Panel A: Effect of early nonemployment in the structural equation:
 
 
OLS

2SLS

Standard errors
^{a}

Robust

Cluster g∗p

Robust

Cluster g∗p

Outcomes:
^{b}
 
(1)

(2)

(3)

(4)

Salaried employment

coeff

–0.00169***

–0.00169***

–0.00199

–0.00199


se

(0.00034)

(0.00041)

(0.00387)

(0.00348)


Pval
 
0.00019
 
0.57020


Bootstrap Pval ^{c}
 
0.00000
 
0.62462


Exogeneity test Pval ^{d}
  
0.937

0.931

Selfemployment

coeff

0.00054*

0.00054

0.00219

0.00219


se

(0.00030)

(0.00041)

(0.00340)

(0.00313)


Pval
 
0.19253
 
0.48668


Bootstrap Pval
 
0.19219
 
0.52853


Exogeneity test Pval
 
0.619
 
0.587

Overall employment

coeff

–0.00115***

–0.00115***

0.00020

0.00020


se

(0.00021)

(0.00025)

(0.00229)

(0.00164)


Pval
 
0.00005
 
0.90232


Bootstrap Pval
 
0.00000
 
0.91291


Exogeneity test Pval
  
0.540

0.386

Log earnings

coeff

–0.0269***

–0.0269***

–0.0947**

–0.0947***


se

(0.0033)

(0.0040)

(0.0447)

(0.0354)


Pval
 
2.78E08
 
0.01051


Bootstrap Pval
 
0
 
0.03003


Exogeneity test Pval
   
0.0361

Log hours worked

coeff

–0.0203***

–0.0203***

–0.0666**

–0.0666***


se

(0.0024)

(0.0029)

(0.0326)

(0.0251)


Pval
 
1.05E08
 
0.01105


Bootstrap Pval
 
0
 
0.03403


Exogeneity test Pval
   
0.0481

Panel B: Effect of the instrument from the first stage (OLS)
  
Outcome:

Standard errors:

Robust

Cluster (g*p)
  
Early nonempl.

Coeff

5.0319***

5.0319***
  

se

(1.6519)

(1.7139)
  

Pval
 
0.0053
  

Bootstrap Pval
 
0.0120
  

F stat
 
8.620
  

Bootstrap F stat ^{e}
 
5.84
  
 Standard errors between parentheses. Panel A reports results from estimating β in Eq. (2), excluding \(\overline {UR}_{p}\) and m
i
n
U
R
_{
pt
}. β is the effect of one pp increase in \(y^{0}_{it_{1}}\), i.e. the % of hours spent in nonemployment at potential experience 0–2 relative to potential total hours if one would work fulltime during the whole period. For clustered standard errors, we report the Pvalue and the wild bootstrap Pvalue. Column 1–2 (3–4) show OLS (2SLS). In 2SLS the provincial unemployment rate at graduation is used as instrument for \(y^{0}_{it_{1}}\!\). Panel B reports the effect of the instrument on \(y^{0}_{it_{1}}\) in the first stage and the corresponding F statistic
 ***p <0.01, **p <0.05, *p <0.1

^{a}Robust indicates heteroskedasticrobust standard errors. Clusters are defined by graduation year g and province of residence at graduation p (G =44 clusters)

^{b}The outcomes are measured at potential experience 6. For continuous outcomes we add value one before taking the log, so that nonsalaried employed at the moment of evaluation are included with outcomes equal to zero after the logarithmic transformation

^{c}Computed according to the wild bootstrap procedure proposed by Davidson and MacKinnon (2010) for 999 repetitions

^{d}With clustered standard errors, this test is defined as the difference between two SarganHansen statistics: one for the equation where \(y^{0}_{it_{1}}\) is treated as endogenous and one for the equation where \(y^{0}_{it_{1}}\) is treated as exogenous. Under the null that \(y^{0}_{it_{1}}\) is exogenous, the statistic is distributed as χ
^{2}(1)

^{e}Bootstrap F statistic is the F statistic corresponding to the Bootstrap Pvalue of the t statistic of the instrument: we rely on the equivalence between F and t distribution: with G=44, t
^{2}(G−1)=F(1,G−1)