How sensitive are matching estimates of active labor market policy effects to typically unobserved confounders?

Journal for Labour Market Research

Table 2 Relevance of typically unobserved confounders

	Baseline	Auxiliary							Extended
	Specification	Specifications							Specification
	(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)
Outcomes of non-participants
Regular employment
\(R^2\)	0.1811	0.1823	0.1889	0.1862	0.1879	0.1818	0.1896	0.1822	0.2014
p-value of joint significance of U		0.1573	0.0000	0.0002	0.0000	0.4080	0.0000	0.4180	0.0000
Real monthly labor earnings
\(R^2\)	0.1790	0.1806	0.1910	0.1851	0.1862	0.1807	0.1886	0.1803	0.2057
p-value of joint significance of U		0.0589	0.0000	0.0000	0.0000	0.0481	0.0000	0.3155	0.0000
Treatment assignment
Pseudo-\(R^2\)	0.1003	0.1009	0.1059	0.1042	0.1027	0.1007	0.1037	0.1015	0.1124
p-value of joint significance of U		0.5120	0.0001	0.0039	0.0051	0.6735	0.0014	0.3579	0.0021
Control variables
Typically observed covariates X	\(\checkmark\)	\(\checkmark\)	\(\checkmark\)	\(\checkmark\)	\(\checkmark\)	\(\checkmark\)	\(\checkmark\)	\(\checkmark\)	\(\checkmark\)
Typically unobserved covariates U
Attitudes towards work		\(\checkmark\)							\(\checkmark\)
Job search behavior			\(\checkmark\)						\(\checkmark\)
Concessions willing to make for a job				\(\checkmark\)					\(\checkmark\)
Satisfaction in different domains					\(\checkmark\)				\(\checkmark\)
Participation, social status and networks						\(\checkmark\)			\(\checkmark\)
(Mental) Health							\(\checkmark\)		\(\checkmark\)
Inter-generational information								\(\checkmark\)	\(\checkmark\)

This table compares summary statistics from OLS and logit regressions to inspect the relevance of (each block of) typically unobserved covariates