Table 2.

FE-IV estimation of total impact of retirement on subsequent health outcomes

	Physical health	Depression	Anxiety
Retirement	− 0.093*** (0.04)	0.220*** (0.04)	0.113*** (0.03)
Observations	1101	1098	1008

Each cell presents the total effect of retirement on corresponding health outcomes using the fixed-effect instrumental variable estimation method. In all analyses, we control for health behaviors, age, marital status, and income levels. Instrument is a dummy variable with value 1 if individual is older than 65 years old, and 0 otherwise. The sample is limited to pre-retirement healthy individuals in all these four health outcomes. The health outcomes are binary variables. Table 7 in Appendix reports the detailed estimation results. Robust individuals-clustered standard errors are in parenthesis. *** indicate the significant at the 1% level