5: HEALTH CARE FINANCING. Issues, Reforms, the Unfinished Agenda

Transcription

1 5: HEALTH CARE FINANCING Issues, Reforms, the Unfinished Agenda John C. Langenbrunner, With contributions from Jan Bultman, Cheryl Cashin, Dominic S. Haazen, Katherine H. Dahlmeier, Dorothee Eckertz, Paul Shaw, and Verdon S. Staines and acknowledgement of Teresa Osicki and Aida Tapalova 146

2

3 I. INTRODUCTION More than 90% of all ECA loan projects in the past decade 40 reflect some funding commitment for health financing reform. Nearly all countries sought the Bank s assistance in improving the health care financing systems in some way, making this area integral with other sector change. More than $152 million, or about 7-8% of the overall funds for HNP projects, were focused on financing reforms. Table 1 suggests lending in absolute terms was highest in Eastern Europe, in countries such as Bulgaria, Poland, Romania, the Baltics and Turkey; it was not as high in FSU countries. The split in funding patterns probably reflects different approaches to reform in the ECA region. In Eastern Europe, an early consensus for reforms nationwide was typically manifested in a lending program for the entire country. In FSU countries, a go slow more cautious approach was found, and pilots or demonstration reforms were typically developed for testing and replication later (Azerbaijan, Kyrgyzstan, Kazakhstan, Armenia, Georgia, Russia). Using Staff Appraisal Reports (SARs) and Project Appraisal Documents (PADs), Table 1 and Table 2 provide a quick summary of lending levels by country and by source and uses of funding looking at government, Bank, and other donors funding, as well as funds targeted to local and foreign inputs. Table 2 shows lending by categories of disbursement civil works, goods and services, technical assistance (TA) and training. Not surprisingly, most funding has gone to goods and services, such as computers for insurance funds, or to purchasers and providers in financing reform programs. Training and TA each constitute about 20% of the overall funding of the component. As in Table 1, the pattern of lending for TA and training in absolute terms was highest in Eastern Europe in countries such as Bulgaria, Poland, Romania, the Baltics and Turkey; it was not as high in FSU countries. In more cases than not, the Bank arrived in ECA countries to find either well-developed plans for financing changes (Poland, Estonia, Kyrgyzstan), or a country in the process of implementing financing reforms (Hungary, Romania, Russia). Changes were based, in part, on the basic features of the Bismarck model found in Western Europe, but significant differences also emerged as the model was adapted to the particular context of the ECA region. The shift resulted in changes in the way money was collected and pooled, creating a new relationship between purchasers of care and providers. Legislative reform was, however, not always matched by concrete change on the ground and the objectives set out in policy were not fully or even partially attained in some cases. Nevertheless, the Bank through both its AAA and lending services did provide a significant input into the transition to new forms of health care financing. 40 A total of 30 Project Appraisal Documents (PADs) and Staff Appraisal Reports (SARs) were found and reviewed. 147

4 Table 1: Health Financing Reform Lending: Summary Table Total Total Local costs by Financiers: Total Foreign costs by Financiers: Total by Financiers ECA Country & Project Title in Projects Local Cost Government* IBRD Others (**)Foreign Cost Government* IBRD Others (**) Government* IBRD Others (**) $US.000 Albania "Health Services Rehabilitation Project" "Health System Recovery and Development Project" Armenia "Health Financing and Primary Health Care Development" , , , , , , , , Azerbaijan "Health Reform Project" Belarus Bosnia and Herzegovina "Essential Hospital Services Project" "Basic Health Project" , , , , Bulgaria "Health Sector Restructuring Project" , , , , , ,447.5 "Health Sector Reform Project" , , , , , , ,438.6 Croatia "Health System Project" , , , ,487.5 Estonia "Health Project" , , , , , , , Georgia "Health Project" , ,064.7 "Primary Health Care Project - Health II" Hungary "Health Services and Management Project" , , , , , , ,631.0 Kazakhstan "Health Reform Program Project" , , , , , ,200.0 Kosovo(***) "Education and Health Project" , , , ,800.0 Kyrgyz Republic "Health Sector Reform Project' "Second Health Sector Reform Project" , , , ,718.0 Latvia 148

5 "Health Reform Project" , , , , , , , ,461.3 Lithuania "Health Project" , , , , , , , , ,885.9 Moldova "Health Investment Fund Project" , , , ,424.0 Macedonia "Health Sector Transition Project" , , , , ,512.6 Poland "Health Services Development Project" , , , , , , , ,832.0 Romania "Health Rehabilitation Project" , , , , , , ,762.0 "Health Sector Reform Project" Russian Federation "Health Reform Pilot Project" , , , , , , , ,952.5 Slovenia "Health Sector Management Project" , , ,563.7 Tajikistan "Primary Health Care Project" Turkey "Second Health Project: Essential Health Services and Management Development" , , , , , , , ,807.7 "Primary Health Care Services Project" MEER Project (Marmara Earthquake Emergency Project) Ukraine Uzbekistan "Health Project" , , , , , , ,678.6 Totals 152, , , , , , , , , , ,297.2 Sources: World Bank, ECA PADs and SARs, * - includes local government ** - others include the following: Country Financier: Armenia Others Bosnia&H Others Bulgaria EU PHARE Estonia Others Kazakhstan USAID Latvia SIDA (Swedish International Development and Cooperation Agency) Lithuania EU PHARE, Government of Sweden, others Moldova Dutch Government Tajikistan Swiss Development Corporation *** TA & Training are not separated 149

6 Table 2: Summary of Lending for Health Financing Reform by Disbursement Category (US$ in 000s) Civil Works Goods Technical Assistance Training ECA Country Total Local Foreign Total Local Foreign Total Local Foreign Total Local Foreign $US.000 Albania "Health Services Rehabilitation Project" "Health System Recovery and Development Project" Armenia "Health Financing and Primary Health Care Development" , , , , , , Azerbaijan "Health Reform Project" Belarus Bosnia and Herzegovina "Essential Hospital Services Project" 1996 "Basic Health Project" , Bulgaria "Health Sector Restructuring Project" , , "Health Sector Reform Project" , , , , ,548.8 Croatia "Health System Project" , , , Estonia "Health Project" , , Georgia "Health Project" , "Primary Health Care Project - Health II" 2001 Hungary "Health Services and Management Project" , , , ,185.0 Kazakhstan "Health Reform Program Project" , , Kosovo "Education and Health Project" , ,422.0 Kyrgyz Republic "Health Sector Reform Project' 1996 "Second Health Sector Reform Project" , ,

7 Latvia "Health Reform Project" , , , , , , Lithuania "Health Project" , , , , , , Moldova "Health Investment Fund Project" , Macedonia "Health Sector Transition Project" , , , Poland "Health Services Development Project" , , , , , , , ,484.0 Romania "Health Rehabilitation Project" , , , , , , ,645.0 "Health Sector Reform Project" 2000 Russian Federation "Health Reform Pilot Project" , , , , , Slovenia "Health Sector Management Project" , Tajikistan "Primary Health Care Project" Turkey "Second Health Project: Essential Health Services and Management Development" , , , , , , , ,323.7 "Primary Health Care Services Project" MEER Project (Marmara Earthquake Emergency Project) Ukraine Uzbekistan "Health Project" , , , Total 5, , , , , , , , , , , ,289.5 Sources: World Bank, ECA PADs and SARs,

8 Definition This financing paper brings together the existing knowledge base, and outlines the gaps in knowledge that need to be addressed. It begins with a review of the old system. The main body of the paper is then organized into three sections: Sources of revenue and revenue collection; 41 Organization, management, and pooling of financial resources; 42 and, Allocation of resources or the purchasing of services. 43 The framework assumes that, for each of these three functions, it is possible to identify related policy issues. These are outlined in Table 3. Decisions on each of these policy issues will shape the overall structure of the health care financing system. For example: The equity of the financing system will depend both on the level and distribution of the contributions; Equity of access will depend on who has access and to what services, as well as, on user charges and informal payments; Efficiency will be influenced largely by the extent of pooling and the methods of allocation of resources, including incentives in the allocation process. Depending on the extent of decentralization and fragmentation in the system, these functions and the associated decisions may be carried out by different bodies. For example, while the central government might decide the contribution rate and the proportion to be paid by the employer and the employee, the collection of the contributions may be the responsibility of regional branches of the health insurance fund (Dixon et al., 2002). Table 3: Policy Issues Related to Different Financing Functions Financing Function Related Policy Issue Collection of Funds How much money to collect and from whom? Pooling of Funds How to pool resources? How to allocate resources to purchasers? Purchasing of Services Who and what to cover? From whom to buy and how to buy? At what price to buy and how to pay? Source: Adapted from Preker et al, 2000 Each section begins with some contextual information, and then moves onto the Bank s work in this area. The latter sections detail the Bank s work and its interaction with policymakers in this area. It draws lessons based on experience and evolving trends in the region. The paper ends with a discussion on training and institutional capacity building for future action. 41 Revenue collection refers to the process of mobilizing resources, usually from households or corporate entities but also external donors. 42 Pooling refers to the spreading of financial risk across the population or a sub-group of the population through the accumulation of prepaid health care revenues. This facilitates solidarity, primarily between the healthy and sick and, depending on the method of funding, between the rich and the poor. 43 Purchasing is the process of obtaining services from providers on behalf of the covered population. Provision of services and how these are delivered and by whom is not within the scope of this cluster paper. 152

9 Methodology All Bank projects undertaken by the HNP group since 1990 were reviewed. All Staff Appraisal Reports (SARs), Project Appraisal Documents (PADs), and Implementation Completion Reports (ICRs) documents for ECA HNP projects from the study period were reviewed for contributions to reforming financing systems in terms of revenue collection, pooling and allocation. Several small papers on specific topic areas were developed: sources of revenues, insurance organizations, management and information systems, benefit packages, and out-of-pocket payments. These papers were utilized as background documents. A literature review was also conducted to complement the paper s scope. Interviews were held with Task Team Leaders to gather more information about experiences in project implementation and management. Several limitations constrain a full analysis. Firstly, donor funding for this component, especially in complement with Bank financing, is generally thought to be key to successful implementation. While donor funding is formally part of the PAD/SAR reporting process, the reporting is very uneven. Many projects with significant donor support for TA and training such as Poland, Kyrgyzstan and the Russia Health Reform Pilot Project do not appear in this documentation. Secondly, this analysis reports expected disbursement at the design and loan approval stage. Countries are not obliged to actually disburse funds for all categories. Good disbursement data are not available through routine reporting systems in the Bank. Thus, only a partial picture of planned funding can be presented, and only an anecdotal picture of actual funding can be presented for the last decade, in this area of effort. Looking Back: The Transition A full discussion can be found in Paper 1, and is briefly summarized here. As early as 1987, the FSU and Eastern bloc countries began testing new organizational and financing models to improve both revenues and efficiency. In the late 1980s and early 1990s, a number of Eastern European countries moved almost immediately to new contractual-based social insurance arrangements (Czech Republic, Hungary) and new resource allocation methods. The New Economic Mechanism (NEM) in the FSU, too, picked a number of geographic demonstration areas, re-organized the polyclinics into family practice groups and initiated fundholding arrangements. The objective was to shift the locus of care to less expensive outpatient and primary services. There were early successes with drops in admissions (Sheiman, 1993; Langenbrunner et al, 1994), and expenditure shifts from approximately 70:30 inpatient to outpatient spending, to levels closer to 50:50 (Schieber, 1993). Samara, a region in Russia, reported closures of 5,500 beds (Meditsinskaya Gazete, 26, June 1996). The results and primary issues arising from these and other changes will be discussed and analyzed to determine the Bank s most effective role in each ECA country s health care system. 153

10 II. ISSUES Sources of Financing and Revenue Collection Worldwide, the main sources of revenue for health care are taxes, social insurance contributions, voluntary insurance premia and user charges (formal and informal). Most countries rely on a mix of these sources. Taxes are compulsory for the whole population and are levied by government. Social insurance contributions are compulsory for all or some of the population; they are kept separate from other government revenues and are usually managed by a fund or funds independent of government. 44 With the economic downturn and dislocation in the late 1980s and early 1990s, all countries in ECA had to cut real public spending for health in the first years of the transition, and did so roughly in proportion with GDP decline (Belli, 2000). The majority of these countries embarked on a search for a more diverse revenue base specifically earmarked for the health sector. This expanded revenue base has typically included: New revenues through small patient co-payments, especially for outpatient pharmaceuticals; Separate employer-based payroll contributions that de-linked revenues from the budgetary process; Private contracting with enterprises for selected services; and A private supplemental insurance sector. Preker et al. (2002) cite several reasons why many ECA countries shifted to payroll taxes, direct user charges and other sources of earmarked funding: To increase or at least stabilize funding streams for health services. Providers, especially physicians, saw it as an instrument to increase salaries; To break the monopoly of government, particularly the Ministry of Finance, over the ownership and financing of health services; To increase the responsibility of individuals for their own health and the financing of health care; To improve efficiency by making health care providers more accountable for the use of resources (Chinitz et al 1998); To give responsibility for health care to organizations independent from government (mainly driven by ideological concerns about the role of the state). 44 In terms of equity, direct taxes (i.e. levied on individuals, households or firms) are usually set progressively, that is the higher the income the higher the proportion paid. In contrast indirect taxes (i.e., levied on goods and services) are regressive because those on lower incomes spend a greater proportion of their income on consumption. Social insurance contributions are usually levied proportionately to income. Where an income ceiling is applied, above which income is exempt from contributions, social health insurance becomes mildly regressive. Furthermore, because contributions are levied only on earned income (not on profits or income from investments and savings) they place a heavier burden on those with lower incomes. In contrast private health insurance and user charges are higher for those in greatest need, thus relating how much you pay to how ill you are (or are likely to be). In terms of efficiency, taxation is associated with strong expenditure control; it draws on a broad revenue base and is administratively efficient. Depending on the organization of social insurance, expenditure control might be strong if there is a single fund or government caps the overall budget or sets contribution rates. Social insurance draws only on earned income and therefore adds to the cost of labor with a potentially negative effect on economic growth. If separate systems of collection are implemented this will add to administrative costs. 154

11 Table 4: Advantages and Disadvantages of Different Methods of Revenue Enhancement Method of Revenue Advantages Disadvantages Collection Direct Taxation Indirect Taxation (e.g., VAT taxes) Social Health insurance Voluntary Insurance Health Wide revenue base (all income) Administratively simple Usually progressive and promotes solidarity Large risk pool Allows flexibility in spending across sectors Universal coverage Visible source of revenue (all transactions) Administratively simple Compliance easy Earmarked for health Separate from other government revenues (may) Link contribution to benefit Low resistance to increases Independent management of funds May allow choice of insurer May allow choice of insurer May relate payment to utilisation. Compliance may be difficult Allocations subject to political negotiation Potential tax distortions Potential tax distortions Allocations rely on consumption levels Usually regressive Compliance difficult Increases costs of labor and may reduce international competitiveness Revenue follows economic cycle Strong regulatory framework Narrow revenue base (only applies to earned income) Strong regulatory framework needed Adverse selection (results in escalating premium) Risk selection (leaves some uninsured) Access related to insurance cover Usually regressive User Charges Relates payment to utilisation May deter access to necessary services Access related to ability to pay Regressive Limited pooling of funds Source: Adapted from Dixon, et al The payroll tax has emerged as a standard part of a diversified source of health care financing in the region. Of the ECA region s countries, 14 have introduced payroll taxes 9 as a predominant mechanism of financing and 5 as a complementary resource to general tax revenues and out-of-pocket payments (Preker et al., 2002). Contribution rates range from 2% in Kyrgyzstan to 18% in Croatia (see Table 5). The contribution rate tended to be higher in Eastern Europe, and where it continues to be the predominant form of financing. In former Soviet states, the tax rate was typically set at less than 5%. In one country, Latvia, the tax was set as a proportion of the personal income tax. In most countries, the contribution rate was not synchronized with a detailed actuarial analysis of expected costs and revenues for the insured population (Ensor 1999). Indeed, there were no actuaries in the ECA region in the early 1990s, and many public insurance organizations such as in Croatia still do not have actuaries on staff today. 155

12 Table 5: ECA Countries: Adoption of New Earmarked Tax for Health: Payroll Tax and Characteristics of Health Insurance Contribution Revenues Payroll Tax Rate, 1999 Year Introduced Albania 1995 Public: 3.4% (1.7:1.7) Private: 3 5% Croatia % (18:0) Czech Republic % (9:3.5) Estonia % (13:0) Hungary % (11:3) plus hypothecated tax of US$170 per employed person Georgia % (3:1) Salaried Employer:Employee Self-employed Non-Active Population 7% of statutory minimum wage Central budget 18% of declared 18% of gross pension and other benefits income plus central budget 13.5% of 35% of Central budget transfer 13.5% of 80% of net pretax income statutory minimum wage 13% of declared Central budget transfer income 14% of declared Central budget. Per capita amount of income but at least transfer is unspecified the minimum wage plus hypothecated tax of US$170 per person 4% income tax Central budget, but amount unspecified Kazakhstan % 3% of declared Per capita oblast contribution for (3:0) income nonworking Kyrgyzstan % 2% of declared Oblast contribution of undetermined level (2:0) income Latvia % of personal income 28.4% of personal General budget transfer tax income tax Macedonia % 12% of declared Central budget, but accounts for only 10% (2:0) income of revenues Poland % 7.5% of declared 7.5% of gross benefits income Romania % 7% of declared 7% income tax based on gross benefits (7:7) income Russia % 3.6% of declared Central budget. Per capita amount of (3.6:0) income transfer is unspecified. Slovakia % 13.7% of declared Central budget. Per capita amount of (10:3.7) income transfer specified as the contribution rate applied to 73% of the statutory minimum wage. Slovenia % 13.25% of Central budget. declared income Azerbaijan No change Belarus Moldova Ukraine Turkmenistan Turkey Tajikistan Uzbekistan No change No change No change No change No change No change No change 156

13 Instead, the rate-setting process reflected a combination of optimistic eye-balling of desired revenues and guesses about the political acceptability of adding to the already heavy tax burden on employers and employees. This has resulted in frequent changes in the setting of rates, or in funds being depleted for periods of time. These changes in sources of funding were typically linked with organizational changes such as the enactment of separate, self-sustaining social (public) health insurance trust funds. These funding and organizational changes have been perceived as catalysts for further system reforms such as changes in purchasing arrangements, provider payment policy changes, and organizational changes (see below) to improve efficiency. Despite the switch to social insurance contributions, general tax revenues continued to play a significant role in health care funding in many countries. Voluntary health insurance was intended to develop as a supplementary source of revenue. However, the market in private health insurance remains small in most countries and does not contribute significantly to health care expenditure. Private funding, in the form of co-payments and informal payments for health services within the public health care sector, is much more significant. However, the level and scope of these payments varies significantly across countries (Lewis 2002). Figure 1 shows the relative importance of taxation and social health insurance in ECA countries towards the end of the 1990s, for countries moving to some mix. 45 The distance from the diagonal represents the share of private funding which is mostly out-of-pocket. In the region, there were: Seven countries that funded health care predominantly from taxation: Albania, Kazakhstan, Latvia, Moldova, Poland, Romania and Russia; Six countries relied predominantly on social insurance contributions: Czech Republic, Croatia, Estonia, Hungary, Slovakia and Slovenia; In four countries -- Armenia, Azerbaijan, Georgia and Kyrgyzstan -- forms of pre-payment almost totally collapsed and health care was predominantly funded by out-of-pocket payments. World Bank and other donor analyses since the late 1990s (see, for example, Kutzin, et al, 2000, Feeley et al, 2000, and Langenbrunner et al., 2002) show that a significant number of additional ECA countries witnessed a partial collapse of pre-payment, and health care is significantly funded by out-of-pocket payments. In Moldova, Macedonia, Uzbekistan, Ukraine, Belarus and Russia, out-of-pocket payment now account for more than 40% of total expenditure on health, and is often the majority form of expenditure (e.g., Russia, Moldova, Uzbekistan). 45 These data are likely to have changed. For example since 1998 Poland has implemented a 7.5% social health insurance contribution. In Kyrgyzstan, more funding is through public insurance and general taxation. 157

14 Figure 1:Percentage of Total Expenditure on Health from Taxation, Social health insurance and other Sources (includes voluntary health insurance and out-of-pocket payments), 1997 (or latest available year) 100 % of total health expenditure from social health insurance SL CZ SK EST HU CR RU AL ROM KAZ GE AZ KY MO PO LAT % of total health expenditure from taxation Key: AL: Albania; AZ: Azerbaijan; CR: Croatia; CZ: the Czech Republic; ES: Estonia; GE: Georgia; HU: Hungary; KAZ: Kazakhstan; KY: Kyrgyzstan; LAT: Latvia; MO: Republic of Moldova; PO: Poland; ROM: Romania; RU: Russia; SK: Slovakia; SL: Slovenia. Note: The distance from the diagonal line represents the proportion of total expenditure from private sources (out-of-pocket expenditure and voluntary health insurance). Source: Preker et al, 2002 Results of Diversified Financing Sources Failure to Increase Overall Levels of Funding Despite the hope of increasing overall levels of funding, the diversified base often failed to bring about expected revenues. In addition to compliance issues mentioned above, Ministries of Finance often slashed contributions from general revenues at a pace roughly equal to new payroll taxes (see Figure 2). This often happened in FSU countries where revenue payroll taxes accounted for less than half of revenues. Another prevailing factor was the overall weak macroeconomic context. Figures 3 and 4 show per capita GDP for selected countries from the region and the change in GDP over the period respectively. The countries have been clustered into three groups by color, groups A, B, and C. There is a high correlation between those countries (C) with low per capita GDP and negative economic growth, and a corresponding high reliance on out of pocket expenditure (Figure 1 above). Except Poland, all countries in second grouping (B) have experienced negative growth. These countries, despite introducing social health insurance, continue to rely on general taxation as the main source of funding for health care. Finally, the countries (A) that have gone to rely the most on earmarked contributions (accounting for more than 60% of total expenditure on health) are also those with the highest levels of per capita GDP. 158

15 Figure 2: Trends in Per Capita Government Health Expenditures (MHI + Budget) in Kazakhstan ( ) MHI introduced Nominal expenditure Real expenditure MHI canceled Source: Agency of the Republic of Kazakhstan on Statistics web site Figure 3: GDP Per Capita in Selected ECA Countries, 1997 (in PPP $US) 16,000 14,000 Per capita GDP (ppp $) 12,000 10,000 8,000 6,000 4,000 2,000 0 CR CZ ES HU SK SV A AL KAZ LAT PO ROM RU Key: A: Group A average; AL: Albania; AZ: Azerbaijan; B: Group B average; C: Group C Average; CR: Croatia; CZ: the Czech Republic; ES: Estonia; GE: Georgia; HU: Hungary; KAZ: Kazakhstan; KY: Kyrgyzstan; LAT: Latvia; MO: Republic of Moldova; PO: Poland; ROM: Romania; RU: Russia; SK: Slovakia; SL: Slovenia.Source: Preker et al, 2002 B AZ GE KY MO C 159

16 Low Compliance Rates Part of the reason for not realizing revenues stemmed from poor compliance. In general, the countries of Eastern Europe and especially the FSU have faced considerable difficulty in collecting payroll taxes for health. Some countries, such as Estonia, Hungary and the Czech Republic, have structural characteristics that increase the likelihood of successful introduction of a payroll tax, including relatively higher per capita income, and a large percentage of the population living in urban areas and working in the formal sector (Ensor and Thompson 1997). Ensor (1999) noted registration was initially made easier given the large number of employees in the government sector or number of large state enterprises in many countries, such as Kazakhstan. But there have been major challenges in many countries. A significant economic burden was created by new health and social insurance taxes (totaling 44% in Hungary). In countries with less developed regulatory systems, a large proportion of unemployed or self-employed workers, and weak tax collection systems (Romania, Albania), establishing and sustaining a payroll tax is more difficult (Belli 2000). Premium collection varied between 9% and 52% of expected revenue in Kazakhstan in 1996 in different oblasts, rising to only 40% in Other countries Romania, Kyrgyzstan, Russia and even Estonia -- have reported similar problems of collection due to avoidance by labor and small businesses. A variety of other issues related to collection mechanisms, indebted enterprises and large numbers of the population outside the system, particularly farmers and the unemployed, contributed to the problem. Low levels of compliance may have been further exacerbated because there is often no link between contribution and benefit. The historical legacy of the socialist era was that many countries had a constitutional right to health care for all, which was retained in the transition period in most countries. Figure 4: Percentage Change in Real GDP in Selected ECA Countries PO SV % change in real GDP C CZ ES HU SK A AL KAZ LAT ROM RU B KY AZ GE C MO Key: A: Group A average; AL: Albania; AZ: Azerbaijan; B: Group B average; C: Group C Average; CR: Croatia; CZ: the Czech Republic; ES: Estonia; GE: Georgia; HU: Hungary; KAZ: Kazakhstan; KY: Kyrgyzstan; LAT: Latvia; MO: Republic of Moldova; PO: Poland; ROM: Romania; RU: Russia; SK: Slovakia; SL: Slovenia. Source: Preker et al,

17 Fostering an Underground Economy? As ECA countries move toward market-based economies dominated by smaller firms in very competitive markets, the payroll tax encourages employers to go underground to avoid payment. A large informal economy developed following transition, and the new increases in payroll tax probably further discouraged compliance and reinforced the flight of employers from the formal sector. In Central Asia, Kazakhstan introduced mandatory social health insurance in Langenbrunner et al (1994) found that compliance rates varied by size of firm and rates increased as size of employer increased. Figure 5: Social insurance Taxes as Share of Total Labor Cost, Mid-1990s Eastern Europe & Former Soviet Union (13) High-Income OECD (17) North Africa & the Middle East (13) Latin America & the Caribbean (19) East Asia & the Pacific (16) Sub-Saharan Africa (29) Note: Number of countries in region in brackets. Source: Palacios, R. and M. Pallares-Miralles, International Patterns of Pension Provision, Social Protection Discussion Paper No. 0009, World Bank, Washington, DC, Over the past decade, the Bank has done some work to help countries assess other types of revenues besides the payroll tax, but the work has been mostly in response to country requests. This policy dialogue has made a difference on many occasions. For example: The Prime Minister in Poland asked for a series of analytic pieces on sources of revenues and models of health financing in the late 1990s as Poland was moving from the Beveridge model to implementing the Bismarckian model of financing. Some recommendations were utilized by the Government; others ignored. In Croatia, Georgia, Kyrgyzstan and Russia, Bank loans financed the building of macroeconomic and/or micro-economic models to assess various sources of revenues and potential impacts. This built capacity as well as the opportunity for assessing options and developing policy dialogue; In other countries, such as Kyrgyzstan and Kazakhstan, mission teams were asked to develop quick assessments for the government or Parliament, and develop alternative policy options. In the first case, the advice was rejected (mid-1990s); in the second instance, (2002) it was accepted. The Bank currently has no favored strategy relating to payroll or other types of taxes for raising revenues, though it has often recommended against the payroll tax, especially in the context of poverty reduction work. 161

18 But by itself, tax and fiscal policy related to health is difficult to change unless managed within the broader context of tax and fiscal policy of the country. The HNP group would be well-advised to work more closely with other tax and fiscal experts in the Bank, and across donors such as the IMF, which does not favor "extrabudgetary" allocations due to its perceived lack of spending flexibility. Lack of Coordinated Collection and Transfers for Health In the prevalent models in Western Europe, some of the revenues of health insurance are transfers received from the Pension Fund and the Labor Fund (unemployment). In addition there are transfers from the budget for special social groups such as students, refugees, mothers during maternity etc. Farmers and the self-employed are also assessed in different ways. Only a few countries adopted a model whereby the payroll contribution of the actively employed population covers health insurance from cradle to grave (lifetime benefit). Similarly in many ECA countries contributions to the health insurance funds on behalf of the non-working population are made through transfers from other social insurance funds, such as unemployment and pension funds, or from government revenues. Annex 1 details the contributions and relationships across funds in ECA countries. In some cases, funds cooperation and the systems of inter-fund transfers have been made to increase revenues and lower administrative costs. In Bulgaria, the decision was made to combine the premium collection with the National Social Security Institute (NSSI), which collected pension and unemployment insurance premiums. The management of both the NSSI and the National Health Insurance Fund (NHIF) were surprised that the actual premium revenue for both institutions increased as a result, and actually exceeded expectations. However, due to chronic deficits with general revenues, and/or across the social security system, or in some cases just plain political bickering, these transfers were in many cases not made and substantial arrears built up. Health insurance funds were often obliged to provide health services to the whole population despite the lack of contributory income. This resulted in large financial deficits in the health insurance funds. As with the payroll tax, others in the donor community, such as the IMF do not favor "extra-budgetary" allocations and they do not recommend fund-to-fund transfers due to the perceived lack of transparency. Related to the ECA HNP, other divisions such as ECA SP, and PREM have not always agreed on advice given to countries about merging and management of funds. There are two potential solutions that the Bank ECA HNP and other stakeholders will need to take into consideration and promote. 1. To find an appropriate legal and regulatory framework to prevent arrears and allow for transparent transfer mechanisms between social security funds (pension, labor, health). This need not be complicated, but merely a basic set of rules governing management and transfer of funds. 2. To find out how countries handle "indexing at the payroll source", where one payroll contribution pays for lifetime health benefits. Increased Reliance on Out-of-Pocket and Informal Payments The Bank has extensively studied this issue, especially in the last 3-5 years, by Lewis (2000), Chawla (see, for example, Poland study in 1999), Shahriari (Georgia, 2002) and others. There is now a clearer sense within ECA that payments made by patients and families to supplement formal coverage are common. Estimated frequencies of informal payments in the region are typically high (Lewis, 2000). The percentage of patients reporting that they had been required to make some payment for a service was: 162

19 60% in Slovakia, 66% in Tajikistan, 70% in the Republic of Moldova, 74% of hospital patients in Russia, 75% in Kyrgyzstan, 78% of inpatients in Poland, 78% in Azerbaijan, and 91% in Armenia. However, informal payments are not high in the Czech Republic where doctors salaries have risen above the rate of inflation of average wages. The level of payments is highest for inpatient care with drugs and outpatient care subject to lower fee levels. In relation to household income, out-of-pocket payments for health care can account for as much as 21% of monthly income in Georgia, 9.1% in Albania and 4.1% in Romania. Further survey data is needed to establish more accurately the level and extent of informal payments. Informal payments take a number of forms and may exist for a number of reasons. They range from the ex post gift in-kind to the ex ante cash payment. These payments or gifts may be part of the culture, or may be due to the lack of a cash economy, a lack of finances to pay health care workers, a lack of drugs and basic equipment to treat patients and/or due to weak governance. At their worst they may be a form of corruption, undermining official payment systems, and reducing access to health services (Ensor 2002). Much less understood by the Bank is what to do about it. Despite the extensive number of studies, and a large volume of recommendations, the Bank and its lending programs have yet to develop a well-formulated strategy for addressing this issue. This will be an important challenge for the Bank over the next five to ten years. How Much Revenue Is Enough? The sustainability of health care systems in the region depends largely on the ability to generate sufficient revenue. This is a key challenge given the number of contextual and structural problems in the region. As in the OECD countries, there is no right answer of how much is enough, but an analysis of overall expenditures in ECA countries suggests that the levels of expenditure are not significantly different from the OECD countries, which average 6-8% as a share of GDP for health services. Total reported expenditure on health in the region in 1999 (the latest available year) ranged from as low as 1.8% in Azerbaijan and 3.3% of GDP in Albania in 1997, to 9.5% in Croatia. As Figure 6 suggests, spending has been relatively stable as a share of GDP over the last decade. The decreases in spending on an absolute basis reflect changes in macroeconomic trends, as discussed above. Expenditure as a share of GDP has been highest in Croatia, Slovakia, Slovenia and the Czech Republic, and lowest in countries such as Georgia, Azerbaijan, and Albania at less than 4% of GDP (World Bank, 2002; Preker et al 2002). Another set of countries Russia, Uzbekistan, Ukraine, Turkey, Turkmenistan, Romania, Kyrgyzstan and Bulgaria report expenditures between 4-6% of GDP. However, the recent studies since 1999 through the Bank funding (e.g., LSMS) and others sources suggest the overall levels of spending are not as low. Table 6 (below) develops adjusted spending levels based on some of these recent studies. In fact, most of the countries originally thought to be low in terms of spending moved considerably higher, once good estimates of private spending informal and formal were integrated into spending totals. Examining these numbers as a share of share of GDP does allow a better understanding and the perceived gap somewhat disappears once private and out-of-pocket spending is better understood. 163

20 Figure 6: Expenditures as a Share of GDP since 1990: ECA Countries 14.0% Albania Armenia Azerbaijan 12.0% Bulgaria Bosnia & Herzegovina Belarus 10.0% Czech Republic Estonia Georgia Croatia 8.0% Hungary Kazakhstan Kyrgyz Republic 6.0% Lithuania Latvia Moldova 4.0% Macedonia, FYR Poland Portugal 2.0% Romania Russian Federation 0.0% Slovak Republic Slovenia Tajikistan Turkmenistan Turkey Table 6 Estimates for Overall Expenditures as a Share of GDP (Once new studies of private spending are included in the estimates) Country Adjusted Expenditure Source (for Estimating Private Expenditures) Levels as a Share of GDP Azerbaijan 6.0 % LSMS, 1998 Bulgaria 6.1 % Balibanova, 1999 Georgia 4.8 % World Bank, 1999 Russia 9.2 % USAID-funded survey (Feeley et al, 2000) Uzbekistan 6.8 % World Bank, Family Budget Survey, 2001 Underfunding, Fragmented Funding, or Lack of Financial Protection? The upshot here is to call into question whether the real issue in ECA is levels of funding or whether the real issue is financial protection, especially of the poor and vulnerable. It is more likely that spending levels as a share of GDP in ECA are similar to the OECD, and the real issue in the years ahead will be to look for ways to increase the pooling of these funds through public or publicly-regulated channels in order to provide adequate financial protection. It is also true that despite levels of financial protection, poorer countries will continue to spend far less than needed, measured on the basis of per capita levels. Indeed, several countries, such as the IDA-eligible Albania, Azerbaijan, Armenia, Georgia, Kyrgyzstan, Moldova, Tajikistan and Uzbekistan spend less than $100 per capita per year. With input process for outside pharmaceuticals and equipment, the levels of spending remain inadequate. 164

21 The Bank will need to consider new and innovative approaches for achieving better levels of financial protection. One option is to pursue new forms of insurance and community financing schemes. For example, in Georgia new Japanese Trust Funds will allow piloting of new community based financing schemes. These new insurance arrangements will need to be backed with government transfers and social re-insurance mechanisms. Community-financing schemes are one option for extending financial protection and also for reaching the poorest of the poor if designed correctly (Preker and Jakab, 2001). Even so, community financing has failed more often than not (ILO, 2002) and has yet to be tested in the ECA region. Looking Ahead: Lessons and Implications In higher income countries with higher levels of formal employment (namely Croatia, the Czech Republic, Estonia, Hungary, Slovakia and Slovenia) the payroll or dedicated tax appears to have been a somewhat effective way of mobilizing resources for the health sector. But for the great majority of countries, the great emphasis in the last decade on new funding did not necessarily increase spending, but did perhaps help stabilize funding, especially in the context of the macro-economic downturn and issues related to collection and transfer. Economic recovery and capacity building in the region will help towards increasing the revenue collected through payroll taxes. As a percentage of GDP, however, combined public and private spending together is not very different today from total spending in OECD countries in the nearby European Union. At the same time, many countries may have lost ground in terms of financial protection, though good baselines do not exist for these countries for the period prior to Lower income countries in the region such as Albania, Romania and Kazakhstan, with little formal employment, or countries with low governance capacity, such as Russia, found that payroll contributions were not viable. Further efforts to ensure compliance may be necessary. The delegation of responsibility for revenue collection to quasi-state agencies or independent insurance funds has created significant challenges for the state in this respect. Another option is to further diversify funding sources, e.g., through subsidies from other forms of taxation or by pooling out-of-pocket payments. Transfers from other public sources do not always occur, though. These mechanisms need to be transparent and need to ensure that funds are not penalized (e.g., by reduced subsidies 46 ) for increasing their revenue and/ or efficiency. Where there is a large informal economy, direct taxation (i.e., taxes levied on income or profits) may face similar problems of compliance. However, it places less of a direct burden on labor costs and may therefore have less negative consequences for the development of the economy. Indirect taxes (i.e., taxes levied on goods and services) are more visible and may be less easily evaded; however, they are more regressive. Experience from low- and middle-income countries outside Europe, with community health insurance, for example, suggests that formalizing out-of-pocket payments and establishing systems of pre-payment (or insurance) will be extremely difficult (Mills and Bennett 2002). Informal payments are partly a response of the health care system, particularly health care providers, to the lack of financial resources, and the response of patients to a system that is unable to provide adequate access to basic services. Governments should ensure that the limited resources are targeted more effectively in order to secure access to basic services, for example by shifting resources from secondary and tertiary care to primary care. Willingness to contribute to a formal system of pre payment should be higher if there are clear benefits and patients are not also expected to pay informally. 46 There is some evidence to suggest that countries that shifted to social health insurance were better able to maintain levels of spending on health care (Preker et al 2002). However, anecdotally social health insurance revenues were simply used to substitute for general revenues by the Ministry of Finance and overall funding for the health sector did not increase as a result of the introduction of social health insurance contributions. 165

22 The Bank has too often been at the periphery of these events and developments. In the past few years, its AAA has focused increasingly on the loss of financial protection, but has done more in the way of diagnostic work than development of targeted solutions. The Bank can continue to play a role and an even broader role -- in advice and analysis of different funding sources, and exposing governments to a broader array of fiscal instruments. It can help governments understand options and potential impacts. It can also help governments understand issues of financial protection of the poor and impoverished. The lessons of new funding sources further suggest a familiar strategy: that ECA countries need to focus less on levels of revenues and more on improved efficiency and financial protection as high priority issues. At the same time, the poorest countries will need help from outside funding sources such as the Bank, to address the absolute levels of spending for health services. The lack of health financing is generally the result of deeper macro-economic problems that cannot be solved solely through interventions in the health sector. Options such as adjustment lending can be considered, but adjustment operations rarely include direct budgetary support to the health system. Adjustment lending directly for Health Insurance (HI) funds needs to be explored, as well as lending that provides more for recurrent expenditures relative to new investments in civil works and equipment. Pooling and Organization/Management of Revenues The second important function of health care financing is to pool and manage the resources collected from various sources and to allocate these to purchasers and/or providers. The extent of pooling will depend on how much of the revenues collected are pooled through a single fund and whether different sources of funding are pooled or remain separate. For example, tax revenues may be pooled together with social insurance contributions to enable funds to purchase health care services on behalf of all citizens. Alternatively pooling may be limited if tax revenues are kept separate to provide public services directly for those who do not make insurance contributions. Pooling enhances efficiency because it reduces the incentives for risk selection and may break historical patterns of allocation. It also increases equity and solidarity principles by sharing risks across a larger population. Voluntary health insurance may, if it is group rated, pool risks amongst the employees of a company or if it is community rated amongst the residents of a particular area. However, voluntary health insurance initially is risk-rated individually (and may be experience-rated subsequently), therefore pooling amongst subscribers is extremely limited. If revenue from user charges are retained by the providers who collect them, little pooling of funds takes place. However, where revenue from user charges are pooled with other revenue to provide services for a specific population, pooling may occur. A well-designed pooling function can be judged on the extent to which multiple revenue streams are integrated or fragmented and the size of the population across which pooling occurs. Pooling and Health Insurance In ECA, the issue of pooling (or lack of it) has been addressed most often in the context of new social insurance organizations, set up to manage and pool funds for health care services. HI reform commanded wide support, but for different reasons: ministries of finance, for example, hoped for increased efficiency and cost control, while doctors expected higher salaries. Nevertheless, the development of HI has been one of the biggest challenges governments have faced in the past decade -- to establish semi-autonomous para-statal organizations and then to implement the new social insurance programs. The main dilemmas included: 166