Table 4.

Coefficients from Heckman Selection Models of Factors Predicting Emigration Cost (standard errors in parentheses)

	Fuzhou-US	Mingxi-Europe
Owning family business	0.18+ (0.10)	0.04 (0.05)
Cadre	−0.40 (0.27)	0.26 (0.19)
Cadre in household	−0.28** (0.07)	0.05 (0.05)
Prior emigrant in household (ref. no prior emigrant)
Prior emigrant within 2 years	−0.16+ (0.09)	−0.05 (0.06)
Prior emigrant 2–5 years	−0.43** (0.08)	−0.07 (0.06)
Prior emigrant over 5 years	−0.69** (0.08)	0.12 (0.09)
Village emigration prevalence ratio	2.74** (0.31)	0.06 (0.16)
Rural area	0.01 (0.12)	0.06 (0.06)
−2 log likelihood	13,444.8	1,168.2
Likelihood ratio test of ρ = 0 (p-value)	< 0.001	0.22
N	6,632	1,296

⁺p <0.1;

^*p <0.05;

^**p<0.01.

Note: We took the midpoints of the emigration cost categories and converted them into the 2003 constant U.S. dollars. They are log transformed and included in the regression. We used the Heckman selection model because emigration costs were observed only for emigrants. The first stage regression (which predicts emigration status) is not shown. The predictors included in the first regression are the same as those included in Table 3.