, and requires careful monitoring and regulation by central government to protect the rights of disenfranchised citizens. 259. Taken together, these arguments drawn make a strong case for maintaining centralized financing of at least a uniform minimum package of cash benefits and social services; capturing the benefits of local knowledge by giving local authorities responsibility for administration (in particular, targeting, delivery, and monitoring); responding to local needs by ensuring scope for local innovation; reaping the gains from competition by diversifying service providers to include non-government agencies; and establishing structures to ensure accountability to the poor, vulnerable and socially disconnected groups. Experience with decentralization of social assistance in Europe and Central Asia 260. Indeed, the wide range of country experiences with decentralizing social assistance in Europe and Central Asia, generally reflect and are consistent with the arguments above. Topinska (1 996) notes that the efforts to restructure social assistance in Europe and Central Asia in the 1990s relied to a great degree on decentralization, and particularly on increasing local participation in the financing and administration of benefits, as well as enlisting greater support for vulnerable groups from families and non-governmental organizations. The experience has been mixed. This is mainly because decentralization in the region in most cases was motivated less by the primary objective of improving service delivery, and more by secondary objectives. These secondary objectives range from coping in times of economic and financial crises; devolution of fiscal responsibility and contingencies; and reactionary rejection of any form of central planning after the fall of socialism (Matkovic, 2006). Therefore, when decentralizing mandates and functions to local authorities, national governments in the region have often paid little attention to the arguments made in the policy literature, especially in the case of social assistance. 261. There are several examples of responsibility for social assistance being fully devolved with little care or thought given to adequate financing arrangements. In Romania decentralization of social assistance and child allowances lead to a deterioration of outcomes. After passing a law on local public finance in 1998 which devolved responsibility for social assistance to local governments, total spending on social assistance and child protection - including benefits and services for the disabled and elderly - fell by nearly 40 percentage points in real terms between 1996 (when fiscal decentralization began) and 1999. This resulted in a 16 percent decline in cash benefits from their level in 1996. Because responsibility for providing social assistance was devolved without adequate concern for fiscal decentralization, a lack of financial resources at the local level forced municipalities to scale back provision of benefits and services, and in some cases, to stop providing services altogether. To repair the damage to poor households left in the wake of Romania's decentralization in the late 1990s, the country recentralized social assistance financing in 2002. 262. In Bosnia and Herzegovina, decentralization of social assistance was plagued by similar problems, and lead to similar negative outcomes. Decentralized financing resulted in substantial inter- regional disparities in coverage, with poorer localities providing the least amount of services. In the face of resource constraints, eligibility criteria for most benefits in Bosnia became ad hoc, as local welfare offices relied on their discretion to ration scarce municipal resources. 263. Regional experience also shows that the performance of social assistance in decentralized environments (e.g., in Hungary and Russia) depends on the whether a country is largely homogenous in social and economic terms with small differences in regional incomes, or one where there are substantial social and income differences. Milanovic (2000) notes that in Russia, soon after the breakdown of the Soviet Union, the central government's financial role was weakened contributing to 120 its inability to allocate adequate financial resources for social assistance. The central government had little influence over the distribution of social assistance across the autonomous republics, oblasts, or cities. As a result, social assistance became entirely a local matter: locally financed, and with benefit eligibility and benefit levels wholly under local control. This led to a wide dispersion of the coverage and levels of social assistance in the country. 264. However, regional experience also illustrates the benefits expected of decentralization, listed at the start of this section. In Albania decentralization was found to have had a positive impact on the delivery of social assistance. Alderman (1998) finds that local officials in Albania had more information on households' welfare than was available to central government authorities, and that the allocation of social assistance among households by local authorities was much better targeted than the allocation of social assistance funds by the central government. After decentralization, Alderman found considerable improvement in targeting with more than half of social assistance was going to the poorest quintile. 265, In Hungary, decentralization of service provision proved more successful than decentralization of cash benefits, and has led to the gains from innovation and competitions between providers, predicted in the policy literature. The Hungarian social assistance system was decentralized following the 1990 Act on Local Self-Government. The Act gave almost full control over social assistance to newly formed local councils, although central government remained in charge of financing. Whereas Romania provides an example of devolution, Hungary's case is an example of de- concentration. The transfer of responsibility for welfare service provision brought greater flexibility, which spawned innovation and the introduction of new services, as well as new service providers, who have proved more responsive to households. By the end of the 1990s, with decentralization of social welfare services well entrenched, there were over 5000 non-government organizations providing services to poor and vulnerable groups, under the oversight and regulation of central government authorities (Andrews and Ringold, 1999). 266. In a recent review of regional experience with the decentralization of social welfare assistance, Matkovic (2006) concludes that there are few theoretical arguments in favor of decentralizing financing, and many concrete examples to illustrate the dangers of fragmenting the larger risk pool provided by the national public finance system, and forgoing the economies of scale from centralized provision. Further, a rich, practical experience with assigning social welfare mandates and functions to local governments in Europe and Central Asia, indicates that the greatest advantages from decentralization are to be had by giving municipal authorities greater administrative responsibilities over cash benefits (in particular, delegating responsibility for targeting and other administrative functions), and much greater autonomy in the provision of welfare services (de-concentration). 267. However, reflecting many of the arguments set out at the start of this section, Matkovic warns that several important pre-conditions must be in place for decentralization - even along these clearly limited lines - to succeed in improving social assistance. These are (i) that the rights and eligibility requirements for social assistance be clearly defined and knowledge of these rights be improved among households; (ii) that structures to encourage innovation and differentiation of services be in place; (iii) that structural reforms, such as the shift away from institutional care, be planned, legislated and well underway prior to the transfer of functions; (iv) that quality standards be set, and mechanisms developed for these standards to be monitored by local and central government authorities; (iv) that government agencies, at the local and national level, be strengthened and ready to take up their new responsibilities. Matkovic concludes that decentralization of mandates and responsibilities without these conditions in place can lead to the development of a myriad of models and levels of provision, the lack of standards, abuse and misuse of resources, and ultimately to a deterioration of social welfare assistance for the poor, vulnerable and socially disconnected. 121 268. Does this imply more spending on the social sectors, and in this particular case, more spending on social welfare assistance? Ditch and Barnes (1996) noted that in countries where there is a mix of state and municipal funding, there was considerable scope for conflict between the different levels of government, although there is perhaps less evidence of this than might be expected. In Sweden, for instance, where social assistance was funded partly by block grants from central government and partly by local taxation, there was considerable political opposition to nationally imposed levels of benefits. Individual municipalities could award benefits at a rate below the suggested minimum. Whilst claimants could have their right to enforce their claim to the minimum level, it emerged that only the articulate and well informed could make use of the benefit provision. There were similar outcomes in Finland, where the abolition, in 1991, of the normative directives obliging municipalities to pay at the national rate left no redress for the claimants. Ditch and Barnes (1996) note that there was widespread recognition that the wide geographical variations in benefit payments were less related to the need for assistance than the ability of municipalities to provide it, and thus there was considerable pressure on the state to increase its share of financing. Preliminary outcomes since decentralization of social assistance in Poland 269. The results presented in the final part of this section are a preliminary attempt at measuring the efficiency impact of the change in how social assistance transfer programs are administered and financed that began in 1999, but that was put into full effect in 2004. It is important to stress the purely indicative nature of this analysis, as the household survey data on which it is based will only allow an assessment of performance at the district (voyvod) level, rather than at the municipal (gmina) level at which social assistance administration actually takes place. The results are intended to motivate a closer examination and consideration of the policy choice to proceed with decentralization, so that further steps in this direction are taken with consideration for the important distinctions raised in the literature and the experience from Poland's neighbors cited above. 270. In Poland the largest concentration of eligible households can be found in mazowieckie - the capital region where average incomes are highest, - even though the incidence of eligible households in the population of the region is lowest. As incomes in this region rose from 2003 to 2007, its share of Poland's eligible households remained constant. Conversely, the voyvods with the highest incidence of eligible people, also have smaller (although not the smallest) concentration of Poland's eligible households. At first glance, this is may seem counterintuitive. However, since poor people have the strongest income motivation to move away from poor places, it shouldn't be surprising to find large concentrations of households eligible for social assistance living in the most economically prosperous places in Poland. 64 27 1, Table 35 presents estimates of the population in each voyvod that is entitled to some form of social assistance but not receiving any benefits (the coverage gap, measured as a percentage of the entitled population in each voyvod). Nationally, the coverage gap has been more or less stable: at first it has fallen from 30.1 in 2003 to 28.1 per cent in 2005 but then it has grown to 31 percent in 2007, indicating a slight national worsening in covering the poor and vulnerable population. In 2003 the coverage gap was largest in Slqskie and Swigtokrzyskie where in each voyvod around 40 per cent of the entitled population was not receiving benefits. 64 The World Bank's WDR 2009 discusses the motivation and importance to area development of internal migration, particularly of poor households. 122 272. The fairly stable coverage gap at national Table 35: The Coverage Gap Across Poland's Voyvods (Percentage of level between 2003 and the population eligible for social assistance but not receiving benefits) 2007 underwent very Entitled without benefits unevenly distributed 2003 2005 2007 Difference 2003-2007 changes across voyvods. Total Population 30.1 28.2 31.0 0.9 The greatest progress was 02 dolnoSlqskie 28.6 25.9 31.0 2.4 made local authorities in mazowieckie, where the 04 kujawsko-pomorskie 21.2 18.5 22.2 1 .o coverage gap was reduced 06 lubelskie 18.3 24.9 29.0 10.7 by 9 percentage points. 08 lubuskie 25.0 16.5 25.3 0.3 Although still progress, 10 t6dzkie 26.9 23.4 29.3 2.4 local authorities in 12 malopolskie 26.5 23.7 29.7 3.2 podlaskie only managed to reduce their coverage gap 14 mazowieckie 37.5 35.5 28.5 -9.0 by 1.3 percentage points. 16 opolskie 21.4 20.2 17.2 -4.2 Altogether the gap was 18 podkarpackie 18.8 16.3 16.1 -2.7 closing in 6 voyvods. In 20 podlaskie 32.2 37.0 30.9 -1.3 contrast to this, in 10 of 22 pomorskie 37.5 25.0 30.6 -6.9 Poland's 16 voyvods, the 24 Slqskie 40.6 43.4 47.2 6.6 share of the entitled population without benefits 26 Swietokrzyskie 39.0 36.0 40.6 1.6 actually increased, the 28 warminsko-mazurskie 31.5 27.8 33.8 2.3 strongest in lubelskie (by 30 wielkopolskie 28.2 29.6 35.4 7.2 10.7 P*P*) and 32 zachodniopomorskie 35.4 28.2 28.0 -7.4 wielkopolskie (by 7.2 p.p.). Source: Authors' estimates using HBS data 273. Table 36 presents the same measure of targeting efficiency that has been used nationally in previous sections to show differences in targeting performance across Poland's districts and how this has changed over time since local governments have had greater administrative responsibility for benefit delivery. As in prior discussions of this measure, it is worth pointing out again that this indicator does not capture targeting efficiency perfectly. This is because the income threshold for means testing family allowances and minimum income assistance does not apply as unambiguously and evenly as the categorization of the population according to consumption quintiles from which the indicator is constructed, If a benefit was distributed evenly across the population, the indicator would be equal to 1. A value higher than 1 shows a progressive outcome from means testing. If all means tested benefits were being received by the poorest quintile, the measure would equal 5 . A measure less than one would indicate a regressive distribution. 274. The first three columns of Table 36 attempt to capture the targeting efficiency of family allowances and complementary benefits across the voyvods and over time. The last three try to do the same for minimum income assistance. In nine of Poland's 16 voyvods municipal authorities have achieved an improvement in the share of family allowances being received by the poorest 20 percent of households. The voyvod where family allowances are most progressively distributed (and have been since 2003) is podlaskie, however warminsko-mazurskie maintained most progressive distribution throughout entire analyzed period. So similar to the measure of closing the coverage gap discussed above, targeting efficiency for family benefits increased in only seven of Poland's 16 voyvods from 2003 to 2007. The number of voyvodships where both measures improved was only two (podkarpackie and podlaskie). 123 275. As a measure of targeting efficiency of minimum income assistance, the indicator should provide a more Table 36: Targeting Efficiency of Family Allowances and Minimum accurate reflection Income Assistance Across Voyvods, (Coady-Grosh-Hoddinott Ratio*) of performance. This Family Allowances Minimum Income is because the income Assistance threshold for 2003 2005 2007 2003 2005 2007 minimum income Total Population 1.7 1.7 1.8 1.7 1.7 2.1 assistance better 02 dolnoSlqskie 2.0 2.0 1.7 2.2 1.9 2.1 corresponds to the threshold for 04 kuj awsko-pomorskie 1.8 2.0 2.1 2.6 2.3 2.1 classifying the poorest 06 lubelskie 1.7 1.8 2.2 1.5 1.5 2.2 20 percent of the 08 lubuskie 1.4 1.3 1.4 0.9 2.5 1.3 population - but still 10 t6dzkie 1.7 1.4 1.3 1.7 2.0 2.7 not exactly. In the last 12 matopolskie 1.5 1.7 1.6 1.9 1.8 2.1 three columns of Table 36 what is 14 mazowieckie 1.8 1.6 1.8 1.4 1.9 1.8 immediately apparent 16 opolskie 1.3 1.2 1.3 1.1 2.0 2.1 is that minimum 18 podkarpackie 1.7 1.8 1.9 2.2 2.0 1.7 income assistance is a 20 podlaskie 1.7 1.5 2.4 1.4 2.4 3.6 more progressively 22 pomorskie 2.2 2.0 2.1 1.3 2.8 1.7 distributed benefit 24 Slqskie 1.7 1.6 1.5 1.9 2.4 2.5 than family allowances. 26 Swietokrzyskie 1.5 1.8 2.0 2.7 1.7 2.6 Furthermore, among 28 warmihsko-mazurskie 2.2 2.5 2.2 1.5 2.8 1.8 the 16 voyvods, local 30 wielkopolskie 1.6 1,6 1.8 1.8 2.0 2.4 authorities in 1 1 have 32 zachodniopomorskie 1.9 1.9 1.5 1.6 2.0 1.6 achieved an improvement in the Source: Estimates using HBS 2007 share of *CGH in this table calculated as received by households in share of benefits consumption Quintile 1 , divided the 20. Equaltotal benefits of benefits across households would by distribution received by the yield a CGH = 1. 100 percent of benefits received by households in the lowest quintile would poorest households. yield a CGH = 5. However, in five voyvods the efficiency with which minimum income assistance is targeted appears to have stagnated and even to have deteriorated. 276. There is nolegal explanation for such strong differences across voyvods in the access to family benefits nor to minimum income assistance, since there are no budget limitations to those benefits. Both types of benefits are guaranteed by the law to all households that meet eligibility criteria.. The disparities in performance across voyvods that can be shown using data from the HBS could be reflecting serious differences in administration capacity between gmina social welfare offices. These results indicate where the MLSP could target additional monitoring efforts to determine the extent of performance problems at the gmina level, and where reparatory measures may be required to ensure social assistance transfers are reaching the households that really need them and are not being paid to those who do not. 124 V. Conclusions 277. Relative to other branches of the social protection system, Poland's spending on social assistance transfers is low, even by regional standards. Of the 15.7 percent of GDP Poland spent on social protection in 2007, only 1.4 percentage points was spent on family allowances and minimum income assistance, compared to 9 percentage points spent on old age pensions. In a region where long- term unemployment and dependence on social assistance has long been a problem, and where means testing social assistance benefits in order to improve efficiency and incentives is still a relatively new and rare approach, low levels of spending are not on their own a cause of concern. 278. However, even among the countries of Central and Eastern Europe that target family and child benefits to poorer households, spending in Poland on these benefit categories is still relatively low. Other than Poland, the countries in the region that means test family and child benefits are the Czech Republic and Slovenia, where governments spent 2.5 and 2.3 percent of GDP on family and child benefits and minimum income assistance, respectively, in 2007, compared to Poland's 1.4 percent of GDP. 279. Furthermore, about 31 percent of the population who fill the income criteria for some form of targeted social assistance may not be receiving benefits. And although in aggregate the share that Poland spends on targeted social protection programs (as a share of all social protection spending) has risen since 2004, this has not yet yielded an improvement in targeting performance sufficient to be captured in Poland's Household Budget Survey - on the contrary, according to HBS we can see slight worsening of this measure. 280. Furthermore, although at the national level the coverage gap has been stable and some improvements are observed in targeting performance, when both measures are regionally dissagregated tocapture differences across voyvods, the impact of the steps taken to decentralize social assistance has been mixed. The evidence presented in this chapter does not point to a clear improvement in performance from giving local governments a greater role in administering and financing social assistance. In several cases the targeting of family allowances and minimum income assistance has actually worsened. 281. Several explanations for this mixed performance can be drawn from the literature and regional experience reviewed in Section V. First, decentralization of social assistance to-date has been primarily a deconcentration and devolution of functions rather than a devolution of financing and decision-making about minimum benefit levels or eligibility criteria. Thus, the results may be on-balance supportive of decentralization contained to a deconcentration and delegation of administrative and benefit delivery functions to local authorities as the literature concludes. Second, Poland's fiscal equalization structures work and the inequities of the local revenues bases are fairly well compensated through federal equalization transfers, leaving gminas across the country in a good position to provide locally financed social assistance. Third, the aggregated results could be hiding local failures, so that we observe decentralization working well in some voyvods (presumably those with high local revenue bases relative to the number of people eligible for social assistance, and good accountability structures) while the real differences in performance are taking place at the gmina level. Finally, during the same period of analysis, there have been two major structural reforms - decentralization and the consolidation of family allowances and child benefits - so the before-and-after comparisons mix the impact of the two. 282. Clearly a more thorough examination of benefit administration is required using data that can capture improvement at the local level where responsibility actually lies. Only then can research help policy makers arrive at a more definitive conclusion about the impact decentralization has had on the coverage and targeting efficiency of social assistance. Out of concern that local authorities face stronger 125 incentives to shirk on their responsibilities to vulnerable groups as the crisis unfolds and as budget constraints tighten, the Ministry could use this period during which much closer monitoring and scrutiny of gmina social welfare offices is in any case required, to consolidate the data needed to conduct a more thorough review of decentralization than has been possible in this report. 283. This said, the disparities in performance across voyvods presented here could reflect serious differences in administration capacity between gmina social welfare offices. Large gaps in coverage typically reflect a lack of outreach to find households that are likely to be eligible but uninformed about their rights. Substantial leakage of benefits to non-poor groups could indicate a need for better, or more frequent cross checking between administrative databases of different social protection programs, and with income and asset administrative data held by other government agencies. These results presented in this chapter indicate where the MLSP could target additional monitoring efforts to determine the extent of performance problems at the gmina level, and where reparatory measures may be required to ensure social assistance transfers are reaching the households that really need them and are not being paid to those who do not. 126 cr 0 I 0 M rl LABOR MARKET PROGRAMS65 Introduction 284. The recent economic crisis has a substantial impact on the Polish labor market. Unemployment is sharply increasing, which creates a strong pressure on the growth in labor market expenditures. Workers who lose their jobs claim unemployment benefit and expect government's assistance in finding a new job. Moreover there is also a pressure to protect the existing jobs, for example by means of government wage subsidies. The growth in labor market expenditures resulting from the increase in unemployment obviously creates an additional fiscal strain. At the same time, however, it acts as an automatic stabilizer, damping the fall in income and demand. 285. This chapter focuses on labor market expenditures and their impact on employment outcomes in Poland. In particular the chapter examines the implications of the current global economic crisis for labor market spending in Poland and the capacity of the existing labor market programs to mitigate the impact of the crisis on poverty. The analysis of labor market institutions and long-term structural labor market issues is beyond the scope of the chapter. Thus it does not address issues such as employment protection legislation, the minimum wage, labor taxation, labor supply incentives, or the skills mismatch. Some of these issues were addressed in the recent World Bank regional labor market studies (World Bank 2005, World Bank 2007) and in the earlier Poland labor market study (World Bank 2001). A discussion of labor market policies and programs in transition economies is contained in Rutkowski (2004) and World Bank (2005). 286. The chapter addresses three broad questions. First, it looks at the level of labor market expenditures in Poland. Second, it examines their efficiency. Finally, it discusses ways of improving the efficiency of labor market programs. 287. It finds that labor market expenditures are modest in Poland by OECD/EU standards, although they are within the same range as in other EU new member states. Active labor market programs vary in terms of their cost-effectiveness, and the impact of some of them is limited. Thus, rather than increasing spending on labor market programs, Poland should strive to improve their efficiency. This can be achieved by revising the program mix so as to maximize the overall employment impact and by better targeting programs to those groups for whom they have the largest impact. Particularly there is scope to improve the quality of job search assistance and counseling. Given the decentralized setting for the provision of employment programs, improving their efficiency may require strengthening the fiduciary role played by the Ministry or Labor. 288. The chapter is divided into five sections. Section I sets the stage by presenting recent labor market developments in Poland. Section I1 analyses the evolution of labor market expenditures in Poland after the EU accession. Section I11 compares labor market expenditures in Poland to those in other OECD V and EU countries. Section I examines the effectiveness and efficiency of labor market programs in Poland. Section V concludes and suggests ways to improve the efficiency of labor market spending in Poland. 65 Prepared by Jan Rutkowski. This chapter benefits from information provided by the staff of the Ministry of Labor and Social Policy, in particular Malgorzata Sarzalska, Iga Magda, Justyna Garbarczyk and Ireneusz Piqtakiewicz. Useful information on the fimctioning of local (poviat) Labor Offices was provided by Jerzy Bartnicki, Director of Kwidzyn Labor Office. I. Recent Labor Market Developments 289. Labor market conditions, particularly the unemployment rate, are an important determinant of expenditures on labor market programs. Thus this section presents the evolution of unemployment in Poland in recent years with the aim of providing background information for the forthcoming analysis. 290. As a result of the global economic crisis labor market conditions have started to deteriorate in Poland since November 2008. The unemployment rate increased by 2.4 percentage points over just four months and reached 1 1.2 percent in March 2009 (Figure 40).66The consecutive fall to 10.8 percent in May reflects primarily the seasonal factor. Despite this sharp increase, the current unemployment rate is only slightly higher (less than 1 percentage point) than it was 12 month earlier. It is thus the trend that is worrisome, rather than the present unemployment level. 291. So far, the increase in the unemployment rate has not been coupled with a fall in the labor force participation Figure 40: A Steep Increase in Unemployment Rate After a Prolonged rate. To the decline contrary, the participation rate in Unemployment rate(%) I the first quarter of 17 III_I______I_______l_________-___---_ ~ __I--__ 2009 was 1 percentage point 15 higher than a year 13 I--__ earlier. This is a 1 good news, because 11 - as a rule the fall in 9 b labor demand I manifests itself in 7 1 both the fall in the I - I rate. The increase in the labor force Source: Ministry of Labor and Social Policy; Bank staff calculations. participation rate Note: Registered unemployment rate may, at least partly, be attributed to recent labor market reforms (especially to the reduction in the tax wedge on labor) which were meant to strengthen labor supply incentives and reduce the cost of labor. As a result of the increase in the labor force participation rate, the employment/working age population ratio - which is a summary measure of the utilization of labor resources -- is presently slightly (close to 1 percentage point) higher 66 Throughout the text we are using the registered unemployment rate rather than the ILO unemployment rate, which is calculated on the basis of the Labor Force Survey. Although the registered unemployment rate provides a less accurate reflection of labor market conditions than the ILO unemployment rate, it is available on a monthly basis (the ILO unemployment rate is available only on the quarterly basis), and in addition it is more closely related to labor market expenditures, which are determined by the number of registered unemployed. Importantly, while the registered and the ILO unemployment rates diverge in levels (the former being higher than the latter) they tend to move in parallel. In fact, both are strongly correlated (r2=0.988)with the underlying relationship of the form: Q(lfs) = 1.3u(reg) - 5.5; where Q(lfs)is the estimated LFS unemployment rate, and u(reg) is the registered unemployment rate. The equation was estimated based on 25 quarterly observations from 2003 to 2009. 132 than a year earlier. Thus the labor market effects of the crisis have not yet fully materialized. By all likelihood, these effects will become visible in the second half of the 2009. 292. The recent surge in unemployment was clearly Figure 41: The Increase in Unemployment Reflected an Upsurge in caused by an adverse Inflow, Especially Due to Lay-Offs demand shock. Inflow into Inflow into Unempioyment unemployment skyrocketed, year over year, %change especially inflow due to lay- 300 250 offs (Figure 41). Inflow into 200 unemployment peaked in 150 March, when it was 56 percent higher than a year 100 50 I _ /- earlier. Currently (May) 0 . . inflow into unemployment is "^ -50 30 percent higher than a year earlier, but inflow due to lay- offs is about twice as high. These two indicators signal the worsening of labor market conditions earlier than the Note: Inflow into unemployment is the number of newly registered unemployed in a commonly used given month. Inflow due to lay-offs is the number of newly registered unemployed whose employment was terminated by the employer. unemployment rate. Source; Ministry of Labor and Social Policy; Bank staff calculations. 293. The fact that the increase in unemployment was driven by the rise in lay-offs implies a growth in unemployment benefit claims and associated expenditures. Figure 42 illustrates this trend. In May 2009 the number of unemployment benefit's recipients over 60 percent higher than a year earlier. This is in contrast to the increase in unemployment that occurred Figure 42: Lay-Offs Led to a Surge in Unemployment Benefit in Poland in the early 2000s' Claims which was largely driven by the Recipients of Unemployment Benefit labor market entry of the baby year over year, %change boom generation (see Figure 70 43). While new labor market 60 ._. entrants are not eligible to 50 . .. .. 40 unemployment benefit, workers . , . " 30 . . . .. ,. . .....I. . . ." . .,.. .___ " . l-..l"l ..I , that are laid-off are eligible. I _ 20 Accordingly the current increase 10 0 . . .. . - .. .. in unemployment is likely to 0 . .xII . -10 lead to a stronger increase in I" -20 i expenditures on unemployment -30 ' benefits than the previous one. Source: Ministry of Labor and Social Policy; Bank staff calculations. 133 294. The recently Figure 43: Job Losers, Who are Eligible to Unemployment accelerated inflow into Benefit, Replaced New Labor Market Entrants, Who are Not unemployment has led to an Eligible, as a Dominant Group Among the Unemployed increase in the share of short- Unemployment by Reason 2007 and2008 term unemployed. Presently the majority (55 percent) of the unemployed are jobless for less than 6 months (Figure 44). Many of them are thus eligible to unemployment benefit.67 This New entry again is in contrast to the high 18 Labor market re-entry unemployment period of the Quit early 2000s, when the majority of the unemployed were long- m Job loss term unemployed and thus not eligible to unemployment benefit. 2007 Q4 2008 Q4 295. Thus the change in the profile of unemployment Source: Labor Force Survey, Central Statistical Office, Bank staff calculations. associated with the current economic crisis entails an increase in spending on unemployment benefit over and above the increase resulting from the growth in the number of the unemployed. Now most of the unemployed are job losers, unemployed for a relatively short term, and thus eligible to unemployment benefit. This implies, that the unemployment benefit coverage ratio - the percentage of the unemployed who receive unemployment benefit - is going to increase and be significantly higher than it was during the earlier period of high unemployment in the early 2000s, when the profile of unemployment was different. 296. What are the workers' groups that face the highest risk of u unemployment? Answering this question is key to the identification of workers vulnerable to poverty and thus target groups for active labor market programs. Table 37 indicates that workers facing the highest risk of unemployment are young, less-educated and residing mainly in urban areas. Women face a somewhat higher risk of unemployment than men. The relationship between the risk of unemployment and educational attainment is particularly strong. High professional and technical skills provide good protection against unemployment, whereas low skills or lack of vocational skills elevates the risk. For example, the unemployment rate among workers with primary education exceeds 11 percent and that among workers with secondary general education approaches 10 percent. This is much higher than the unemployment rates among workers with tertiary or secondary technical education, which are 3.6 and 6.5 percent, respectively. These data suggest that active labor market interventions should be primarily targeted at less educated workers lacking specific vocational or technical skills, as it is this particular group that faces the poorest employment prospects. 61 Eligibility criteria for unemployment benefit are presented in Section IV below. 134 297. 'Ikre is a strong link Figure 44: Most of the Unemployed are Jobless for Less than 6 between unemployment and Months poverty in Poland. Unemployment by Duration Unemployment elevates the 2008 Q4 risk of poverty, and the poor are much more frequently unem loyed than the non- poor! For illustration, the unemployed are twice as likely to be poor as the employed (the poverty rate being 20.6 and 41.6 percent, respectively).6' And those labor force participants who arc poor are 2.5 times more likely to be unemployed than those who are non-poor, This Vote The 2008 data on the durational structure of unemployment are non-comparable strong relationship between with those for earlier period due to the change in definition ofjob search duration, unemployment and low which was introduced in 2008 income implies that policy Source. Labor Force Survey, Central Statistical Office, Bank staff calculations measures aimed at supporting the unemployed (either through income support or through enhancing their employment chances) have a significant positive effect on poverty. 11. Evolution of Labor Market Expenditures 298. This section shows that labor market expenditures in Poland are countercyclical and thus act as automatic stabilizer. During the downturn the share of passive labor market program increases and that of active programs decreases. We argue that this shift towards passive income support programs in response to the economic crisis is justified given that active programs have limited net impact, especially when job opportunities are scarce. 299. The expenditures on labor market programs decreased in Poland during the period of 2005- 2008 along with the decrease in the unemployment rate. In 2005, with the registered unemployment rate at 18.2 percent, labor market expenditure accounted for 0.56 percent of GDP. In 2008, with the unemployment rate at 9.8 percent, labor market expenditure accounted for 0.45 percent of GDP (Figure 45). Accordingly, 1 percentage point reduction in the unemployment rate was associated with a 0.01 percentage point decrease in the share of labor market expenditures in GDP. The elasticity of labor market expenditures to unemployment was thus rather low. 08 Causality runs priinarilq from unemployment to poverty However, persons raised in poor families may face worse labor market prospects, for instance due to lower investments in human capital. 69 Poverty is defined here in relative terms as living in a household with per capita consumption in the bottom quintile (20 percent) of the consumption distribution. Population in the above examples i s limited to the labor force. 135 Table 37: Unemployment Incidence and Profile by Socio-demographic Characteristics, 2008 (4th quarter) Unemployment incidence Unemployment share percent All workers 6.7 100.0 Gender Men 6.0 48.4 Women 7.6 51.6 Age 15-24 17.1 25.9 25 - 34 6.4 31.7 35 - 44 5.0 17.0 45 + 5.2 25.4 Location Urban areas 6.9 63.5 Rural areas 6.4 36.5 Education Tertiary 3.6 13.7 Secondary technical 6.5 25.2 Secondary general 9.6 12.3 Basic vocational 7.0 33.9 Primary or less 11.4 15.0 Source; Labor Force Survey, Central Statistical Office; Bank staff calculations. Figure 45: Labor Market Expenditures are Countercyclical and Act as an Automatic Stabilizer Revenues and Expenditures of Employment Fund 2005-2008 __ - 08 i 70 - 07 06 Revenues, %of GDP 05 I Expenditures,% of 04 GDP I-. --- + Surplus, % o f expenditures (rhs) O3 02 01 r - --- oo 2005 2006 2007 2008 Source: Ministry of Labor and Social Policy; Bank staff calculations. 300. In contrast, the revenues of the Labor Fund, out of which labor market programs are financed, hardly respond to changes in the unemployment rate. They remained roughly stable throughout 2005-2008 and accounted for around 0.7 percent of GDP (Figure 45). As a result of the stable revenues and declining expenditures the Labor Fund amassed a substantial surplus, which in 2008 represented close to 60 percent of labor market expenditures. This surplus created fiscal space to finance 136 t h e growing demand for labor market programs during the current period of economic downturn (see Box 4 for the projections of t h e L a b o r Fund?s deficit under different unemployment scenarios). Box 4: L a b o r F u n d deficit U n d e r Different Unemployment Scenarios The current economic crisis and rising unemployment is puts a strain on the Labor Fund, which finances both passive and active labor market programs in Poland. Will the Fund?s resources be sufficient to finance the increasing demands for unemployment benefit payments and active labor market programs? Simulation results suggest that the Fund will be able to cover the increased program expenditures under different scenarios regarding the dynamics of unemployment until 20 11. This box presents simulation results under two scenarios. Scenario I is based on the unemployment rate that was endogenously determined within a structural macroeconomic model used by the Institute for Structural Research. Scenario 2 is based on the unemployment rate consistent with the government?s projections (14 percent registered unemployment rate in 2009 which approximately corresponds to 11.2 percent LFS unemployment rate). The evolution of the unemployment rate under both the scenarios is presented in Table A . Table A. The dynamics of the LFS unemployment rate under different scenarios. 2008 2009 2010 201 1 Scenario I 6.7 8.5 8.8 8.9 Scenario 2 6.7 11.2 11.7 10.6 Source: Institute for Structural Research. Under Scenario 1 the Fund runs a surplus through 201 1 (although a declining one), despite the projected fall in revenues (due to lower employment) and an increase in expenditures (see Figure A). Under Scenario 2 the Fund starts to run a deficit in 2010 which persists in 201 I . However, the Fund still will be able to cover the increased expenditures because of the substantial reserve build during the 2005-2008 period, when the Fund was running a substantial surplus. Detailed data on projected Labor Fund revenues and expenditures is provided in Annex Table A3. Figure A. Labor Fund deficit under different unemployment scenarios Source: Institute f o r Structural Research; Bank staffcalculations. It should be noted that the simulations were carried out under the assumption that the increase in unemployment will cause an increase in expenditures on both passive and active programs. This means that the increase in expenditures might be overestimated because international and historical evidence shows that expenditures on active programs are treated as residual ones (Le. revenues - unemployment benefit payments), and accordingly their share in total labor market expenditures tends to fall during the period of high unemployment. Source: Bank staflanalysis. 137 301. As to the response of labor market expenditure to the growing unemployment it seems likely that it will be stronger than the earlier one associated with the declining unemployment. In other words, the elasticity of labor market,expenditures with respect to the unemployment rate is likely to be higher when unemployment is rising than when it is falling. The reason is that while the expenditure per unemployed increased quite substantially during the recent period of upturn, it probably will not decrease proportionately during the current period of downturn. 302. Figure 46 shows that the decrease in the unemployment rate during 2005-2008 was associated with a strong increase in expenditures per unemployed. There are two reasons for this increase. First, the availability of funding. The growing surplus of the Labor Fund created room to spend more on active labor market programs, such as training, and to increase the coverage of the programs. Second, increase in the unit cost of the most of active labor programs (this issue is discussed in more detail in the ensuing section). For example, the unit cost of training increased by 18 percent in 2007 over the previous year, and this increase was not accompanied by a higher job placement rate after program completion. Apparently, service providers were adjusting the price of their services to the growing demand, which led to cost inflation. In short, the growing surplus of the Labor Fund was encouraging higher spending on active labor market programs. Figure 46: Expenditure Per Unemployed Increased with the Fall in the Unemployment Rate Expenditures on Unemployment and Unemployment Rate 2005 - 2008 50 20 -Standardized expenditures on unemployment Unemployment rate (rhs) 2005 2006 2007 2008 Note: Standardized expenditure = expenditure per unemployed relative to GDP per economically active person. The indicator is unitless. It shows how much a country spends per one unemployed person relative to labor productivity. For example, the value of 4.5 means that in 2008 Poland spend per one unemployed 4.5 times as much as the value of output produced by one economically active person. Source: Ministry of Labor and Social Policy; Bank staff calculations. 303. The increase in spending per unemployed was thus associated with an increase in the share of spending on active labor market programs in total labor market expenditure. More generally, there is an inverse relationship between the unemployment rate and spending on active programs as a share of total labor market expenditure. When unemployment is low the share of spending on active programs is high. For illustration, spending on active programs accounted for only 34 percent of total spending in 2005 when the unemployment rate was over 18 percent, and for as much as 55 percent in 2008 when the unemployment rate was less than 10 percent (Figure 47). This suggests that spending on active programs is a residual (Le. revenues minus spending on unemployment benefit), given that 138 spending on passive programs (unemployment benefit) is mandatory and its amount is determined by legal provisions and the number of the eligible unemployed. If so, then the current growth in unemployment is most likely to crowd out expenditure on active programs and lead to an increase in the share of expenditures on unemployment benefit. 304. Is the projected shift of spending away from active programs toward passive programs a reason for concern? After all, the common wisdom is that spending on active programs is a more effective way of fighting unemployment than spending on passive programs (a conviction reflected in the terminology). Although regarded as a matter of principle, this is in fact an empirical question. Active programs are a preferred means of combating unemployment if additional spending on those programs is justified by an adequately significant increase in (re)employment probability of program participants compared with similar non-participants receiving unemployment benefit. In other words, given their cost, active programs must have a substantial net impact in order to be an efficient tool for combating unemployment. We examine the net impact of active labor market programs in Poland below. Most of the programs have a limited net impact on employment, and some - for example public works - have a negative impact. 305. During the period of crisis the reliance on unemployment Figure 47: The Increase in Labor Market Expenditure was benefit is thus a rational policy to Driven by the Growing Spending on Active Labor Market help the unemployed by providing Programs, Whose Share in Total Expenditure Markedly temporary income support. Increased Furthermore, a greater use of active Expendituresof Employment Fund on Active and Passive Programs programs would probably not Percentage Distribution produce significant efficiency gains 2005 and 2008 and, in particular, would not ' significantly improve employment outcomes. For example, expanding 80% . - training programs does not seem to 70% . - be an adequate policy response under 60% the current economic conditions, when the main cause of job loss and unemployment is weak product demand and attendant scarcity of job opportunities, rather than the skills mismatch. Put simply, people become unemployed not because they have inadequate skills, but because jobs 2005 2008 are being destroyed. Under these circumstances increasing Source: Ministry of Labor and Social Policy: Bank staffcalculations. - programs expenditures on training . - would most probably produce little effect. Therefore in our view the reallocation of Labor Fund resources toward income support, i.e. passive programs and the adjustment of expenditure on active programs to available resources is an appropriate policy response to the current economic crisis. The imminent increase in the share of passive programs will reflect the deteriorating labor demand conditions, and as such should not be of concern. 306. Active labor market programs are more effective during the periods of economic growth, when the it is structural factors (the skills mismatch in particular), rather than demand deficiency, that are the main cause of unemployment. Moreover, active labor market programs may play an 139 important role in facilitating labor market integration of persons who are not eligible to unemployment benefit, such as new labor market entrants (youth in particular) and the long-term unemployed. 111. Labor Market Expenditures in Poland in Comparative Perspective 307. This section compares labor market expenditures in Poland to those in other OECD countries. It argues that the level of labor market expenditures in Poland is modest by OECD standards, although Poland spends on labor programs a higher percentage of GDP than most middle income countries. It concludes that rather than increasing spending on labor market programs Poland should aim at increasing their efficiency through better targeting and expanding programs that are most cost-effective while scaling-down programs that have little impact. Figure 48: Poland Spends on Labor Market Programs a Higher Percentage of GDP than Most Middle-Income OECD Countries but Less than High-Income Countries ExpendituresonLaborMarket Progmnsas%ofGDP Selected OECD Countries, 2006 - - r- - l_-----l-- n _ -- _ I I -I ~~~ ~~ hote: Unemployment rates for OECD countries are shown in Annex Table A4 Source OECD, Employment Outlook, Bank staff calculations 308. Are labor market expenditures in Poland low or high? One way of answering this question is to benchmark Poland against other countries, controlling for differences in income level. Figure 48 shows the results of such a comparison. The results indicate that labor market expenditures as a percentage of GDP are modest in Poland by OECD standards. However, labor market programs are a luxury (superior) good, with income elasticity well exceeding one. Richer OECDIEU countries spend on labor market programs disproportionately more than poorer ones. Poland being a middle-income country spends a substantially lower percentage of GDP on labor market programs than high income OECD countries. For example, Spain spends on labor market programs 1 percent of GDP more than Poland. But Poland, in turn, spends on labor market programs 0.5 percent of GDP more than the Slovak Republic, a country with a similar unemployment rate." ' O The discrepancy between data on labor market expenditures as a share of GDP in Poland presented in Figure 48 and in Figure 49 is due to the fact that data come from different sources, which use different definitions. 140 309. Figure 49 presents a more a more detailed international comparison of spending on active labor market programs. The data refer to expenditure per unemployed relative to GDP per person of working age, and thus one controls not only for the income level (labor productivity) but also for the unemployment rate. Again, such standardized expenditures on active labor market program are in Poland well below the EU average. In units of labor productivity, Poland spends on an average unemployed substantially less than not only all EU15 countries but also less than a number of EU12 countries (Bulgaria, Lithuania, Slovenia and Hungary). Still, it spends more than the neighboring Slovakia and the same as the Czech Republic. In interpreting these data it should be borne in mind that they refer to the period of high unemployment in Poland (2006). We have shown earlier that expenditures on active labor market programs increased in Poland along with the fall in the unemployment rate. Accordingly, Poland moved somewhat closer to the EU average (see Figure 46). Still, this recent move up the ladder does not change the conclusion that expenditure on active labor market programs is modest in Poland by EU standards, including new EU member states. Figure 49: Expenditures on Active Programs are Modest in Poland Expenditure on Active Programs per Unemployed Relative toGDP per Person of WorldngAge EU Countries, 2006 Netherlands Sweden Ireland Norway Austria Belgium France Finland Luxembourg Spain I Germany Italy UK Portugal Bulgaria I Lithuania Slovenia Hungary i ~ Poland i 1 I Czech R . ! Latvia I I j ~ Slovakia Romania I Estonia i 15 20 2s 30 35 Note: The indicators used shows how much a country spends per one unemployed person relative to an indicator of labor productivity. The indicator is unitless. For example, its value of five (as in Poland) means that a country spends per one unemployed five times as much as the value of output produced by a person of working age. Source: Eurostat; Bank staff calculations. 310. Should Poland spend more on labor market programs, particularly on active ones? The above comparison obviously does not provide an unambiguous answer. But it shows that among middle income countries Poland is not an outlier. It is high income EU countries that drive the EU average level of spending up, Should then middle income EU countries, and Poland in particular, follow the example of their richer neighbors? Considerable caution is warranted in this respect. Poland may decide to increase labor market expenditure as it grows richer and converges to the average EU income level. But at this juncture a conservative approach, whereby labor market expenditures per unemployed do not increase relative to labor productivity seems more appropriate. This is because the opportunity cost of increasing labor market expenditures is substantial, while the benefits are uncertain (see the next Section). That is, 141 given that the net impact of active labor market programs is limited, social welfare would be enhanced by channeling scarce resources to projects with a higher social rate of return. 3 11. International evidence does not support the view that in general countries which spend more on labor market programs have better employment outcomes. High labor market expenditures do not guarantee low unemployment. Moreover, micro econometric evidence points to limited net impact of most active labor market programs across countries (Betcherman et al. 2004, Boone and van Ours 2004, Kluve 2006). The disappointing performance active labor market programs implemented in Germany - a country with high administrative capacity -- is particularly illustrative. Wunsch and Lechner 2007, in a paper under a telling title ""What Did All the Money Do? On the General Ineffectiveness of Recent West German Labor Market Programs" conclude that "on average, all programs fail to improve their participants' chances of finding regular, unsubsidized employment". 312. Given that the rate of return for most active labor market programs tends to be low the rational way for Poland to proceed is not to expand active labor market programs and increase spending but rather to focus on improving the efficiency of the programs. This requires rigorous monitoring and evaluation of the existing programs and using the evaluation results to improve program design, increase targeting efficiency and to reallocated resources to programs that are most cost-effective and to worker groups that benefit most from given intervention. Such an approach is likely to enhance the programs' impact while not increasing their cost. To put it simply, rather than spending more on active labor market programs Poland should strive to improve the efficiency of spending.' IV. Labor Market Programs and Their Efficiency 3 13. This section examines the efficiency of labor market programs in Poland. It first looks at passive programs (unemployment benefit). Then it discusses the cost-effectiveness of active programs. Finally it presents new policy initiatives in response to the crisis. It finds that unemployment benefit system is not overly generous and per se does not create significant labor supply disincentives. However there is room to improve the cost-effectiveness of active labor market programs, by improving targeting and by expanding programs that are most cost-effective and reducing the scope of programs which are costly but have little impact. Unemployment benefit 3 14. The main features of the unemployment benefit system in Poland are summarized in Box 5. The system is not overly generous by OECD standards. The replacement rate, at 45 percent of the gross minimum wage, is relatively I o w . ~ ' The duration of benefit receipt depends on regional unemployment rate, with a maximum of 12 months in high unemployment regions or for some categories of the unemployed (e.g. older workers).72 The actual coverage rate at around 15 percent was low in the last years, which reflects the eligibility criteria and the structure of unemployment (new labor market entrants and long-term unemployed are not eligible to unemployment benefit). However, with the accelerated inflows into unemployment due to the lay-offs associated with the current economic slow-down, the coverage rate sharply increased, and presently (May 2009) reached 22 percent. 71 The replacement rate for the net benefit is 60 percent of the net minimum wage, and is higher than for the gross benefit because unemployment benefit is not subject to social insurance contribution. (Net benefit/wage - gross benefitiwage -social insurance contribution - health insurance contribution - PIT). 72 Until February 2009 the maximum duration was 18 months. 142 3 15. To strengthen the job search incentives the government plans to introduce a sliding scale for unemployment benefit. The amount will be higher during the first three months of unemployment, and then reduced by 20 percent afterwards. Such benefit design is used in many countries and is assumed to promote more active job search by the unemployed. The change is planned to come into force as of January 2010. The expected result of the introduction a sliding scale is shorter average duration of unemployment spells, and thu,s lower unemployment rate. However it should be borne in mind that job search incentives tend to have a stronger effect when the labor market is buoyant than when is depressed. 3 16. The regional differentiation of unemployment benefit duration is intended to provide longer income support to the unemployed in regions (poviats) where job opportunities are few and joblessness is high. Such an approach is used by some countries experiencing substantial regional variation in labor market conditions (e.g. Canada), and Poland falls into this category. The (registered) unemployment rate ranges from the less than 8 percent in the Wielkopolskie voivodship to nearly 19 percent in the Warminsko-mazurskie voivodship, with the variation across counties (poviats) being still larger. However, such an approach .has its downside, because longer benefit duration in higher unemployment regions leads to weaker job search incentives, which may ossify the regional differences in unemployment rates. Furthermore, counties (poviats) are considered here as separate and isolated labor markets, which is not warranted because they are relatively small territorial units. After all, the unemployed living in one county can seek employment in the neighboring counties, where job opportunities are more plentiful. Thus the design of the system is such that it discourages labor mobility or commuting. More research is needed to determine whether the benefits of the system (income support in high unemployment regions) outweigh its costs (`job search and mobility disincentives), and whether the system should be re-designed so that to limit labor market distortions. 3 17. Given the modest generosity of the unemployment benefit system in Poland, the benefit by and of itself it does not create major labor supply disincentives. However, there is a range of other non-work benefits, which jointly contribute to the low labor force participation in Poland. These benefits include early retirement pensions, disability pensions and pre-retirement benefits. Significant reforms were undertaken in recent years to limit benefit availability. These include the most recent elimination of early retirement provisions, and earlier reforms that restricted the eligibility to disability pensions and phased out pre-retirement benefits. All in all, Poland has made substantial strides in reforming its unemployment and related benefits system in order to strengthen labor supply incentives and increase the labor force participation rate. However, the effects of these recent reforms still remain to be seen. 3 18. Unemployment benefit goes mainly to the poor. Figure 50 shows that the bottom quintile (the poor) receives the largest share of the spending on unemployment benefit (28 per~ent).'~ Moreover, targeting efficiency, measured by the share of unemployment benefit going to the bottom quintile, somewhat increased from 2005 to 2007. Still, unemployment benefit is only weakly targeted to the poor since the bottom quintile receives only 28 percent of total expenditure. In other words, the "middle class" receives a substantial proportion of the spending on unemployment benefit, too. However, unemployment benefit is an insurance type benefit, contingent on the realization of the risk of unemployment, and as such it is not ex ante targeted to the poor. In this context the expost targeting is quite remarkable. This reflects the fact that unemployment is quite closely associated with poverty in Poland, and thus unemployment benefit is an effective poverty mitigation instrument. 73 Data refer to both unemployment benefit and unemployment assistance. Unemployment benefit is an insurance type benefit, while unemployment assistance is a means-tested benefit paid to the unemployed clients of social assistance offices. The bulk of the expenditure is on unemployment benefit; unemployment assistance accounts for only a small percentage of the expenditure. 143 Active labor market programs Active labor market programs (ALMP) are intended to actively assist the unemployed in their job search, to improve their employment chances and thus to facilitate their (re)employment. T h e y include j o b search assistance, training, hiring subsidies, public w o r k s and support for business start-ups. Box 6 briefly discusses their potential and limitations in addressing unemployment. Box 5: Unemployment Benefit System This box summarizes the main features of the unemployment benefit system in Poland. Eligibility 12 months of covered employment in the 18 months preceding unemployment. Benefit amount Flat rate benefit of PLN 575 (USD 310 at PPP).a) The standard amount is modified depending on the duration of covered employment: 80 percent of the standard amount if the duration of covered employment is less than 5 years; 120 percent of the standard amount if the duration of covered employment is 20 or more years. Starting with 2010 the benefit will be paid on a sliding scale, with benefit amount reduced by 20 percent after the first three months o f j o b search. Replacement ratehi The benefit accounts for: 18 percent of the gross and 24 percent of the net average wage; 45 percent of the gross and 60 percent of the net minimum wage. Benefit duration 6 months if the local (powiat) unemployment rate is up to 150 percent of the national average; 12 months if the local unemployment rate is above the 150 percent of the average unemployment rate; 12 months for persons aged 50 or more; 12 months for persons with the duration of covered employment of 20 years or more; 12 months for workless families with dependent children. a) Since June 2009, when the benefit was raised from PLN 5 5 1, b) The replacement rate i s higher when expressed in net terms, rather than gross terms because unemployment benefit is not subject to social insurance contribution. Source: Law on Employment Promotion (as of January, 2009) 3 19. The institutional setup for delivery of active labor market programs is highly decentralized in Poland. Active labor market programs are financed from t h e Labor Fund (LF) administered b y the Ministry of Labor a n d Social Affairs. T h e Ministry allocates t h e Fund's resources to districts (voivodships) according to a n algorithm, whereby t h e amount of resources received by a district is mainly the function of the number of registered unemployed. Then t h e district level Labor Offices (known as W U P ) distribute L F resources t o local (poviat) level Labor Offices (known as PUP), using their o w n allocation algorithms and according t o the Regional Employment Plan.74 These algorithms vary b y 74 Regional Employment Plan prepared by WUP in consultations with PUPs is a general document, which does not have much impact on labor market policy carried out by PUPS. It is PUP's director who determines the mix of active programs to be implemented at the local labor market. PUP's Director reports to the head of the local koviat) government, and makes decisions on labor market policy in consultation with Local Employment Council. The Council plays an advisory role and does not intervene in the management of PUP. The major problem faced by the PUP's Director is the unpredictability of the actual amount and timing of the financial resources that PUP receives. 144 district, but as a rule PUPs receive LF resources in proportion to the number of registered unemployed. It is PUPs who determine the mix of active labor market programs and who are responsible for their implementation. PUPs are accountable only to local (poviat) government, with WUPs role limited to general oversight and technical support. The Ministry's of Labor fiduciary role is limited to setting the standards for delivery of active labor market programs. It plays no direct supervisory role and has no control function over the use of labor funds by either WUPs or PUPs. Figure 50: Unemployment Benefit is Received Mainly by the Poor Distributionof Expendlturer on Unemployment Benefit by Consumption Decile 2005 and 2007 16 ~ ___.---_.__-._.I_.c_.___-..----~...__1_1.____1_____ 14 , ~ 12 10 . 8 I2005 6 - I2007 4 2 ' 0 - D1 D2 03 D4 D5 06 D7 D8 D9 D10 Note: Data refer to both unemployment insurance benefit and unemployment assistance. Source: Household Budget Survey, Central Statistical Office; Bank staff calculations. 320. The existing mechanism for the regional allocation Labor Fund's resources and for determining program mix has some apparent weaknesses. First, the regional allocation mechanism that is used provides little incentives for voivodships to reduce the number of registered unemployed because by doing so they would automatically reduce the amount of funds received. There is no reward for increasing the the job placement rate (outflows from unemployment to employment), which is the main indicator of PUP's success. Second, the decentralized mechanism for the determination of program mix may, with virtually no central oversight, limits the scope for reallocation of funds across programs so as to increase the allocative efficiency. This is important because, as we will see later, programs vary substantially in terms of their cost-effectiveness and net impact. Finally, the decentralized system limits the scope for using performance based budgeting, that is to linking the resources received by PUPs to their results (see Ivory and Thomas, 2007). Within the existing setting, the PUPs performance can be compared and rewarded only within districts but not between them. And again, we will see that some regions are more successful in delivering effective employment services than others. In this context rewarding good performance and increasing funding to most efficient PUPs has the potential for enhancing the overall efficiency of the utilization of the Labor Fund. This is because a significant proportion of PUP's revenues comes from the special reserve administered by the Minister of Labor, which allocation is discretionary, and thus the resources that will actually available for PUP cannot be adequately planned for. In addition, the fact that a substantial proportion of resources allocated to PUPs goes through WUPs causes substantial delays in actual availability of resources (the recent revision to the Law on Employment, which will come into force in 2010, reduced the role of WUP in transferring resources to PUPs). 145 lox 6: Labor Market Programs: Their Potential and Limitations in Addressing Unemployment during Economic Crisis There is a controversy among economists as well as policy makers concerning the potential of active labor market programs, such as training, hiring subsidies and public works, to address unemployment. This chapter argues that this potential is limited. Evidence suggest that well designed active labor market programs (ALMP) can improve employment prospects of disadvantaged workers groups (e.g. workers lacking work experience or skills). However, their net impact - the increase in employment chances thanks to program participation - tends to be relatively small. Their effectiveness in addressing mass unemployment, resulting from large demand shocks, is questionable. Only large scale public works can compensate for the deficient labor demand and provide temporary employment opportunities for the most needy workers. There are examples of successful implementation of public works in countries such as Argentina, India and S . Korea (during the recent Asia economic crisis). But in ECA such programs have never been implemented on a large scale. They are usually implemented on a small scale by local governments and targeted at low-skilled workers. As such they play the role of temporary income support rather than employment generation. Nonetheless, public works are in principle the only ALMP instrument to be used and expanded during the period of economic crisis in order to prevent unemployment and provide temporary work opportunities. Some economists advocate expanding training programs during the crisis to prevent lay-offs and to prepare the unemployed workers for the future job opportunities. However the logic behind such policy prescription does not seem convincing. Training programs can address structural unemployment associated with the skills mismatch, but are not an efficient response to demand deficiency unemployment. During the crisis workers lose their jobs not because they lack the necessary skills but because firms have to cut costs in the face of shrinking demands for their products. Under the crisis conditions the first priority is to support job creation, either by public investment in infrastructure or through public works, rather than addressing potential skill gaps. And training for the future skill needs may be a risky investment. Future skill needs are difficult to predict, and evidence indicates that training programs are successful only if they respond to well identified needs of both workers and employers. A more general argument against expanding training and other job matching programs in response to the demand shock is that there is an inverse relationship between matching efficiency and the unemployment/vacancies (U/V) ratio. When the U N ratio goes up - and that is what happens during the crisis -- then it is increasingly difficult to place more unemployed to fewer j o b openings. Thus an increase in the U/V ratio limits the efficiency of job matching programs, including training. Given the limited effectiveness of job-matching programs under the conditions of weak labor demand, this chapter argues that during the crisis the first priority is to provide the unemployed with income support. This can be done either by paying unemployment benefit to those who are eligible, or by providing conditional cash transfers to those, who are not. Public works and workfare programs - when the receipt of benefits is conditioned on doing socially useful work - are examples of the latter. Source. World Bank Staff analysis. 146 Box 7: Active Labor Market Programs in Poland: Key Features This box provides a brief description of active labor market programs in Poland." Training Target group: not specified (all unemployed) Benefit: stipend amounting to 120 percent of unemployment benefit paid to the trainee. Program duration: 6 months (1 2 month in special cases). Program expenditure limit: up to 3 times the average national wage per trainee. Programs targeted at disadvantaged workers groups The Law on the Promotion of Employment specifies the following disadvantaged worker groups: Youth (less than 25 years old) Long-term unemployed (registered as unemployed for 12+ months in last two years) Older workers (over 50 years old) Low-skilled workers (less than secondary education, no vocational qualifications, no labor market experience) Disabled persons Ex-prisoners Single parents Women reentering the labor market after maternity. Employment subsidy (known as "interventionworks") Objective: to promote employment of disadvantaged worker groups by subsidizing their wages Target group: all disadvantaged worker groups Wage subsidy: equal to unemployment benefit, paid to the employer; in addition a special premium of up to 150 percent of the national average wage if the employer continues to employ the worker after the subsidy expires and employs him/her for at least 12 months. Program duration: maximum 6 months Apprenticeships Objective: to smooth the school-to-work transition for young workers, and to facilitate labor market entry of new entrants Target category: all unemployed, but preference given to unemployed youth Benefit: stipend amounting to 120 percent of unemployment benefit, paid to the trainee. Program duration: 12 months for the unemployed youth, 6 months for other unemployed On-the-job training Objective: to provide on-the job vocational training for the unskilled unemployed and to award them with vocational qualifications certificate. Target group: unskilled workers Benefit: stipend amounting to 120 percent of unemployment benefit paid to the trainee. Program duration: (a) 12 to 18 months in the case of vocational training, or (b) 3-6 months in the case of basic vocational program. Public works Objective: to provide temporary employment and income for disadvantaged worker groups. Target group: all above except youth Wage subsidy of up to 50 percent of the national average wage, paid to the employer Program duration: up to 6 months Workfare (known as "socially useful works `7 Objective of the program is to maintain labor market attachment of marginalized worker groups by subsidizing their temporary part-time employment. Target group: long-term unemployed claiming social assistance benefits Employment: 10 hours of "socially useful'' work per week 147 Wage: at least USD 3.3 (at PPP) per hour, paid partly by the Labor Fund and partly by municipal (gmina) government. Business support Objective: To support business start-ups by the unemployed. and job creation by employers. Target group: the unemployed who want to start their own business, employers n h o create job for the unemployed. Subsidy: a one-time lump-sum pakment of up to 6 times the national average wage (35 times the unemployment benefit). a) According to the 2008 revision to the Law on the Promotion of Employment; some of the provisions will come into force in January 2010 Source: Law on the Promotion of Empioyment; Bank stafanuiysis. 32 1 . All major types of active labor market programs are being implemented in Poland. These include direct job creation (public works), subsidized employment, (re)training, and small business support (see Box 7 for a brief description of the programs). In terms of expenditures, the two biggest programs, which jointly absorb 64 percent of expenditures on active measures are business support and apprenticeships/on-the-job training (Figure 51, Panel A). On the other hand, the majority of the unemployed are enrolled in training and apprenticeships programs (Figure 5 1, Panel B). Business support, while being the most expensive program (36 percent of total labor market expenditures), benefits few of the unemployed (only 12 percent). This naturally raises the question of its impact, an issue to which we return later. Figure 51: Labor Fund Spends Most of its Budget on Business Support Program, but the Majority of Unemployed is Enrolled in Training and Apprenticeships Programs. Panel A Panel B Percentage distribution of expenditures by p r o g r a m Percentage distribution of beneficiaries by program 3% *Training c Tralning Employmentsubsidies * Employmentsubsidies Public w o r k Public w o r k h Businesssupport 1 Workfare Apprenticeships On-the-job tralrung O n t he-jo b training Apprenticeships Other Bustness support Source Ministr) o f Labor and Social Polics; Bank staffcalculations. Cost-effectiveness 322. Apparently the programs are quite successful in placing the unemployed to jobs. According to the official data, the job placement rate varies from over 70 percent for employment subsidy programs to 45 percent for training (Figure 52). However, information on job status covers the period immediately following program participation (up to 3 months), and is thus not necessarily representative of longer term effects, In addition, in some cases job placement is equated with the deregistration from the Labor Office. Thus the actual longer term employment effects of program participation are probably considerably smaller. In order to properly measure the programs' employment effect one needs to carry out a special 148 follow-up survey that determines the beneficiaries labor force status at a certain period after program participation (usually after 6 to 12 months). Figure 52: Officially Reported Job Placement Rates are High, but They are a Weak Measure of Programs' Effects Job Placement Rateby Prcgram 2007 . . . . . . . . . . . . . . . . . . . . . . . ..................................... Employment Apprenticeships On-the-job Public w o r k Training Workfare subsidies training Source; Ministry of Labor and Social Policy; Bank staff calculations. 323. Training, which is the largest labor market program, compares unfavorably to other programs with respect to the job placement rate. As many as 56 percent of participants registered again as the unemployed after completing the training program. This is a rather disappointing results, especially given that the labor market was buoyant in 2007 and job opportunities were numerous. This implies that in majority of cases training program was not well suited either to the needs of the unemployed, or to that of the employer. Although, as we will see later, training has a positive net impact on employment chances of the unemployed, it is at best only moderately effective in getting the unemployed back to jobs. International evidence also indicates that training is less effective when labor market conditions are deteriorating and job vacancies are few (Dar and Tsannatos 1999). Accordingly, expanding the scope of the training programs does not seem to be an effective response to the current crisis. Workers are losing their jobs not because they have inadequate skills but because product demand is falling and there is less jobs. Hence training workers will hardly improve their employment prospects. 324. There is substantial variation in cost-effectiveness across programs. The program unit cost varies from USD 80 for training programs to USD 240 for public works to USD 640 for business start- u p ~ The variation in cost-per job placement is equally large. It costs about USD 200 to place one . ~ ~ unemployed into job by providing training, but it costs USD 540 to provide employment by means of public Thus training is the most cost-effective program despite the relatively low job-placement rate. It is cheaper to provide a job to the unemployed by means of training them than by means of any other program. However, the programs are not perfect substitutes. As a rule they are targeted at different groups of the unemployed, and have different objectives (see Box 6 ) . Accordingly, expanding training programs at the cost of reducing the size of other programs is not necessarily the way to improve the efficiency of labor market expenditure. 75These costs are expressed in international dollars at the PPP exchange rate of IUSD = 1.85 PLN. 76 These figures are lower-bound estimates of the cost-per job placement, given that the official job placement rates are overestimated. 149 325. Program costs also vary across regions. Take for example training (Figure 54). The cost per in the high-cost Figure 53: Substantial Variation in Cost-Effectiveness Across region (Zachodniopomorskie) Pro - is 80 percent higher than in the low-cost region Cost-effectiveness of Active Labw Market Programs 2007 (thous PLN) (Wielkopolskie). Expectedly, the variation in the cost per = Cost per beneficiary 9 Cost per job placement 14 placement is still higher. In 12 the high-cost region (Opolskie) the cost per placement is 2.4 times as high as in the low cost region (Wielkopolskie). Apparently cost effectiveness of training in some regions is * i M E VI e E s 4 s 5 P substantially higher then in 1 n c i x 5 M 3 a E 2 n a u -u 2 other regions. Other programs n n f VI E E : : E W exhibit similar or even greater a . P 7 W s W regional variation in costs. In 6 - m m the case of public works, the Y E cost per job placement in the high-cost region Source Ministry of Labor and Social Policy; Bank staff calculations (Mazowieckie) is 4.3 times higher than in the low cost region (Lodzkie). 326. Presumably some portion of the regional variation in the cost per job placement is explained by differences in local labor market Figure 54: Program Costs Vary Across regi0ns.R conditions. In general it Training is easier to place the Unit Cost and Cost perJobPlacement byRegion unemployed to jobs in 2007 (NationalAverage=lW) regions where labor =Cost p e r placement . Unit cost I 160 markets are strong than in 140 a - - - _- - _ _ -- . I _I____^__- those where they are I ~ weak. However, we found no statistically significant correlation between the regional unemployment rates and job placement rates for a range of programs. Therefore the regional differences in program costs are likely to reflect mainly differences in the Source: Ministry of Labor and Social Policy; Bank staff calculations. efficiency of regional 150 Labor Offices." Some regional Labor Offices use public resources more efficiently than others. There is then scope for improving the efficiency in those regions, where program costs are significantly higher than the national average. A benchmarking exercise to identify Labor Offices whose cost-effectiveness is particularly low should be made part of regular monitoring of Labor Fund expenditures. Such monitoring coupled with setting performance standards whereby every intervention has to have a minimum job placement rate can be expected to improve the overall efficiency of the utilization of the Labor Fund resources (Ivory and tho ma^).'^ This issue is addressed in more detail in section V below. 327. Unit cost of most programs increased quite substantially in the recent period. This increase was not accompanied by an increase in the job placement rates, which led to a fall in cost-effectiveness of most of the programs. For example, the unit cost of training programs increased by 18 percent and the cost per placement increased by 20 percent in 2007 (Figure 5 5 ) . These increases occurred during the period of falling and low unemployment, therefore they cannot be accounted for by the lack of job opportunities and more difficult job placements. Instead the two likely drivers of cost increase are (a) the availability of financial resources due to the substantial surplus of the Labor Fund and the attendant pressure to use the resources, and (b) the selection effect whereby it is mainly hard-to-place workers who remain jobless during the period of high labor demand. We discussed the role of the first factor above in Section 111. Unfortunately, no evidence other than anecdotal is available to substantiate the second hypothesis. Still, the selection effect might have contributed to the increase in the increase in the cost per placement, but was less likely to give rise to the increase in unit cost. Thus, the cost inflation should be attributed primarily to the Figure 55: Cost of Most Programs Increased Over Last Year budget constraint faced by Labor Offiies. Inasmuch this was the case, Percentagechange i n Program Cost 2007 t o 2006 this points to inefficiency in the use of public funds. 25 Net impact 20 15 10 328. In assessing the efficiency 5 of active labor market programs 0 one key issues is their net impact. -5 l E Net impact is the change in labor -10 market outcomes of program participants that is attributed to the program intervention. The assess the net impact of a program one needs to compare labor market outcomes (such as employment or wages) of w program participants with those of a control group, which consists of the Source Ministry of Labor and Social Policy; Bank staff calculations. unemployed with similar characteristics (such as age, gender, education, duration of unemployment, local labor market conditions) who did not participate in the program. Most often the net impact is measured in terms of the increase in 77 The nature of the factors accounting for the differences in efficiency of regional Labor Offices, whether they are exogenous (e.g. labor market structure, market for service provision) or endogenous (e.g. quality of management and staff, incentives structure) is not well recognized and warrants further more in depth research. '' In some countries there are targets set for performance and standard measurements (examples include the UK and the US). Local offices have reporting requirements and are supervised by either the central office or the relevant ministry. 151 the probability of obtaining employment. Figure 56 presents the results of an evaluation of the net impact of main active labor market programs in Poland in 2006. 329. Evaluation results show that the largest programs - business support, training and apprenticeships have a positive and significant net impact on employment prospects of the unemployed. For example, the net impact of the training programs was found to be around 13 percent (Figure 56). This means that 13 out of 100 unemployed who were enrolled in training courses obtained a job thanks to the training received. This is a relatively large effect by international standards (Dar and Tsannatos 1999, Betcherman and others 2004). However, a significant net impact does not necessarily imply that the rate of return is high, that is that the benefits of the program justify its cost. The net effect of 13 percent implies that the cost per additional (net) job placement (Le. job placement attributable to program participation) is 7.7 (100/13) times higher than the cost per beneficiary (unit cost). Thus using training to place the unemployed to jobs is costly notwithstanding the positive and significant net effect. It should also be emphasized that the positive net impact of training (or any other) programs in Poland does not imply that additional jobs were created. This is because of the so called substitution effect. Some of the graduates of training programs get jobs at the expense of other jobseekers not getting them. The substitution effect is hard to measure, but the available evidence indicates that it can be substantial and thus should be taken into account when evaluating the impact of training and other active labor market programs. 330. Business support programs seem to have the largest net impact on employment. Econometric analysis suggests that the participation in this program increased one's employment chances by one-third (Figure 56). However, the impact seems to be overestimated. This is because econometric techniques used to estimate the net impact control only for the observable factors (such as age, gender, education, et^.).'^ At the same time business success very much depends on factors that are not measured and thus not included in the econometric model: motivation, entrepreneurship, and other soft skills. If these key factors are not controlled for then the results are biased upwards. Moreover, to fully assess the net impact of business support programs one needs to factor in the business survival rates. If they are low, then the net impact of the program is obviously reduced. But again, due to data limitations, this factor was not accounted for in the estimation. Thus the actual impact of business support programs is most likely to be much smaller. Accordingly, the results should not be interpret as suggesting that business support program are the most effective way of addressing unemployment, and they do not warrant the expansion of the program. We saw earlier that business support is the most costly program. It creates jobs, but at a high cost. The popularity of the program does not imply it has a high rate of return. In addition, this program - probably more than other programs -- is subject to diminishing returns. Only a fraction of the unemployed have entrepreneurial skills, and the extension of the program to those who lack the necessary skills would be futile. 19 The fact that unobservable characteristics of program participants are not controlled for is a generic weakness of semi-experimental methods (as opposed to experimental methods based on random assignment to treatment and control groups). This weakness is particularly pronounced in the case of programs, where unobservable characteristics play a particularly important role. They include business support and training programs, where unobservable skills (entrepreneurship, innate abilities) are key for success and influence the participant selection process. By ignoring the effect of unobservable we tend to overestimate the effect of program participation. This risk i s less substantial in the case of programs where unobservable participant characteristics tend to play a smaller role (wage subsidies, public works). Therefore the evaluation results for wage subsidies and public works seem more reliable than for training and especially business support programs. 152 331. Public works have a negative net impact. This Figure 56: Most Active Labor Market Programs in Poland Improve means that the unemployed the Employment Chances of the Unemployed who were employed on Net Impact ofActiw Labor Market Programs public works programs were 2006 less likely to have a job after participation in the program Business support than the similar unemployed Training who did not participate. This ~ I i 1 I results may seem surprising Apprenticeship I I I but is not unexpected. ~ International evidence indicates that public works rarely have a positive impact Publicworb , 1 I I I on future employment Jobsearch assistance chances, and often have a ~ ! 1 L * I negative impact (Dar and -10 -5 0 5 10 15 20 25 30 35 Tsannatos 1999, Betcherman and others 2004). One reason Source: Zatrudnienie w Polsce 2007. Bank staff calculations for the negative effect is that participants are locked in the program and do not look for regular jobs. Another possibility is stigma I associated with public works participation. Given that public works have no positive effect on future employment chances, they should be regarded as mainly an temporary employment and income support program, rather than an employment promotion program. 332. Employment subsidies do not improve employment prospects of the unemployed. This means that employers tend to employ the unemployed for only as long as they receive the subsidy. Once the subsidy expires, the unemployed loose their jobs. The program does not promote sustainable employment and is thus not a particularly efficient way of combating unemployment. This suggests that .the program should be reviewed and redesigned with the aim of improving its impact, or should be scaled down. 333. Surprisingly, job search assistance was found to have a negative impact on employment chances. This result is at odds with the results of other evaluations (for a wide range of countries) that find job search assistance to be the most cost-effective labor market intervention (Dar and Tsannatos 1999, Betcherman and others 2004). The result reported here is likely to reflect methodological problems with quantifying job search assistance and with constructing the adequate control group. Obviously, it should not be interpreted as implying that job search assistance is harmful; instead the result suggests that the quality of job search assistance provided by PES in Poland is rather low and that there is considerable scope for improvement. 334. To sum up, the net impact evaluation results indicate that the largest active labor market programs have a positive and significant net impact. This refers to training and apprenticeship programs, and to business support. At the same time other programs: public works and employment subsidies, do not improve job prospects of the unemployed, although they provide temporary employment and income support. Given the uneven effects of different programs there is scope for improving efficiency by (a) redesigning the programs that have little impact, and (b) reallocating fund from worse to better performing programs. However, the latter exercise cannot be done in a mechanical way, given that different programs have different objectives. More thorough program evaluation is necessary to determine which programs work best for different client groups. 153 Targeting efficiency 335. Another issue relevant to the assessment of efficiency of active labor market programs is targeting. Are those workers who are at the highest risk of long-term unemployment dominant among the participants? The evidence is limited but suggests that the targeting efficiency of most programs is not particularly high, especially when it comes to targeting by education. We focus here on targeting by education and by urbanhural location as these are the only two dimensions of targeting for which the relevant data is available. 336. While the incidence of unemployment is the highest among workers with less than secondary education, public works is the only program where the unemployed with less than secondary education (including basic vocational training) represent a clear majority among the participants (78 percent, compared with 57 percent in the general population of the unemployed). In contrast, large programs such as training, apprenticeships and business support appear to be targeted primarily at the unemployed with secondary or in some cases even tertiary education (Figure 57, Panel A). For example, about one quarter of beneficiaries of apprenticeship and business support programs have tertiary education. Earlier we have shown (Section 11, Table 1) that workers with tertiary education face good employment prospects. Targeting programs at this groups thus carries a risk of deadweight loss. That is, many program participants with tertiary education would have found jobs anyway, without program intervention. Targeting programs at workers with secondary general education (as opposed to secondary technical) is justified as the incidence of unemployment among this group is relatively high, and workers without any vocational or technical skills tend to face poor employment prospects. Figure 57: Targeting of Active Labor Market Programs by Education and Urbanmural Location Panel A Panel B Program Participants by Education Program Participants by Location2007 2007 m Rural areas Small town (< 25,000) 9 Basicvocationalor less 9 Secondary . i. Tertiary Town (< 100,000) City (over 100,000) 100% 100% 80% 80% - 60% - 60% 40% - 40% 20% 20% 0% - 0% - Source; Zatrudnienie w Polsce 2007. 337. Most programs beneficiaries come from larger cities. The exceptions are public works and employment subsidies, which benefit predominantly inhabitant of rural areas and small towns (Figure 57, Panel B). For example, over 50 percent of training participants come from large cities. In contrast, 80 percent of public works participants come from rural areas and small towns. Unemployment in Poland is largely an urban phenomenon (see Table 1 in Section 11) and it is thus not surprising that program beneficiaries are mainly urban residents. However, it is rural areas and small towns where job opportunities are most scarce, whereas in majority of large cities job opportunities are plentiful. In this context public works and employment subsidies are well targeted and compensate for the weak labor 154 demand in rural areas and small towns. But as we have shown earlier, these two programs have no positive impact on participants' future employment prospects, and thus should be regarded as temporary employment cum income support programs. Overall targeting by urbanhural location seems reasonable. Job creation programs (public works, employment subsidies) are run in rural areas and small towns where jobs are few. Training and apprenticeship programs are run in larger cities where job opportunities are better, but it is the skill mismatch that prevents the unemployed from getting jobs. 338. To sum up, there appears to be room to improve targeting efficiency of active labor market programs, especially with respect to targeting programs at different (educational) skill categories of workers. Presently some programs (most notably training) seem to be at least partially targeted at workers who are relatively well educated and face good employment prospects. At the same less educated workers, whose employment prospects are weak, are underrepresented in some programs. This may be because they do not benefit from those programs, but if this is the case then there are grounds for changing the program mix so to help the most disadvantaged workers, including the low skilled ones. Improving targeting efficiency requires better information, in particular the evaluation of the net impact of different programs for different categories of the unemployed (low skilled, youth, long-term unemployed, etc.). Only once this information is available targeting efficiency can be fully assessed and measures defined to increase it. V. Labor Market Policies in Response to the Crisis 339. To mitigate the effects of the current economic crisis on the labor market, the government initiated a process of tripartite negotiations which recently (June 2009) lead to the agreement on the so called anti-crisis package. The key elements of the package are a salary compensation scheme for employees forced to work fewer hours, referred to as short-time working subsidy and enhanced working time elasticity, that allows employers to redistribute working hours over 12 months rather than the current 3 months to encourage adjustment of working time and discourage layoffs. The details of both programs are presented in Box 8. 340. In addition, governments in counties (poviats) that are particularly severely affected by the crisis can implement special employment programs (in addition to regular programs) that are financed from a special reserve within the Labor Fund that is at the disposal of the Ministry of Labor. These special programs are defined by the local government so as to best respond to local needs and labor market conditions. 341. The measures envisaged by the anti-crisis package are in line with the recent recommendations on policies to soften the impact of the crisis issued by the European Commission as well as the World Bank (Carone and others 2009, World Bank 2008). Specifically, both the European Commission and the World Bank endorse financial support to temporary flexible working-time arrangements, arguing that temporary adjustment of working hours ("short-time") in line with production needs can be an important source of labor input flexibility, and thus a preferred alternative to lay-offs. They also both argue that such action needs to be combined with measures supporting employability and guiding people towards new jobs, empowering workers to take advantage of new opportunities when the economy recovers (Carone 2009, p.3; World Bank 2008, p, 12-15). 155 Box 8: Special Programs to Mitigate the Effects of the Economic Crisis on Employment There a r e two new labor market programs intended to mitigate the employment impact of the current economic crisis. First, the short-time working subsidy, and second the flexible working time arrangements. Short-time working subsidy is a wage subsidy paid to employers who experienced a substantial fall in sales revenues and who therefore put workers on reduced work hours or temporarily laid them off (put on mandatory leave). The government will also finance the cost of training workers who are put on mandatory leave. Flexible working time arrangements allow employers to redistribute working hours over 12 months to encourage adjustment of working hours and consequently discourage lay-offs. Specifically, workers can be requested to work shorter hours during the period of economic downturn and then work longer hours during the period of upturn, without being paid overtime premium. Both measures were agreed during the course of tripartite negotiations between trade unions, employers' representatives and the government and constitute two main elements of the so called "anti-crisis package". The provisions of the anti-crisis package are planned to be in place until December 201 1 , Of the two above mentioned programs the short-term working time subsidy is expected to have a more significant labor market effect. Here is its more detailed description. The program is available only to employers who experienced a year-over-year fall in sales revenues of at least 25 percent during three consecutive mpnths after July 1, 2008. Moreover, to be eligible, the employers need to present a recovery plan and cannot have tax arrears. Employees who are temporarily laid-off a r e to be paid the statutory minimum wage (PLN 1276), which is in part financed by the government subsidy amounting to unemployment benefit (PLN 552 o r 43 percent of the total amount), with the remaining part covered by the employer. Wages of employees working short hours will be subsidized up to 70 percent of unemployment benefit. The wage subsidy is to be paid for a period of maximum 6 months. It is estimated that the program will save about 250 thousand jobs. It is also planned that out of that number 50 thousand employees will receive government co-financed training. The total fiscal cost of the program is estimated to amount to PLN 1.46 billion, of which wage subsidies will amount to PLN 0.96 billion and the cost of training to PLN 0.5 billion. The program is to be financed by the Labor Fund (although formally the benefits will be paid out of the Fund of Guaranteed Employee Benefits). Programs of short-term working subsidy, similar to that recently adopted in Poland, are in place in a few EU countries, including Germany, France, Spain, the Netherlands and Italy. However, the U.K. government has recently resisted calls for a program to subsidize short-time working. The short-time working program has its pros and cons. On the positive side, as the proponents of the program argue, the program saves jobs that otherwise would be destroyed. Workers who would lose these jobs would end up claiming unemployment benefit, so the program is fiscally neutral. But the social benefits of keeping people employed rather than unemployed are obvious, making the program a preferred alternative to the unemployment benefit program. On the negative side, the program carries with it a risk of deadweight loss, that is the government subsidizing salaries of workers who would not actually be made redundant, which implies wasting public resources. Furthermore, there is a risk that the program would support jobs that would ultimately be lost regardless. The subsidy may go to unviable firms, inhibiting enterprise restructuring and thus leading to inefficient allocation of resources. Whether the benefits of the program outweigh its cost is ultimately an empirical question, and the verdict is still out. The controversial feature of the Polish program is a high benefit replacement rate for low-wage workers. In the extreme case, minimum wage workers on temporary lay-off will still receive their full wage and do not bear the cost of adjustment. Source: Draft Act on "Mitigation of the Effects of Economic Crisis on Employees and Employers", July I , 2009 342. At the same time the European Commission and the World Bank suggest temporary expansion of unemployment insurance, in particular the extension of benefit duration. A t present the government d o e s not plan to extend t h e duration of unemployment benefit. The Polish system is flexible 156 in that it allows for longer benefit duration (12 months instead of standard 6 months) in high unemployment regions and for some disadvantaged worker groups. However, "high unemployment" is defined in relative terms, that is in relation to the national unemployment rate. The advantage of such a system design is that it provides for extended income support in regions where labor market conditions are particularly bad without unduly weakening job search incentives in regions where labor market conditions are better. Such a design also limits the fiscal cost of the unemployment benefit scheme. But on the other hand such a system is less well suited to a situation of an adverse systemic shock that hits the whole country, as it is the case with the current global economic crisis. Under such conditions a case can be made for defining "high unemployment" in absolute rather than relative terms, and consequently for temporarily extending potential unemployment benefit duration whenever the national unemployment rate exceeds certain threshold. 343. Should Poland consider temporarily switching to a system with benefit duration extended in all regions where the unemployment rate exceeds certain absolute rather than relative threshold? The current (2009) threshold for extended unemployment benefit is the 14.4 percent unemployment rateego In 2010 the threshold will be raised to 16 percent. This is seems pretty high. However, the registered unemployment rate tends to be about 3 percentage points higher than the actual (ILO/LFS) unemployment rate. Accordingly, in 20 10 the unemployed will be entitled to extended unemployment benefit in regions where the actual unemployment rate exceeds about 13 percent. This seems pretty reasonable, and therefore there is no immediate case for switching to a different system, especially given that the associated fiscal cost would be non-negligible. Alternatively, the government may consider introducing a more gradual scale, with benefit duration extended to 9 months in regions where the unemployment rate is up to 50 percent above the national average. But the benefits of such a change would need to be weighed against its fiscal cost. 344. Finally, should the traditional active labor market programs, such as public works or training, be expanded to cope with the crisis? Definitely there is scope for the adjustment of the scale of selected programs at the local level. Regions that are most affected by the crisis may need to expand some of the programs - for example public works - to provide temporary employment and income support to poor and jobless workers. Such a possibility is provided for by the "special" regional employment programs mentioned above. 345. However, there is no convincing reason for substantially scaling up expenditures on active labor market programs at the national level. Mainly because the net impact of these programs tends to be at best limited (see above). Accordingly, the marginal benefit of program expansion would be small, while the cost would be substantial. And such cost would be particularly difficult to justify in a time of fiscal stringency. When budget deficit is high there is no space for additional programs that are characterized by a low rate of return. 346. In particular, the calls for using large scale training programs as a policy to cope with the crisis should be resisted. The increase in unemployment in the wake of the crisis is caused by the collapse of demand, not by the skills gap. Workers loose their jobs not because they lack adequate skills but because of the plunge in product demand. Thus training or retraining will not address the underlying cause of growing unemployment. And training is effective only when it is demand driven. That is when there are job vacancies that cannot be filled in because workers lack the required skills, stnd training programs address the problem by providing the unemployed with skills that they need to take available jobs. Moreover, accurately predicting the skill needs that will emerge once the economy recovers is a hard task. There is thus a risk of costly investments in skills that will soon prove redundant. This does not That is 150 percent of the national unemployment rate in June 2008. 157 mean that training should not be used during the period of crisis. It can still be a useful tool of improving employment prospects of workers with well identified skill needs. And indeed short-time working can be used by workers as an opportunity to upgrade their skills, as envisaged by the program. The point is that training is not a panacea for unemployment, especially when it is caused by deficient demand. VI. Conclusions and Policy Implications 347. There are three key findings emerging from the analysis done in this chapter. First, the level of labor market expenditures in Poland is moderate by OECD standards. Although they absorb a lower share of GDP than in high income OECD countries, they absorb a higher share than in most middle- income OECD countries. Expenditures on active labor market programs per one unemployed (relative to labor productivity) are in Poland similar as in other EUlO countries, but lower than EU15 countries. There is thus no evidence that labor market expenditures in Poland are too high and should be reduced.' Neither there is evidence that they are too low and should be increased. The key issue is thus the efficiency of labor market spending in Poland, rather than the level of spending. 348. Second, there are sufficient resources secured to cope with the labor market effects of the current global economic crisis. The Labor Fund, out of which both passive and active labor market programs are financed, amassed a substantial surplus during the earlier period of falling unemployment. Simulation results suggest that current revenues (coming from a payroll tax) plus the accumulated surplus will be sufficient to finance unemployment benefits and active programs under the falling employment and high unemployment scenario until at least 201 1. Specifically there is no imminent risk of insufficient funding for unemployment benefit. This is important given that unemployment benefit goes mainly to poorer households and is thus an effective poverty mitigation instrument.8' At the same time the unemployment benefit system is not overly generous and thus does not create significant labor supply disincentives. 349. Third, the analysis suggests that there is room to improve the efficiency of labor market spending, that is to achieve better outcomes given the expenditure level. In particular the following issues stand out. 0 Governance structure. Highly decentralized structure for employment programs delivery, which limits the scope for increasing the overall efficiency through reallocation of Labor Fund resources across regions and programs; 0 Zncreasingprogram costs. The unit cost of most programs increased well above inflation in recent years. This was not coupled by an increase in the job placement rates, which led to the fall in the cost-effectiveness of most of the programs. One possible reason for the cost inflation is the accumulation of the Labor Fund's surplus, which softened the budget constraint faced by local Labor Offices and led them to use the available resources more freely. 0 Regional allocation o funds. Allocation of Labor Fund resources by regions according to f the formula whereby the size of the transfer depends on the absolute number of the unemployed in the region. This limits the incentives to reduce registered unemployment, as lower unemployment implies lower transfers; 81 Obviously only for those households where poverty results from a job loss by a family member 158 Substantial variation in cost-effectiveness across programs. Some programs are costly while their benefits - measured in terms of improved chances of sustainable employment - are either low or uncertain. For example, business support is a costly program that absorbs a large share of total resources but benefits few unemployed and its actual net effect is difficult to ascertain. Lack of impact of someprograms Programs also differ substantially with respect to their net impact; while some do improve employment chances of the unemployed, others do not. Most disturbingly, the recent net impact evaluation study has found that job search assistance and counseling have little effect on the unemployed chances to find a job. This is worrisome because in general job search assistance is found to be the most cost-effective program. But apparently in Poland job search assistance is not effective. This suggests that the quality of services provided is low, and thus that there is substantial room to improve it. This is an important issue which merits further, more thorough research to identify specific reasons for the inefficiency of public employment services. Regional variation in programs ' cost-effectiveness, In some regions program unit cost and cost per job placement are much higher than in others. And these differences are not accounted for by differences in regional labor market conditions. This implies that some local Labor Offices are more efficient than others. If so then there is scope for enhancing the overall efficiency in program delivery by improving incentives for good performance by local labor offices, for instance by setting performance standards. Mixed evidence on program targeting. Public works and wage subsidies appear to be well targeted to low-skilled unemployed and to rural areas and small towns, where job opportunities are few. However these programs hardly improve future employment chances. The two largest programs: training and apprenticeships, are only partially targeted to the less-skilled workers, whose employment prospects are significantly worse than those of better skilled workers. Business support, which is the most expensive program, is not well targeted as its beneficiaries tend to be relatively -well educated. There is thus room for improving targeting efficiency and possibly changing program mix so as to reach to the most vulnerable workers. 350. Some of the above issues are of structural nature and cannot be resolved without a fundamental overhaul of the whole system. This particularly refers to governance issues arising from the decentralized structure of the provision of labor market programs. Such a structure limits the Ministry's of Labor fiduciary role and its ability to influence program design and mix and results in the limited accountability regarding the use of public funds. However, the discussion of such long-term systemic issues is beyond the scope of this report. We are focusing here on short to medium-term measures to improve the efficiency of labor market spending. Such measures can be grouped under two headings: ( 1 ) measures to improve the overall efficiency of labor market expenditures, and (2) measures to improve efficiency of labor market programs. We discuss them in turn. Measures to improve the overall efficiency o labor market expenditures include: f In the medium term, consider lowering the contribution rate to the Labor Fund. The contribution rate should be set so as to secure resources sufficient to finance labor market 159 programs under the steady-state unemployment rate. That is, it should be set so that revenues equal expenditures Over the business cycle. Put differently, Labor Fund should generate surplus during the period of economic upturn, which then can be used to finance the increased expenditures during the period of downturn. The current contribution rate of 2.45 percent seems to high in the light of the substantial surplus generated by the Labor Fund even when the (registered) unemployment rate exceeded 15 percent. Such persistent surplus gives rise to a softer budget constraint faced by labor offices, allows them to finance programs that are less effective, and leads to increase in program cost and consequently to the fall in their cost-effectiveness. In short, persistent surplus contributes to the risk of less efficient use of public funds. Reducing the surplus by adjusting revenues improves the incentives foraefficiency. An important additional benefit of lowering the contribution rate is a lower tax wedge on labor, which encourages formal employment. 0 Revise the formula for allocating Labor Fund's resources across regions so as to strengthen incentives for the reduction in registered unemployment. One possibility is that the relevant formula includes job placement rate (adjusted for the regional unemployment rate), so that better performing regions are awarded additional funds. 0 Develop the methodology and encourage regional (voivodship) Labor Offices to use performance standards to improve the effectiveness of local (poviat) Labor Offices. Measures to improve the efficiency of labor market programs include: 0 Introduce a system for monitoring and evaluation of active labor market programs. Whereas monitoring should be carried out on a regular basis, program evaluation should be done periodically. The objective of the system is to determine cost-effectiveness, targeting efficiency and net impact of different programs for different groups of the unemployed. Use the results to reallocated funds to programs that are most efficient, and to target programs to those unemployed who benefit most from a given intervention. Develop a more efficient system of job search assistance and activation policies by (a) strengthening the capacity of public employment services, and (b) contracting out specialized employment services to non-public service providers. 0 Identify main factor accounting for the low effectiveness of public employment services, particularly job search assistance, and develop and implement a reform program to enhance their effectiveness . 3 5 1 . It should be underlined that the above recommendations are consistent with the Governments priorities for labor market reforms presented in Rutkowski (2008). They are alos consistent with the Government's plans to develop modem and efficient employment services and activation policies in Poland (FISE 2009). 160 STATISTICAL ANNEX Table A l . Monthly uneniploymcnt ratr (registered) percent Month 2007 2008 2009 January 15.1 11.7 10.5 February 14.9 1.5 10.9 March 14.4 1.1 11.2 April 13.7 0.5 11.0 May 13 10 10.8 June 12.4 9.6 July 12.2 9.4 August 12 9.3 September 11.6 8.9 October 11.3 8.8 November 11.2 9.1 December 11.4 9.5 Source: Ministry of Labor and Social Policy 161 0 0 N 00 0 c! 0. Q\ N m N I 0 0 - r-4 2 m N o\ 0 0 N 00 0 9 0 N ?i I O 0 N - x m o\ 0 0 N 00 0 0 N z I O 0 N - x Q\ 0 0 cy 00 0 0 c1 z m N I 0 0 N - 2 N N Q\ 0 0 N 00 0 0 N 2 m 13 I 0 - P 0 N 2 r- e B L z V E $ z Table A3. Labor Fund: Projected revenues and expenditures of the under different unemployment scenarios (2009-2011) Scenario 1: optimistic Unemployment rate, LFS ( percent) 6.7 8.5 8.8 8.9 Unemployment rate, registered ( 9.5 12.1 12.5 12.7 percent) Labor Fund (PLN million) Revenues 9131 4 95 19.7 9897.6 10348.4 Expenditures 6460.3 7861.4 9228.7 9927.6 of which: Active programs 3500.0 4000.0 4500.0 5000.0 Passive programs 2460.3 3361.4 4228.7 4427.6 Balance 2671.1 1658.3 668.9 420.8 Scenario 2: pessimistfc Unemployment rate, LFS ( percent) 6.7 11.2 11.7 10.6 Unemployment rate, registered ( 9.5 14.0 14.0 12.7 percent) Labor Fund (PLN million) Revenues 9131.4 9264.1 96 19.6 10177.8 Expenditures 6460.3 8939.7 10598.7 10753.7 of which: Active programs 3500.0 4000.0 4500.0 5000.0 Passive programs 2460.3 4439.7 5598.7 5253.7 Balance 2671.1 324.4 -979.1 -575.8 a) Actual data Source Instytut Badan Strukturalnych T;rhle A4. H a r m o n i d irnemploj iiicnt rates for OECD cotiiitrics, 2005-2008, Country 2005 2006 2007 2008 Australia 5.0 4.8 4.4 4.2 Austria 5.2 4.7 4.4 3.9 Belgium 8.5 8.3 7.5 7.0 Canada 6.8 6.3 6.0 6.1 Czech Republic 7.9 7.1 5.3 4.4 Denmark 4.8 3.9 3.8 3.4 Finland 8.3 7.7 6.8 6.4 France 9.3 9.2 8.3 7.8 Germany 10.6 9.8 8.4 7.3 Greece 9.9 8.9 8.3 7.7 Hungary 7.2 7.5 7.4 7.8 Ireland 4.4 4.5 4.6 6.3 Italy 7.7 6.8 6.1 6.8 Japan 4.4 4.1 3.9 4.0 Korea 3.7 3.5 3.2 3.2 Luxembourg 4.6 4.6 4.2 4.9 Netherlands 4.7 3.9 3.2 2.8 New Zealand 3.8 3.8 3.7 4.2 Norway 4.5 3.4 2.6 2.5 Poland 17.8 13.9 9.6 7.2 Portugal 7.7 7.8 8.1 7.8 Slovak Republic 16.2 13.4 11.2 9.6 Spain 9.2 8.5 8.3 11.4 Sweden 7.3 7.0 6.2 6.1 Switzerland 4.4 4.0 3.6 3.5 United Kingdom 4.8 5.4 5.3 5.6 United States 5.1 4.6 4.6 5.8 European Union 8.9 8.2 7.1 7.0 OECD - Europe 9.8 8.9 7.9 7.9 OECD - Total 6.8 6.2 5.7 6.0 Source: OECD. 164 MEDIUM-TERM BUDGETING AND PERFORMANCE BUDGETING'~ Introduction 352. The economic crisis has put great emphasis on the need for governments to spend well scarce public resources. In this context, the reform of public financial management practices to ensure efficiency and effectiveness of public spending has become even more important than before the crisis. Two major reforms- medium-term expenditure frameworks and performance budgeting- have been central elements in improving the management of public finances. The reforms have been developed and refined by many countries to help build fiscal policy credibility and predictability via a more strategic, multi-year, budget planning perspective. At the same time, performance budgeting concepts and methods have been especially important in many countries in bringing a greater focus on the results from government spending. 353. Poland has made substantial progress in the last three years in moving towards performance budgeting, and is now moving to introduce a stronger multi-annual focus to budget preparation. An indicative performance budget-including a programmatic breakdown of expenditures, program objectives, and key performance indicator- is now presented to the parliament for information alongside the traditional annual budget. A number of steps will be required to move from this promising start to a position where the budget is appropriated on a program basis and budgeting becomes truly performance-informed, including: 0 The reform of the accounting and financial management information systems to support budgeting by program. 0 Further development of performance indicators, the introduction of a program evaluation mechanism, and the further refinement of the program structure and definition of program objectives. 354. Improved institutional collaboration in the design and implementation of the reforms will also be crucial to their success. 3 5 5 . The key element of the government's medium-term budgeting initiative is a new Multi-Year State Financial Plan (MYSFP), to be voted on by the parliament. The MYSFP is expected to be a four-year fixed pian covering the period of each government's term in office. It is not expected to set binding multi-year aggregate fiscal limits, but will instead focus on improved fiscal projections and a statement of the government's medium-term fiscal policy. For this to make a difference to macro-fiscal policy, the required statement of medium-term fiscal policy should be specific (e.g. including numerical fiscal targets) rather than general in nature. However, the most important requirement for improved medium-term budgeting in Poland is a substantial improvement in the quality of projections of medium- term expenditures and revenues. Only on the basis of much better projections will it be possible to properly assess, and then take action to ensure, the compatibility of existing expenditure and revenue policies with MT aggregate fiscal policy. 82 The chapter was prepared by Marc Robinson. It benefited from comments from Leszek Kasek and Kaspar Richter. 356. Both performance budgeting and medium-term budgeting require the development of improved expenditure prioritization processes which are able to make full use of performance information and medium-term program expenditure projections. 357. The success of performance budgeting and medium-term budgeting in Poland requires substantial additional capacity building, extending the thematic scope of the major performance budgeting training effort which is currently underway with EU support. Also of benefit would be further technical assistance in a number of areas, including accounting/financial management information systems, evaluation and performance indicator development. 358. This chapter examines cross-cutting reforms in the area of public financial management. It focuses on the two main areas of reform currently underway in Poland- medium-term budgeting and performance budgeting. I. Medium-Term Budgeting 359. The development of medium term budgeting in Poland is in its early days. The centerpiece of the planned new framework is the development of a new Multiyear State Fiscal Plan. The other notable development has been the increasing use in the budget of multi-year programs in respect particularly to major capital projects. Before examining these developments, it is useful to provide context by outlining the nature and benefits of medium term expenditure frameworks (MTEFs). Key Features and Benefits of Medium Term Expenditure Frameworks 360. The core elements of a M T E F a r e 0 A clearly stated aggregate MT fiscal framework (MTFF), including a statement of desired medium-term outcomes for key fiscal aggregates (especially deficits and debt)-that is, medium-term fiscal targets-and of the aggregate expenditure envelope and aggregate revenue levels consistent with that fiscal policy. 0 A process to ensure that expenditure and revenue policy are consistent with the MTFF-that is, to maximize the probability that the government's expenditure and revenue policies will, over the medium-term, result is levels of aggregate expenditure and revenue consistent with the fiscal targets set in the MTFF. 0 Resulting from this process, a set of medium-term sectoral or ministry expenditure projections or ceilings which both embody government expenditure policies and are consistent with its MT fiscal policy. In most countries, these are indicative (Le. the government makes no advance commitment to maintaining sectoral/ministry expenditure at the projected levels), but in some countries (e.g. the UK prior to the current crisis) the government commits itself to funding sectors or ministries at these projected MT levels. 36 1. A fundamental tool for ensuring the compatibility of expenditure, revenue and aggregate fiscal policy a r e what are often referred to as "forward estimates". These are MT estimates of expenditure and revenue on a "current policy" basis-that is, projections which indicate what expenditure and revenue will be in each of the next three or four years if there are no new spending initiatives, no changes to tax laws, and all commitments made to future expenditure (including political promises) are taken into account. Armed with good forward estimates, ensuring the consistency of "bottom up'' 166 expenditure and tax policy with "top down" aggregate expenditure limits driven by fiscal policy is a two stage process, involving: Firstly, an assessment via the forward estimates of whether unchanged expenditure policy is consistent with desired levels of aggregate expenditure. If the forward estimates indicate that existing expenditure policy will result in an excessive level of aggregate expenditure over the medium-term horizon, then adjustments are made to current spending policy (or, potentially, to tax policy). If, on the other hand, current expenditure policy implies expenditure below the aggregate envelopes, then there is room for new spending initiatives (or tax cuts). Secondly, a similar assessment of the impact of potential new spending initiatives, using estimates of their MT costs to determine whether they can be accommodated without breaching the aggregate expenditure envelope, or whether they must either be scaled back or room found for them by cutting elsewhere. The key benefits of an effective MTEF are: Improved aggregate fiscal discipline: the MTEF significantly reduces the chances of sound fiscal policy being undermined by higher-than-expected expenditure arising from expenditure policies and commitments which the government made in the past without fully considering their future implications. 0 Reduced forward funding uncertainty for spending ministries, leading to better planning and management. This reduced uncertainty is greatest when government makes commitments about funding over the MT time horizon. However, even where the MT sectoral/ministry projects are only indicative, an effective MTEF can greatly reduce uncertainty. This is because, to the extent that the MTEF ensures that all expenditure policy decisions are fully consistent with aggregate fiscal policy over the medium-term, it will substantially reduce the need for government to make ad hoc spending cuts because of unexpected trends in the level of aggregate spending. 0 Improved expenditure prioritization: the process of reconciling expenditure policy with top- down aggregate expenditure envelopes encourages a more systematic consideration of expenditure priorities. For this reason, MTEFs in many countries are closely linked with improved expenditure review and prioritization processes (such as the Spending Reviews in the UK, the Re`vision ge`ntrale des politiques publiques in France, and work of the Expenditure Review Committee of Cabinet in Australia). These processes provide an excellent means by which to feed performance information more systematically into budget preparation, providing an important link between the MTEF and performance budgeting. 362. The current severe economic crisis has introduced a degree of uncertainty into public finances across the world which might make one question the relevance of medium-term budgeting. In the short term, this uncertainty has adversely impacted on the reliability of medium-term estimates. However, what makes budgeting from a medium (and long) term perspective more important than ever is the fact that, as the international economy emerges from crisis, governments will face a major challenge in restoring the sustainability of public finances. It is crucial under these circumstances to further strengthen understanding of the future aggregate fiscal implications of present expenditure and revenue 167 policy decisions. The purpose of a credible multi-year fiscal plan is also to assure market participants that the current deterioration in fiscal position is addressed over a medium term. Medium-Term Budgeting in Poland 363. The need to strengthen the medium-term perspective in Polish budget preparation has been the subject of public discussion, and of comment from the European Commission, for some time. In response to this, the government proposed in 2008 revisions to the Public Finance Act which would make mandatory a new Multiyear State Financial Plan ( M S F P ) , to be adopted by the Council of Ministers. The M S F P would include a statement of the government's MT fiscal policy, MT projections of expenditure and revenue (on a function/tasMsub-task basis) and aggregate fiscal projections together with macroeconomic assumptions. 364. At time of writing, the amendments to the Public Finance Act, including those covering medium-term budgeting, were still under consideration by the parliament, and their final shape not yet determined.*' 365. Although the government's initial proposal was for a four-year rolling plan (Le. with an additional year being added as the program is updated each year), it is understood that the final form of the legislation is likely to require a four-year f l e d plan. This means that each new government would, on coming into office after an election office, place before the parliament a financial plan covering its full four-year term in office. Although the plan would be updated each year during the government's term in office, it would not be extended to cover additional years, but continue to cover only the term of life of the government. The M S F P would represent the medium-term financial plan of each government, with the four-year time frame being synchronized with the electoral cycle. 366. It appears that the revised Public Finance Act may not change the position in relation to medium term aggregate fiscal policy substantiallys4. The government's proposed amendments included a provision (Art. 120) which would have made the medium-term debt projections contained in the M S F P limits which the relevant annual budgets would have been obliged to respect, although this was heavily qualified by a stipulation that this limit could be breached in (undefined) "particularly justified cases". It is understood, however, that this provision is no longer likely to feature in the final legislation. If this is the case, the macro-fiscal content of the M S F P will be confined to the statement of MT fiscal policy combined with MT fiscal projections. If the statement of fiscal policy is general rather than specific in nature, this would not be very different from the current position, in which MT fiscal projections are made public in the a number of documents (the Justzjication of the Budget Act, the Convergence Programs and the Debt Management Strategv) and the Convergence Program contains a quite general statement of MT fiscal policy which mainly emphasizes Poland's determination to meet the Stability and Growth Pact criteria so as to facilitate early Euro adoption. The government's proposed amendments to the Act stipulate that the M S F P is to include a statement of the "directions" of fiscal policy, but this requirement could presumably be satisfied with a statement no more specific than that contained in recent Convergence Programs. 367. It would arguably have been useful if the proposed amendments had been much more specific about the required contents of the MYSFP fiscal policy statement, and had included a '` 83 Although a copy of the amended Act proposed to the parliament in 2008 is available, there was no documentary version of the probable modifications of the government's draft emerged during the process of review by the ar I i ament. Note, however, the separate provisions which aim to tighten up the enforcement mechanisms in respect to Poland's debt limits. 168 requirement to specify numeric M T fiscal targets for the deficit and debt. Even in the absence of this, however, the government could develop and make public a statement identifying precisely the proposed content of the fiscal policy element of the MYSFP, including quantitative fiscal targets. 368. The possibility that Poland might adopt the approach of countries like the Netherlands and Sweden in setting binding medium-term ceilings for aggregate expenditure was the subject of some discussion prior to the crisis. Proponents of such ceilings viewed them as an instrument for reducing excessive levels of aggregate expenditure, and making expenditure policy less pro-cyclical in the upswing of the business cycle. In formulating its proposed amendments to the Public Finance Act, the government considered and rejected this possibility. In the light of current crisis conditions, this is understandable. Whatever the merits of such ceilings may have been in the pre-crisis period, a persuasive argument can be mounted that they would be inappropriate in the present highly uncertain fiscal environment. The same can be said of binding forward commitments in respect to sectoral or ministerial expenditure ceilings, on the UK model. This is an option which Poland should consider further only in some years time, when the foundations of medium-term budgeting are well established and macroeconomic and fiscal conditions have stabilized. In the latest Article IV Report, the Fund mentions that by focusing on nominal deficits, the pro-cyclical bias in fiscal policy remains and emphasizes benefits stemming from binding multi-year expenditure ceilingsEs. 369. What is certainly a novel element in the proposed MYSFP is the inclusion of MT projections of revenue and expenditure (aggregate and by program). To the extent that these represent credible projections of the medium term path of financial flows implicit in current policy, the MYSFP has the potential to shed new light on the compatibility of current expenditure and revenue policy with sound aggregate policy. This could be of considerable value as a step in the direction of an effective MTEF. 370. The most fundamental challenge in Poland will be to ensure that the MYSFP matters, and that it does not become simply a document which looks impressive but has little impact on decision making either in executive government o r in the parliament. Two issues which are of particular important in determining the outcome are the quality of the forward estimates, and the development of a good budget processes for reconciling expenditure policy with fiscal policy over the MT time horizon. The Quality of the Forward Estimates 371. The success of an MTEF in reconciling expenditure and revenue policy with sound fiscal policy over the M T depends critically on the quality of the forward estimates-that is, on the extent to which they provide an accurate forecast of the levels of aggregate expenditure and revenue which will result over the medium term from current expenditure and revenue policies. Only to the extent that the forward estimates are accurate will they provide advance warning that changes in expenditure policy are needed to meet aggregate fiscal constraints. The MoF acknowledges that there is a need for major improvement in this area, particularly in respect to expenditure estimates (revenue estimates would appear to be considerably better). While there are certain sectors-such as social security benefits-where sophisticated and reliable medium-term expenditure projections are prepared, in general this is not the case. In many sectors, the projection methodology used to date has apparently been the indexation of base year expenditure using expected inflation. 372. Good forward expenditure estimates require, by contrast, a methodology which identifies the key causal factors which will determine the future expenditure levels implied by current *'IMF, Republic of Poland: Staff Report for the 2009 Article IV Consultation, Country Report No. 091266, August 2009 169 policy-what are often known as the "cost drivers". For example, in the school education area, a key factor would be the expected decrease in (and geographical location of) the school age population. The implications of demographic changes on education spending are extensively discussed in the education chapter. In relation to health services, both demographic demand factors and expectations of sector- specific movements in remuneration will be relevantg6. Good forward estimates involve, in short, the construction of a simplified model to project the costs of each key area of expenditure with a reasonable degree of accuracy. The model needs to be agreed between the MoF and relevant spending ministry, rather than simply left to the discretion of the latter. As part of this, steps need to be taken to ensure that all spending ministries use the same assumptions about key macroeconomic variables (e.g. the inflation rate and rate of GDP growth). The development of improved forward estimates will require a major effort in capacity building within the MoF, as well as in the spending ministries themselves. 373. Once the forward estimates have become sufficiently reliable, it will become possible to use them as the basis for the preparation of each year's budget. This would mean that in preparing, say, the 201 2 annual budget, the initial envelope given to each ministry would be based on the estimate of the costs of its existing expenditure policies provided by the forward estimates. This envelope would then only be changes to the extent that the government decided to make changes in expenditure policy affecting the ministry concerned. The use of forwards estimates in this manner would be a major step forward in ensuring that medium-term budgeting is a matter of substance rather than appearance. Improving the Budget Process 374. Like performance budgeting, good medium-term budgeting requires major improvements in the expenditure prioritization processes during budget preparation. Such improved prioritization processes will be essential to ensure that, when the forward estimates and the aggregate expenditure envelope are brought together and it becomes apparent that changes in expenditure policy are required, the government is able to make the most appropriate decisions about where to make those changes. Good expenditure review processes, as discussed above in relation to performance budgeting, have an important role to play. To support this, change within the MoF will be required. In particular, strengthened policy analysis skills will be needed to enable the MoF to play, in conjunction with the Chancellery, a more active role in advising the minister of finance and the Council of Ministers on expenditure policy. Multi-Annual "Programs" 375. A significant development over recent years, separate from the planned MYSFP, is the increased role of multi-annual "programs" in the budget. These "programs" are included as one of the Attachments (no. 12) to the budget appropriation act. The majority are of a capital nature (a typical example being the multi-annual program is that which allocates funds for identified capital works at the Adam Mickiewicz University in Poznan over the period 2004-201 1). The objective of this arrangement is to give greater certainty of future funding levels for major capital projects and major multi-year EU funded projects (both of a current and capital nature), and thereby improve their execution. Prior to the adoption of this arrangement, the problem in relation to major capital projects was that, with funding provided only on an annual basis, ministries could never assume that they would receive funding to continue work on major projects already initiated. This led to work being slowed down or stopped in the transition period between financial years, until funding for the coming year was settled. In respect to EU funds, a similar problem arose from the fact that, while the EU providing multi-year funding, Polish project co-financing had to be approved on an annual basis. Under the new arrangement, with parliament 86 A majority of public expenditures on health in Poland is executed off-state budget by the National Health Fund, and are financed from health contributions - see more in the heath chapter. 170 voting on multi-year funding, uncertainty about future funding of these multi-annual projects is greatly reduced, even though parliament could in principle alter the multi-year allocations in future years. The expansion of these multi-annual programs is a positive development, even in a context where Poland is not yet ready to consider multi-year appropriations more generally. 376. The question has arisen of the relationship between there multi-year "programs" and the performance budgeting program structure. The solution adopted so far has been to treat each of these multi-year projects as a specific sub-task within the performance budgeting structure. This is not, however, an appropriate solution, as each of these projects is in reality part of a more general task or sub- task. The multi-year projects should therefore be subsumed within those tasks or sub-tasks, perhaps as "actions" with multi-year funding. For example, a project for capital works at a particular university would be more appropriately be treated as part of an overall university education task rather than as a task or sub-task in its own right. Implementation of Medium-Term Budgeting 377. The development of medium-term budgeting in Poland is currently being led by the Public Finance Reform Department, which is the same group which is responsible for the implementation of performance budgeting reforms. This is an appropriate arrangement insofar as there are major synergies between performance budgeting and medium-term budgeting. Some indication of this can already be seen in the fact that the forward expenditure estimates to be included in the M S F P are to be presented programmatically. At the same time, however, it would seem that there may be a need to further strengthen the medium-term budgeting work of the Public Finance Reform Department, which has understandably been dominated principally by the performance budgeting effort. As indicated above, it is in the expenditure prioritization process that medium-term budgeting and performance budgeting come together most closely in practice. This suggests that the initiation of a project to create strengthened expenditure review processes (as recommended in the Action Plan below) would help greatly to strengthen coordination of the performance budgeting and medium-term budgeting reforms. 11. Performance Budgeting The Performance Budgeting Framework 378. Poland has been working since 2006 on the implementation of a performance bud eting system, with the encouragement and substantial financial support of the European Union?` The performance budgeting model being implemented is a program budgeting system. This is a system in which programs - categories of spending based on intended results (outcomes and outputs) - are intended to be a central focus of budgeting and management. The Polish authorities view allocative efficiency - improved expenditure prioritization - as the main objective of the new systemg8, although they also anticipate productive efficiency gains. Improved expenditure prioritization is important to Poland in dealing with the fiscal challenges which it faces arising from the current crisis as well as from longer-term pressures. If expenditure can be prioritized effectively so as to target necessary expenditure reductions at the lowest priority areas, fiscal consolidation becomes more effective and sustainable. In the longer term, 87 For an overview and critical analysis of these reforms, see E. Korczyc, "Poland", in L.Kasek (ed), Performance- based Budgeting and Medium Term Expenditure Frameworks in Emerging Europe, World Bank (forthcoming). As the 2009 Polish program budget document presented to Parliament puts it: "An analysis of the performance budget is to indicate the expenses that do not fulfil the purposefulness and effectiveness requirements, i.e. they do not really contribute to the achievement of assumed objectives and the performance of specific tasks and subtasks, and therefore should obviously become state budget economies. The economies that need to be used for achieving real effects in the areas of the most important needs of the state and society." 171 performance budgeting can, if combined with a broader set of results-oriented public management reforms, also help deliver sustained gains in the quality of public expenditure. 379. The Polish program budgeting framework is based on a four-level hierarchy the top level of which is "functions", each of which is broken into multiple "tasks", which a r e in turn comprised of multiple "sub-tasks", which are then (or will be from the 2010 on) decomposed into "actions". Functions (of which there are 22) represent high-level groups of programs covering areas such as health (function 20) and labor market (function 14).89They are similar to "missions" in the French performance budgeting framework. Being highly aggregated, they are not intended to serve as the primary focus of expenditure prioritization decisions. It is at the level of tasks (of which there are over one hundred) that the government would make key decision about expenditure reallocations. Tasks can therefore be considered the equivalent of "programs" in the sense in which the word is usually used internationally. Sub-tasks (of which there are hundreds) and actions represent a detailed break-down of expenditure more relevant to ministry-level decision-making. Implementation Progress and Plans 380. The first major step in the performance budgeting reforms has been the development of an indicative program budget-that is, a presentation of the budget in programmatic terms which is provided to parliament for information alongside a traditional budget law which appropriates money on the basis of a combination of economic classification, organizational unit and projects. The indicative program budget presents function, task and sub-task expenditure for the coming budget year and the subsequent three years, as well as stating task and sub-task objectives and identifying key performance indicators. Articles 124 and 158 of the Public Finance Act now make this presentation mandatory. The first indicative program budget, covering selected pilot agencies, was presented to parliament with the 2008 budget. The version presented with the 2009 budget extended coverage considerably to cover all ministries. It is hoped to extend the coverage of the indicative budget in the 2010 program budget to cover most of the government sector of national government, with the support of amendments to the Public Finance Act currently before the parliament which will require agencies and "targeted funds" (such as the social insurance fund) to present their budgets in program terms. Although considerable work will be required to refine program structure, objectives and indicators, the development of these indicative program budgets covering the whole of government is a considerable achievement. 38 1, The government's implementation schedule identifies the 2013 budget as the tentative point for the crucial transition from a purely indicative program budget to program-based budget appropriations. The intention, however, is to make a judgment about this closer to the event, in the light of progress made in building the foundations for this move. Program Structure 382. Some further work is required to improve the program classification. Certain expenditure appears, for example, to be classified to the wrong function, task or sub-task. For example, grants from the national government to local governments for school education are not included in the education function (Function 3), but in another function. Certain multi-year "programs"-see the discussion below under medium-term budgeting-have been grafted somewhat artificially into the program structure as specific sub-tasks. 89 More precisely, "organization of healthcare and health policy". 172 383. A major weakness of the current program structure is the presence of too many tasks and sub-tasks which are based on support activities rather than outcomes and outputs. For example, the Ministry of Interior and Administration has a task devoted to the sanitary inspection of its various facilities. In a more results-oriented program structure, this sanitary inspection would be seen as a support activity and its costs would be attributed to other tasks (such as border guards and the police service) defined in terms of outputs delivered to the public. Moreover, most overhead costs (e.g. the costs of services such as ministry human resources and financial management, and IT support) of Polish ministries are allocated at present to Function 22-Policy Development and Coordination. This means, for example, that the health function excludes a significant element of the costs of running the health ministry. 384. The reason this,has happened is the inability of the accounting system at present to allocate indirect costs between multiple tasks or functions. Under these circumstances, the approach taken to date has some justification. Although second-best, the creation of specific overhead cost programs is a reasonable practical compromise in the early days of implementing a program budgeting system, when the capacity of the accounting system to allocate indirect costs is yet to be developedg0.However, it does underline the importance of accounting reform to support performance budgeting, a topic further discussed below. 385. An unusual design feature of the Polish system is that tusks cross ministry/agency boundaries. Such an approach may substantially complicate effective budget management. For example, task 13.1 (Social Support and Integration) incorporates sub-tasks which are the responsibilities respectively of the Ministry of Finance (MoF), the Ministry of Labor and Social Policy, the Ministry of Interior and Administration, and the Polish Academy of Sciences. Internationally, most program budgetin systems avoid multi-ministry programs, so that each program is the responsibility of only one ministry!? The Polish approach creates potential issues in moving to program-based appropriations. This arises from the need in any effective budget appropriation system for there to be a clear statement of the total budget funding being provided to each ministry. If each program is specific to a ministry, appropriating funds by program also means clear appropriations to each ministry. In Poland?s case, however, to appropriate money at the level of tasks would leave unclear how the task funding was to be divided between the ministries sharing the task, creating a risk of confusion and conflict during budget execution. Only if parliament were to appropriate funds at the sub-task level would the budget law give a clear statement of each ministry?s budget allocations. However, sub-tasks represent a very disaggregated level of expenditure, and generally speaking, parliaments and governments leave the allocation of resources at this detailed level to the discretion of individual ministries, focusing themselves only at the more aggregated strategic level. Appropriation at the sub-task level might also create an inappropriate degree of rigidity in resource allocation during budget execution. 386. This problem could only be completed resolved with a redesign of the program structure. Even without such a redesign, the scale of the problem would be greatly reduced if the number of tasks and sub-tasks based on support activities, as discussed above were reduced. To take the example referred to above, the classification of sanitary and similar support services of the Ministry of Interior and Administration as part of the cost of delivering security services to the public would remove them from the health function and transfer them to the security function (function 11). 90 See M. Robinson ?Cost Information?, in M. Robinson (ed.), Performance Budgeting: Linking Funding and Results, PalgraveiIMF, 2007, pp. 49-50. 9? In France, missions cross ministry boundaries, but programs do not. 173 Program Objectives 387. In order to realize the performance budgeting objective of results-oriented budgeting, it is necessary that task and sub-task objectives, and their associated key performance indicators, should be results-oriented-and, more particularly, should be defined as far as possible in terms of outcomes. A considerable amount of further work is required to achieve this in Poland. The 2009 performance budgeting certainly contains many examples of good practice in the outcome-oriented definition of taskhub-task objectives. At the same time, however, there are many cases where objectives have been defined in terms of the provision of a service (output), without any reference to the outcome which that service aims at delivering.'* In other cases, taskhub-task objectives have been formulated in terms of bureaucratic activities, in a confusion of ends and meansSg3 388. As part of this review, a detailed examination of the program structures, objectives and performance indicators of the key social policy functions-Functions 3 (education), 13 (social insurance and support from families), 14 (labor market) and 20 (healthbwas undertaken. The findings may be found in an appendix to this chapter, and provide many examples of tasks and sub-tasks the objectives of which could be redefined to focus them on results. 389. The MoF is well aware of the need to improve the specification of objectives, and is giving a high priority during 2009 to working with line ministries to achieve this. This is one of the key roles of the new position of national performance budgeting coordinator. The production of guidelines and a supporting manual to assist spending ministries would also be useful (see below). Performance Information 390. Good, results-oriented performance information is crucial to the success of performance budgeting, This means, firstly, good performance indicators, particularly focused on outcomes, output quality and quantity. Poland has a long way to go in this respect. There are at present few outcome or output quality indicators. As can be seen from the Appendix, there are, for example, no measures of student learning outcomes in Function 3, no measures based`on poverty rates in Function 13, and almost no measures based on rates of mortality and morbidity or incidence of disease in Function 20. The lack of an outcome focus in respect to performance indicators is to some extent understandable. Outcome and output quality measurement is often not easy, and takes considerable time to develop. It is therefore quite reasonable in the early stages of building a performance budgeting system to have a preponderance of output quantity and activity measures, and gradually move more into outcome and quality measurement. Nevertheless, there is significant scope even in the short term for Poland to increase the number of outcome indicators used in its performance budgeting framework, drawing on outcome information which is already publicly available. And in the medium-term, an effort needs to be made to considerably increase the number of outcome indicators. 39 1. Good performance information also requires the development of program evaluation, because indicators alone are not generally sufficient to assess how well a program is performing and whether, if it is performing badly, the program can be addressed by redesign and management changes, or whether the program should be cut. 92 A typical example is the specification of the objective of the emergency medical services sub-task (20.1.3) as "provision of medical assistance to persons in an emergency", when a more outcome-oriented formulation might have been "reducing death and illness arising from medical emergencies". 93 As exemplified by the Agricultural Social Insurance Fund's formulation its task 13.3 objective as the "implementation of tasks in accordance with the act on social insurance of farmers" (an outcome-focused objective would in this case be something like "alleviating poverty in the farming community"). 174 392. The MoF has certain plans for improving performance information. External advice has been commissioned on an improved set of performance indicators for five key Functions. The Ministry?s implementation timetable also indicates an intention to start work on methodological guidelines for program evaluation in 2013. Nevertheless, more needs to be done, and more quickly, to guarantee the success of performance budgeting and of results-oriented management more generally. Each ministry and agency should be required to make a major effort to build its performance measurement systems, including additional data collection, the development of an IT platform for performance measures, and indicator verification. As part of indicator verification, external quality assurance processes need to be established (a function in relation to which the supreme audit institution should play an important part). The preparation of guidelines and a supporting manual to assist spending ministries would, once again, be valuable. 393. The MoF must also build its own capacity in performance measurement and evaluation, because in a performance budgeting system neither the selection of key performance indicators nor program evaluation can be left to the exclusive discretion of each spending ministry. Since a core aim is to provide useful performance information to central decision-makers, including ministers and the parliament, the MoF must be demanding in relation to the information it receives from spending ministries. At present, the Public Finance Act gives spending ministries exclusive power to select performance indicators, and MoF is restricted to persuasion in its efforts to encourage the development of better indicators. The legal provisions should be changed to require that the key program performance indicators be subject to agreement between spending ministries and MoF. Cost Information 394. The other key information requirement is program costing. It is essential that government and the public know what programs cost as well as what results they deliver. Moreover, when the budget is appropriated on a program basis, it will be essential to monitor the execution of the budget by program continuously during the financial year, and to do this all expenditure will need to be recorded in the accounting system on a program basis at the time it is undertaken. 395. To achieve this, a major government-wide effort will be required to modify the chart of accounts, develop new accounting procedures and train staff for the job. Major change in the financial management information system (FMIS) will also be required. Polish ministries and agencies operate a diverse range of financial management information systems, information from which is transferred and consolidated centrally for budget execution monitoring purposes by MoF?s Trezor system. These systems are designed around the need to record expenditure by ?paragraph?-roughly speaking, by economic classification (salaries, capital expenditure e t c t a n d organizational unit. They are not at present configured to the task of recording expenditure by sub-task and task, which requires the allocation of, say, salaries to the relevant task and sub-task. In those ministries which have a modern FMIS, modifications to permit accounting for programs will not generally be difficult. In a significant number of organizations, however, there will be a need for replacement of unsuitable older systems to achieve this functionality. As mentioned above, one of the greatest challenges will be the development of a capacity to handle overhead costs, where the costs of support services such as IT will eventually need to be appropriately allocated between the results-oriented tasks to which they support. Dealing with overhead costs appropriately requires more than the right financial management information technology. Considerable work to define the appropriate allocation principles (?cost drivers?) will also be needed. 396. The MoF is broadly aware of this challenge. It has commissioned work to assess the modification to accounting systems required to support performance budgeting, and plans in conjunction with this to prepare a statement of required functionality. Because this work is absolutely critical to the full implementation of performance budgeting by the tentative target date of 2013, it needs to be pursued aggressively. It would be necessary to have new/modified financial management information systems 175 fully operational in ministries at least one full year prior to the financial year in which the move to full program appropriations takes place-which means two year from now. The implementation of a satisfactory FMIS alone could be expected to take at least two years, starting from a point where the required new functionality has already been clearly specified. 397. The transformation of the FMIS is a task beyond the limited and overstretched resources of the MoF performance budgeting team. Moreover, if the FMIS is to be overhauled, this should be done addressing the full range of current and potential user requirements, which go beyond the implementation of performance budgeting. It is therefore highly desirable that a distinct FMIS reform project be established, with its own implementation team, but working in close coordination with the performance budgeting team. Such a coordinated approach could raise the quality and consistency of forward estimates of public expenditures through their common use of policy-based expenditure programs and program- and activity-based costing methodologies. 398. It is also important to recognize that the specification of required functionality for the financial management information system depends upon decisions about the new budget classification-that is, about how expenditure will be appropriated by parliament when the performance budgeting system is fully operational. It will therefore be necessary to make those decisions now rather than to defer them for several years. This is discussed in the next section. The Appropriation Framework 399. Whether the move to program-based budget appropriations comes in 2013 o r later, there will need to be a prior decision as to what will then happen to the existing budget appropriation framework. The current appropriation structure-which is based on "parts", "chapters" and "sub- chapters" (mainly organizational units, but also some projects), as well as on a type of economic classification ("paragraphs")-has limited relationship to the program structure. The current tentative view of the MoF is that, when program-based appropriations are introduced, it will probably be in "parallel" with the present appropriations. Both a traditional budget and a performance budget would, in other words, be passed by the parliament. Whether this "parallel" appropriation system would be temporary or ongoing has not yet been decided. 400. A dual appropriation system would, however, be unduly complex, and would raise the risk of the program budgeting being treated as subordinate to the traditional budgeting and, consequently, not being taken sufficiently serious. It would be preferable for the traditional budget to be replaced by a single, new program-based program appropriation structure which would include, as well as the program hierarchy, certain of economic classification appropriation categories (notably, for capital, current and wages and salaries expenditure). The chapter and sub-chapter elements of the present appropriation system would, however, be dispensed with. 401. An excellent model of good practice on program appropriations to which Poland might have regard is that adopted by France in 2006. One reason why the French appropriation model is particularly relevant is that (as mentioned above) it includes overarching "function"-style groups of .~~ programs (called " m i ~ s i o n s " ) The French model also provides a good example of how to integrate a limited number of crucial economic classification "control totals" in a program based appropriation system. 94This system is described in detail in B. Lannaud, "Performance in the new French budget system", in M. Robinson (ed.), Performance Budgeting: Linking Funding and Results, PalgraveiIMF,2007. 176 402. Poland has one considerable advantage in this respect over most other countries which have moved from a traditional to program budget framework-that it does not at present have an unduly detailed set of input appropriation categories which would need to be simplified. Spending ministries are essentially free to shift money between current expenditure categories (subject to a reasonable requirement not to increase the total remuneration bill), and may even reallocate between capital and current with MoF approval. This flexibility is precisely what is required to "let the managers manage" under a performance budgeting system. Budget Process Change 403. The focus of the reform effort to date has been technical-the development of the program classification and the supporting performance information. The ultimate success of performance budgeting will depend, however, on the eventual use of the performance framework to make budget preparation more performance-informed. This requires that budget decision-makers at all levels- Council of Ministers, Finance Minister, MoF and spending ministries-make tangible use of the program structure and performance information when they decide how to allocate scarce taxpayer resources between competing programs. 404. International experience suggests that such change is not easy to bring about, and that all too often the introduction of performance budgeting systems brings about little change in budget decision-making. To ensure the success of the system, it is crucial to introduce systematic central spending review processes which are designed in such a way as to facilitate the use of performance information. Performance-informed spending review needs to cover two distinct areas: (1) review of all new spending initiatives; (1) continuous review of ongoing "base" expenditure. Poland has limited track record in both of these areas. There is still no process for review of base expenditure, and this is a particular weakness which will need to be remedied if performance budgeting is to work. In respect to the review of new spending proposals, the Chancellery of the Prime Minister has over the past two years developed a system which requires spending ministries to report and provide information about the costs and benefits of proposed new initiatives (called also "tasks"). These tasks are then subject to analysis by the Chancellery, under the leadership of a powerful state secretary (minister) who has responsibility for government-wide prioritization. It is significant that this Chancellery system has been developed independently of the performance budgeting initiative, as a reflection of which the "tasks" upon which this system are based have no relation to the "tasks" in the program structure of the performance budgeting system. 405. Spending review is closely related to medium term budgeting, and is a key area in which performance budgeting and medium-term budgeting intersect. The matter is therefore discussed in the section about medium-term budgeting. 406. Another valuable step which will help to ensure that performance information is used in the budget process will be to require spending ministries to justify their budget proposals with performance information and it a programmatic format. Institutional Arrangements for Implementation 407. Within the MoF, the development of the performance budgeting system is the responsibility of the Public Finance Reform Department, rather than of the department responsible for budget preparation (the Budget Department). This is a reasonable arrangement during the reform 177 implementation period.g5 However, even with such an approach, it remains essential that the Budget Department interest itself closely in the design of the reforms, because the success of performance budgeting will depend in substantial measure on the Budget Department using the program framework and associated performance information to prepare the budget. It is, for example, highly desirable that the Budget Department officials who are responsible for certain functional areas (such as education) should be involved in the process of negotiating appropriate key performance indicators for the function. It is not clear, however, that the Budget Department has to date been sufficiently involved in the reform process. Such involvement could be promoted by the creation of an internal steering group for the reforms, but also depends on close informal cooperation between the Budget Department and the Public Finance Reform Department. 408. Institutional arrangements for implementation of the reforms could also be enhanced in spending ministries, where in some cases the reform has been treated as a matter primarily for the ministry budget ("economic") department, with limited involvement on the part of the service delivery and planning departments. The MoF is well aware of this problem, and is actively seeking to broaden participation in the reform process within the spending ministries. 409. It is of relevance that a major institutional player in expenditure prioritization in Poland- the Chancellery of the Prime Minister-has at present little involvement in the development of the performance budgeting system and appears to make no use of it. This is notwithstanding the fact that until early 2008 it was a department within the Chancellery which was responsible for the development of the performance budgeting system. In the Polish system, the Prime Minister plays a particularly powerful role in budgetary matters, and the minister and ministry of finance do not have the same degree of power that it tends to have in, say, Anglophone parliamentary systems. The role of the Chancellery is particularly marked at the present time, when a state secretary (a ministerial position) of the Chancellery has an explicit government-wide prioritization mandate, as part of which he heads a committee of state secretaries which as part of its duties considers the draft state budget each year. The mechanism for the review of new spending proposals, referred to above, is her/his responsibility. 4 10. If the performance budgeting system is to work effectively, it will need to be seen by all of the key players in expenditure prioritization as offering the information and the analytic framework most useful to the task. It would help to ensure this if these key players were involved properly in the development of the system. MoF has recently taken the important step of creating a formal consultation forum bringing together state secretaries of the spending ministries and chaired by the national performance budgeting coordinator (and which is supported by a technical level subcommittee). But more direct consultation and coordination with the Chancellery would be useful. Substantial engagement by the Chancellery in the development of the performance budgeting system will be helped if the Chancellery can be persuaded that the performance information which the system will yield will be of major benefit to it in its role of bringing government priorities to bear more effectively on spending ministry activity . Support for Spending Ministries 41 1. MoF coordination and leadership is critical to support the implementation of the performance budgeting framework at the spending ministry level. The MoF is well aware of this, and has commissioned a large-scale training program to build capacity within ministries. 95 It is, for example, similar to the approach taken in France during the run-up to the full implementation of its performance budgeting system. 178 4 12. What has been missing to date has, however, been guidelines (regulations) and supporting manuals to assist ministries in implementing key steps including the formulation of objectives and the selection o performance indicators. In many countries (e.g. New Zealand or France) where PB reforms f have been successfully implemented, the lead agency for reform has issued methodological guidelines that are designed to provide general assistance across all ministries. There are currently no such methodological guidelines available to ministries in Poland, and this is something which spending ministries have indicated they would like to see made available. Two specific areas were such guidance would be particularly valuable are mentioned above: performance indicators and the specification of objectives. These guidelines and supporting manuals, together with PB training resources for ministries, could usefully be made available on the Ministry of Finance website. 111. Future Directions for Medium-Term Budgeting and Performance Budgeting Training and Capacity Building Needs 413. The training and capacity building requirements for the successful implementation of the new performance budgeting system in Poland a r e extensive, and a r e focused in the following areas: Results-oriented program classification and objective specification, Performance indicator development, Program evaluation methodology, Development of a new budget classification and chart of accounts, Transformation of the financial information technology system. 414. A large-scale training effort is currently underway, supported by substantial EU funding. To date, this training has focused principally on the first two of the above areas. 415. In respect to the other three areas, Poland could benefit not only from an appropriate training program, but also from targeted capacity-building focused on developing the skills of the core teams which will manage change in the areas concerned. 416. In respect to medium-term budgeting, the most immediate capacity building requirement would appear to be in respect to the improvement of forward estimates, particularly on the expenditure side. IV. Action Plan 417. The following is an action plan identifying key steps required to meet the goal of program- based budget appropriations in financial year 2013, as well as to ensure that the new Multi-Year State Financial Plan also to ensure that budgeting becomes increasingly performance-informed. 179 Y s \ \ \ \. \ h 2 M C 3 E .- S U 0 L3 a B L Lo CII b P EL a v) x 7 P L * v1 8 U a m U Y E k 5 t t: E a E C a s E - . a Y > E a c a C (F z V $ c 2 a c E a ? 2 E L c E G 5 v 6 w 0 3 c Qi C 6 e- E x I t Ef u i t E a E sa S E a i i a i & c E u : a t e c ; 0 i e E .- 0 m U 1 .- > - . c' m V a C E 0 .I vl 2 G P a 5 ? r E i 00 00 rl BIBLIOGRAPHY Betcherman, Gordon, Karina Olivas and Amit Dar (2004), "Impact of Active Labor Market Programs: New Evidence from Evaluations with Particular Attention to Developing and Transition Countries, Social Protection Discussion Paper Series, 29 142, World Bank, Washington, DC. Boone, Jan and Jan C. van Ours (2004), "Effective Active Labor Market Policies", IZA Discussion Paper No. 1335. Bukowski, Maciej ed. (2008), Zatrudnienie w Polsce 2007. Bezpieczenstwo na elastycznym rynku pracy, Warszawa, Ministerstwo Pracy i Polityki Spolecznej. Carone, Giuseppe, Gert Jan Koopman and Karl Pichelmann (2009), Labour market prospects and policies to soften the impact of the financial crisis, European Commision, ECFIN Economic Brief, Issue 1, May. Available at http://ec.europa.eu/economy finance/publications. Central Statistical Office (GUS) (2008). "Szkoly wyzsze i ich finanse w 2007 r." Dar, Amit and Zafiris Tsannatos (1999), "Active Labor Market Programs: A Review of the Evidence from Evaluations", Social Protection Discussion Paper Series, 990 1, World Bank, Washington, DC. EIB and the World Bank (2004). "Tertiary Education in Poland," July 2004. Eurydice (2007). "National summary sheets on education systems in Europe and ongoing reforms: Poland." FISE (2009), Jak w dobie kryzysu usprawnic funkcjonowanie rynku uslug zatrudnieniowych, Warszawa. (How to improve employment services in a time of crisis). Herbst, M. (2004). Human Capitasl Formation in Poland: Where does quality come from?" DISES Working Paper #99. University of Trieste. Italy. Herbst, M. (2008). "Bottlenecks in the decentralization of education funding in Poland," EUROREG. Herbst, M., J. Herczynski, and A. Levitas (2009). ,,Finansowanie oiwiaty w P o k e - diagnoza, dylematy, mozliwoici," EUROREG. Ivory, Thomas and Cristina Thomas (2007), "A Review of Best Practices Regarding Performance Based Financing of Regional Employment Offices in OECD Countries", World Bank, Washington DC, photocopy. Jakubowski, M. and I. Topinska (2006). "Impact of Decentralization on Public Service Delivery and Equity," Centre for Social and Economic Research, Warsaw, 2006. Kluve, Jochen (2006), "The Effectiveness of European Active Labor Market Policy", IZA Discussion Paper No. 201 8. OECD (2006). "Assessing Scientific, Reading and Mathematical Literacy: A Framework for PISA 2006." Paris. OECD (2007). "PISA 2006: Science Competencies for Tomorrow's World." Paris. 191 OECD (2007). Education at a Glance. Paris OECD (2008). Education at a Glance. Paris Paciorek, J. WiSniewski, P. Szyrmer (2007). ,,Co wie i potrafi 15-letni uczen - sprawdzita OECD w badaniach PISA," CASE, 2007. Rutkowski, Jan (2004), "Labor Market Interventions during the Transition in ECA", Spectrum, Summer, The World Bank, Washington, DC. Rutkowski, Jan (2008), Priorytety Rynku Pracy. Synteza. Ministerstwo Pracy i Polityki Spolecznej, Warszawa (Priorities for Labor Market Refroms. Synthesis Report). Strategic Advisors Team (2008). "Report on the Intellectual Capital of Poland." Prime Minister's Office. World Bank (2001), Poland's Labor Market. The Challenge of Job Creation. Washington, DC. World Bank (2003). "Poland: Towards a Fiscal Framework for Growth- A Public Expenditure and Institutional Review." Report 25033-POL. Washington, DC. World Bank (2005), Enhancing Job Opportunities: Eastern Europe and the Former Soviet Union, Washington, DC. World Bank (2007), "Labor Markets in EU8+2: From the Shortage of Jobs to the Shortage of Skilled Workers, World Bank EU8+2 Regular Economic Report, Special Topic, September 2007. World Bank (2007). "Higher Education Financing in the New EU Member States." World Bank Working Paper No. 112. Washington, DC. World Bank (2007). "Vocational Education in the New EU Member States." World Bank Working Paper No. 116. Washington, DC. World Bank (2008), Labor Market Programs that Make a Difference in a Time of Crisis, HDNSP, Washington, DC, processed. World Bank (2008). "Linking Education Policy to Labor Market Outcomes." Directions in Development. Human Development Network. Washington, DC. Wunsch, Conny and Michael Lechner (2007) "What Did All the Money Do? On the General Ineffectiveness of Recent West German Labour Market Programmes" IZA Discussion Paper No. 2800. 192 IBRD 33467R PO L A N D SELECTED CITIES AND TOWNS PROVINCE (WOJEWÓDZTWO) CAPITALS NATIONAL CAPITAL POLAND RIVERS MAIN ROADS MAIN RAILROADS PROVINCE (WOJEWÓDZTWO) BOUNDARIES INTERNATIONAL BOUNDARIES 14°E 16°E 18°E 20°E 22°E 24°E B a l t ic Sea RUSSIAN To Gulf of Gusev FEDERATION Gdansk LITHUANIA Gdynia To Slupsk ´ Gdansk Vilnius To 54°N Stralsund Koszalin POMORSKIE Elblag Suwalki WA R M I N S K O - 54°N MAZURSKIE ZACHODNIOPOMORSKIE Wisla To Neubrandenburg Olsztyn Szczecin GERMANY K U J AW S K O - PODLASKIE BELA RUS To Berlin POMORSKIE , Lomza Pila Bydgoszcz Ostroleca Bialystok Notec Torun Od Gorzów Ciechanów ra Wielkopolski ew Warta ar N Wloclawek Bu g To To Baranavichy , Plock Berlin Poznan Wisla MAZOWIECKIE Konin WARSAW LUBUSKIE War ta Siedlce To Pinsk 52°N 52°N Zielona Góra WIELKOPOLSKIE Biala Skierniewice Podlaska To To Leszno Ny Berlin Kalisz Lódz Kovel' sa s O Sieradz dr LUBELSKIE a a DOLNOSLASKIE DOLN OSL ASKIE LÓDZKIE To Piotrków Radom Lublin To Dresden Trybunalski Kovel' Legnica Jelenia Wroclaw Chelm Góra Kielce ,, Walbrzych OPOLSKIE Czestochowa Zamosc Opole SWIETOKRZYSKIE To SLASKIE Tarnobrzeg S Prague a an tula Vis Katowice To Prague PODKARPACKIE 50°N Kraków 50°N Tarnów Rzeszów To , L'viv C Z E C H R EPU B LI C LIC MALOPOLSKIE Przemysl Bielsko- Krosno Biala Nowy Sacz To Brno 0 25 50 75 100 Kilometers To Rysy (2,499 m) To UK R UKR AI N E Zvolen Kosice 0 25 50 75 Miles This map was produced by the Map Design Unit of The World Bank. The boundaries, colors, denominations and any other information SLOVAK REPUBLI C SLO VA K REPUBLIC shown on this map do not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or acceptance of such boundaries. 18°E 20°E 22°E 24°E MARCH 2007