H N P D I S C U S S I O N P A P E R The Impact of the Degree of Risk-Sharing in Health Financing on Health System Attainment Guy Carrin, Riadh Zeramdini, Philip Musgrove, Jean-Pierre Poullier, Nicole Valentine and Ke Xu September 2001 THE IMPACT OF THE DEGREE OF RISK-SHARING IN HEALTH FINANCING ON HEALTH SYSTEM ATTAINMENT Guy Carrin, Riadh Zeramdini, Philip Musgrove, Jean-Pierre Poullier, Nicole Valentine and Ke Xu September 2001 Health, Nutrition and Population Discussion Paper This series is produced by the Health, Nutrition, and Population Family (HNP) of the World Bank's Human Development Network (HNP Discussion Paper). The papers in this series aim to provide a vehicle for publishing preliminary and unpolished results on HNP topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For free copies of papers in this series, please contact the individual authors whose names appear on the paper. Inquiries about the series and submissions should be made directly to the Editor in Chief. Submissions should have been previously reviewed and cleared by the sponsoring department which will bear the cost of publication. No additional reviews will be undertaken after submission. The sponsoring department and authors bear full responsibility for the quality of the technical contents and presentation of material in the series. Since the material will be published as presented, authors should submit an electronic copy in a predefined format as well as three camera-ready hard copies (copied front to back exactly as the author would like the final publication to appear). Rough drafts that do not meet minimum presentational standards may be returned to authors for more work before being accepted. The Editor in Chief of the series is Alexander S. Preker (apreker@worldbank.org); For information regarding this and other World Bank publications, please contact the HNP Advisory Services (healthpop@worldbank.org ) Tel (202) 473-2256; and Fax (202) 522-3234. ISBN 1-932126-01-5 © 2001 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 All rights reserved. ii Health, Nutrition and Population Discussion Paper The Impact of the Degree of Risk sharing in Health Financing on Health System Attainment Guy Carrina, Riadh Zeramdinia, Philip Musgroveb, Jean-Pierre Poulliera, Nicole Valentinea and Ke Xua aWorld Health Organization, Geneva, Switzerland bWorld Bank, Washington D.C., U.S.A. Prepared for Working Group 3 of the WHO Commission on Macroeconomics and Health Version: September, 2001 Abstract: A simple econometric analysis is undertaken concerning the impact of the degree of risk sharing in countries' health financing organization on the goals of the health system, as defined in the World Health Report 2000, i.e., the level of health and its distribution across the population, the level of responsiveness and its distribution across the population, and fair financing. The degree of risk sharing varies according to whether countries have a universal coverage system, financed via social health insurance or general taxation, or systems with less well-developed coverage including variants of social health insurance and/or general taxation benefiting specific population groups. We undertook a classification of countries according to the degree of risk sharing, based primarily on the health care financing legislation of the World Health Organization's 191 member states and on its data base of Health System Profiles. The results obtained give empirical support to the hypothesis that the degree of risk sharing in health financing organizations impacts positively on health system attainment, as measured by the five goals indicators. The effects found prove to be quite robust, after introducing the GINI index among the set of explanatory variables in the models for the distributional measures. Keywords: risk sharing, health system goals, health financing system Disclaimer : The findings, interpretations, and conclusions expressed in the paper are entirely those of the authors, and do not necessarily represent the views of the World Bank, its Executive Directors, or the countries they represent. Correspondence Details: G. Carrin, World Health Organization, 1211 Geneva 27, Switzerland, Tel. +41 22 791 2780. Fax +41 22 791 4328. carring@who.ch iii iv Contents PREFACE.................................................................................................................................. VII ACKNOWLEDGMENTS............................................................................................................IX ACRONYMS................................................................................................................................ X II. HEALTH SYSTEM GOALS AND FUNCTIONS IN A NUTSHELL.....................................1 III. THE ORGANIZATIONAL FORM OF HEALTH FINANCING AND ITS LINK TO GOAL ACHIEVEMENT...........................................................................................................................2 D. ORGANIZATION OF HEALTH FINANCING IN THE WORLD.........................................3 V. MODELING THE IMPACT OF THE ORGANIZATIONAL FORM OF HEALTH FINANCING ON HEALTH ATTAINMENT...............................................................................5 A. DESCRIPTIVE DATA ANALYSIS...................................................................................................5 B. SPECIFICATION OF THE BASIC MODEL........................................................................................8 Impact on the level of health and on responsiveness ...............................................................8 Impact on the distributional measures of the goals................................................................10 C. SPECIFICATION OF ENLARGED MODELS ................................................................................... 11 The GINI index of income inequality in the equations for the distributional measures...........11 D. RESULTS................................................................................................................................ 13 Estimation results for the basic model....................................................................................13 Estimation results with the GINI index as an explanatory variable in the equations for the distributional measures..........................................................................................................15 Estimation results when using interaction terms with the ratio of public health expenditure to total health expenditure.........................................................................................................16 Key conclusions.....................................................................................................................16 Preliminary analysis with updated data .................................................................................17 D. COMMUNITY RISK-SHARING ARRANGEMENTS: FURTHER NEED TO MEASURE THEIR IMPACT.........................................................................................................................17 VI. CONCLUDING REMARKS.................................................................................................18 VII. APPENDIXES ......................................................................................................................21 APPENDIX A ........................................................................................................................... 21 APPENDIX B............................................................................................................................ 31 APPENDIX C............................................................................................................................ 39 APPENDIX D ........................................................................................................................... 49 APPENDIX E............................................................................................................................ 55 APPENDIX F............................................................................................................................ 63 APPENDIX G............................................................................................................................ 68 APPENDIX H ........................................................................................................................... 79 APPENDIX I............................................................................................................................. 82 VIII. BIBLIOGRAPHY.............................................................................................................105 v vi PREFACE In January 2000, Dr. Gro Harlem Bruntland, Director General of the World Health Organization (WHO), established a Commission on Macroeconomics and Health (CMH) to provide evidence on the importance of health to economic development and poverty alleviation. This HNP Discussion Paper is based on a report on community financing submitted in September 2001 to Working Group 3 of the CMH. The mandate of Working Group 3 was to examine alternative approaches to domestic resources mobilization, risk protection against the cost of illness, and resource allocation. The working group was chaired by Professor Alan Tait (Former Deputy Director of Fiscal Affairs, International Monetary Fund, and currently Honorary Fellow at University of Kent at Canterbury and Honorary Fellow at Trinity College, Dublin) and Professor Kwesi Botchewey (Director of Africa Research and Programs at the Harvard Center for International Development). Professor Jeffery D. Sachs (Chairman of the Commission and Director of the Harvard Center for International Development) presented the findings of the CMH in a report submitted to WHO on December 20, 2001--Macroeconomics and Health: Investing in Health for Economic Development. The CMH report recommended a six-pronged approach to domestic resource mobilization at low income levels: "(a) increased mobilization of general tax revenues for health, on the order of 1 percent of GNP by 2007 and 2 percent of GNP by 2015; (b) increased donor support to finance the provision of public goods and to ensure access for the poor to essential health services; (c) conversion of current out-of-pocket expenditure into prepayment schemes, including community financing programs supported by public funding, where feasible; (d) a deepening of the HIPC (Highly Indebted Poor Countries) initiative, in country coverage and in the extent of debt relief (with support form the bilateral donor community); (e) effort to address existing inefficiencies in the way in which government resources are presently allocated and used in the health sector; and (f) reallocating public outlays more generally from unproductive expenditure and subsidies to social- sector programs focused on the poor." Most community financing schemes have evolved in the context of severe economic constraints, political instability, and lack of good governance. Usually government taxation capacity is weak, formal mechanisms of social protection for vulnerable populations absent, and government oversight of the informal health sector lacking. In this context of extreme public sector failure, community involvement in the financing of health care provides a critical, though insufficient, first step in the long march toward improved access to health care by the poor and social protection against the cost of illness. The CMH stressed that community financing schemes are no panacea for the problems that low-income countries face in resource mobilization. They should be regarded as a complement to--not as a substitute for--strong government involvement in health care financing and risk management related to the cost of illness. Based on an extensive survey of the literature, the main strengths of community financing schemes are the degree of outreach penetration achieved through community participation, their contribution to financial protection against illness and increase in access to health care by low-income rural and informal sector workers. Their main weaknesses are the low level of revenues that can be mobilized from poor vii communities, the frequent exclusion of the poorest from participation in such schemes without some form of subsidy, the small size of the risk pool, the limited management capacity that exists in rural and low- income contexts, and their isolation from the more comprehensive benefits often available through more formal health financing mechanisms and provider networks. The work by the CMH proposed concrete public policy measures that governments can introduce to strengthen and improve the effectiveness of community involvement in health care financing. This includes: (a) increased and well targeted subsidies to pay for the premiums of low-income populations; (b) use of insurance to protect against expenditure fluctuations and use of reinsurance to enlarge the effective size of small risk pools; (c) use of effective prevention and case-management techniques to limit expenditure fluctuations; (d) technical support to strengthen the management capacity of local schemes; and (e) establishment and strengthening of links with the formal financing and provider networks. The report presented in this HNP Discussion Paper has made a valuable contribution to our understanding of some of the strengths, weaknesses, and policy options for securing better access for the poor to health care and financial protection against the impoverishing effects of illness, especially for rural and informal sector workers in low-income countries. Alexander S. Preker Chief Economist Health, Nutrition, and Population viii ACKNOWLEDGMENTS The authors of this report are grateful to the World Health Organization (WHO) for having provided an opportunity to contribute to the work of the Commission on Macroeconomics and Health and to the World Bank for having published the report as an HNP Discussion Paper. Discussions with and suggestions from Alex Preker, Melitta Jakab, David Evans, Kei Kawabata, and Ajay Tandon are gratefully acknowledged. During the writing of the paper, Philip Musgrove was on secondment to the World Health Organization. The views expressed in this document are solely the responsibility of the authors. ix ACRONYMS CO Sub-index of responsiveness DALE Disability adjusted life expectancy DARS Dummy variable = 1 when the country has an advanced risk-sharing system; otherwise 0 DMRS Dummy variable = 1 when the country has a medium risk-sharing system; otherwise 0 DMRS1 Dummy variable = 1 when the country has health insurance schemes whereby only employees are covered; otherwise 0 DMRS2 Dummy variable = 1 when the country has health insurance schemes that cover specific groups only; otherwise 0 DSHI Dummy variable = 1 when the country has a social health insurance scheme; otherwise 0 EDU Enrolment in primary education of the relevant age group GINI Gini index of income inequality GDP Gross Domestic Product GT General taxation HEC Health expenditure per capita (in U.S. dollars) IECS Index of equality of child survival" IFFC Index of fairness of financial contribution IR Index of level of responsiveness IRD Index of distribution of responsiveness PHE% Share of public health expenditure in total health expenditure RESPECT Sub-index of responsiveness "Respect for persons" SHI Social health insurance WHR World Health Report x I. INTRODUCTION There are important linkages between what health systems can achieve in terms of pre-set goals and the functions that they undertake. The World Health Report (WHR) 2000 has designed a coherent framework for analyzing these linkages.1 In this paper, we specifically address the health financing function of pooling of resources and how it influences health systems attainment. One essential question is whether health financing organizations provide sufficient financial risk protection for the population. People's access to health services depends on this protection. Health financing organizations that do not include the low- income population groups, for instance, will lead to many individuals' being unable to pay for care. The extent to which these population groups are effectively included in risk-sharing arrangements is therefore likely to affect a goal such as the equality of health status. Health financing organizations may also be more or less engaged in purchasing an adequate package of health services for the entire population. In this sense, they may affect the average level of access to good care and therefore indirectly have an impact upon the average level of health. Apart from the level and distribution of health status, other goals may be considered. In the next section, we give an overview of the goals of health systems as proposed by the WHR 2000, and discuss how they relate to the functions of these systems. The main purpose of this paper is to undertake a simple econometric analysis of the impact of the degree of risk sharing in countries' health financing organization on the goals of the health system. The degree of risk sharing will vary according to whether countries have a universal coverage system, financed via social health insurance or general taxation, or systems with less well-developed coverage, including variants of social health insurance and/or general taxation benefiting specific population groups. Risk sharing via community health financing schemes could not be considered for lack of data at the national level. In preparation of the econometric analysis, we turn to the specific linkage between the goals and the health financing function in section 3. Then in section 4 we classify the health financing organization of 191 countries by the degree of risk sharing. This classification will help in defining the variables that measure risk sharing, and which will be used in the econometric analysis. We examine the available data on public health expenditure and health expenditure by nongovernmental organizations and communities. The specification of the econometric models and estimation results are presented in sections 5 and 6, respectively. We conclude in section 7. II. HEALTH SYSTEM GOALS AND FUNCTIONS IN A NUTSHELL The framework, as presented in the WHR 2000, defines a set of goals or objectives and includes ways to measure the achievement toward these goals. Of course, to obtain these achievements, health systems do need to carry out a number of functions. Below, we address both goals and functions. The goals considered are good health, responsiveness, and fair financing. Good health is approached in two ways. One is by striving for the best attainable average level for the entire population. The other is by minimizing the differences in health status among individuals and groups. Health is measured via disability- 1WHO (2000). See also Murray and Frenk (2000). 1 adjusted life expectancy,2 whereby account is taken of time lived with a disability. Second, responsiveness measures how the health system performs relative to non-health aspects of provided health services. Responsiveness captures the extent to which the health system is client-oriented and treats people with respect. Respect for people includes the following aspects: respect for the dignity of the person, confidentiality, and autonomy. Within client orientation, we consider prompt attention, the quality of the amenities, the access to social support networks, and the choice of provider. Note that the distinction between overall level and distribution across the population also applies to responsiveness. Third, fair financing requires that health expenditure of households be distributed according to ability to pay rather than to individual risk of illness. In a fairly financed system, everyone should be financially protected. It is crucial therefore that health systems rely as fully as possible on prepaid contributions that are unrelated to individual illness or utilization. It is clear that when analyzing fair financing, we are concerned with distributive aspects only. We thus obtain five objectives: the level and distribution of health, the level and distribution of responsiveness, and fair financing. Measurements have been designed so as to quantify the achievement with respect to each of these objectives.3 We further consider four main functions of the health system: the delivery of health services; the creation of resources for health (investment in people, buildings and equipment); health financing (raising, pooling, and allocating the revenues to purchase health services); and stewardship. The latter refers to a government's responsibility for the general health of its population. The stewardship function is of special importance, as it will have an impact on the way the other three functions are carried out. Work is currently underway at WHO to define indicators for the various functions, so that their possible impact on goal achievement can be measured. This paper can be seen as an element of this particular work, in that it focuses on the nature of risk sharing in the world's different health financing systems and its possible impact on the goals as defined above. III. THE ORGANIZATIONAL FORM OF HEALTH FINANCING AND ITS LINK TO GOAL ACHIEVEMENT A crucial concept in health financing is that of pooling. Pooling is defined as the "accumulation and management of revenues in such a way as to ensure that the risk of having to pay for health care is borne by all members of the pool and not by each contributor individually."4 The larger the degree of pooling, the less people will have to bear the financial consequences of their own health risks. 2This summary measure of population health adjusts life expectancy at birth for the burden of disability. Disability weights are used to convert years lived in disability into equivalent years lived in good health. See further Mathers et al. (2000). 3 See WHO (2000) for a summary of the methods. For further details, we refer to http://www- nt.who.int/whosis/statistics/discussion_papers/discussion_papers.cfm?path=statistics,discussion_papers 4WHO (2000, p.96). 2 Health financing systems encompass various degrees of risk sharing. There are two major ways to ensure financial risk protection for a nation's entire population. One is a system whereby general taxation (GT) is the main source of financing health services. Services are usually provided by a network of public and contracted private providers, often referred as a national health service. The second is social health insurance (SHI), whereby workers, enterprises, and government pay financial contributions. The base for workers' and enterprises' contributions is usually the worker's salary. Social health insurance either owns its own provider networks, works with accredited private providers, or combines both approaches. In principle, both systems pool all of the population's risks, with contributions that are delinked from individual risks. This approach theoretically avoids exposing individuals to no or insufficient access to the health care they need. These systems are often denoted as universal coverage systems, but financial protection may still be judged inadequate in a number of these systems. There are also systems with no explicit reference to overall coverage of the population. These include mixed health financing systems, with some part of the population partially covered via general taxation, and another part covered by health insurance schemes. The latter may address specific groups only. Still, they may practice full pooling among their members and define health insurance contributions according to capacity to pay, rather than according to individual health risks. In other words, these schemes may apply community rating such as in a social health insurance scheme, but for specific groups only. Such schemes may include voluntary private insurance arrangements, mutual health funds, enterprise-based and community health insurance. Finally, some countries do finance health services via general taxation but offer only incomplete coverage. For the purpose of this paper, we will say that countries that aim at universal coverage and use either general taxation or social health insurance enjoy systems with advanced risk sharing. Such schemes allow for a more equal access among individuals to health services. In addition, such schemes generally better define an adequate package of health services to which citizens are entitled. Countries with mixed health financing systems will be associated with medium risk sharing. The countries with general taxation systems that incompletely cover the population are then associated with low risk sharing. In this paper we will investigate whether larger degrees of risk sharing have a beneficial impact on the five indicators of goal achievement. D. ORGANIZATION OF HEALTH FINANCING IN THE WORLD In Table 1 of Appendix A, countries are classified according to the criterion of risk sharing as defined above, based on health care financing legislation of the 191 member states of the World Health Organization (WHO). Our main source for this revision was the publication Social Security Programs throughout the World provided by the U.S. Social Security Administration (1999). However, for 52 countries, no or insufficient information was given. For the latter group, and in order to identify the category of health financing system, WHO's data base of Health System Profiles5 and other selected publications were used.6 5 These can be found on www.who.int/country_profiles/main.cfm/ 6These include Nolan and Turbat (1995) and the website of the Center for International Health Information www.cihi.com 3 In Table 1 approximately 40 percent of the countries are characterized as advanced risk-sharing systems they have either a general taxation system (50 countries) or a social health insurance scheme (30 countries), covering nearly the entire population. The 61 countries with medium risk sharing are further classified into three main variants. In the first variant, health insurance covers all employees and self- employed, though subject to a number of exclusions.7 The second variant covers only employees and the third covers specific groups only, for instance through mutual health funds and enterprise-based health insurance for particular categories of workers. In these three variants, there are 9, 20, and 32 countries, respectively. Finally, 50 countries are classified among those with low risk sharing. These countries are generally characterized by under-financed health systems as compared to the health needs of the population. The names of a number of countries are printed in boldface italics. For those countries, the proposed classification is uncertain, due to incomplete or absent information on the size and structure of the eligible population that is effectively covered by the health financing system. The overall classification allows us now to define the two main organizational dummy variables: DARS = 1 when a country belongs to the set of advanced risk-sharing systems and 0 otherwise; DMRS = 1 when a country belongs to the set of medium risk-sharing systems and 0 otherwise. In Table 2 of Appendix A, we rank countries according to the category of risk sharing and the percentage share of public8 health expenditure in total health expenditure. The three categories considered are a share between 75 and 100 percent, between 50 and 75 percent, and below 50 percent. We use the latter ratio as a simple quantitative indicator of the system's degree of financial risk protection. In fact, of the countries with advanced risk sharing, 74 out of 80 have a ratio above 50 percent; 41 have a ratio above 75 percent. Of the countries with medium risk sharing, only 3 out of 61 have a ratio above 75 percent. For countries with low risk sharing, a tilt toward low ratios would be expected. However, for 9 out of 50 countries with low risk sharing, ratios above 75 percent are reported, which is surprising. However, it is recognized in the WHR 2000 that quite a number of countries have incomplete data and mixed degrees of reliability,9 which may partly explain this finding. It is also interesting to rank countries according to the category of risk sharing and to the income level, as measured by 1998 gross domestic product per capita (in U.S. dollars). According to Table 3, Appendix A, of the 80 countries with advanced risk sharing, 20 belong to the category of upper middle-income countries, and 34 to the high-income category. Most countries with low to medium risk sharing belong to the low- income and lower middle income categories. In this set of countries, only Andorra and the United States belong to the upper middle-income or high-income category. 7For instance, the agricultural self-employed population may not be covered. Or workers in small enterprises with fewer than 10 workers may not be insured. 8Note that social insurance expenditure is included in public health expenditure. 9 For these nine countries, the data are either incomplete with low reliability (two countries out of nine) or are incomplete with high to medium reliability. 4 V. MODELING THE IMPACT OF THE ORGANIZATIONAL FORM OF HEALTH FINANCING ON HEALTH ATTAINMENT A. DESCRIPTIVE DATA ANALYSIS As a prelude to the econometric analysis, descriptive statistics for the five health attainment indices are computed. The health attainment indices are the disability-adjusted life expectancy (DALE), the index of level of responsiveness (IR), the index of fairness of financial contribution (IFFC), the index of distribution of responsiveness (IRD) and the index of equality of child survival (IECS). All data used originate from the Statistical Appendix of WHR 2000. In Table 1a, statistics are presented related to all countries that have observations on the indices. In Table 1b, however, countries whose risk-sharing classification is uncertain are removed from the samples. Appendix B presents the histograms associated with the five indicators for the full and restricted samples.10 The indices are classified according to the category of risk sharing of countries' health financing organizations. We present the mean, coefficient of variation, minimum and maximum. A first general tendency is that the means of the indicators are larger, the greater the degree of risk sharing. One exception is in Table 1a where the mean fair financing index for countries with advanced and medium risk sharing is smaller than that of the countries with low risk sharing. However, in Table 1b, the mean IFFC for countries with advanced risk sharing exceeds that for countries with low risk sharing. Second, using the restricted samples (Table 1b), the coefficients of variation (CV) indicate that, except in the case of IR, there is a lower relative dispersion around the mean in countries with advanced risk sharing than in countries with medium risk sharing. The latter is consistent with the fact that we have defined 3 sub-groups with different degrees of risk sharing within the set of countries with medium risk sharing. Notice also that in three cases (fair financing, distribution of responsiveness and distribution of health), countries in the low risk sharing category show lower coefficients of variation than those for the countries with medium risk sharing. It stands to reason that the low risk sharing category of countries is likely to be more homogeneous than the group of countries with medium risk sharing. Except for the value related to IR, the coefficients of variation are higher, however, when compared with the CV of countries with advanced risk sharing. 10 The samples for IR and IRD do not contain countries whose risk-classification is uncertain. In other words, for those variables, only the full samples are considered. 5 Table 1a: Descriptive Statistics (full samples) Disability Index of Level of Index of fairness of Index of distribution Index of equality Statistics adjusted life- Responsiveness financial contribution of responsiveness of child survival expectancy (IR) (IFFC) (IRD) (IECS) (DALE) Total sample Mean 56.8262 0.5165 0.8730 0.8967 0.6659 CV1 0.21650 0.1542 0.1203 0.0969 0.2878 Min 25.9000 0.3740 0.6230 0.7230 0.2450 Max 74.5000 0.6880 0.9920 0.9999 0.9990 Number of observations 191 30 21 33 58 Countries with advanced risk sharing (DARS = 1) Mean 66.0725 0.5849 0.8732 0.9772 0.9296 CV 0.07550 0.1272 0.0643 0.0252 0.1490 Min 52.3000 0.4430 0.8020 0.9180 0.6320 Max 74.5000 0.6880 0.9390 0.9999 0.9990 Number of observations 80 8 5 9 7 Of which countries with Social Health Insurance (DSHI = 1) Mean 68.5267 0.5452 0.8945 0.9715 0.9990 CV 0.05520 0.1150 0.0704 0.0290 0 Min 62.2000 0.4430 0.8500 0.9180 0.9990 Max 74.5000 0.6120 0.9390 0.9960 0.9990 Number of observations 30 5 2 6 4 Of which countries with General Taxation (DSHI = 0) Mean 64.6000 0.6510 0.8590 0.9886 0.8370 CV 0.07850 0.0492 0.0694 0.0128 0.2237 Min 52.3000 0.6320 0.8020 0.9750 0.6320 Max 73.0000 0.6880 0.9210 0.9999 0.9990 Number of observations 50 3 3 3 3 Countries with medium risk sharing (DMRS = 1) 52.9033 0.5153 0.8623 0.8846 0.6792 Mean 0.21520 0.1109 0.1463 0.0932 0.2320 CV 29.1000 0.4180 0.6230 0.7230 0.2610 Min 72.3000 0.6230 0.9920 0.9860 0.9660 Max 61 16 11 17 34 Number of observations Countries with low risk sharing (DARS = 0 and DMRS = 0) 1: CV is the coefficient of variation 6 Mean 46.8180 0.4285 0.8962 0.8227 0.5309 CV 0.24110 0.1165 0.1183 0.0847 0.2816 Min 25.9000 0.3740 0.7140 0.7280 0.2450 Max 66.7000 0.4940 0.9610 0.9490 0.7850 Number of observations 50 6 5 7 17 Table 1b: Descriptive Statistics (restricted samples) Disability Index of Level of Index of fairness of Index of distribution Index of equality Adjusted life- Responsiveness financial contribution of responsiveness of child survival Statistics expectancy (IR) (IFFC) (IRD) (IECS) (DALE) Total sample Mean 58.0588 0.5165 0.8721 0.8967 0.6843 CV1 0.20840 0.1542 0.1233 0.0969 0.2636 Min 25.9000 0.3740 0.6230 0.7230 0.2610 Max 74.5000 0.6880 0.9920 0.9999 0.9990 Number of observations 160 30 19 33 52 Countries with advanced risk sharing (DARS = 1) Mean 67.1179 0.5849 0.8910 0.9772 0.9378 CV 0.06450 0.1272 0.0513 0.0252 0.1598 Min 56.3000 0.4430 0.8500 0.9180 0.6320 Max 74.5000 0.6880 0.9390 0.9999 0.9990 Number of observations 67 8 4 9 6 Of which countries with Social Health Insurance (DSHI = 1) Mean 68.5267 0.5452 0.8945 0.9715 0.9990 CV 0.06460 0.1150 0.0703 0.0290 0 Min 62.2000 0.4430 0.8500 0.9180 0.9990 Max 74.5000 0.6120 0.9390 0.9960 0.9990 Number of observations 30 5 2 6 4 Of which countries with General Taxation (DSHI = 0) Mean 65.9757 0.6510 0.8875 0.9886 0.8155 CV 0.06740 0.0492 0.0534 0.0128 0.3183 Min 56.3000 0.6320 0.8540 0.9750 0.6320 Max 73.0000 0.6880 0.9210 0.9990 0.9990 Number of observations 37 3 2 3 2 Countries with medium risk sharing (DMRS = 1) 53.7596 0.5153 0.8623 0.8846 0.6849 1CV is the coefficient of variation 7 1) 0.20810 0.1109 0.1464 0.0932 0.2282 29.1000 0.4180 0.6230 0.7230 0.2610 Mean 72.3000 0.6230 0.9920 0.9860 0.9660 CV 57 16 11 17 33 Min Max Number of observations Countries with low risk sharing (DARS = 0 and DMRS = 0) Mean 48.0056 0.4285 0.8800 0.8227 0.5655 CV 0.24520 0.1165 0.1307 0.0847 0.2258 Min 25.9000 0.3740 0.7140 0.7280 0.3360 Max 66.7000 0.4940 0.9590 0.9490 0.7850 Number of observations 36 6 4 7 13 B. SPECIFICATIONOFTHEBASIC MODEL Impact on the level of health and on responsiveness (i) The level of health is measured by the Disability Adjusted Life Expectancy11 (DALE). We propose the following basic specification: Ln (80--DALE) = a1 + b1 Ln HEC + c1 Ln EDU + d1 DARS (1), where HEC refers to health expenditure per capita (in US$). EDU refers to the educational attainment in society, and is measured by enrolment in primary education of the relevant age group. The dependent variable is the logarithm of the difference between the observed DALE and a maximum of 80. With this specification, we say that these differences depend first upon overall resources for health. However, health status is not dependent only upon the activities in the health system. The variable EDU is therefore included among the determinants in equation (1) and is meant to capture the impact of overall social development on health. Both HEC and EDU are expected to raise DALE and so to be negatively related to the distance of DALE from the maximum. The last explanatory variable, DARS, is also expected to have a negative impact on the distance between the maximum of 80 and the observed DALE. We reason that generally health financing schemes with advanced risk sharing better define an adequate benefit package of health services to which citizens are entitled. The latter should increase the overall level of health in society. We submit that a better definition of the benefit package is the result of a greater stewardship role exercised by governments in view of the national importance of the health financing schemes. Alternative models are also tested. One tests whether social health insurance has a specific impact on the health level. A dummy variable DSHI, equals 1 when the country has a social health insurance scheme and 0 otherwise, will be added to the explanatory variables of equation (1). If we reason that, on average, general taxation and social health insurance schemes cover similar population groups with similar health 11 This summary measure of population health adjusts life expectancy at birth for the burden of disability. Disability weights are used to convert years lived in disability into equivalent years lived in good health. See further Mathers et al. (2000). 8 interventions,12 social health insurance should not do better or worse than general taxation; hence, we expect an effect that is not statistically different from zero. The second alternative model studies the marginal impact of a mixed health financing scheme. A dummy variable DMRS, equal to 1 when the country has a mixed health financing system and 0 otherwise, is included next to DARS. Our hypothesis is that the marginal impact of DMRS on Ln(80-DALE) is negative. Mixed health financing schemes also include health insurance schemes applying risk sharing and therefore should have a beneficial impact on health level attainment. In a third alternative model, we test whether certain groups of schemes within the overall set of mixed health financing systems would have an additional net effect on the level of health. We select the group of mixed systems that encompass health insurance schemes whereby only employees are covered (DMRS1 = 1 and 0 otherwise) and health insurance schemes that cover other specific groups only (DMRS2 = 1 and 0 otherwise). As these health insurance schemes offer a lower degree of financial risk protection, as compared with schemes that cover all employees and self-employed, the expected sign of the impact of DMRS1 and DMRS2 is positive. Fourth, we add both DSHI and DMRS to the explanatory variables of equation (1). Finally, we bring DSHI, DMRS, DMRS1 and DMRS2 together into the equation. (ii) The level of responsiveness is measured by an index (IR) that varies between 0 and 1, with 1 being the maximum. Two alternative functional forms are adopted: Ln [IR/(1- IR] = a21 + b21 HEC + c21 EDU + d21 DARS (2a) and Ln (1--IR) = a22 + b22 Ln HEC + c22 Ln EDU + d22 DARS (2b). Equation 2a has a logistic specification and ensures that the predicted values for IR stay within the 0­1 interval. In equation 2a, the impact of HEC is presumed to be positive, as more resources are likely to facilitate the responsiveness of health systems. In particular, the "client orientation" elements of responsiveness such as the quality of amenities and choice of provider can be expected to be especially resource-dependent. In the present case, EDU can be understood as capturing the positive effect of a literate and more developed society on the "respect for persons"; the autonomy of persons is especially likely to improve with a better education status. We hypothesize that advanced risk-sharing systems are associated with a larger degree of stewardship. The latter is likely to positively influence the mechanisms and incentives that entail a greater responsiveness. The coefficient of DARS is therefore expected to be positive. In equation 2b, the dependent variable is measured as the logarithm of the distance of IR from the maximum. In this specification, all coefficients but the intercept are expected to be negative. 12 See also Musgrove (1996, p.51) for a discussion of this issue. 9 As in the case of the health level, alternative models can be estimated. Using either type of functional forms, DSHI is expected to be neutral vis-à-vis responsiveness; we therefore expect a coefficient that is not statistically different from zero. In the logit form of the equation, DMRS is expected to exert a positive effect, whereas a negative impact is expected to be associated with DMRS1 and DMRS2. When using the second functional form for the dependent variable, the signs of the coefficients associated with DMRS, DMRS1 and DMRS2 are expected to be opposite that of the coefficients in the logit specification. Impact on the distributional measures of the goals The three measures considered are the index of fairness of financial contribution13 (IFFC), the index of distribution of responsiveness14 (IRD) and the index of equality of child survival (IECS).15 All indices vary between 0 and 1, with 1 corresponding to complete equality. The functional forms adopted for these dependent variables ensure that the predicted indices stay within the 0-1 interval. We first formulate models focusing on the effects of the degree of risk-sharing only. In the simplest equation we estimate the impact of the dummy variable (DARS). We have adopted the same functional forms as in equations 2a and 2b: Ln [Ij/(1- Ij)] = a31 + b31 DARS (3a) and Ln (1--Ij) = a32 + b32 DARS (3b). where Ij (j = 1,...,3) refers to the three above-mentioned indices, respectively. The effect of DARS on the indicator of fair financing is expected to be positive when using the logit form of the equation. In countries with advanced risk sharing, more so than in other systems, people pay financial contributions according to their capacity to pay. This then should be associated with a higher IFFC. Second, universal coverage systems are presumed to pay more attention to the objective of equal treatment for equal need. It is therefore assumed that such systems also respond to people's expectations as to the non-medical aspects of health care in a more equal way. Hence, the effect of DARS on the 13 This index measures how the health financing contribution (HFC) is distributed across households. HFC is composed of contributions that are implicitly paid via taxes (e.g., income taxes, value-added tax) for health, of explicit social health insurance contributions, premiums for private health insurance and of out-of-pocket payments. The IFFC is constructed in such a way that households that spend a very large share of income above subsistence are weighted more heavily. See further Murray et al. (2000). 14 The responsiveness inequality index is based on an assessment of the disadvantage with respect to responsiveness as experienced by different groups including poor people, women, old people and indigenous groups or minorities. The index accounts for the relative importance of these groups into total population. See further Valentine et al. (2000). 15 This index is based on data of expected survival time under age 5, themselves derived from child mortality distributions. In this index, the survival of each child under 5 is compared with that of all others. This index is used in the WHR 2000 as a measure of the distribution of health, pending more information on health inequality in the population at large. See further Gakidou and Murray (2000). 10 distribution of responsiveness is anticipated to be positive as well. Third, we postulate also that universal coverage systems are more apt than other systems to provide people with a similar benefit package, irrespective of their socioeconomic background. The variable DARS is therefore expected to exert a positive effect on the equality of child survival. When considering the second functional form, it stands to reason that the coefficients of DARS are expected to have the opposite sign. For alternative models, we first include DSHI as an additional dummy variable in equations 3a and 3b. The sign of the coefficients of DSHI is uncertain, however. Whether social health insurance is inferior or superior to general taxation in terms of fair financing, depends on a host of factors. The latter include the way health insurance contributions are levied (with an earnings ceiling or not), the progressivity of income taxes, the level of copayments and/or user fees, and the types of health services that are excluded from coverage and their prices. In general, when adding DMRS to the explanatory variables, we expect its effect to be positive and negative in the two functional forms, respectively. The effects of DMRS1 and DMRS2 are anticipated to be negative and positive in the case of the two functional forms, respectively. C. SPECIFICATIONOFENLARGEDMODELS The GINI index of income inequality in the equations for the distributional measures In one enlarged model, the GINI index measuring the distribution of income is included among the explanatory variables: Ln [Ij/(1- Ij)] = a41 + b41 GINI + c41 DARS (4a) and Ln (1--Ij) = a42 + b42 GINI + c42 DARS (4b). where Ij (j = 1,...,3) refers again to the three indices, respectively. Income inequality in society, as measured by the GINI, is expected to be mirrored, at least partially, in the distribution of the health financing burden on the various households. For instance, in equation 4a, it is expected that the larger the income inequality, the lower is the degree of fair financing. The coefficient b41 is therefore expected to be negative. In the case of equation 4b, a positive coefficient is predicted. We further anticipate that countries with advanced risk sharing are apt to counteract the initial effect of overall income inequality by introducing better financial risk protection for all of the population. Hence, we expect that the impact of DARS is maintained. Further variants of the basic equations 4a and 4b are investigated, via the inclusion of DSHI, DMRS, DMRS1 and DMRS2. In principle, there should be no change in the supposed direction of the effects 11 already commented upon earlier. In addition, the impact of interaction variables, combining the GINI index with the organizational dummy variables, can be studied. The coefficients of the interaction variables are expected to show that the larger the degree of risk sharing, the more the impact of the GINI index is offset. For instance, the coefficient of the interaction term between GINI and DARS is anticipated to be positive and negative, respectively. The impact of the ratio of public health expenditure to total health expenditure on the health system attainment indicators The various models considered so far measure the average impact of the different risk-sharing schemes on the attainment indicators. Enlarged models with the inclusion of interaction variables between the ratio of public health expenditure to total health expenditure (PHE%) and the organizational dummy variables among the determinants can also be considered. We expect that a higher PHE% would reinforce the effect of the organizational variables in the earlier models. The more health expenditure is managed through the public sector, and thus the higher the degree of risk pooling, the larger the equality of people within the health system is presumed to be. The basic equations are the following: Ln (80--DALE) = a51 + b51 Ln HEC + c51 Ln EDU + d51 DARS + e51 DARS*PHE% (5a) Ln [IR/(1- IR] = a52 + b52HEC + c52 EDU + d52 DARS + e52 DARS*PHE% (5b) and Ln (1--IR) = a53 + b53 Ln HEC + c53 Ln EDU + d53 DARS + e53 DARS*PHE% (5c). Ln [Ij/(1- Ij)] = a54 + b54 DARS + c54 DARS*PHE% (5d) and Ln (1--Ij) = a55 + b55 DARS + c55 DARS*PHE% (5e). where Ij (j = 1,...,3) refers to the three equality indices, respectively. The coefficients e51 and e53 are expected to be negative. The coefficient e52 is anticipated to have a positive sign. The coefficients c54 and c55 are expected to be positive and negative, respectively. Note that in alternative equations, we also investigate the interaction of PHE% with DSHI, DMRS, DMRS1 and DMRS2. 12 D. RESULTS Estimation results for the basic model The equations have been estimated with the ordinary least squares method, using data for the explanatory variables HEC, EDU and PHE% that pertain to the year 1997. The GINI index pertains to specific years, depending upon the country, within the period 1986-1999. The data and their sources are presented in Appendix I. Different sample sizes were used: the full samples (using all available observations), restricted samples (deleting observations of countries with uncertain risk-sharing classification), and more restricted samples (previously defined restricted samples but with additional deletion of influential data16). The results of the regressions run with the different sample sizes are presented in Appendixes C, D, and E, respectively. The results concerning the level of health (DALE) with the full sample are presented in Table 1 of Appendix C. In all models, the effects of DARS, HEC and EDU are as expected and are statistically significant at the 1 percent significance level. The other organizational dummy variables do not show a significant impact. Using the adjusted R2 , regression 2 is the best. However, using the Akaike criterion,17 regression 1 is preferred. When using the restricted sample (Table 1 of Appendix D), we obtain similar results with DARS, HEC and EDU showing statistically significant coefficients. Regression 1 is the best according to the Akaike criterion. From the estimates related to the level of responsiveness (IR), in Tables 2a and 2b of Appendix C, we see that HEC and EDU do not have a statistically significant impact. One major reason is likely to be that the index of responsiveness contains both elements of respect for persons and client orientation, and that these are influenced differently by HEC and EDU. For instance, HEC may be important in explaining client orientation, whereas it may not be when explaining respect for persons. Therefore, when analyzing the determinants of the overall index of responsiveness, the effect of HEC may disappear. The results also show mixed results for the statistical significance of the coefficient of DARS. The adjusted R2 and the Akaike criterion point each time at regression 5 as the best one. This regression includes DARS, DSHI and DMRS as explanatory variables. Both the coefficients of DARS and DMRS have the expected sign in both sets of equations. In regression 5 of Table 2a, the coefficient of DSHI is not statistically significant, as expected. Still, this particular coefficient becomes significant when using the other functional form for the dependent variable. The number of countries with universal coverage in the sample is small (8), and values of IR for specific countries may well heavily influence the regression results. For example, the deletion of data for Bulgaria, which has SHI and is characterized by a relatively low level of IR, renders the coefficient of DSHI statistically insignificant at the 10 percent level in both functional forms. We refer to the regression results presented in Tables 1a and 1b in Appendix E. Using this particular restricted sample, and the logit specification, regression 5 is preferred according to the adjusted R2 and Akaike criteria. In those regressions, the coefficients of DARS and DMRS are significant at the 1 percent and 5 percent level, respectively. In the case of the second functional form, regression 5 is preferred according to both the 16Mukherjee, White and Wuyts (1998, p.138) refer to influential data as points that pull the regression line in their direction. Influential data are not necessarily associated, however, with outliers (large residuals). 17See for instance Greene (2000, p.306). 13 adjusted R2 and Akaike criterion. In this regression, the coefficients of DARS and DMRS are also significant at the 1 percent and 5 percent level, respectively. An additional regression analysis was undertaken with the subresponsiveness indices "respect for persons" (RESPECT) and "client orientation" (CO) as dependent variables. We present only the best equations (according to the Akaike criterion) in Table 1 of Appendix H. There are no statistically significant effects of HEC and EDU in the equations for "respect for persons". However, in the logit regression for "client orientation", HEC becomes statistically significant. The coefficients of DARS and DRMS are statistically different from zero, except for the coefficient of DMRS in regression 2 for client orientation. In regression 2 for RESPECT and CO, the coefficient of DSHI proves to be statistically significant. However, the latter result is no longer maintained after deleting data for Bulgaria from the sample; see Table 2 of Appendix H where the best results are presented. In addition, the impact of HEC now becomes statistically insignificant in all four regressions. The full sample results related to the index of fair financing (IFFC) are presented in Tables 3a and 3b of Appendix C. The explanatory power of the regressions is minimal: none of the explanatory variables has a statistically significant impact on the IFFC. The same results are obtained when using the restricted samples (Tables 2a and 2b of Appendix D). We submit that the major reason for these unsatisfactory results is the relatively small sample size. Moreover, the sample did not include sufficient data on countries with advanced and with low-risk sharing. For instance, the (full sample) data on advanced risk sharing are those of Bulgaria, Jamaica, Kyrgyzstan, Romania and Russia, and do inadequately reflect the experience of high-income countries with either social health insurance or general taxation financing. Estimates for the distribution of responsiveness (IRD) with the full sample are presented in Tables 4a and 4b of Appendix C. In both sets of equations the coefficients of DARS and DSHI are statistically significant. The impact of DSHI is against our expectations. The number of countries with universal coverage in the sample is quite small (9), and values of IRD for specific countries may influence the regression results. For example, when we delete data for Chile and Poland, that have SHI, and that are characterized by relatively low IRD, the coefficient of DSHI becomes statistically insignificant at the 10 percent level in both functional forms. Still, the coefficients of DARS all remain significant at the 1 percent level. These regression results are presented in Tables 2a and 2b in Appendix D. The full sample results for the index of equality of child survival (IECS), in Tables 5a and 5b of Appendix C show that both DARS and DMRS have statistically significant impacts in several of the regressions. We also notice that the coefficient of DSHI is statistically significant in regressions 2, 5 and 6. Similar results are obtained when using the restricted sample; see Tables 3a and 3b of Appendix D. Again the number of countries with universal coverage in both the full and restricted samples is small, namely 7. One country, Uzbekistan (with a GT health financing system), has a particularly low value for IECS18. When we delete this country's data from the sample, the statistically significant effect of DSHI disappears; we refer to the regression results in Tables 3a and 3b of Appendix E. According to the Akaike criterion and the adjusted R2, regression 4 is the best for both functional forms. The coefficients all have the expected sign. DARS and DMRS are statistically significant at the 1 percent and 5 percent level, respectively. 18The IECS of Uzbekistan is 0.632. 14 Estimation results with the GINI index as an explanatory variable in the equations for the distributional measures For the estimation of the enlarged model (equations 4a and 4b), we have used the restricted samples only. We will only present the "best" equations according to the adjusted R2 and/or the Akaike criterion. We first refer to Table 1 of Appendix F. In both functional forms of the fair financing equation (IFFC), the coefficients of the GINI index19 have the anticipated sign but are not statistically significant. The coefficients of DARS are also not statistically significant. Both equations have very low explanatory power. Related to the distribution of responsiveness (IRD), both functional forms show significant impacts of both DARS and DMRS, as well as of the GINI index. All coefficients have the expected sign. One can conclude that these risk-sharing arrangements are efficient in counterbalancing the overall effect of income inequality. A threshold for the GINI indices can be computed, indicating the value above which risk sharing is no longer able to counteract the effect of overall income inequality. In the case of a country with an advanced risk-sharing scheme, the threshold value is between 56.920 and 57.9.21 In the case of medium risk-sharing schemes, the threshold is between 25.622 and 26.323. From these estimates, one can infer that advanced risk-sharing schemes are more effective in counteracting the effects of overall income inequality in society. For example, let us assume that a country has a GINI of 35. If this country has an advanced risk-sharing scheme, its effect will outweigh the impact of income equality. Using the regression estimates for the first functional form, the combined effect will be +0.8588 . However, if the country has a medium- 24 risk sharing arrangement, the combined effect will be ­0.3252 25. Note that these results are sensitive, however, to the exclusion of values for specific countries. For instance, using the more restricted sample (thereby excluding the data for Poland and Chile26), the coefficients of DARS and DMRS remain statistically significant. However, the coefficients of GINI are no longer statistically significant at the 10 percent level. In the regression results related to the inequality of child survival (IECS), the sign of the GINI coefficients is against our expectations. Surprisingly, the coefficient of GINI is also statistically significant at the 10 percent level, at least in the first functional form. In the second functional form, the coefficient of GINI is not statistically different from zero, however. The coefficients of DARS have the anticipated sign, however, and are both statistically significant at the 1 percent level. 19In the regressions, these were entered in percentage terms. 20Derived from the equation (2nd functional form): 0.0352*GINI - 2.0025 = 0. 21Derived from the equation (logit specification): -0.0375*GINI + 2.1713 = 0. 22Derived from the equation (2nd functional form): 0.0352*GINI - 0.8994 = 0. 23Derived from the equation (logit specification): -0.0375*GINI + 0.9873 = 0. 24+0.8588=2.1713--0.0375*35 25-0.3252=0.9873--0.0375*35 26Chile and Poland have lower values for IRD than other countries, namely 0.918 and 0.970, respectively. 15 Specifications were tested with interaction terms between the GINI and the organizational dummy variables. There is no general improvement in the regression results. In most of the equations, the coefficient associated with the GINI index loses its statistical significance. In addition, the coefficients associated with the interaction between GINI and the organizational variables frequently have opposite signs to what is expected. These results are therefore not presented or commented upon further. Estimation results when using interaction terms with the ratio of public health expenditure to total health expenditure Inclusion of the interaction variables with PHE% in equations 5a to 5e, and using the restricted samples, did not result in a general improvement of the estimation results. For instance, in a number of cases, the coefficients of DARS have the correct sign but are statistically insignificant. In other instances, the coefficient of DARS has a negative sign. One reason is likely to be multicollinearity; the correlation coefficient between DARS and DARS*PHE% was 0.9678, whereas the correlation between DMRS and DMRS*PHE% was 0.9165. The subsequent use of DARS*PHE% together with DARS, GINI and GINI*DARS gave unattractive results as well. Further estimations were done with transformed interaction variables. In the case of the interaction between DARS and PHE%, the variable constructed was DARS*(PHE%-0.5). The coefficient associated with this variable reveals the impact of the difference between PHE% and a threshold of 50 percent. The advantage of using this variable was that it reduced the correlation with DARS; the correlation coefficient now becomes 0.7545. The results for IR, IFFC, IRD and IECS are not satisfactory: the coefficient of the new interaction variable has a wrong sign, is not statistically significant, or both. Only in the case of DALE did we obtain a satisfactory result: both the coefficients of DARS and the interaction variable have the expected sign and are statistically significant. The latter is presented in Table 2 in Appendix F. In other words, for those advanced risk-sharing systems with a PHE% above 50 percent, the level of PHE% reinforces the "average" effect of DARS. For instance, in the case of Oman with a PHE% of 63.31 percent, the combined impact of DARS and DARS*(PHE% - 0.50) becomes -0.2694. For those countries with a PHE% below 50 percent (Chile, Republic of Korea, Brunei Darussalam and United Arab Emirates), the initial effect of DARS is weakened. For instance, for Chile with a PHE% of 40.10 percent, the combined effect of DARS and DARS*(PHE% - 0.50) on the dependent variable becomes ­0.1637. Key conclusions A first conclusion from the estimates is that the degree of advanced risk sharing, as measured by the dummy variable DARS, is significant in the equations for four of the five goal measurements. No impact could be found in the case of the index of fair financing, but we submit this is due to the inadequate sample. In addition, in at least two of these measurements (level of responsiveness, distribution of health), the variable DMRS also has been shown to have a statistically significant impact. Second, when enlarging the set of explanatory variables in the models for the distributional measures with the GINI index, DARS remains statistically significant in the equations for IRD and IECS. In addition, DMRS has a statistically significant impact in the equations for IRD. An additional interpretation emerges from the results, namely that risk sharing corrects for, or may even outweigh, the negative effect of overall income inequality on the fair financing index and the index of distribution of responsiveness. 16 Third, using interaction terms with PHE% leads to plausible results for DALE only: the level of PHE% reinforces the average positive effect of advanced risk sharing. Preliminary analysis with updated data Since publication of the WHR 2000, WHO has developed updated estimates for the level (HEC) and share of public health expenditure in total health expenditure (PHE%). When using updated data for HEC in the equations for DALE and IR, results (in terms of explanatory power, sign and statistical significance of coefficients) similar to those presented here are obtained. The use of the updated PHE% does not significantly change the estimates for the equations with the interaction terms. Estimates of the index of fair financing (IFFC) were also obtained for an additional 30 countries. Reestimation of the equations using an enlarged sample of 50, now leads to two interesting results: (i) the advanced risk-sharing dummy variable DARS exerts a statistically significant effect on the fair financing index; (ii) the GINI index has a statistically significant impact on IFFC but is counterbalanced by a health financing system characterized by advanced risk-sharing. These preliminary results prove to be more in line with those obtained for the other distributional measures. D. COMMUNITY RISK-SHARING ARRANGEMENTS: FURTHER NEED TO MEASURE THEIR IMPACT Community-risk sharing arrangements are increasingly recognized as an intermediate response to the constraints that many countries experience to rapidly extending financial risk protection to the national population. A body of research exists with respect to community financing arrangements and their functioning within communities, districts or regions. Information at the national level is clearly lacking. We have made an attempt to scan the literature and other sources,27 to see whether community risk-sharing organizations exist at country level. We refer to Table 4 of Appendix I, where we divided countries into an "information" and "no information" subcategory. Only countries with low to medium risk sharing will be considered, as countries with advanced risk sharing in principle do not need to be complemented by community risk-sharing schemes. We recorded that in the set of countries with a public health expenditure ratio of 50 to 75 percent, 25 out of 44 countries have community risk-sharing schemes operating. In the countries with a ratio below 50 percent, 42 out of 58 are reported to have such schemes. This is not unexpected, as we would expect community risk-sharing schemes to be established where governments are not able to make sufficient advance in risk protection. However, these data are insufficient for econometric analysis. Further work is needed about the quantitative importance of community risk-sharing arrangements at the country level. The latter could be measured by the number of risk-sharing schemes and the percentage of population covered by such schemes. Alternatively, one could measure the ratio of the expenditures incurred by such schemes to overall private health expenditure. The higher this ratio, the greater is the effort to share risks. Current work on National Health Accounts at WHO goes into this direction, by attempting to collect data on 27Especially Atim (1998), Bennett, Creese and Monasch (1998), Carrin, De Graeve and Devillé (1999), ILO and PAHO (1999) and Ginneken van (1999). 17 expenditure by nongovernmental institutions and communities. Further work is needed on identifying the part of this expenditure that is spent within the framework of risk-sharing arrangements. VI. CONCLUDING REMARKS The results presented give empirical support for the hypothesis that the degree of risk sharing in health financing organizations matters for health system attainment, as measured by the five indicators. Especially the categorical variables indicating whether a country has a health financing organization with advanced or medium risk-sharing categories, are seen to have a significant impact. These effects prove to be quite robust, after introducing the GINI index among the explanatory variables in the models for the distributional measures. We noted that the plausibility of the results improves when using the restricted samples, deleting data for those countries whose classification was considered uncertain. Further information will be necessary to address this uncertainty. In general, final data for larger samples of countries are welcome for four of the health system attainment indices, especially for the index of fair financing contribution (IFFC), so that these better reflect the patterns of risk sharing in the world. In the current samples, some of the risk-sharing schemes are underrepresented, which has entailed sensitivity of the results to specific data points. Further work could also be done on designing much more refined quantitative measures for the degree of risk sharing. Indeed, within each of the categories of health financing organization that we considered, a further variety in the degree of financial protection of different population subgroups may well be present. In addition, more work needs to be undertaken to measure the quantitative importance of risk-sharing schemes for communities and the informal sector at the country level as well as their depth of risk sharing. Only then can further econometric analysis be undertaken. In the meantime, given the empirical results obtained so far, one can clearly hypothesize beneficial impacts of these schemes on the health system attainment indicators. 18 19 VII. APPENDIXES APPENDIX A Classification Tables 21 Table 1: classification of Countries by degree of risk sharing in the health financing system Advanced risk sharing Medium risk sharing Low risk sharing Social Health General taxation All employees and self-employed All employees covered by Specific groups only covered Afghanistan insurance (with some exclusions) covered health insurance by health insurance Angola (SHI) by health insurance Armenia Australia Albania Colombia Algeria Botswana Bahamas Austria Antigua-Barbuda Ecuador Andorra Brazil Bangladesh Belgium Azerbaijan El Salvador Argentina Burkina Faso Benin Bulgaria Bahrain Equatorial Guinea Bolivia Burundi Bhutan Chile Barbados Libya Cape Verde Cameroon Cambodia Costa Rica Belarus Mongolia Congo China Central African Republic Croatia Belize Peru Egypt Côte d'Ivoire Czech Republic Bosnia and Tunisia Gabon Dominican Republic Chad Estonia Herzegovina Uruguay Guinea Guatemala Comoros France Honduras Guinea-Bissau Germany Brunei Darussalam Lebanon Haiti D. R. of Congo Greece Canada Mali India Hungary Cook Islands Mexico Indonesia Djibouti Israel Cuba Namibia Iran Eritrea Japan Cyprus Panama Iraq Latvia D. P.'s R. of Korea Paraguay Jordan Ethiopia Lithuania Denmark Philippines Kenya Fiji Luxembourg Dominica Senegal Lesotho Monaco Finland Turkey Madagascar Gambia Netherlands Iceland Venezuela Mauritania Georgia Norway Ireland Morocco Poland Italy Mozambique Ghana Republic of Korea Jamaica Myanmar Romania Grenada Kazakhstan Nicaragua San Marino Kuwait Niger Guyana Slovakia Kyrgyzstan Pakistan Slovenia Malaysia South Africa Kiribati Switzerland Malta Thailand Lao People's D. R. The F. Y. of Mauritius Trinidad and Tobago Macedonia New Zealand United States of America Liberia Yugoslavia Niue Viet Nam Malawi Oman Yemen Palau Maldives Portugal Marshall Islands Qatar Micronesia Nauru Nepal Nigeria Papua New Guinea Rwanda Sao Tome and Principe Sierra Leone Solomon Islands 22 Table 1 (continued): Classification of Countries by degree of risk sharing in the health financing system Advanced risk sharing Medium risk sharing Low risk sharing Social Health General taxation All employees and self- All employees Specific groups only covered by Somalia insurance employed (with some covered by health health insurance Sri Lanka (SHI) exclusions) covered by insurance Sudan health insurance Suriname Republic of Moldova Swaziland Russia Syrian Arab Republic Saint Kitts and Nevis Togo Saint Lucia Tonga Saint Vincent A. T. G. Tuvalu Samoa Uganda Saudi Arabia United Republic of Tanzania Seychelles Vanuatu Singapore Zambia Spain Zimbabwe Sweden Tajikistan Turkmenistan Ukraine United Arab Emirates United kingdom Uzbekistan 23 Table 2: Classification of countries by type of health financing system and by the share of public health expenditure in total health expenditure1 Advanced risk sharing Medium risk sharing Low risk sharing Public health All employees and self- All employees covered by Specific groups only covered expenditure as a Social Health insurance General Taxation employed (with some health insurance by health insurance percentage of total (SHI) exclusions) covered by health expenditure health insurance Belgium Albania Mongolia Andorra Guinea-Bissau Chad 75% to 100% Bulgaria Azerbaijan Guyana Costa Rica Belarus Kiribati Croatia Bosnia and Herzegovina Micronesia Czech Republic Cook Islands Nauru Estonia Cuba Papua New Guinea France D. P.'s R. of Korea Sao Tome and Principe Germany Denmark Solomon Islands Hungary Ice land Tuvalu Israel Ireland Japan Kuwait Lithuania Niue Luxembourg Palau Norway Republic of Moldava Slovakia Russia Slovenia The F. Y. of Macedonia Samoa Saudi Arabia Sweden Seychelles Tajikistan Turkmenistan Ukraine United Kingdom Uzbekistan Notes: Shares of public health expenditure in total health expenditure are taken from the World Health Report (WHO, 2000). 1 24 Table 2 (continued): Classification of countries by type of health financing system and by the share of public health expenditure in total health expenditure Advanced risk sharing Medium risk sharing Low risk sharing All employees and self- All employees covered by Specific groups only covered Public health Social Health insurance General Taxation employed (with some health insurance by health insurance expenditure as a (SHI) exclusions) covered by percentage of total health insurance health expenditure Australia Antigua-Barbuda Colombia Algeria Botswana Bhutan 50% to 75% Austria Bahrain Ecuador Argentina Guatemala Central African Republic Greece Barbados Equatorial Guinea Bolivia Iraq Comoros Latvia Belize Cape Verde Jordan Eritrea Monaco Canada Libya Gabon Kenya Fiji Netherlands Dominica Guinea Lesotho Grenada Poland Finland Namibia Madagascar Lao people's D. R. Romania Italy Panama Mozambique Liberia San Marino Jamaica Senegal Nicaragua Malawi Switzerland Kazakhstan Turkey Trinidad and Tobago Maldives Yugoslavia Kyrgyzstan Venezuela Marshall Islands Malaysia Rwanda Malta Somalia Mauritius Swaziland New Zealand United Rep. of Tanzania Oman Vanuatu Portugal Qatar Saint Kitts and Nevis Saint Lucia Saint Vincent Spain 25 Table 2 (continued): Classification of countries by type of health financing system and by the share of public health expenditure in total health expenditure Advanced risk sharing Medium risk sharing Low risk sharing Public health All employees and self- All employees covered by Specific groups only covered expenditure as a Social Health insurance General Taxation employed (with some health insurance by health insurance percentage of total (SHI) exclusions) covered by health expenditure health insurance Chile Brunei Darussalam El Salvador Congo Brazil Afghanistan <50% Republic of Korea Cyprus Peru Egypt Burkina Faso Angola Singapore Tunisia Honduras Burundi Armenia United Arab Emirates Uruguay Lebanon Cameroon Bahamas Mali China Bangladesh Mexico Côte d'Ivoire Benin Paraguay Dominican Republic Cambodia Philippines Haiti D. R. of Congo India Djibouti Indonesia Ethiopia Iran Gambia Mauritania Georgia Morocco Ghana Myanmar Nepal Niger Nigeria Pakistan Sierra Leone South Africa Sri Lanka Thailand Sudan United States of America Suriname Viet Nam Syrian Arab Republic Yemen Togo Tonga Uganda Zambia Zimbabwe 26 Table 3: Classification of countries by type of health financing system and by income1 group Advanced risk sharing Medium risk sharing Low risk sharing Income level All employees and self- All employees covered by Specific groups only covered Social Health insurance General Taxation employed (with some health insurance by health insurance (SHI) exclusions) covered by health insurance Low Income Azerbaijan Mongolia Congo Burkina Faso Afghanistan Bosnia and Herzegovina Guinea Burundi Angola $ 760 or less Kyrgyzstan Honduras Cameroon Armenia Republic of Moldova Mali China Bangladesh Senegal Côte d'Ivoire Benin Guinea-Bissau Bhutan Haiti Cambodia India Central African Republic Indonesia Chad Kenya Comoros Lesotho Democratic Rep of Congo Madagascar Djibouti Mauritania Eritrea Mozambique Ethiopia Myanmar Gambia Nicaragua Ghana Niger Kiribati Pakistan Lao people's D. R. Viet Nam Liberia Yemen Malawi Micronesia Nepal Nigeria Rwanda Sao Tome and Principe Sierra Leone Solomon Islands Somalia Notes: Income groups are defined according to 1998 GDP per capita in US dollars (World Bank, 2000) 1 27 Table 3 (continued): Classification of countries by type of health financing system and by income group Advanced risk sharing Medium risk sharing Low risk sharing Income level All employees and self- All employees covered by Specific groups only covered Social Health insurance General Taxation employed (with some health insurance by health insurance (SHI) exclusions) covered by health insurance Low Income Tajikistan Sudan Turkmenistan Togo $ 760 or less Tonga Uganda United Republic of Tanzania Zambia Zimbabwe Lower-middle income Bulgaria Albania Colombia Cape Verde Dominican Republic Fiji Costa Rica Belarus Ecuador Algeria Guatemala Georgia $ 761 to $ 3030 Lithuania Belize El Salvador Bolivia Iran Guyana Romania Cuba Equatorial Guinea Paraguay Iraq Maldives The F. Y. of Macedonia D. P.'s R. of Korea Peru Egypt Jordan Marshall Islands Yugoslavia Dominica Tunisia Namibia Morocco Papua New Guinea Jamaica Philippines Thailand Sri Lanka Kazakhstan Suriname Niue Syrian Arab Republic Russia Swaziland Saint Vincent and the Grenadines Vanuatu Samoa Ukraine Uzbekistan 28 Table 3 (continued): Classification of countries by type of health financing system and by income group Advanced risk sharing Medium risk sharing Low risk sharing Income level All employees and self- All employees covered by Specific groups only covered Social Health insurance General Taxation employed (with some health insurance by health insurance exclusions) covered by health insurance Upper-middle income Chile Antigua-Barbuda Libya Argentina Botswana Grenada Croatia Bahrain Uruguay Gabon Brazil Nauru $ 3031 to $ 9630 Czech Republic Barbados Lebanon Trinidad and Tobago Estonia Cook Islands Mexico South Africa Hungary Malaysia Panama Poland Mauritius Turkey Republic of Korea Oman Venezuela Slovakia Palau Saint Kitts and Nevis Saint Lucia Saudi Arabia Seychelles High Income Australia Brunei Darussalam Andorra United States of America Bahamas Austria Canada Tuvalu $ 9361 or more Belgium Cyprus France Denmark Germany Finland Greece Iceland Israel Ireland Japan Italy Latvia Kuwait Luxembourg Malta Monaco New Zealand Netherlands Portugal Norway Qatar San Marino Singapore Slovenia Spain Switzerland Sweden United Arab Emirates United Kingdom 29 Table 4: Classification of countries by health financing system and by the share of public expenditure in total health expenditure1, and by the information on community financing Medium risk sharing Low risk sharing All employees and self- All employees covered by Specific groups only covered Public health employed (with some health insurance by health insurance expenditure as exclusions) covered by health a percentage of insurance Community financing Community total health Information financing expenditure Community Community Community Community Community Community No financing financing financing financing financing financing Information Information No Information No Information No Information Information Information Mongolia Andorra Guinea Bissau Chad Kiribati Guyana Micronesia 75% to 100% Papua New Guinea Nauru Sao Tome and Principe Solomon Islands Tuvalu Notes: Shares of public health expenditure in total health expenditure are taken from the WHR 2000. 1 30 Table 4 (continued): Classification of countries by health financing system and by the share of public expenditure in total health expenditure and by the information on community financing Medium risk sharing Low risk sharing All employees and self- All employees covered by Specific groups only covered by Public health employed (with some health insurance health insurance expenditure as exclusions) covered by health a percentage of insurance Community financing Community total health Information financing expenditure Community Communit Community Community Community Community No financing y financing financing financing financing financing Information Information No Information No Information No Informatio Information Information n Ecuador Libya Argentina Algeria Guatemala Botswana Comoros Bhutan Equatorial Bolivia Cape Verde Jordan Iraq Grenada 50% to 75% Central African R. Guinea Colombia Gabon Kenya Lesotho Malawi Guinea Namibia Madagascar U. R. of Tanzania Eritrea Panama Turkey Mozambique Fiji Senegal Nicaragua Lao People's D. R. Venezuela Trinidad and Liberia Tobago Maldives Marshall Islands Rwanda Somalia Swaziland Vanuatu 31 Table 4 (continued): Classification of countries by health financing system and by the share of public expenditure in total health expenditure, and by the information on community financing Medium risk sharing Low risk sharing All employees and self-employed All employees covered by health Specific groups only covered by health Public health (with some exclusions) covered insurance insurance expenditure as by health insurance a percentage of Community Community total health Community Community Community Community Community financing Communit financing financing expenditure financing financing financing financing Information y financing Information No Information No Information No No Information Information Information Informatio n <50% El Salvador Tunisia Congo Egypt Brazil Mauritania Bahamas Afghanistan Peru Honduras Lebanon Burkina Faso Morocco Bangladesh Angola Uruguay Mali Burundi Pakistan Benin Armenia Mexico Cameroon Yemen Cambodia Gambia Paraguay China D. R. of Congo Georgia Philippines Côte d'Ivoire Djibouti Sierra Leone Dominican Republic Ethiopia Sudan Haiti Ghana Syrian A. R. India Nepal Tonga Indonesia Nigeria Iran Sri Lanka Myanmar Suriname Niger Togo South Africa Uganda Thailand Zambia United States of Zimbabwe America Viet Nam 32 APPENDIX B Histograms and Descriptive Statistics of Health System Attainment Indicators 31 25 20 Series: DALE Sample1191 Observations191(fullsample) 15 Mean 56.82618 Median 60.50000 Maximum 74.50000 10 Minimum 25.90000 Std. Dev. 12.30564 Skewness -0.722048 Kurtosis 2.376146 5 Jarque-Bera 19.69375 Probability 0.000053 0 25 30 35 40 45 50 55 60 65 70 75 32 6 5 Series: IR Sample14191 Observations30(fullsample) 4 Mean 0.516500 Median 0.519000 3 Maximum 0.688000 Minimum 0.374000 Std. Dev. 0.079667 2 Skewness 0.098158 Kurtosis 2.325257 1 Jarque-Bera 0.617273 Probability 0.734448 0 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 33 8 Series:IFFC 6 Sample14190 Observations21(fullsample) Mean 0.872952 Median 0.903000 4 Maximum 0.992000 Minimum 0.623000 Std. Dev. 0.105047 Skewness -1.173836 2 Kurtosis 3.418909 Jarque-Bera 4.976166 Probability 0.083069 0 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00 34 10 8 Series: IRD Sample 14 191 Observations 33 (full sample) 6 Mean 0.896724 Median 0.914000 Maximum 0.999900 Minimum 0.723000 4 Std. Dev. 0.086934 Skewness -0.545552 Kurtosis 1.992005 2 Jarque-Bera 3.034020 Probability 0.219367 0 0.70 0.75 0.80 0.85 0.90 0.95 1.00 35 8 Series: IECS Sample14191 6 Observations 58 (full sample) Mean 0.665948 Median 0.648000 4 Maximum 0.999000 Minimum 0.245000 Std.Dev. 0.191684 Skewness -0.065030 Kurtosis 2.522016 2 Jarque-Bera 0.593013 Probability 0.743411 0 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 36 6 Series: IFFC Sample 14 190 5 Obs. 19 (restricted sample) 4 Mean 0.872053 Median 0.903000 3 Maximum 0.992000 Minimum 0.623000 Std.Dev. 0.107470 2 Skewness -1.210402 Kurtosis 3.406341 1 Jarque-Bera 4.770114 Probability 0.092084 0 0.60 0.65 0.70 0.75 0.80 0.85 0.90 0.95 1.00 37 8 Series: IECS Sample 14 191 Obs. 52 (restricted sample) 6 Mean 0.684269 Median 0.657000 4 Maximum 0.999000 Minimum 0.261000 Std.Dev. 0.180362 Skewness 0.022189 2 Kurtosis 2.528239 Jarque-Bera 0.486477 Probability 0.784084 0 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 38 APPENDIX C Regression Results with Full Samples 39 Table 1: Regression results on DALE1 Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant 4.9490 4.9321 4.9548 4.8910 4.9379 4.8725 (0.2964) (0.2956) (0.2972) (0.3043) (0.2964) (0.3035) (16.6978) (16.6825) (16.6699) (-9.4826) (16.6570) (16.0564) HEC -0.1936 -0.1929 -0.1907 -0.1884 -0.1900 -0.1876 (0.0191) (0.0190) (0.0197) (0.0199) (0.0196) (0.0198) (-10.1390) (-10.1365) (-9.6951) (-9.4826) (-9.6873) (-9.4727) EDU -0.2121 -0.2087 -0.2102 -0.1967 -0.2068 -0.1928 (0.0758) (0.0756) (0.0761) (0.0774) (0.0759) (0.0772) (-2.7968) (-2.7598) (-2.7637) (-2.5412) (-2.7258) (-2.4993) DARS -0.2969 -0.2554 -0.3291 -0.3418 -0.2883 -0.3008 (0.0633) (0.0699) (0.0819) (0.0831) (0.0868) (0.0879) (-4.6922) (-3.6531) (-4.0161) (-4.1109) (-3.3217) (-3.4230) DSHI -0.1031 -0.1038 -0.1052 (0.0749) (0.0751) (0.0753) (-1.3769) (-1.3834) (-1.3966) DMRS -0.0377 -0.0831 -0.0390 -0.0858 (0.0609) (0.1043) (0.0607) (0.1039) (-0.6197) (-0.7968) (-0.6428) (-0.8260) DMRS1 0.0004 0.0012 (0.1071) (0.1067) (0.0039) (0.0113) DMRS2 0.0790 0.0811 (0.1027) (0.1024) (0.7687) (0.7921) R-squared 0.7995 0.8023 0.8000 0.8019 0.8029 0.8048 Adjusted R-squared 0.7949 0.7963 0.7939 0.7926 0.7954 0.7942 S.E. of regression 0.2599 0.2590 0.2605 0.2613 0.2596 0.2603 Ak. Info criterion 0.1717 0.1720 0.1834 0.2037 0.1835 0.2033 Sample size 136 136 136 136 136 136 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 40 Table 2a: Regression results1 on the level of responsiveness (Logit) Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -0.3865 -0.4238 -0.4509 -0.4575 -0.4893 -0.4985 (0.2514) (0.2456) (0.2392) (0.2523) (0.2312) (0.2442) (-1.5374) (-1.7255) (-1.8851) (-1.8133) (-2.1169) (-2.0413) HEC 0.0004 0.0003 0.0004 0.0003 0.0002 0.0002 (0.0003) (0.0003) (0.0003) (0.0003) (0.0003) (0.0003) (1.4829) (1.0452) (1.2662) (1.1605) (0.8084) (0.7077) EDU 0.0040 0.0046 0.0026 0.0027 0.0031 0.0033 (0.0029) (0.0028) (0.0028) (0.0030) (0.0027) (0.0029) (1.3868) (1.6103) (0.9178) (0.8985) (1.1479) (1.1273) DARS 0.1380 0.3946 0.3696 0.3708 0.6328 0.6381 (0.1469) (0.2201) (0.1814) (0.1905) (0.2351) (0.2475) (0.9395) (1.7930) (2.0370) (1.9661) (2.6911) (2.5780) DSHI -0.3397 -0.3452 -0.3492 (0.2217) (0.2067) (0.2171) (-1.5321) (-1.6697) (-1.6081) DMRS 0.2517 0.2272 0.2543 0.2245 (0.1275) (0.1892) (0.1226) (0.1821) (1.9743) (1.2008) (2.0744) (1.2331) DMRS1 0.0151 0.0129 (0.1908) (0.1836) (0.0792) (0.0700) DMRS2 0.0361 0.0465 (0.1678) (0.1617) (0.2151) (0.2876) R-squared 0.3342 0.3984 0.4344 0.4359 0.5007 0.5034 Adjusted R-squared 0.2473 0.2890 0.3315 0.2666 0.3818 0.3205 S.E. of regression 0.2519 0.2448 0.2374 0.2486 0.2283 0.2393 Ak. Info criterion 0.2163 0.1890 0.1273 0.2728 0.0767 0.2193 Sample size 27 27 27 27 27 27 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 41 Table 2b: Regression results1 on the level of responsiveness Log [1­IR] Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant 0.2387 -0.1483 -0.2327 -0.2033 -0.1397 -0.1027 (0.4554) (0.4341) (0.4359) (0.4651) (0.4091) (0.4368) (-0.5243) (0.3417) (0.5338) (-0.4371) (-0.3414) (-0.2351) HEC -0.0299 -0.0237 -0.0141 -0.0119 -0.0073 -0.0045 (0.0267) (0.0255) (0.0271) (0.0291) (0.0254) (0.0274) (-1.1205) (-0.9302) (-0.5226) (-0.4082) (-0.2873) (-0.1631) EDU -0.0780 -0.1038 -0.0716 -0.0800 -0.0979 -0.1085 (0.1122) (0.1071) (0.1074) (0.1152) (0.1010) (0.1085) (-0.6954) (-0.9690) (-0.6663) (-0.6938) (-0.9693) (-1.0001) DARS -0.1004 -0.2486 -0.2257 -0.2292 -0.3782 -0.3840 (0.0732) (0.1031) (0.0986) (0.1038) (0.1178) (0.1240) (-1.4159) (-2.4117) (-2.2899) (-2.2084) (-3.2107) (-3.0957) DSHI 0.2075 0.2131 0.2150 (0.1092) (0.1029) (0.1077) (1.9010) (2.0708) (1.9961) DMRS -0.1232 -0.1119 -0.1269 -0.1134 (0.0700) (0.1010) (0.0653) (0.0942) (-1.7608) (-1.1080) (-1.9445) (-1.2030) DMRS1 0.0006 -0.0275 (0.1011) (0.0943) (0.0055) (0.0120) DMRS2 -0.0225 -0.0275 (0.0896) (0.0836) (-0.2515) (-0.3290) R-squared 0.3193 0.4153 0.4034 0.4066 0.5045 0.5095 Adjusted R-squared 0.2305 0.3090 0.2949 0.2286 0.3866 0.3288 S.E. of regression 0.1316 0.1247 0.1259 0.1317 0.1175 0.1229 Ak. Info criterion -1.0826 -1.1606 -1.1403 -0.9977 -1.2521 -1.1140 Sample size 27 27 27 27 27 27 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 42 Table 3a: Regression results1 on the fairness of financial contribution to health systems (Logit) Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant 2.3447 2.3447 2.4902 2.4902 2.4902 2.4902 (0.2632) (0.2694) (0.4820) (0.4887) (0.4941) (0.5027) (8.9079) (8.7026) (5.1666) (5.0954) (5.0399) (4.9538) DARS -0.3267 -0.4710 -0.4723 -0.4723 -0.6165 -0.6165 (0.5394) (0.6780) (0.6816) (0.6912) (0.8069) (0.8209) (-0.6057) (-0.6946) (-0.6929) (-0.6833) (-0.7641) (-0.7510) DSHI 0.3605 0.3605 0.3605 (0.9838) (1.0086) (1.0261) (0.3665) (0.3575) (0.3514) DMRS -0.2117 0.6288 -0.2117 0.6288 (0.5813) (0.9143) (0.5959) (0.9404) (-0.3642) (0.6877) (-0.3553) (0.6686) DMRS1 -0.9005 -0.9005 (0.9976) (1.0261) (-0.9026) (-0.8776) DMRS2 -1.0907 -1.0907 (0.8923) (0.9178) (-1.2224) (-1.1885) R-squared 0.0189 0.0262 0.0261 0.1099 0.0334 0.1172 Adjusted R-squared -0.0327 -0.0820 -0.0821 -0.1126 -0.1372 -0.1771 S.E. of regression 1.0529 1.0777 1.0777 1.0928 1.1048 1.1241 Ak. Info criterion 3.0313 3.1191 3.1192 3.2197 3.2069 3.3067 Sample size 21 21 21 21 21 21 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 43 Table 3b: Regression results1 on the fairness of financial contribution to health systems Log [1­IHFC] Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -2.4903 -2.4903 -2.6060 -2.6060 -2.6060 -2.6060 (0.2332) (0.2387) (0.4273) (0.4311) (0.4380) (0.4434) (-10.6801) (-10.4341) (-6.0992) (-6.0450) (-5.9496) (-5.8773) DARS 0.3351 0.4630 0.4508 (0.4508) 0.5787 0.5787 (0.4779) (0.6006) (0.6043) (0.6097) (0.7153) (0.7241) (0.7012) (0.7708) (0.7661) (0.7395) (0.8091) (0.7992) DSHI -0.3197 -0.3197 -0.3197 (0.8715) (0.8941) (0.9051) (-0.3668) (-0.3576) (-0.3532) DMRS 0.1684 -0.6255 0.1684 -0.6255 (0.5153) (0.8065) (0.5283) (0.8295) (0.3267) (-0.7756) (0.3187) (-0.7541) DMRS1 0.9010 0.9010 (0.8799) (0.9051) (1.0239) (0.9955) DMRS2 1.0049 1.0049 (0.7871) (0.8095) (1.2768) (1.2414) R-squared 0.0252 0.0325 0.0310 0.1231 0.0382 0.1303 Adjusted R-squared -0.0261 -0.0750 -0.0767 -0.0961 -0.1315 -0.1595 S.E. of regression 0.9327 0.9547 0.9554 0.9640 0.9794 0.9915 Ak. Info criterion 2.7889 2.8767 2.8782 2.9688 2.9659 3.0557 Sample size 21 21 21 21 21 21 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 44 Table 4a: Regression results1 on the distribution of responsiveness of health systems (Logit) Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant 2.1440 2.1440 1.6327 1.6327 1.6327 1.6327 (0.2640) (0.2488) (0.4843) (0.4900) (0.4548) (0.4588) (8.1220) (8.6176) (3.3712) (3.3324) (3.5903) (3.5591) DARS 2.4428 3.7145 2.9540 2.9540 4.2257 4.2257 (0.5055) (0.7464) (0.6458) (0.6533) (0.8303) (0.8376) (4.8328) (4.9767) (4.5745) (4.5218) (5.0895) (5.0452) DSHI -1.9075 -1.9075 -1.9075 (0.8618) (0.8508) (0.8583) (-2.2133) (-2.2420) (-2.2225) DMRS 0.7217 0.0493 0.7217 0.0493 (0.5755) (0.8946) (0.5403) (0.8376) (1.2542) (0.0552) (1.3357) (0.0589) DMRS1 1.0846 1.0846 (0.9467) (0.8864) (1.1457) (1.2237) DMRS2 0.6675 0.6675 (0.8642) (0.8092) (0.7724) (0.8249) R-squared 0.4297 0.5097 0.4581 0.4824 0.5382 0.5624 Adjusted R-squared 0.4113 0.4771 0.4220 0.4084 0.4904 0.4814 S.E. of regression 1.2932 1.2188 1.2814 1.2963 1.2032 1.2137 Ak. Info criterion 3.4108 3.3201 3.4203 3.4957 3.3210 3.3883 Sample size 33 33 33 33 33 33 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 45 Table 4b: Regression results1 on the distribution of responsiveness of health systems Log [1--IRD] Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -2.2917 -2.2917 -1.8309 -1.8309 -1.8309 -1.8309 (0.2527) (0.2368) (0.4651) (0.4717) (0.4343) (0.4392) (-9.0671) (-9.6763) (-3.9365) (-3.8819) (-4.2159) (-4.1686) DARS -2.3185 -3.5783 -2.7792 -2.7792 -4.0390 -4.0390 (0.4840) (0.7105) (0.6201) (0.6289) (0.7989) (0.8019) (-4.7905) (-5.0363) (-4.4816) (-4.4194) (-5.0940) (-5.0369) DSHI 1.8897 1.8897 1.8897 (0.8204) (0.8125) (0.8217) (2.3034) (2.3259) (2.2998) DMRS -0.6505 -0.0531 -0.6505 -0.0531 (0.5526) (0.8611) (0.5160) (0.8019) (-1.1771) (-0.0617) (-1.2606) (-0.0663) DMRS1 -0.9857 -0.9857 (0.9113) (0.8486) (-1.0816) (-1.1615) DMRS2 -0.5807 -0.5807 (0.8319) (0.7747) (-0.6980) (-0.7496) R-squared 0.4254 0.5117 0.4508 0.4728 0.5371 0.5592 Adjusted R-squared 0.4068 0.4792 0.4141 0.3975 0.4892 0.4776 S.E. of regression 1.2382 1.1602 1.2306 1.2479 1.1490 1.1620 Ak. Info criterion 3.3239 3.2216 3.3393 3.4195 3.2289 3.3012 Sample size 33 33 33 33 33 33 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 46 Table 5a: Regression results1 on the equality of child survival (Logit) Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant 0.6246 0.6246 0.1291 0.1291 0.1291 0.1291 (0.1715) (0.1459) (0.2883) (0.2901) (0.2413) (0.2415) (3.6431) (4.2809) (0.4480) (0.4452) (0.5351) (0.5346) DARS 4.6707 2.5221 5.1662 5.1662 3.0175 3.0175 (0.4935) (0.6190) (0.5338) (0.5371) (0.6231) (0.6236) (9.4641) (4.0742) (9.6790) (9.6186) (4.8431) (4.8385) DSHI 3.7601 3.7601 3.7601 (0.7958) (0.7599) (0.7606) (4.7276) (4.9482) (4.9435) DMRS 0.7432 1.2244 0.7432 1.2244 (0.3530) (0.6646) (0.2955) (0.5534) (2.1052) (1.8423) (2.5148) (2.2124) DMRS1 -0.2691 -0.2691 (0.7324) (0.6099) (-0.3675) (-0.4413) DMRS2 -0.6458 -0.6458 (0.6501) (0.5413) (-0.9934) (-1.1930) R-squared 0.6153 0.7264 0.6440 0.6526 0.7551 0.7637 Adjusted R-squared 0.6084 0.7164 0.6310 0.6264 0.7414 0.7410 S.E. of regression 1.2244 1.0420 1.1885 1.1960 0.9949 0.9959 Ak. Info criterion 3.2766 2.9705 3.2336 3.2781 2.8942 2.9273 Sample size 58 58 58 58 58 58 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 47 Table 5b: Regression results1 on the equality of child survival Log [1­IECS] Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -1.1287 -1.1287 -0.8065 -0.8065 -0.8065 -0.8065 (0.14356) (0.1149) (0.2452) (0.2479) (0.1936) (0.1949) (-7.8631) (-9.8199) (-3.2895) (-3.2534) (-4.1655) (-4.1382) DARS -4.2511 -2.2139 -4.5733 -4.5733 -2.5361 -2.5361 (0.4132) (0.4877) (0.4540) (0.4590) (0.4999) (0.5032) (-10.2885) (-4.5398) (-10.0743) (-9.9638) (-5.0733) (-5.0400) DSHI -3.5652 -3.5652 -3.5652 (0.6269) (0.6097) (0.6137) (-5.6868) (-5.8476) (-5.8092) DMRS -0.4834 -0.8033 -0.4834 -0.8033 (0.3003) (0.5680) (0.2371) (0.4465) (-1.6097) (-1.4142) (-2.0384) (-1.7988) DMRS1 0.1780 0.1780 (0.6259) (0.4921) (0.2844) (0.3618) DMRS2 0.4297 0.4297 (0.5555) (0.4368) (0.7734) (0.9837) R-squared 0.6540 0.7821 0.6696 0.6745 0.7977 0.8026 Adjusted R-squared 0.6478 0.7742 0.6576 0.6499 0.7864 0.7836 S.E. of regression 1.0251 0.8208 1.0108 1.0221 0.7983 0.8035 Ak. Info criterion 2.9214 2.4934 2.9098 2.9638 2.4537 2.4981 Sample size 58 58 58 58 58 58 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 48 APPENDIX D Regression Results with Restricted Samples 49 Table 1: Regression results on DALE1 Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant 4.9423 4.9208 4.9638 4.8203 4.9426 4.7958 (0.3328) (0.3324) (0.3346) (0.3468) (0.3341) (0.3463) (14.8493) (14.3324) (14.8372) (13.8982) (14.7946) (13.8505) HEC -0.1919 -0.1914 -0.1883 -0.1841 -0.1878 -0.1835 (0.0197) (0.0196) (0.0203) (0.0204) (0.0202) (0.0204) (-9.7498) (-9.7509) (-9.2935) (-9.0153) (-9.2911) (-9.0114) EDU -0.2141 -0.2096 -0.2141 -0.1833 -0.2094 -0.1780 (0.0834) (0.0832) (0.0835) (0.0858) (0.0834) (0.0856) (-2.5684) (-2.5175) (-2.5631) (-2.1377) (-2.5121) (-2.0798) DARS -0.2963 -0.2546 -0.3411 -0.3640 -0.3000 -0.3221 (0.0654) (0.0730) (0.0875) (0.0886) (0.0930) (0.0939) (-4.5321) (-3.4880) (-3.9001) (-4.1075) (-3.2260) (-3.4310) DSHI -0.0982 -0.0989 -0.1016 (0.0774) (0.0775) (0.0774) (-1.2700) (-1.2769) (-1.3139) DMRS -0.0520 -0.3640 -0.0530 -0.1380 (0.0673) (0.0886) (0.0671) (0.1129) (-0.7726) (-4.1075) (-0.7891) (-1.2220) DMRS1 0.0168 0.0173 (0.1145) (0.1141) (0.1468) (0.1512) DMRS2 0.1298 0.1324 (0.1095) (0.1092) (1.1853) (1.2120) R-squared 0.7874 0.7902 0.7885 0.7927 0.7913 0.7957 Adjusted R-squared 0.7821 0.7832 0.7813 0.7820 0.7825 0.7834 S.E. of regression 0.2639 0.2632 0.2643 0.2639 0.2636 0.2631 Ak. Info criterion 0.2049 0.2076 0.2161 0.2283 0.2185 0.2296 Sample size 124 124 124 124 124 124 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 50 Table 2a: Regression results1 on the fairness of financial contribution to health systems (Logit) Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant 2.2874 2.2874 2.3117 2.3117 2.3117 2.3117 (0.2786) (0.2871) (0.5561) (0.5668) (0.5741) (0.5880) (8.2099) (7.9678) (4.1571) (4.0784) (4.0266) (3.9317) DARS -0.1146 -0.1762 -0.1390 -0.1390 -0.2005 -0.2005 (0.6072) (0.8370) (0.7864) (0.8016) (0.9944) (1.0184) (-0.1888) (-0.2105) (-0.1767) (-0.1734) (-0.2017) (-0.1969) DSHI 0.1231 0.1231 0.1231 (1.1118) (1.1482) (1.1759) (0.1107) (0.1072) (0.1047) DMRS -0.0332 0.8074 -0.0332 0.8074 (0.6494) (0.9818) (0.6704) (1.0184) (-0.0511) (0.8224) (-0.0495) (0.7928) DMRS1 -0.9005 -0.9005 (1.0349) (1.0735) (-0.8702) (-0.8389) DMRS2 -1.0907 -1.0907 (0.9256) (0.9601) (-1.1784) (-1.1360) R-squared 0.0021 0.0029 0.0023 0.0930 0.0030 0.0937 Adjusted R-squared -0.0566 -0.1218 -0.1225 -0.1662 -0.1964 -0.2548 S.E. of regression 1.0791 1.1118 1.1122 1.1336 1.1482 1.1759 Ak. Info criterion 3.0894 3.1939 3.1945 3.3097 3.2990 3.4141 Sample size 19 19 19 19 19 19 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 51 Table 2b: Regression results1 on the fairness of financial contribution to health systems Log [1­IHFC] Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -2.4400 -2.4400 -2.4465 -2.4465 -2.4465 -2.4465 (0.2470) (0.2545) (0.4931) (0.5000) (0.5091) (0.5186) (-9.8769) (-9.5861) (-4.9613) (-4.8929) (-4.8058) (-4.7172) DARS 0.1509 0.2088 0.1574 0.1574 0.2153 0.2153 (0.5384) (0.7421) (0.6974) (0.7071) (0.8817) (0.8983) (0.2803) (0.2814) (0.2257) (0.2225) (0.2441) (0.2396) DSHI -0.1158 -0.1158 -0.1158 (0.9858) (1.0181) (1.0373) (-0.1174) (-0.1137) (-0.1116) DMRS 0.0088 -0.7851 0.0088 -0.7851 (0.5758) (0.8660) (0.5945) (0.8983) (0.0153) (-0.9065) (0.0148) (-0.8739) DMRS1 0.9010 0.9010 (0.9129) (0.9469) (0.9870) (0.9515) DMRS2 1.0049 1.0049 (0.8165) (0.8469) (1.2308) (1.1866) R-squared 0.0046 0.0055 0.0046 0.1045 0.0055 0.1054 Adjusted R-squared -0.0540 -0.1189 -0.1198 -0.1513 -0.1934 -0.2387 S.E. of regression 0.9568 0.9858 0.9862 1.0000 1.0181 1.0373 Ak. Info criterion 2.8488 2.9532 2.9541 3.0588 3.0585 3.1631 Sample size 19 19 19 19 19 19 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 52 Table 3a: Regression results1 on the equality of child survival (Logit) Explanatory Regression 1 Regression Regression 3 Regression 4 Regression 5 Regression 6 variables 2 Constant 0.7248 0.7248 0.2798 0.2798 0.2798 0.2798 (0.1677) (0.1507) (0.3097) (0.3097) (0.2761) (0.2747) (4.3223) (4.8106) (0.9037) (0.9035) (1.0134) (1.0189) DARS 5.1209 2.9990 5.5659 5.5659 3.4439 3.4439 (0.4937) (0.7381) (0.5511) (0.5512) (0.7562) (0.7522) (10.3730) (4.0629) (10.0999) (10.0980) (4.5540) (4.5786) DSHI 3.1830 3.1830 3.1830 (0.8850) (0.8623) (0.8576) (3.5966) (3.6915) (3.7114) DMRS 0.6203 1.0737 0.6203 1.0737 (0.3656) (0.6385) (0.3260) (0.5662) (1.6965) (1.6815) (1.9025) (1.8963) DMRS1 -0.1079 -0.1079 (0.7000) (0.6207) (-0.1542) (-0.1739) DMRS2 -0.6458 -0.6458 (0.6070) (0.5383) (-1.0638) (-1.1997) R-squared 0.6827 0.7490 0.7003 0.7125 0.7666 0.7787 Adjusted R-squared 0.6764 0.7388 0.6881 0.6880 0.7520 0.7547 S.E. of regression 1.1374 1.0219 1.1166 1.1168 0.9956 0.9903 Ak. Info criterion 3.1330 2.9372 3.1144 3.1500 2.9029 2.9265 Sample size 52 52 52 52 52 52 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 53 Table 3b: Regression results1 on the equality of child survival Log [1­IECS] Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -1.1863 -1.1863 -0.8758 -0.8758 -0.8758 -0.8758 (0.1392) (0.1209) (0.2593) (0.2613) (0.2235) (0.2243) (-8.5222) (-9.8145) (-3.3780) (-3.3518) (-3.9178) (-3.9049) DARS -4.7368 -2.7675 -5.0473 -5.0473 -3.0779 -3.0779 (0.4098) (0.5921) (0.4614) (0.4650) (0.6122) (0.6142) (-11.5594) (-4.6737) (-10.9400) (-10.8550) (-5.0277) (-5.0112) DSHI -2.9540 -2.9540 -2.9540 (0.7099) (0.6980) (0.7003) (-4.1610) (-4.2321) (-4.2182) DMRS -0.4328 -0.7339 -0.4328 -0.7339 (0.3061) (0.5387) (0.2639) (0.4624) (-1.4138) (-1.3625) (-1.6398) (-1.5874) DMRS1 0.0694 0.0694 (0.5905) (0.5068) (0.1175) (0.1369) DMRS2 0.4297 0.4297 (0.5121) (0.4395) (0.8390) (0.9775) R-squared 0.7277 0.7988 0.7384 0.7451 0.8095 0.8162 Adjusted R-squared 0.7223 0.7906 0.7277 0.7234 0.7976 0.7962 S.E. of regression 0.9441 0.8198 0.9348 0.9421 0.8060 0.8086 Ak. Info criterion 2.7605 2.4964 2.7590 2.8098 2.4803 2.5213 Sample size 52 52 52 52 52 52 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 54 APPENDIX E Regression Results with Restricted Samples (Additional Deletion of Influential Data) 55 Table 1a: Regression results1 on the level of responsiveness 2 (Logit) Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -0.3943 -0.4209 -0.4631 -0.4729 -0.4896 -0.5010 (0.2361) (0.2355) (0.2181) (0.2300) (0.2161) (0.2281) (-1.6698) (-1.7872) (-2.1230) (-2.0561) (-2.2663) (-2.1964) HEC 0.0002 0.0001 0.0001 0.0001 0.0000 0.0000 (0.0003) (0.0003) (0.0003) (0.0003) (0.0003) (0.0003) (0.6413) (0.4386) (0.3299) (0.2395) (0.1150) (0.0251) EDU 0.0043 0.0047 0.0028 0.0030 0.0032 0.0034 (0.0027) (0.0027) (0.0026) (0.0027) (0.0026) (0.0027) (1.5903) (1.7265) (1.1011) (1.0918) (1.2540) (1.2431) DARS 0.3002 0.4688 0.2261 0.5618 0.7244 0.7337 (0.1596) (0.2154) (0.1836) (0.1932) (0.2244) (0.2364) (1.8814) (2.1760) (3.0280) (2.9086) (3.2275) (3.1038) DSHI -0.2524 -0.2521 -0.2562 (0.2186) (0.1987) (0.2084) (-1.1543) (-1.2687) (-1.2294) DMRS 0.2674 0.2372 0.2673 0.2340 (0.1164) (0.1724) (0.1148) (0.1702) (2.2971) (1.3759) (2.3294) (1.3753) DMRS1 0.0061 0.0056 (0.1739) (0.1716) (0.0349) (0.0624) DMRS2 0.0507 0.0565 (0.1531) (0.1511) (0.3314) (0.3739) R-squared 0.4156 0.4505 0.5330 0.5370 0.5678 0.5729 Adjusted R-squared 0.3359 0.3458 0.4440 0.3908 0.4597 0.4068 S.E. of regression 0.2365 0.2348 0.2164 0.2266 0.2134 0.2236 Ak. Info criterion 0.0952 0.1106 -0.0520 0.0932 -0.0525 0.0895 Sample size 26 26 26 26 26 26 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 2The data for Bulgaria were excluded from the "full" samples. 56 Table 1b: Regression results1 on the level of responsiveness 2 Log [1­IR] Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -0.1246 -0.0801 -0.1051 -0.0555 -0.0611 -0.0092 (0.4224) (0.4152) (0.3862) (0.4108) (0.3761) (0.4004) (-0.2950) (-0.1929) (-0.2721) (-0.1350) (-0.1624) (-0.0230) HEC -0.0107 -0.0097 0.0097 0.0140 0.0105 0.0150 (0.0260) (0.0255) (0.0253) (0.0273) (0.0246) (0.0265) (-0.4115) (-0.3820) (0.3820) (0.5115) (0.4285) (0.5646) EDU -0.1204 -0.1313 -0.1176 -0.1320 -0.1284 -0.1435 (0.1050) (0.1032) (0.0960) (0.1001) (0.0935) (0.1001) (-1.1468) (-1.2728) (-1.2251) (-3.4119) (-1.3740) (-1.4334) DARS -0.1830 -0.2734 -0.3333 -0.3417 -0.4221 -0.4313 (0.0760) (0.0991) (0.0952) (0.1001) (0.1096) (0.1151) (-2.4087) (-2.7580) (-3.4999) (-3.4119) (-3.8512) (3.7480) DSHI 0.1501 0.1486 0.1496 (0.1087) (0.0985) (0.1026) (1.3810) (1.5087) (1.4587) DMRS -0.1429 -0.1280 -0.1423 -0.1264 (0.0620) (0.0886) (0.0602) (0.0861) (-2.3070) (-1.4447) (-2.3658) (-1.4684) DMRS1 0.0060 0.0056 (0.0885) (0.0860) (0.0682) (0.0646) DMRS2 -0.0339 -0.0355 (0.0785) (0.0763) (-0.4315) (-0.4657) R-squared 0.4257 0.4735 0.5418 0.5509 0.5886 0.5984 Adjusted R-squared 0.3474 0.3732 0.4546 0.4091 0.4858 0.4422 S.E. of regression 0.1212 0.1187 0.1108 0.1153 0.1075 0.1120 Ak. Info criterion -1.2427 -1.2528 -1.3917 -1.2579 -1.4226 -1.2927 Sample size 26 26 26 26 26 26 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 2The data for Bulgaria were excluded from the "full" samples. 57 Table 2a: Regression results1 on the distribution of responsiveness of health systems 2 (Logit) Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant 2.1440 2.1440 1.6327 1.6327 1.6327 1.6327 (0.2524) (0.2469) (0.4617) (0.4663) (0.4507) (0.4547) (8.4928) (8.6829) (3.5367) (3.5018) (3.6228) (3.5908) DARS 2.9117 3.7145 3.4229 3.4229 4.2257 4.2257 (0.5313) (0.7408) (0.6529) (0.6594) (0.8228) (0.8302) (5.4807) (5.0144) (5.2428) (5.1910) (5.1355) (5.0901) DSHI -1.4049 -1.4049 -1.4049 (0.9239) (0.9107) (0.9188) (-1.5206) (-1.5427) (-1.5290) DMRS 0.7217 0.0493 0.7217 0.0493 (0.5485) (0.8513) (0.5355) (0.8302) (1.3158) (0.0580) (1.3478) (0.0594) DMRS1 1.0846 1.0846 (0.9009) (0.8786) (1.2040) (1.2346) DMRS2 0.6675 0.6675 (0.8224) (0.8020) (0.8116) (0.8323) R-squared 0.5088 0.5463 0.5374 0.5618 0.5749 0.5993 Adjusted R-squared 0.4919 0.5139 0.5044 0.4944 0.5276 0.5192 S.E. of regression 1.2367 1.2097 1.2214 1.2336 1.1924 1.2030 Ak. Info criterion 3.3252 3.3103 3.3297 3.4044 3.3097 3.3796 Sample size 31 31 31 31 31 31 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 2The data for Chile and Poland were excluded from the "full" samples. 58 Table 2b: Regression results1 on the distribution of responsiveness of health systems 2 Log [1--IRD] Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -2.2917 -2.2917 -1.8309 -1.8309 -1.8309 -1.8309 (0.2409) (0.2347) (0.4420) (0.4476) (0.4298) (0.4349) (-9.5135) (-9.7639) (-4.14259 (-4.0906) (-4.2594) (-4.2103) DARS -2.7774 -3.5783 -3.2382 -3.2382 -4.0390 -4.0390 (0.5069) (0.7041) (0.6251) (0.6330) (0.7848) (0.7939) (-5.4790) (-5.0819) (-5.1806) (-5.1157) (-5.1467) (-5.0872) DSHI 1.4015 1.4015 1.4015 (0.8782) (0.8686) (0.8787) (1.5959) (1.6135) (1.5949) DMRS -0.6505 -0.0531 -0.6505 -0.0531 (0.5252) (0.8172) (0.5107) (0.7939) (-1.2387) (-0.0650) (-1.2736) (-0.0669) DMRS1 -0.9857 -0.9857 (0.8648) (0.8402) (-1.1398) (-1.1731) DMRS2 -0.5807 -0.5807 (0.7895) (0.7670) (-0.7356) (-0.7571) R-squared 0.5086 0.5496 0.5342 0.5564 0.5751 0.5974 Adjusted R-squared 0.4917 0.5174 0.5009 0.4881 0.5279 0.5168 S.E. of regression 1.1801 1.1498 1.1694 1.1842 1.1373 1.1505 Ak. Info criterion 3.2314 3.2089 3.2426 3.3227 3.2150 3.2903 Sample size 31 31 31 31 31 31 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 2The data for Chile and Poland were excluded from the "full" samples. 59 Table 3a: Regression results1 on the equality of child survival 2 (Logit) Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant 0.7248 0.7248 0.2798 0.2798 0.2798 0.2798 (0.1171) (0.1183) (0.2092) (0.2038) (0.2115) (0.2061) (6.1898) (6.1263) (1.3375) (1.3729) (1.3234) (1.3579) DARS 6.1819 6.1819 6.6269 6.6269 6.6269 6.6269 (0.3740) (0.8111) (0.3970) (0.3868) (0.7912) (0.7711) (16.5298) (7.6215) (16.6922) (17.1343) (8.3759) (8.5937) DSHI 0.0000 0.0000 0.0000 (0.8972) (0.8524) (0.8308) (0.0000) (0.0000) (0.0000) DMRS 0.6203 1.0737 0.6203 1.0737 (0.2470) (0.4202) (0.2497) (0.4249) (2.5108) (2.5550) (2.4845) (2.5271) DMRS1 -0.1079 -0.1079 (0.4607) (0.4658) (-0.2343) (-0.2318) DMRS2 -0.6458 -0.6458 (0.3995) (0.4039) (-1.6165) (-1.5989) R-squared 0.8479 0.8479 0.8656 0.8778 0.8656 0.8778 Adjusted R-squared 0.8448 0.8416 0.8600 0.8671 0.8570 0.8642 S.E. of regression 0.7942 0.8024 0.7544 0.7350 0.7624 0.7431 Ak. Info criterion 2.4155 2.4547 2.3313 2.3149 2.3705 2.3541 Sample size 51 51 51 51 51 51 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 2The data for Uzbekistan were excluded from the "restricted" samples. 60 Table 3b: Regression results1 on the equality of child survival 2 Log [1--IECS] Explanatory Regression 1 Regression 2 Regression 3 Regression 4 Regression 5 Regression 6 variables Constant -1.1863 -1.1863 -0.8758 -0.8758 -0.8758 -0.8758 (0.0829) (0.0837) (0.1483) (0.1453) (0.1499) (0.1469) (-14.3162) (-14.1693) (-5.9043) (-6.0269) (-5.8425) (-5.9610) DARS -5.7215 -5.7215 -6.0320 -6.0320 -6.0320 -6.0320 (0.2646) (0.5740) (0.2814) (0.2757) (0.5609) (0.5497) (-21.6201) (-9.9685) (-21.4327) (-21.8778) (-10.7546) (-10.9728) DSHI 0.0000 0.0000 0.0000 (0.6348) (0.6043) (0.5922) (0.0000) (0.0000) (0.0000) DMRS -0.4328 -0.7339 -0.4328 -0.7339 (0.1751) (0.2996) (0.1770) (0.3029) (-2.4712) (-2.4500) (-2.4453) (-2.4232) DMRS1 0.0694 0.0694 (0.3284) (0.3320) (0.2114) (0.2091) DMRS2 0.4297 0.4297 (0.2848) (0.2879) (1.5087) (1.4922) R-squared 0.9051 0.9051 0.9158 0.9226 0.9158 0.9226 Adjusted R-squared 0.9032 0.9012 0.9123 0.9159 0.9105 0.9140 S.E. of regression 0.5620 0.5678 0.5348 0.5239 0.5405 0.5297 Ak. Info criterion 1.7238 1.7630 1.6432 1.6380 1.6824 1.6772 Sample size 51 51 51 51 51 51 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 2The data for Uzbekistan were excluded from the "restricted" samples. 61 APPENDIX F Regression Results for Enlarged Models 63 Table 1: Selected regression results1 for enlarged models (with GINI index as explanatory variable in the equation for distributional measures) IFFC IRD IECS Explanatory Variables Logit specification Log[(1--IFFC)] Logit Log[(1--IRD)] Logit specification Log[(1--IECS)] specification Constant 2.8260 -2.8794 3.061 -3.1853 -0.7471 -0.1186 (1.3698) (1.2155) (0.7956) (0.7334) (0.9164) (0.7754) (2.0630) (-2.3689) (3.8539) (-4.3432) (-0.8153) (-0.1530) GINI -0.0119 0.0097 -0.0375 0.0352 0.0355 -0.0258 (0.0296) (0.0262) (0.0180) (0.0166) (0.0206) (0.0174) (-0.4020) (0.3696) (-2.0853) (2.1287) (1.7240) (-1.4803) DARS -0.2568 0.2669 2.1713 -2.0025 5.3537 -4.9042 (0.7162) (0.6355) (0.5222) (0.4814) (0.5531) (0.4680) (-0.3586) (0.4200) (4.1577) (-4.1597) (9.6789) (-10.4788) DMRS 0.9873 -0.8994 (0.4637) (0.4275) (2.1291) (-2.1039) R-squared 0.0121 0.0130 0.5191 0.5229 0.7053 0.7397 Adjusted R-squared -0.1114 -0.1103 0.4590 0.4632 0.6906 0.7267 S.E. of regression 1.1067 0.9821 0.9320 0.8592 1.1912 1.0079 Ak. Info criterion 3.1846 2.9456 2.8286 2.6659 3.2550 2.9208 Sample size 19 19 28 28 43 43 The first and second coefficient in brackets refer to the standard error and t-statistic, respectively 65 Table 2 : Selected regression results1 for enlarged models (with the interaction term DARS*[PHE%- 0.5] as explanatory variable) DALE Explanatory Variables Log[(80--DALE)] Constant 4.9446 (0.3306) (14.9580) HEC -0.1897 (0.0196) (-9.6837) EDU -0.2166 (0.0828) (-2.6155) DARS -0.2088 (0.0843) (-2.4774) DARS*[PHE%-0.5] -0.4556 (0.2798) (-1.6284) R-squared 0.7920 Adjusted R-squared 0.7850 S.E. of regression 0.2621 Ak. Info criterion 0.1990 Sample size 124 1 The first and second coefficient in brackets refer to the standard error and t-statistic, respectively 66 2 . 67 APPENDIX G Graphs of Health Attainment Indicators vs. Share of Public Health Expenditure to Total Health Expenditure 68 Graph 1a: Disability-Adjusted Life Expectancy (DALE) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=1; restricted samples) 75 70 65 DALE 60 55 0.2 0.4 0.6 0.8 1.0 PHE% 69 Graph 2a: Index of level of reponsiveness (IR) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=1) 0.70 0.65 0.60 0.55 IR 0.50 0.45 0.40 0.2 0.4 0.6 0.8 1.0 PHE% 70 Graph 3a: Index of fairness of financial contribution (IFFC) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=1; restricted samples) 0.94 0.92 0.90 IFFC0.88 0.86 0.84 0.5 0.6 0.7 0.8 0.9 PHE% 71 Graph 4a: Index of distribution of reponsiveness (IRD) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=1) 1.00 0.98 0.96 IRD0.94 0.92 0.90 0.2 0.4 0.6 0.8 1.0 PHE% 72 Graph 5a: Index of equality of child survival (IECS) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=1; restricted samples) 1.0 0.9 0.8 IECS 0.7 0.6 0.4 0.5 0.6 0.7 0.8 0.9 1.0 PHE% 73 Graph 1b: Disability-Adjusted Life Expectancy (DALE) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=0; restricted samples) 80 70 60 50 DALE 40 30 20 0.0 0.2 0.4 0.6 0.8 1.0 PHE% 74 Graph 2b: Index of level of reponsiveness (IR) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=0) 0.7 0.6 0.5 IR 0.4 0.3 0.0 0.2 0.4 0.6 0.8 1.0 PHE% 75 Graph 3b: Index of fairness of financial contribution (IFFC) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=0; restricted samples) 1.0 0.9 0.8 IFFC 0.7 0.6 0.0 0.2 0.4 0.6 0.8 PHE% 76 Graph 4b: Index of distribution of reponsiveness (IRD) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=0) 1.0 0.9 IRD 0.8 0.7 0.0 0.2 0.4 0.6 0.8 1.0 PHE% 77 Graph 5b: Index of equality of child survival (IECS) vs. share of public health expenditure in total health expenditure (PHE%) (DARS=0; restricted samples) 1.0 0.8 0.6 IECS 0.4 0.2 0.0 0.2 0.4 0.6 0.8 PHE% 78 APPENDIX H Selected Regression Results with "Respect for Persons" and "Client Orientation" as Dependent Variables 79 Table 1: Selected regression results1 on the level of "Respect for Persons" and "Client Orientation" Respect for Persons Client Orientation Explanatory Variables Regression 1 Regression 2 Regression 1 Regression 2 Logit specification Log[(1--RESPECT)] Logit specification Log[(1--CO)] Constant -0.4603 -0.0237 -0.6571 -0.1014 (0.2358) (0.4048) (0.2639) (0.4792) (-1.9518) (-0.0586) (-2.4900) (-0.2116) HEC -0.0001 0.0165 0.0006 -0.0305 (0.0003) (0.0252) (0.0003) (0.0298) (-0.1799) (0.6540) (2.0721) (-1.0235) EDU 0.0034 -0.1456 0.0035 -0.0778 (0.0028) (0.0999) (0.0031) (0.1183) (1.2224) (-1.4571) (1.1075) (-0.6577) DARS 0.6065 -0.3473 0.4088 -0.4024 (0.2399) (0.1166) (0.2002) (0.1380) (2.5281) (-2.9793) (2.0422) (-2.9161) DSHI -0.3540 0.1950 0.2193 (0.2109) (0.1018) (0.1206) (-1.6785) (1.9150) (1.8189) DMRS 0.2541 -0.1417 0.2521 -0.1103 (0.1251) (0.0646) (0.1406) (0.0765 (2.0380) (-2.1937) (1.7923) (-1.4429) R-squared 0.3892 0.4136 0.5352 0.5420 Adjusted R-squared 0.2437 0.2740 0.4507 0.4330 S.E. of regression 0.2329 0.1162 0.2619 0.1376 Ak. Info criterion 0.1167 -1.2732 0.3239 -0.9355 Sample size 27 27 27 27 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 80 Table 2: Selected regression results1 on the level of "Respect for Persons" and "Client Orientation" 2 Respect for Persons Client Orientation Explanatory Variables Regression 1 Regression 2 Regression 1 Regression 2 Logit specification Log[(1--RESPECT)] Logit specification Log[(1--CO)] Constant -0.4606 0.0473 -0.6695 -0.0175 (0.2240) (0.3799) (0.2455) (0.4500) (-2.0564) (0.1244) (-2.7271) (-0.0388) HEC -0.0002 0.0326 0.0004 -0.0114 (0.0003) (0.0249) (0.0003) (0.0294) (-0.8055) (1.3114) (1.1915) (-0.3885) EDU 0.0035 -0.1732 0.0037 -0.1104 (0.0027) (0.0944) (0.0029) (0.1118) (1.3104) (-1.8347) (1.2757) (-0.9876) DARS 0.6921 -0.3869 0.5979 -0.4493 (0.2327) (0.1107) (0.2067) (0.1311) (2.9747) (-3.4951) (2.8927) (-3.4270) DSHI -0.2670 0.1367 0.1504 (0.2060) (0.0995) (0.1178) (-1.2963) (1.3742) (1.2763) DMRS 0.2663 -0.1556 0.2680 -0.1268 (0.1190) (0.0608) (0.1310) (0.0720) 2.2384) (-2.5608) (2.0458) (-1.7617) R-squared 0.4539 0.4919 0.6066 0.6094 Adjusted R-squared 0.3174 0.3648 0.5317 0.5117 S.E. of regression 0.2212 0.1086 0.2436 0.1287 Ak. Info criterion 0.0194 -1.4026 0.1843 -1.0642 Sample size 26 26 26 26 1The first and second coefficient in brackets refer to the standard error and t-statistic, respectively. 2 The data for Bulgaria were excluded from the "full" samples. 81 APPENDIX I The Data 82 Table 1: Data on DALE and explanatory variables Countries DALE1 HEC2 EDU3 DARS DSHI DMRS DMRS1 DMRS2 Afghanistan 37.70 2.000000 NA 0 0 0 0 0 Albania 60.00 26.00000 NA 1 0 0 0 0 Algeria 61.60 44.00000 96.00 0 0 1 1 0 Andorra 72.30 1368.000 NA 0 0 1 1 0 Angola 38.00 NA 34.70 0 0 0 0 0 Antigua-Barbuda 65.80 775.0000 NA 1 0 0 0 0 Argentina 66.70 676.0000 99.90 0 0 1 1 0 Armenia 66.70 36.00000 NA 0 0 0 0 0 Australia 73.20 1730.000 99.90 1 1 0 0 0 Austria 71.60 2277.000 99.90 1 1 0 0 0 Azerbaijan 63.70 20.00000 NA 1 0 0 0 0 Bahamas 59.10 785.0000 94.60 0 0 0 0 0 Bahrain 64.40 478.0000 98.20 1 0 0 0 0 Bangladesh 49.90 13.00000 75.10 0 0 0 0 0 Barbados 65.00 596.0000 97.40 1 0 0 0 0 Belarus 61.70 78.00000 NA 1 0 0 0 0 Belgium 71.60 1918.000 99.90 1 1 0 0 0 Belize 60.90 176.0000 99.90 1 0 0 0 0 Benin 42.20 12.00000 67.60 0 0 0 0 0 Bhutan 51.80 14.00000 13.20 0 0 0 0 0 Bolivia 53.30 59.00000 97.40 0 0 1 1 0 Bosnia and Herzegovina 64.90 77.00000 NA 1 0 0 0 0 32.30 132.0000 80.10 0 0 1 0 1 Botswana 59.10 319.0000 97.10 0 0 1 0 1 Brazil 64.40 NA 87.90 1 0 0 0 0 Brunei Darussalam 64.40 59.00000 97.90 1 1 0 0 0 Bulgaria 35.50 8.000000 32.30 0 0 1 0 1 Burkina Faso 34.60 6.000000 35.60 0 0 1 0 1 Burundi 45.70 21.00000 99.90 0 0 0 0 0 Cambodia 1Source: WHO (2000), Statistical Annex Table 5 2Source: WHO (2000), Statistical Annex Table 8 3Source: UNDP (2000) 83 Table 1 (continued): Data on DALE and explanatory variables Countries DALE HEC EDU DARS DSHI DMRS DMRS1 DMRS2 Cameroon 42.20 31.00000 61.70 0 0 1 0 1 Canada 72.00 1783.000 99.90 1 0 0 0 0 Cape Verde 57.60 34.00000 99.90 0 0 1 1 0 Central African R. 36.00 8.000000 46.20 0 0 0 0 0 39.40 7.000000 47.90 0 0 0 0 0 Chad 68.60 315.0000 90.40 1 1 0 0 0 Chile 62.30 20.00000 99.90 0 0 1 0 1 China 62.90 247.0000 89.40 0 0 1 0 0 Colombia 46.80 14.00000 50.10 0 0 0 0 0 Comoros 45.10 58.00000 78.30 0 0 1 1 0 63.40 389.0000 NA 1 0 0 0 0 Congo 66.70 226.0000 91.80 1 1 0 0 0 Cook Islands 42.80 23.00000 58.30 0 0 1 0 1 Costa Rica 67.00 352.0000 99.90 1 1 0 0 0 Côte d'Ivoire 68.40 131.0000 99.90 1 0 0 0 0 Croatia 69.80 648.0000 NA 1 0 0 0 0 Cuba 68.00 391.0000 99.90 1 1 0 0 0 Cyprus 36.30 NA 58.2 0 0 0 0 0 Czech Republic 52.30 37.00000 NA 1 0 0 0 0 Democratic R. of Congo 69.40 2574.000 99.90 1 0 0 0 0 37.90 23.00000 31.90 0 0 0 0 0 Democratic R. of Korea 69.80 282.0000 NA 1 0 0 0 0 Denmark 62.50 91.00000 91.30 0 0 1 0 1 Djibouti 61.00 75.00000 99.90 0 0 1 0 0 Dominica 58.50 44.00000 95.20 0 0 1 1 0 61.50 182.0000 89.10 0 0 1 0 0 Dominican Republic 44.10 40.00000 79.30 0 0 1 0 0 Ecuador 37.70 6.000000 29.30 0 0 0 0 0 Egypt 63.10 204.0000 99.90 1 1 0 0 0 El Salvador 33.50 4.000000 35.20 0 0 0 0 0 Equatorial Guinea 59.40 115.0000 99.90 0 0 0 0 0 Eritrea 70.50 1789.000 99.90 1 0 0 0 0 73.10 2369.000 99.90 1 1 0 0 0 Estonia 47.80 138.000 NA 0 0 1 1 0 Ethiopia Fiji Finland 84 France Gabon Table 1 (continued): Data on DALE and explanatory variables Countries DALE HEC EDU DARS DSHI DMRS DMRS1 DMRS2 85 Gambia 48.30 12.00000 65.90 0 0 0 0 0 Georgia 66.30 45.00000 89.00 0 0 0 0 0 Germany 70.40 2713.000 99.90 1 1 0 0 0 Ghana 45.50 11.00000 43.40 0 0 0 0 0 72.50 905.0000 99.90 1 1 0 0 0 Greece 65.50 305.0000 NA 0 0 0 0 0 Grenada 54.30 41.00000 73.80 0 0 1 0 1 Guatemala 37.80 19.00000 45.60 0 0 1 1 0 Guinea 37.20 13.00000 52.30 0 0 1 0 1 Guinea-Bissau 60.20 45.00000 92.80 0 0 0 0 0 Guyana 43.80 18.00000 19.40 0 0 1 0 1 61.10 59.00000 87.50 0 0 1 1 0 Haiti 64.10 236.0000 97.50 1 1 0 0 0 Honduras 70.80 2149.000 99.90 1 0 0 0 0 Hungary 53.20 23.00000 77.20 0 0 1 0 1 Iceland 59.70 18.00000 99.20 0 0 1 0 1 India 60.50 108.0000 90.00 0 0 1 0 1 Indonesia 55.30 251.0000 74.60 0 0 1 0 1 Iran 69.60 1326.000 99.90 1 0 0 0 0 Iraq 70.40 1385.000 NA 1 1 0 0 0 Ireland 72.70 1855.000 99.90 1 0 0 0 0 Israel 67.30 149.0000 95.60 1 0 0 0 0 Italy 74.50 2373.000 99.90 1 1 0 0 0 Jamaica 60.00 59.00000 NA 0 0 1 0 1 Japan 56.40 62.00000 NA 1 0 0 0 0 Jordan 39.30 17.00000 65.00 0 0 1 0 1 Kazakhstan 55.30 122.0000 NA 0 0 0 0 0 Kenya 63.20 572.0000 65.20 1 0 0 0 0 Kiribati 56.30 15.00000 99.50 1 0 0 0 0 Kuwait Kyrgyzstan 86 Table 1 (continued): Data on DALE and explanatory variables Countries DALE HEC EDU DARS DSHI DMRS DMRS1 DMRS2 Lao People's Dem. Rep.. 46.10 13.00000 73.00 0 0 0 0 0 62.20 140.0000 99.90 1 1 0 0 0 Latvia 60.60 461.0000 76.10 0 0 1 1 0 Lebanon 36.90 28.00000 68.60 0 0 1 0 1 Lesotho 34.00 31.00000 NA 0 0 0 0 0 Liberia 59.30 296.0000 99.90 0 0 1 0 0 Libya 64.10 167.0000 NA 1 1 0 0 0 Lithuania 71.10 2580.000 NA 1 1 0 0 0 Luxembourg 36.60 5.000000 58.70 0 0 1 0 1 Madagascar 29.40 15.00000 98.50 0 0 0 0 0 Malawi 61.40 110.0000 99.90 1 0 0 0 0 Malaysia 53.90 107.0000 NA 0 0 0 0 0 Maldives 33.10 10.00000 38.10 0 0 1 1 0 70.50 551.0000 99.90 1 0 0 0 0 Mali 56.80 253.0000 NA 0 0 0 0 0 Malta 41.40 24.00000 62.90 0 0 1 0 1 Marshall Islands 62.70 129.0000 96.50 1 0 0 0 0 Mauritania 65.00 240.0000 99.90 0 0 1 1 0 Mauritius 59.60 242.0000 NA 0 0 0 0 0 Mexico 72.40 1264.000 NA 1 1 0 0 0 Micronesia 53.80 16.00000 85.10 0 0 1 0 0 59.10 66.00000 76.60 0 0 1 0 1 Monaco 34.40 5.000000 39.60 0 0 1 0 1 Mongolia 51.60 100.0000 99.30 0 0 1 0 1 Morocco 35.60 153.0000 91.40 0 0 1 1 0 Mozambique 52.50 593.0000 NA 0 0 0 0 0 Myanmar 49.50 8.000000 78.40 0 0 0 0 0 Namibia Nauru Nepal 87 Table 1 (continued): Data on DALE and explanatory variables Countries DALE HEC EDU DARS DSHI DMRS DMRS1 DMRS2 Netherlands 72.00 2041.000 99.90 1 1 0 0 0 New Zealand 69.20 1416.000 99.90 1 0 0 0 0 Nicaragua 58.10 35.00000 78.60 0 0 1 0 1 Niger 29.10 5.000000 24.40 0 0 1 0 1 Nigeria 38.30 30.00000 NA 0 0 0 0 0 Niue 61.60 91.00000 NA 1 0 0 0 0 Norway 71.70 2283.000 99.90 1 1 0 0 0 Oman 63.00 370.0000 67.70 1 0 0 0 0 Pakistan 55.90 17.00000 NA 0 0 1 0 1 Palau 59.00 552.0000 NA 1 0 0 0 0 Panama 66.00 238.0000 89.90 0 0 1 1 0 Papua New Guinea 47.00 36.00000 78.90 0 0 0 0 0 Paraguay 63.00 106.0000 96.30 0 0 1 1 0 Peru 59.40 149.0000 93.80 0 0 1 0 0 Philippines 58.90 40.00000 99.90 0 0 1 1 0 Poland 66.20 229.0000 99.40 1 1 0 0 0 Portugal 69.30 845.0000 99.90 1 0 0 0 0 Qatar 63.50 1042.000 83.30 1 0 0 0 0 Republic of Korea 65.00 700.0000 99.90 1 1 0 0 0 Republic of Moldova 61.50 35.00000 NA 1 0 0 0 0 Romania 62.30 59.00000 99.90 1 1 0 0 0 Russia 61.30 158.0000 99.90 1 0 0 0 0 Rwanda 32.80 13.00000 78.30 0 0 0 0 0 Saint Kitts and Nevis 61.60 404.0000 NA 1 0 0 0 0 Saint Lucia 65.00 211.0000 NA 1 0 0 0 0 Saint Vincent and the G. 66.40 211.0000 NA 1 0 0 0 0 Samoa 60.50 47.00000 96.50 1 0 0 0 0 San Marino 72.30 2257.000 NA 1 1 0 0 0 88 Table 1 (continued): Data on DALE and explanatory variables Countries DALE HEC EDU DARS DSHI DMRS DMRS1 DMRS2 Sao Tome and Principe 53.50 13.00000 NA 0 0 0 0 0 Saudi Arabia 64.50 260.0000 60.10 1 0 0 0 0 Senegal 44.60 23.00000 59.50 0 0 1 1 0 Seychelles 59.30 424.0000 NA 1 0 0 0 0 Sierra Leone 25.90 11.00000 44.00 0 0 0 0 0 Singapore 69.30 876.0000 91.40 1 0 0 0 0 Slovakia 66.60 311.0000 NA 1 1 0 0 0 Slovenia 68.40 857.0000 NA 1 1 0 0 0 Solomon Islands 54.90 19.00000 NA 0 0 0 0 0 Somalia 36.40 11.00000 NA 0 0 0 0 0 South Africa 39.80 268.0000 99.90 0 0 1 0 1 Spain 72.80 1071.000 99.90 1 0 0 0 0 Sri Lanka 62.80 25.00000 99.90 0 0 0 0 0 Sudan 43.00 13.00000 NA 0 0 0 0 0 Suriname 62.70 114.0000 99.90 0 0 0 0 0 Swaziland 38.10 49.00000 94.6 0 0 0 0 0 Sweden 73.00 2456.000 99.90 1 0 0 0 0 Switzerland 72.50 3564.000 99.90 1 1 0 0 0 Syrian Arab Republic 58.80 151.0000 94.7 0 0 0 0 0 Tajikistan 57.30 11.00000 NA 1 0 0 0 0 Thailand 60.20 133.0000 88.00 0 0 1 0 1 The F. Y. of Macedonia 63.70 120.0000 NA 1 1 0 0 0 Togo 40.70 9.000000 82.30 0 0 0 0 0 Tonga 62.90 141.0000 NA 0 0 0 0 0 Trinidad and Tobago 64.60 197.0000 99.90 0 0 1 0 1 Tunisia 61.40 111.0000 99.90 0 0 1 0 0 Turkey 62.90 118.0000 99.90 0 0 1 1 0 Turkmenistan 54.30 24.00000 NA 1 0 0 0 0 89 Table 1 (continued): Data on DALE and explanatory variables Countries DALE HEC EDU DARS DSHI DMRS DMRS1 DMRS2 Tuvalu 57.40 813.0000 NA 0 0 0 0 0 Uganda 32.70 14.00000 NA 0 0 0 0 0 Ukraine 63.00 54.00000 NA 1 0 0 0 0 United Arab Emirates 65.40 900.0000 82.00 1 0 0 0 0 United Kingdom 71.70 1303.000 99.90 1 0 0 0 0 United R. of Tanzania 36.00 12.00000 47.40 0 0 0 0 0 United States of America 70.00 4187.000 99.9 0 0 1 0 1 Uruguay 67.00 660.0000 94.30 0 0 1 0 0 Uzbekistan 60.20 24.00000 NA 1 0 0 0 0 Vanuatu 52.80 47.00000 71.30 0 0 0 0 0 Venezuela 65.00 150.0000 82.50 0 0 1 1 0 Viet Nam 58.20 17.00000 99.90 0 0 1 0 1 Yemen 49.70 12.00000 NA 0 0 1 0 1 Yugoslavia 66.10 127.0000 NA 1 1 0 0 0 Zambia 30.30 27.00000 72.40 0 0 0 0 0 Zimbabwe 32.90 46.00000 93.1 0 0 0 0 0 90 Table 2: Data on IR and explanatory variables Countries IR1 HEC2 EDU3 DARS DSHI DMRS DMRS1 DMRS2 Bangladesh 0.4070 13.00000 75.10 0 0 0 0 0 Bolivia 0.4580 59.00000 97.40 0 0 1 1 0 Botswana 0.5320 132.000 80.10 0 0 1 0 1 Brazil 0.4810 319.000 97.10 0 0 1 0 1 Bulgaria 0.4430 59.0000 97.90 1 1 0 0 0 Burkina Faso 0.4180 8.00000 32.30 0 0 1 0 1 Cyprus 0.6880 648.000 NA 1 0 0 0 0 Ecuador 0.5320 75.0000 99.90 0 0 1 0 0 Egypt 0.5060 44.0000 95.20 0 0 1 1 0 Georgia 0.4330 45.0000 89.00 0 0 0 0 0 Ghana 0.4800 11.0000 43.40 0 0 0 0 0 Guatemala 0.4970 41.0000 73.80 0 0 1 0 1 Hungary 0.5470 236.000 97.50 1 1 0 0 0 Indonesia 0.5460 18.0000 99.20 0 0 1 0 1 Malaysia 0.6320 110.000 99.90 1 0 0 0 0 Mongolia 0.5790 16.0000 85.10 0 0 1 0 0 Nepal 0.3830 8.00000 78.40 0 0 0 0 0 Peru 0.4240 149.000 93.80 0 0 1 0 0 Philippines 0.5750 40.0000 99.90 0 0 1 1 0 Poland 0.5730 229.000 99.40 1 1 0 0 0 Republic of Korea 0.6120 700.000 99.90 1 1 0 0 0 Senegal 0.4960 23.0000 59.50 0 0 1 1 0 Slovakia 0.5510 311.000 NA 1 1 0 0 0 South Africa 0.5350 268.000 99.90 0 0 1 0 1 Thailand 0.6230 133.000 88.00 0 0 1 0 1 Trinidad and Tobago 0.4730 197.000 99.90 0 0 1 0 1 Uganda 0.3740 14.0000 NA 0 0 0 0 0 United Arab Emirates 0.6330 900.000 82.00 1 0 0 0 0 Viet Nam 0.5700 17.0000 99.90 0 0 1 0 1 Zimbabwe 0.4940 46.0000 93.10 0 0 0 0 0 1Source: WHO (2000), Statistical Annex Table 6 2Source: WHO (2000), Statistical Annex Table 8 3Source: UNDP (2000) 89 Table 3: Data on IFFC and explanatory variables countries IFFC1 DARS DSHI DMRS DMRS1 DMRS2 Bangladesh 0.9560 0 0 0 0 0 Brazil 0.6230 0 0 1 0 1 Bulgaria 0.8500 1 1 0 0 0 Colombia 0.9920 0 0 1 0 0 Guyana 0.9610 0 0 0 0 0 India 0.9620 0 0 1 0 1 Jamaica 0.9210 1 0 0 0 0 Kyrgyzstan 0.8540 1 0 0 0 0 Mexico 0.9030 0 0 1 1 0 Nepal 0.7140 0 0 0 0 0 Nicaragua 0.8740 0 0 1 0 1 Pakistan 0.9490 0 0 1 0 1 Panama 0.9400 0 0 1 1 0 Paraguay 0.8420 0 0 1 1 0 Peru 0.8050 0 0 1 0 0 Romania 0.9390 1 1 0 0 0 Russia 0.8020 1 0 0 0 0 Thailand 0.9520 0 0 1 0 1 United R. of Tanzania 0.9590 0 0 0 0 0 Viet Nam 0.6430 0 0 1 0 1 Zambia 0.8910 0 0 0 0 0 1Source: WHO (2000), Statistical Annex Table 7 90 Table 4: Data on IRD and explanatory variables Countries IRD1 DARS DSHI DMRS DMRS1 DMRS2 Bangladesh 0.7280 0 0 0 0 0 Bolivia 0.7450 0 0 1 1 0 Botswana 0.9050 0 0 1 0 1 Brazil 0.9440 0 0 1 0 1 Bulgaria 0.9960 1 1 0 0 0 Burkina Faso 0.7990 0 0 1 0 1 Chile 0.9180 1 1 0 0 0 Cyprus 0.9910 1 0 0 0 0 Ecuador 0.7230 0 0 1 0 0 Egypt 0.9790 0 0 1 1 0 Georgia 0.8550 0 0 0 0 0 Ghana 0.8470 0 0 0 0 0 Guatemala 0.8120 0 0 1 0 1 Hungary 0.9800 1 1 0 0 0 Indonesia 0.9610 0 0 1 0 1 Malaysia 0.9750 1 0 0 0 0 Mexico 0.9090 0 0 1 1 0 Mongolia 0.9340 0 0 1 0 0 Nepal 0.7920 0 0 0 0 0 Peru 0.8080 0 0 1 0 0 Philippines 0.9860 0 0 1 1 0 Poland 0.9700 1 1 0 0 0 Republic of Korea 0.9920 1 1 0 0 0 Senegal 0.9140 0 0 1 1 0 Slovakia 0.9730 1 1 0 0 0 South Africa 0.8440 0 0 1 0 1 Thailand 0.9490 0 0 0 0 0 Sri Lanka 0.9820 0 0 1 0 1 Trinidad and Tobago 0.9090 0 0 1 0 1 Uganda 0.7960 0 0 0 0 0 United Arab 0.9999 1 0 0 0 0 Emirates 0.8840 0 0 1 0 1 Viet Nam 0.7920 0 0 0 0 0 Zimbabwe 1Source: WHO (2000), Statistical Annex table 6 91 Table 5: Data on IECS and explanatory variables countries IECS1 DARS DSHI DMRS DMRS1 DMRS2 Bangladesh 0.6920 0 0 0 0 0 Benin 0.6800 0 0 0 0 0 Bolivia 0.7250 0 0 1 1 0 Botswana 0.6240 0 0 1 0 1 Brazil 0.7620 0 0 1 0 1 Burkina Faso 0.6540 0 0 1 0 1 Burundi 0.5990 0 0 1 0 1 Cameroon 0.5930 0 0 1 0 1 Central African Republic 0.3010 0 0 0 0 0 Chile 0.9990 1 1 0 0 0 Colombia 0.9120 0 0 1 0 0 Comoros 0.6330 0 0 0 0 0 Côte d'Ivoire 0.4720 0 0 1 0 1 Dominican Republic 0.7890 0 0 1 0 1 Ecuador 0.6790 0 0 1 0 0 Egypt 0.6430 0 0 1 1 0 Ghana 0.6100 0 0 0 0 0 Guatemala 0.7640 0 0 1 0 1 Haiti 0.6020 0 0 1 0 1 India 0.6010 0 0 1 0 1 Indonesia 0.5990 0 0 1 0 1 Japan 0.9990 1 1 0 0 0 Kazakhstan 0.8800 1 0 0 0 0 Kenya 0.6600 0 0 1 0 1 Liberia 0.2450 0 0 0 0 0 Madagascar 0.5440 0 0 1 0 1 Malawi 0.3780 0 0 0 0 0 Mali 0.4890 0 0 1 1 0 Mexico 0.8580 0 0 1 1 0 Morocco 0.748 0 0 1 0 1 Mozambique 0.2610 0 0 1 0 1 Namibia 0.5290 0 0 1 1 0 Nepal 0.5860 0 0 0 0 0 Nicaragua 0.7960 0 0 1 0 1 Niger 0.4570 0 0 1 0 1 Nigeria 0.3360 0 0 0 0 0 Norway 0.9990 1 1 0 0 0 Pakistan 0.4600 0 0 1 0 1 Paraguay 0.8710 0 0 1 1 0 Peru 0.7790 0 0 1 0 0 Philippines 0.8920 0 0 1 1 0 Poland 0.9990 1 1 0 0 0 Rwanda 0.4370 0 0 0 0 0 Senegal 0.7730 0 0 1 1 0 Somalia 0.4950 0 0 0 0 0 Sudan 0.5950 0 0 0 0 0 Thailand 0.8450 0 0 1 0 1 Togo 0.5350 0 0 0 0 0 Trinidad and Tobago 0.8440 0 0 1 0 1 Tunisia 0.7440 0 0 1 0 0 Uganda 0.6530 0 0 0 0 0 United Kingdom 0.9990 1 0 0 0 0 United Republic of Tanzania 0.5300 0 0 0 0 0 1Source: WHO (2000), Statistical Annex Table 5 92 United States of America 0.9660 0 0 1 0 1 Uzbekistan 0.6320 1 0 0 0 0 Yemen 0.5580 0 0 1 0 1 Zambia 0.5350 0 0 0 0 0 Zimbabwe 0.7850 0 0 0 0 0 93 Table 6: Data on IFFC and explanatory variables for enlarged model (with GINI as explanatory variable) Countries IHFC1 DARS GINI2 Bangladesh 0.9560 0 33.60 Brazil 0.6230 0 60.00 Bulgaria 0.8500 1 28.30 Colombia 0.9920 0 57.10 Guyana 0.9610 0 NA India 0.9620 0 37.80 Jamaica 0.9210 1 36.40 Kyrgyzstan 0.8540 1 40.50 Mexico 0.9030 0 53.70 Nepal 0.7140 0 36.70 Nicaragua 0.8740 0 50.30 Pakistan 0.9490 0 31.20 Panama 0.9400 0 48.50 Paraguay 0.8420 0 59.10 Peru 0.8050 0 46.20 Romania 0.9390 1 28.20 Russia 0.8020 1 48.70 Thailand 0.9520 0 41.40 United R. of Tanzania 0.9590 0 38.20 Viet Nam 0.6430 0 36.10 Zambia 0.8910 0 49.80 1Source: WHO (2000), Statistical Annex Table 8 2Source: World Bank (2000 / 2001), Annex Table 5 94 Table 7: Data on IRD and explanatory variables for enlarged model (with GINI as explanatory variable) Countries IRD1 DARS DMRS GINI2 Bangladesh 0.7280 0 0 33.60 Bolivia 0.7450 0 1 42.00 Botswana 0.9050 0 1 NA Brazil 0.9440 0 1 60.00 Bulgaria 0.9960 1 0 28.30 Burkina Faso 0.7990 0 1 48.20 Chile 0.9180 1 0 56.50 Cyprus 0.9910 1 0 NA Ecuador 0.7230 0 1 43.70 Egypt 0.9790 0 1 28.90 Georgia 0.8550 0 0 NA Ghana 0.8470 0 0 32.70 Guatemala 0.8120 0 1 59.60 Hungary 0.9800 1 0 30.80 Indonesia 0.9610 0 1 36.50 Malaysia 0.9750 1 0 48.50 Mexico 0.9090 0 1 53.70 Mongolia 0.9340 0 1 33.20 Nepal 0.7920 0 0 36.70 Peru 0.8080 0 1 46.20 Philippines 0.9860 0 1 46.20 Poland 0.9700 1 0 32.90 Republic of Korea 0.9920 1 0 31.60 Senegal 0.9140 0 1 41.30 Slovakia 0.9730 1 0 19.50 South Africa 0.8440 0 1 59.30 Thailand 0.9490 0 0 34.40 Sri Lanka 0.9820 0 1 41.40 Trinidad and Tobago 0.9090 0 1 NA Uganda 0.7960 0 0 39.20 United Arab 0.9999 1 0 NA Emirates 0.8840 0 1 36.10 Viet Nam 0.7920 0 0 56.80 Zimbabwe 1Source: WHO (2000), Statistical Annex table 6 2Source: World Bank (2000 / 2001), Annex Table 5 95 Table 8: Data on IECS and explanatory variables for enlarged model (with GINI as explanatory variable) countries IECS1 DARS GINI2 Bangladesh 0.6920 0 33.60 Benin 0.6800 0 NA Bolivia 0.7250 0 42.00 Botswana 0.6240 0 NA Brazil 0.7620 0 60.00 Burkina Faso 0.6540 0 48.20 Burundi 0.5990 0 33.30 Cameroon 0.5930 0 NA Central African Republic 0.3010 0 61.30 Chile 0.9990 1 56.50 Colombia 0.9120 0 57.10 Comoros 0.6330 0 NA Côte d'Ivoire 0.4720 0 36.70 Dominican Republic 0.7890 0 48.70 Ecuador 0.6790 0 43.70 Egypt 0.6430 0 28.90 Ghana 0.6100 0 32.70 Guatemala 0.7640 0 59.60 Haiti 0.6020 0 NA India 0.6010 0 37.80 Indonesia 0.5990 0 36.50 Japan 0.9990 1 24.90 Kazakhstan 0.8800 1 35.40 Kenya 0.6600 0 44.50 Liberia 0.2450 0 NA Madagascar 0.5440 0 46.00 Malawi 0.3780 0 NA Mali 0.4890 0 50.50 Mexico 0.8580 0 53.70 Morocco 0.748 0 39.50 Mozambique 0.2610 0 39.60 Namibia 0.5290 0 NA Nepal 0.5860 0 36.70 Nicaragua 0.7960 0 50.30 Niger 0.4570 0 50.50 Nigeria 0.3360 0 50.60 Norway 0.9990 1 25.80 Pakistan 0.4600 0 31.20 Paraguay 0.8710 0 59.10 Peru 0.7790 0 46.20 Philippines 0.8920 0 46.20 Poland 0.9990 1 32.90 Rwanda 0.4370 0 28.90 Senegal 0.7730 0 41.30 Somalia 0.4950 0 NA Sudan 0.5950 0 NA Thailand 0.8450 0 41.40 Togo 0.5350 0 NA Trinidad and Tobago 0.8440 0 NA Tunisia 0.7440 0 40.20 Uganda 0.6530 0 39.20 United Kingdom 0.9990 1 36.10 United Republic of Tanzania 0.5300 0 38.20 United States of America 0.9660 0 40.80 Uzbekistan 0.6320 1 33.30 1Source: WHO (2000), Statistical Annex Table 5 2Source: World Bank ( 2000 / 2001), Annex Table 5 96 Yemen 0.5580 0 39.50 Zambia 0.5350 0 49.80 Zimbabwe 0.7850 0 56.80 97 Table 9: Data on DALE and explanatory variables Countries DALE1 HEC2 EDU3 DARS PHE%2 Afghanistan 37.70 2.000000 NA 0 0.406000 Albania 60.00 26.00000 NA 1 0.777000 Algeria 61.60 44.00000 96.00 0 0.508000 Andorra 72.30 1368.000 NA 0 0.867000 Angola 38.00 NA 34.70 0 0.596000 Antigua-Barbuda 65.80 775.0000 NA 1 0.573000 Argentina 66.70 676.0000 99.90 0 0.575000 Armenia 66.70 36.00000 NA 0 0.415000 Australia 73.20 1730.000 99.90 1 0.720000 Austria 71.60 2277.000 99.90 1 0.673000 Azerbaijan 63.70 20.00000 NA 1 0.793000 Bahamas 59.10 785.0000 94.60 0 0.499000 Bahrain 64.40 478.0000 98.20 1 0.585000 Bangladesh 49.90 13.00000 75.10 0 0.460000 Barbados 65.00 596.0000 97.40 1 0.625000 Belarus 61.70 78.00000 NA 1 0.826000 Belgium 71.60 1918.000 99.90 1 0.832000 Belize 60.90 176.0000 99.90 1 0.516000 Benin 42.20 12.00000 67.60 0 0.472000 Bhutan 51.80 14.00000 13.20 0 0.462000 Bolivia 53.30 59.00000 97.40 0 0.591000 Bosnia and Herzegovina 64.90 77.00000 NA 1 0.926000 32.30 132.0000 80.10 0 0.610000 Botswana 59.10 319.0000 97.10 0 0.487000 Brazil 64.40 NA 87.90 1 0.406000 Brunei Darussalam 64.40 59.00000 97.90 1 0.819000 Bulgaria 35.50 8.000000 32.30 0 0.309000 Burkina Faso 34.60 6.000000 35.60 0 0.356000 Burundi 45.70 21.00000 99.90 0 0.094000 Cambodia 42.20 31.00000 61.70 0 0.201000 Cameroon 72.00 1783.000 99.90 1 0.720000 Canada 57.60 34.00000 99.90 0 0.638000 Cape Verde 36.00 8.000000 46.20 0 0.689000 Central African R. 39.40 7.000000 47.90 0 0.793000 Chad 68.60 315.0000 90.40 1 0.490000 62.30 20.00000 99.90 0 0.249000 Chile 62.90 247.0000 89.40 0 0.545000 China 46.80 14.00000 50.10 0 0.682000 Colombia 45.10 58.00000 78.30 0 0.366000 Comoros 63.40 389.0000 NA 1 0.767000 Congo 66.70 226.0000 91.80 1 0.771000 Cook Islands 42.80 23.00000 58.30 0 0.384000 67.00 352.0000 99.90 1 0.797000 Costa Rica 68.40 131.0000 99.90 1 0.875000 Côte d'Ivoire 69.80 648.0000 NA 1 0.388000 Croatia 68.00 391.0000 99.90 1 0.923000 Cuba 36.30 NA 58.2 0 0.837000 Cyprus 52.30 37.00000 NA 1 0.009000 Czech Republic 69.40 2574.000 99.90 1 0.843000 Democratic R. of Congo 37.90 23.00000 31.90 0 0.729000 Democratic R. of Korea 69.80 282.0000 NA 1 0.650000 1Source: WHO (2000), Statistical Annex Table 5 2Source: WHO (2000), Statistical Annex Table 8 3Source: UNDP (2000) 98 Denmark 62.50 91.00000 91.30 0 0.385000 Djibouti 61.00 75.00000 99.90 0 0.528000 58.50 44.00000 95.20 0 0.270000 Dominica 61.50 182.0000 89.10 0 0.372000 Dominican Republic 44.10 40.00000 79.30 0 0.572000 Ecuador 37.70 6.000000 29.30 0 0.557000 Egypt 63.10 204.0000 99.90 1 0.789000 El Salvador Equatorial Guinea Eritrea Estonia 99 Table 9 (continued): Data on DALE and explanatory variables Countries DALE HEC EDU DARS PHE% Ethiopia 33.50 4.000000 35.20 0 0.362000 59.40 115.0000 99.90 0 0.692000 Fiji 70.50 1789.000 99.90 1 0.737000 Finland 73.10 2369.000 99.90 1 0.769000 France 47.80 138.0000 NA 0 0.665000 Gabon 48.30 12.00000 65.90 0 0.459000 Gambia 66.30 45.00000 89.00 0 0.086000 Georgia 70.40 2713.000 99.90 1 0.775000 Germany 45.50 11.00000 43.40 0 0.470000 Ghana 72.50 905.0000 99.90 1 0.658000 65.50 305.0000 NA 0 0.466000 Greece 54.30 41.00000 73.80 0 0.625000 Grenada 37.80 19.00000 45.60 0 0.572000 Guatemala 37.20 13.00000 52.30 0 0.756000 Guinea 60.20 45.00000 92.80 0 0.791000 Guinea-Bissau 43.80 18.00000 19.40 0 0.336000 Guyana 61.10 59.00000 87.50 0 0.360000 64.10 236.0000 97.50 1 0.849000 Haiti 70.80 2149.000 99.90 1 0.838000 Honduras 53.20 23.00000 77.20 0 0.130000 Hungary 59.70 18.00000 99.20 0 0.368000 Iceland 60.50 108.0000 90.00 0 0.428000 India 55.30 251.0000 74.60 0 0.589000 Indonesia 69.60 1326.000 99.90 1 0.773000 Iran 70.40 1385.000 NA 1 0.750000 Iraq 72.70 1855.000 99.90 1 0.571000 Ireland 67.30 149.0000 95.60 1 0.565000 Israel 74.50 2373.000 99.90 1 0.802000 Italy 60.00 59.00000 NA 0 0.672000 Jamaica 56.40 62.00000 NA 1 0.636000 Japan 39.30 17.00000 65.00 0 0.641000 Jordan 55.30 122.0000 NA 0 0.993000 Kazakhstan 63.20 572.0000 65.20 1 0.874000 Kenya 56.30 15.00000 99.50 1 0.696000 Kiribati 46.10 13.00000 73.00 0 0.627000 Kuwait 62.20 140.0000 99.90 1 0.610000 Kyrgyzstan 60.60 461.0000 76.10 0 0.296000 Lao People's Dem. Rep. 36.90 28.00000 68.60 0 0.726000 34.00 31.00000 NA 0 0.667000 Latvia 59.30 296.0000 99.90 0 0.542000 Lebanon 64.10 167.0000 NA 1 0.757000 Lesotho 71.10 2580.000 NA 1 0.914000 Liberia 36.60 5.000000 58.70 0 0.538000 Libya 29.40 15.00000 98.50 0 0.592000 Lithuania 61.40 110.0000 99.90 1 0.576000 Luxembourg 53.90 107.0000 NA 0 0.639000 Madagascar 33.10 10.00000 38.10 0 0.458000 Malawi 70.50 551.0000 99.90 1 0.589000 56.80 253.0000 NA 0 0.743000 Malaysia 41.40 24.00000 62.90 0 0.303000 Maldives 62.70 129.0000 96.50 1 0.529000 Mali 65.00 240.0000 99.90 0 0.410000 Malta 59.60 242.0000 NA 0 0.923000 Marshall Islands 72.40 1264.000 NA 1 0.625000 53.80 16.00000 85.10 0 0.820000 Mauritania 59.10 66.00000 76.60 0 0.407000 Mauritius 34.40 5.000000 39.60 0 0.713000 100 Mexico 51.60 100.0000 99.30 0 0.126000 Micronesia 35.60 153.0000 91.40 0 0.517000 52.50 593.0000 NA 0 0.990000 Monaco 49.50 8.000000 78.40 0 0.260000 Mongolia Morocco Mozambique Myanmar Namibia Nauru Nepal 101 Table 9 (continued): Data on DALE and explanatory variables Countries DALE HEC EDU DARS PHE% Netherlands 72.00 2041.000 99.90 1 0.707000 New Zealand 69.20 1416.000 99.90 1 0.717000 Nicaragua 58.10 35.00000 78.60 0 0.533000 Niger 29.10 5.000000 24.40 0 0.466000 Nigeria 38.30 30.00000 NA 0 0.282000 Niue 61.60 91.00000 NA 1 0.876000 Norway 71.70 2283.000 99.90 1 0.820000 Oman 63.00 370.0000 67.70 1 0.545000 Pakistan 55.90 17.00000 NA 0 0.229000 Palau 59.00 552.0000 NA 1 0.900000 Panama 66.00 238.0000 89.90 0 0.740000 Papua New Guinea 47.00 36.00000 78.90 0 0.776000 Paraguay 63.00 106.0000 96.30 0 0.356000 Peru 59.40 149.0000 93.80 0 0.397000 Philippines 58.90 40.00000 99.90 0 0.485000 Poland 66.20 229.0000 99.40 1 0.716000 Portugal 69.30 845.0000 99.90 1 0.575000 Qatar 63.50 1042.000 83.30 1 0.575000 Republic of Korea 65.00 700.0000 99.90 1 0.378000 Republic of Moldova 61.50 35.00000 NA 1 0.751000 Romania 62.30 59.00000 99.90 1 0.603000 Russia 61.30 158.0000 99.90 1 0.768000 Rwanda 32.80 13.00000 78.30 0 0.501000 Saint Kitts and Nevis 61.60 404.0000 NA 1 0.515000 Saint Lucia 65.00 211.0000 NA 1 0.651000 Saint Vincent and the G. 66.40 211.0000 NA 1 0.665000 Samoa 60.50 47.00000 96.50 1 0.889000 San Marino 72.30 2257.000 NA 1 0.735000 Sao Tome and Principe 53.50 13.00000 NA 0 0.750000 Saudi Arabia 64.50 260.0000 60.10 1 0.802000 Senegal 44.60 23.00000 59.50 0 0.557000 Seychelles 59.30 424.0000 NA 1 0.762000 Sierra Leone 25.90 11.00000 44.00 0 0.097000 Singapore 69.30 876.0000 91.40 1 0.358000 Slovakia 66.60 311.0000 NA 1 0.818000 Slovenia 68.40 857.0000 NA 1 0.808000 Solomon Islands 54.90 19.00000 NA 0 0.993000 Somalia 36.40 11.00000 NA 0 0.714000 South Africa 39.80 268.0000 99.90 0 0.465000 Spain 72.80 1071.000 99.90 1 0.706000 Sri Lanka 62.80 25.00000 99.90 0 0.454000 Sudan 43.00 13.00000 NA 0 0.209000 Suriname 62.70 114.0000 99.90 0 0.340000 Swaziland 38.10 49.00000 94.6 0 0.723000 Sweden 73.00 2456.000 99.90 1 0.78000 Switzerland 72.50 3564.000 99.90 1 0.693000 Syrian Arab Republic 58.80 151.0000 94.7 0 0.336000 Tajikistan 57.30 11.00000 NA 1 0.878000 Thailand 60.20 133.0000 88.00 0 0.330000 The F. Y. of Macedonia 63.70 120.0000 NA 1 0.848000 Togo 40.70 9.000000 82.30 0 0.428000 Tonga 62.90 141.0000 NA 0 0.460000 Trinidad and Tobago 64.60 197.0000 99.90 0 0.586000 Tunisia 61.40 111.0000 99.90 0 0.417000 Turkey 62.90 118.0000 99.90 0 0.740000 Turkmenistan 24.00000 NA 1 0.860000 102 103 Table 9 (continued): Data on DALE and explanatory variables Countries DALE HEC EDU DARS PHE% Tuvalu 57.40 813.0000 NA 0 0.915000 Uganda 32.70 14.00000 NA 0 0.351000 Ukraine 63.00 54.00000 NA 1 0.755000 United Arab Emirates 65.40 900.0000 82.00 1 0.354000 United Kingdom 71.70 1303.000 99.90 1 0.969000 United R. of Tanzania 36.00 12.00000 47.40 0 0.607000 United States of America 70.00 4187.000 99.9 0 0.441000 Uruguay 67.00 660.0000 94.30 0 0.203000 Uzbekistan 60.20 24.00000 NA 1 0.809000 Vanuatu 52.80 47.00000 71.30 0 0.643000 Venezuela 65.00 150.0000 82.50 0 0.674000 Viet Nam 58.20 17.00000 99.90 0 0.200000 Yemen 49.70 12.00000 NA 0 0.379000 Yugoslavia 66.10 127.0000 NA 1 0.648000 Zambia 30.30 27.00000 72.40 0 0.382000 Zimbabwe 32.90 46.00000 93.1 0 0.434000 104 Table 10: Data on RESPECT and explanatory variables Countries RESPECT 1 HEC2 EDU3 DARS DSHI DMRS Bangladesh 4.07940 13.00000 75.10 0 0 0 Bolivia 4.67300 59.00000 97.40 0 0 1 Botswana 5.34000 132.000 80.10 0 0 1 Brazil 4.85980 319.000 97.10 0 0 1 Bulgaria 4.49690 59.0000 97.90 1 1 0 Burkina Faso 4.22990 8.00000 32.30 0 0 1 Cyprus 6.71040 648.000 NA 1 0 0 Ecuador 5.36550 75.0000 99.90 0 0 1 Egypt 4.88720 44.0000 95.20 0 0 1 Georgia 4.71070 45.0000 89.00 0 0 0 Ghana 4.87820 11.0000 43.40 0 0 0 Guatemala 5.38930 41.0000 73.80 0 0 1 Hungary 5.40800 236.000 97.50 1 1 0 Indonesia 5.45550 18.0000 99.20 0 0 1 Malaysia 6.21670 110.000 99.90 1 0 0 Mongolia 5.88050 16.0000 85.10 0 0 1 Nepal 3.89460 8.00000 78.40 0 0 0 Peru 4.08240 149.000 93.80 0 0 1 Philippines 6.13080 40.0000 99.90 0 0 1 Poland 5.71600 229.000 99.40 1 1 0 Republic of Korea 5.51080 700.000 99.90 1 1 0 Senegal 5.11180 23.0000 59.50 0 0 1 Slovakia 5.31630 311.000 NA 1 1 0 South Africa 5.50670 268.000 99.90 0 0 1 Thailand 5.85100 133.000 88.00 0 0 1 Trinidad and Tobago 4.69610 197.000 99.90 0 0 1 Uganda 3.88280 14.0000 NA 0 0 0 United Arab Emirates 5.90460 900.000 82.00 1 0 0 Viet Nam 6.02800 17.0000 99.90 0 0 1 Zimbabwe 4.93550 46.0000 93.10 0 0 0 1Source: WHO / GPE / FAR data on sub-index responsiveness `respect for persons' as used to establish the index of responsiveness (IR) for WHO (2000). 2Source: WHO (2000), Statistical Annex Table 8 3Source: UNDP (2000) 105 Table 11: Data on CO and explanatory variables Countries CO1 HEC2 EDU3 DARS DSHI DMRS Bangladesh 3.80680 13.00000 75.10 0 0 0 Bolivia 4.28340 59.00000 97.40 0 0 1 Botswana 5.08950 132.000 80.10 0 0 1 Brazil 4.59850 319.000 97.10 0 0 1 Bulgaria 4.25350 59.0000 97.90 1 1 0 Burkina Faso 3.81010 8.00000 32.30 0 0 1 Cyprus 6.87150 648.000 NA 1 0 0 Ecuador 5.16410 75.0000 99.90 0 0 1 Egypt 5.06150 44.0000 95.20 0 0 1 Georgia 3.84230 45.0000 89.00 0 0 0 Ghana 4.46690 11.0000 43.40 0 0 0 Guatemala 4.34370 41.0000 73.80 0 0 1 Hungary 5.42740 236.000 97.50 1 1 0 Indonesia 5.37320 18.0000 99.20 0 0 1 Malaysia 6.31140 110.000 99.90 1 0 0 Mongolia 5.51120 16.0000 85.10 0 0 1 Nepal 3.54630 8.00000 78.40 0 0 0 Peru 4.21200 149.000 93.80 0 0 1 Philippines 5.12290 40.0000 99.90 0 0 1 Poland 5.44660 229.000 99.40 1 1 0 Republic of Korea 6.67550 700.000 99.90 1 1 0 Senegal 4.43520 23.0000 59.50 0 0 1 Slovakia 5.58860 311.000 NA 1 1 0 South Africa 5.00030 268.000 99.90 0 0 1 Thailand 6.47370 133.000 88.00 0 0 1 Trinidad and Tobago 4.59140 197.000 99.90 0 0 1 Uganda 3.26030 14.0000 NA 0 0 0 United Arab Emirates 6.54430 900.000 82.00 1 0 0 Viet Nam 5.27620 17.0000 99.90 0 0 1 Zimbabwe 4.68570 46.0000 93.10 0 0 0 1Source: WHO / GPE / FAR data on sub-index of responsiveness `client orientation' as used to establish the index of responsiveness (IR) for WHO (2000). 2Source: WHO (2000), Statistical Annex Table 8 3Source: UNDP (2000) 106 104 VIII. BIBLIOGRAPHY Atim C. (1998). Contribution of mutual health organizations to financing, delivery, and access to health care. Maryland: Abt Associates, Partnerships for health reform, Technical report no.18. Bennett S., Creese A. and Monasch R. (1998). Health Insurance Schemes for People Outside Formal Sector Employment. Geneva: World Health Organization, WHO/ARA/CC/98.1. Carrin G., De Graeve D. and Devillé L.(eds.) (1999). The Economics of Health Insurance in Low and Middle-Income Countries. Social Science and Medicine (special issue), vol.48 Gakidou E. & Murray C.J.L. (2000). Estimates of the distribution of child survival in 191 countries. Geneva: WHO, GPE discussion paper no.19. Ginneken van W. (ed.) (1999). Social Security for the Excluded Majority-Case studies of developing countries. Geneva: International Labour Office. Greene W.H. (2000). Econometric Analysis. New Jersey: Prentice Hall Inc, 4th ed. ILO and PAHO (1999). Synthesis of case studies of micro-insurance and other forms of extending social protection in health in Latin America and the Caribbean. Meeting on Extension of social protection in health to excluded groups in Latin America and the Caribbean, Mexico, 29 November-1 December 1999 (http://oitopsmexico99.org.pe) Mathers C., Sadana R., Salomon J., Murray C.J.L. and Lopez A.D. Estimates of DALE for 191 countries: methods and results. Geneva: WHO, GPE discussion paper no.15. Mukherjee C., White H. and Wuyts M. (1998). Econometrics and Data analysis for Developing Countries. London: Routledge. Murray C.J.L. & J. Frenk (2000). A framework for assessing the performance of health systems. Bulletin of the World Health Organization, vol.78, no.6, pp.717-731. Murray C.J.L., Knaul F., Musgrove Ph., Ke Xu and Kei Kawabata (2000). Defining and measuring fairness in financial contribution to the health system. Geneva: WHO, GPE discussion paper no.24. Musgrove Ph. (1996). Public and Private Roles in Health--Theory and Financing Patterns. Washington DC: World Bank Discussion Paper no. 339. Nolan B. and Turbat V. (1995). Cost Recovery in Public Health Services in Sub-Saharan Africa. Washington: Economic Development Institute of the World Bank. Social Security Administration (1999). Social Security Programs throughout the World-1999. Washington: U.S. Government Printing Office. UNDP (2000). Human Development Report 2000. New York: Oxford University Press. Valentine N., de Silva A. & Murray C.J.L. (2000). Estimates of responsiveness level and distribution for 191 countries: methods and results. Geneva: WHO, GPE discussion paper no.22. WHO (2000). The World Health Report 2000. Health systems : improving performance. Geneva : WHO. World Bank (1999). World Development Report 1999 / 2000. New York: Oxford University Press. 105 World Bank (2000). World Development Report 2000 / 2001. New York: Oxford University Press. 106 About this series... This series is produced by the Health, Nutrition, and Population Family (HNP) of the World Bank's Human Development Network. The papers in this series aim to provide a vehicle for publishing preliminary and unpolished results on HNP topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For free copies of papers in this series please contact the individual authors whose name appears on the paper. Enquiries about the series and submissions should be made directly to the Editor in Chief Alexander S. Preker (apreker@worldbank.org) or HNP Advisory Service (healthpop@worldbank.org, tel 202 473-2256, fax 202 522-3234). For more information, see also www.worldbank.org/hnppublications. THE WORLD BANK 1818 H Street, NW Washington, DC USA 20433 Telephone: 202 477 1234 Facsimile: 202 477 6391 Internet: www.worldbank.org E-mail: feedback@worldbank.org