본문 바로가기 주메뉴 바로가기
국회도서관 홈으로 정보검색 소장정보 검색

목차보기

Title page

Contents

Abstract 3

1. Introduction 4

2. Background and study design 9

2.1. Setting 9

2.2. Participants 11

2.3. Intervention and randomization 12

2.4. Measurement and timing 13

2.5. Sample size and power 14

3. Estimation 15

3.1. Effects on enrollment, utilization and heath 15

3.2. Adverse selection 18

4. Results 20

4.1. Insurance enrollment 20

4.2. Insurance and hospital utilization 22

4.3. Health outcomes 25

4.4. Mechanisms through which insurance generates value 26

4.5. Adverse selection 27

5. Welfare analysis 29

5.1. Does offering (free) public insurance improve welfare? 29

5.2. What is the optimal public insurance premium? 38

6. Conclusion 41

References 43

Online Appendix 64

A. Methods 64

B. Welfare 68

C. Figures 73

D. Tables 80

Table 1. Characteristics of sample villages 54

Table 2. Household characteristics in different access-to-insurance conditions 55

Table 3. Trial design in each district 56

Table 4. Effects of premium & subsidy combinations on insurance enrollment 57

Table 5. Effect of premium & subsidy combinations on use of insurance and on hospital utilization (intent to treat estimates) 58

Table 6. Effect of insurance enrollment on use of insurance and on hospital utilization (complier average treatment effect) 59

Table 7. Number of significant health outcomes per category 60

Table 8. Effect of insurance access on total and out-of-pocket (OOP) medical expenditures (ITT) 61

Table 9. Enrollment by predicted spending 62

Table 10. Welfare calculations using four different measures of cost 63

Figure 1. Illustration of adverse selection 51

Figure 2. Inverse demand, cost and welfare calculations at midline 52

Figure 3. Inverse demand, cost and welfare calculations at endline 53

Table A.1. Outcomes measured by topic at midline (18 months post-intervention) and endline (3.5 years post-intervention) surveys 80

Table A.2. Effects of different premium & subsidy combinations on insurance enrollment 82

Table A.3. Enrollment by health (education, raven scores and risk aversion controlled) 83

Table A.4. Effect of insurance access on total and out-of-pocket (OOP) medical expenditures (CATE) 84

Table A.5. Effect of insurance enrollment on use of insurance for hospital care and on hospital utilization (CATE), by wealth of household 85

Table A.6. Reasons given for respondents' inability to use RSBY insurance card, by survey and utilization status 86

Table A.7. Effect of different premium & subsidy combinations on midline (18 mos.) health outcomes (intent to treat estimate) 87

Table A.8. Effect of different premium & subsidy combinations on endline (3.5 yrs.) health outcomes (intent to treat estimate) 90

Table A.9. Effect of insurance enrollment on midline (18 mos.) health outcomes (complier average treatment effect estimates) 95

Table A.10. Effect of insurance enrollment on endline (3.5 yrs.) health outcomes (complier average treatment effect estimates) 97

Table A.11. Number of significant health outcomes per category, without multiple testing adjustments 100

Table A.12. Estimates of demand and average cost using multiple different measures of cost 102

Table A.13. Sample size per insurance access group and response rates by survey and insurance access group 103

Table A.14. Respondent and response rate by module and insurance access group 104

Figure A.1. Balancing test result 73

Figure A.2. Estimates of total effects of insurance (CATE) on a randomly-selected set of health outcomes from midline (18 months) 74

Figure A.3. Estimates of total effects of insurance (CATE) on a randomly-selected set of health outcomes from endline (3.5 years) 75

Figure A.4. Quantile treatment effects on OOP spending 76

Figure A.5. Timeline of study 77

Figure A.6. Subject flowchart, with exclusions and attrition by endline 78

Figure A.7. Predicted medical expenditures 79