The effect of vocational rehabilitation on the employment outcomes of disability insurance beneficiaries: new evidence from Canada

Table 7 Estimates of the ATET of VR on individual outcomes, men

	Matching estimator		Inverse probability weighting	Genetic matching
	Kernel	Local linear	Inverse probability weighting	Genetic matching
	Sample restricted to propensity score values [0.001, 0.40]
Leaving disability rolls	0.057	0.073	0.039	−0.036
Leaving disability rolls	(0.082)	(0.113)	(0.076)	(0.096)
Gainful employment	0.062	0.070	0.030	−0.073
Gainful employment	(0.080)	(0.099)	(0.077)	(0.096)
Substantial gainful employment	0.025	0.035	0.001	−0.055
Substantial gainful employment	(0.078)	(0.097)	(0.074)	(0.098)
Sample size	786	786	786	786
	Sample restricted to propensity score values [0.001, 0.35]
Leaving disability rolls	0.014	0.015	0.003	−0.067
Leaving disability rolls	(0.083)	(0.082)	(0.084)	(0.111)
Gainful employment	0.022	0.025	0.007	0.000
Gainful employment	(0.081)	(0.079)	(0.083)	(0.109)
Substantial gainful employment	0.009	0.026	0.013	−0.022
Substantial gainful employment	(0.083)	(0.078)	(0.083)	(0.111)
Sample size	774	774	774	774
	Sample restricted to propensity score values [0.001, 0.50]
Leaving disability rolls	0.033	0.044	0.018	0.000
Leaving disability rolls	(0.082)	(0.085)	(0.072)	(0.101)
Gainful employment	0.024	0.027	−0.008	−0.017
Gainful employment	(0.081)	(0.088)	(0.073)	(0.097)
Substantial gainful employment	−0.008	−0.005	−0.033	−0.049
Substantial gainful employment	(0.078)	(0.085)	(0.073)	(0.100)
Sample size	805	805	805	805

Notes: ATET denotes average treatment effect on the treated. Observations with propensity score values outside the range in square brackets are omitted from the sample used to estimate the ATET. Standard errors in parentheses. The kernel and local linear matching estimates were obtained with the Epanechnikov kernel, with bandwidths computed using cross-validation; see Table 1 for the bandwidths. Genetic matching estimates are based on single nearest neighbor matching with replacement. Standard errors for kernel and local linear matching and inverse probability weighting are obtained by bootstrapping with 1000 replications. Standard errors for genetic matching are based on Abadie and Imbens (2006).