Table 5 Vocational vs. academic upper secondary education.

From: Dynamics of returns to vocational education in China: 2010–2017

 

Model1

Model2

Model3

Model4

 

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

Panel A: IV estimates

F statistic of first stage

22.06

 

1.083

 

2.623

 

196.2

 

Durbin p-value

0.056

 

0.332

 

0.001

 

0.827

 

βvus

 

−0.734* (0.412)

 

1.420 (2.050)

 

2.532* (1.351)

 

0.031 (0.132)

N

 

3983

 

1700

 

4143

 

4143

Panel B: Lewbel estimates

Chi-sq of Breusch-Pagan test

116.54

 

9.97

 

86.95

 

139.08

 

p-value

0.000

 

0.002

 

0.000

 

0.000

 

βvus

 

−0.193** (0.089)

 

−0.031 (0.231)

 

0.019 (0.103)

 

0.047 (0.079)

N

 

4015

 

1715

 

4277

 

4416

Adj R-squared

 

0.276

 

0.253

 

0.280

 

0.279

Panel C: PSM estimates

 

N.treat

N. control

 

ATT

 

Std. Err.

 

t-statistic

 

(1)

(2)

 

(3)

 

(4)

 

(5)

 

2612

2103

 

−0.036

 

0.035

 

−1.009

  1. Panel A and Panel B: Model 1 is instrumented by the proportion of university graduates over year. Model 2 is instrumented by the enrollment ratio. Model 3 is instrumented by parental education. Model 4 is instrumented by a dummy which equals one if the enrollment year is after 1985 and zero otherwise. The control variables are the same for all four models, except that Model 3 does not directly control for parental education. Panel C: ATT stands for Average Treatment effect on the Treated. Bootstrapped standard errors with 200 repetitions are reported.
  2. ***p < 0.01; **p < 0.05; *p < 0.1.