Title page 1

Contents 1

Abstract 2

1. Introduction 3

2. Institutional Setting and Data 8

2.1. The Danish Pension System 8

2.2. The 2006 Welfare Agreement and the 2011 Retirement Reform 9

2.3. Data Sources 11

2.3.1. Health and Healthcare Utilization Outcomes 12

2.3.2. Labor Market Outcomes 14

3. Empirical Strategy 14

3.1. Window Choice 17

4. Results 18

4.1. The Increased ERA's Effect on Labor Force Participation 18

4.2. The Increased ERA's Effect on Health and Healthcare Utilization 22

4.2.1. Alternative Specifications and Robustness Checks 27

4.3. Additional Health and Healthcare Utilization Outcomes 29

4.4. Cutoff-Specific Effects 30

5. Mechanisms 31

5.1. Characterization of Compliers 31

5.2. Treatment Effect Heterogeneity using Causal Forest 35

5.3. Marginal Health and Healthcare Utilization Effects of Increased ERA 40

5.3.1. Estimated MTEs 41

6. Discussion and Conclusion 43

Bibliography 45

Appendix 54

A. Overview of Selected Literature 54

B. Details on Data and Variables 55

C. Alternative Specifications and Robustness 61

D. Additional Results 72

E. Heterogeneous Treatment Effects Estimated by Causal Forest 84

F. Marginal Treatment Effects 92

Tables 19

Table 1. Effects on Labor Force Participation 19

Table 2. Effects on Health and Healthcare Utilization Outcomes 23

Table 3. Effects on Health and Healthcare Utilization Outcomes 26

Table 4. Estimated Treatment Effect Parameters 42

Figures 3

Figure 1. Public Debt and Retirement Ages 3

Figure 2. The 2011 Reform's Effect on Retirement Ages 10

Figure 3. Employment Degree for Cohorts Across Age 11

Figure 4. Effects on Labor Force Participation 20

Figure 5. Dynamic Effects on Labor Force Participation 21

Figure 6. Effects on Health and Healthcare Utilization Outcomes 24

Figure 7. Effects on Health and Healthcare Utilization Outcomes Over Time 25

Figure 8. Characterization of ERA Compliers 34

Figure 9. Estimated Group Average Treatment Effects (1/2) 37

Figure 10. Estimated Group Average Treatment Effects (2/2) 38

Figure 11. Estimated Marginal Treatment Effects 43

Appendix Tables 54

Table A.1. Literature Overview 54

Table B.1. Sample Descriptives 56

Table C.1. Effects on LFP - Conditional on Employment in Previous Period 62

Table C.2. The Continuity Framework: Effects on Health and Healthcare Utilization Outcomes 63

Table C.3. Effects on Health and Healthcare Utilization Outcomes 67

Table C.4. RD-DD: Effects on Health Outcomes 69

Table C.5. Pooled RD-DD: Effects on Health Outcomes: ERA payment 71

Table D.1. Effects on Health and Healthcare Utilization Outcomes 73

Table D.2. Effects on Mortality 74

Table D.3. Effects on Health and Healthcare Utilization Outcomes: 1954:Q2-1954:Q3 78

Table D.4. Effects on Health and Healthcare Utilization Outcomes: 1955:Q2-1955:Q3 79

Table D.5. Effects on Health and Healthcare Utilization Outcomes: 1953:Q4-1954:Q1 80

Table D.6. Effects on Health and Healthcare Utilization Outcomes: 1954:Q4-1955:Q1 81

Table D.7. Effects on Health and Healthcare Utilization Outcomes: 1956:Q2-1956:Q3 82

Table E.1. GP Visits: Characteristics for Individuals with High vs. Low Predicted Treatment Effects 88

Table E.2. Painkillers: Characteristics for Individuals with High vs. Low Predicted Treatment Effects 89

Table E.3. Antidepressants: Characteristics for Individuals with High vs. Low Predicted Treatment Effects 90

Table E.4. CVD Medicine: Characteristics for Individuals with High vs. Low Predicted Treatment Effects 91

Table F.1. Estimating MTEs: First Stage 95

Table F.2. Marginal Treatment Effects - Different Specification Choices 96

Appendix Figures 55

Figure B.1. Frequency of Birth Dates 55

Figure B.2. Prevalence of CVD Medicine, Painkillers, and Antidepressants 57

Figure B.3. Prevalence of All Medicine Categories (ATC 1st level) 57

Figure B.4. Decomposition of Selected Medicine Categories 58

Figure B.5. GP Visits by Month 58

Figure B.6. Transfers Excl. Retirement Benefits 59

Figure B.7. Ordinary Transfers and Health-Related Transfers 59

Figure B.8. Labor Force Participation Over Time 60

Figure C.1. Choosing the Bandwidth, w 61

Figure C.2. Continuity Assumption: Effects on Health Outcomes 64

Figure C.3. Robustness to Bandwidth Choice 65

Figure C.4. Donut Hole Specifications 66

Figure C.5. RD-DD: Effects on Health Outcomes 69

Figure C.6. Pooled RD-DD: Effects on Health Outcomes: ERA payment 71

Figure D.1. Dynamic Effects on Additional Health and Healthcare Utilization Outcomes 74

Figure D.2. Effects on Health and Healthcare Utilization Outcomes Over Time: 1954:Q2-1954:Q3 76

Figure D.3. Effects on Health and Healthcare Utilization Outcomes Over Time: 1955:Q2-1955:Q3 77

Figure D.4. Characterization of ERA Compliers 86

Figure F.1. Common Support and Treatment Parameter Weights 94

Figure F.2. Marginal Treatment Effect Curves - Different Specification Choices 97