Table 3.

Difference-in-difference estimates for multinomial logistic regression (unemployment reason)

Control group	Treatment group	Unemployment reason
			Relative Risk Ratio Pre-treatment: 2014, 6 ~ 7		Relative Risk Ratio Time trends control
			Voluntary	Involuntary	Voluntary	Involuntary
15–49	Over 50	Treated × post	1.0379b (0.0017)	1.0512b (0.0022)	1.0586b (0.0017)	1.0727b (0.0023)
		post			0.9518b (0.0008)	1.0616b (0.0015)
	50s	Treated × post	1.0067b (0.0020)	1.1536b (0.0029)	1.0490b (0.0022)	1.1694b (0.0030)
		post			0.9519b (0.0008)	1.0623b (0.0016)
30–49	Over 50	Treated × post	1.0410b (0.0019)	1.0644b (0.0026)	1.0263b (0.0020)	1.0966b (0.0027)
		post			0.9836b (0.0013)	1.0388b (0.0020)
	50s	Treated × post	1.0087b (0.0016)	1.1001b (0.0027)	0.9605b (0.0015)	1.1193b (0.0028)
		post			0.9849b (0.0013)	1.0442b (0.0020)
15–29	Over 50	Treated × post	1.0287b (0.0018)	1.0363b (0.0029)	1.0900b (0.0020)	1.0468b (0.0031)
		post			0.9241b (0.0011)	1.0865b (0.0027)
	50s	Treated × post	0.9973 (0.0022)	1.1355b (0.0036)	1.0823b (0.0024)	1.1415b (0.0037)
		post			0.9239** (0.0011)	1.0862b (0.0027)

^a and ^b indicate that the estimate is significant at the 0.05 and 0.01 levels, respectively. Standard errors are reported in parentheses. Reported values are relative risk ratio and the reference group is the employed. The unemployed due to personal affairs, family affairs, raising children, house affairs, physical and mental disorder, retirement, age and working condition complaints are defined as the voluntary unemployment and involuntary unemployment is defined as the loss of jobs due to company suspension, closure, early∙honorary retirement∙ layoff, being out of work or sluggish business