Skip to main content

Table 12 Robustness checks, first stage estimates

From: Intergenerational transmission of human capital in Europe: evidence from SHARE

Dep. Var.: Parent’s education

(1)

(2)

(3)

(4)

Sample

Overall

Overall

Sons

Daughters

Panel A: Post-WWII sample of parents (1935-1956)

Compulsory education

0.228***

0.241***

0.321***

0.151

 

(0.082)

(0.083)

(0.104)

(0.102)

Compulsory educ*female (parent)

 

-0.034***

-0.039**

-0.029

  

(0.013)

(0.017)

(0.018)

Observations

5,247

5,247

2,639

2,608

Panel B: w/o cohort F.E. for children

Compulsory education

0.217***

0.226***

0.301***

0.153

 

(0.074)

(0.077)

(0.102)

(0.094)

Compulsory educ*female (parent)

 

-0.053***

-0.055***

-0.048***

  

(0.012)

(0.016)

(0.017)

Observations

6,184

6,184

3,117

3,067

Panel C: Parent’s years of schooling<11

Compulsory education

0.145***

0.148***

0.166**

0.110

 

(0.047)

(0.047)

(0.071)

(0.079)

Compulsory educ*female (parent)

 

-0.007

-0.004

-0.008

  

(0.008)

(0.011)

(0.011)

Observations

2,829

2,829

1,407

1,422

Panel D: Narrow windows around the pivotal cohorts (+/- 6 years)

Compulsory education

0.222**

0.236**

0.306**

0.173

 

(0.094)

(0.096)

(0.124)

(0.129)

Compulsory educ*female (parent)

 

-0.034**

-0.040*

-0.028

  

(0.016)

(0.022)

(0.025)

Observations

2,804

2,804

1,382

1,422

Panel E: Binary instrument

First reform

0.429***

0.477***

0.701***

0.242

 

(0.160)

(0.169)

(0.225)

(0.231)

Second reform

0.620***

0.799***

0.853**

0.710**

 

(0.204)

(0.228)

(0.343)

(0.307)

First reform*female (parent)

 

-0.107

-0.178

-0.030

  

(0.129)

(0.163)

(0.174)

Second reform*female (parent)

 

-0.473*

-0.476

-0.407

  

(0.273)

(0.453)

(0.564)

Observations

6,184

6,184

3,117

3,067

  1. Notes: All specifications include controls for country dummies, birth cohort dummies for parents and children (in 1-year intervals), and socio-demographic characteristics. In Panel D, for the countries with more than one reform, I consider only the most recent reform. Standard errors clustered at the parents’ country and cohort level are reported in parentheses. * Significant at 10%; ** significant at 5%; *** significant at 1%.