Fathers matter! – choices of occupations of parents and children

Study on link between occupation of parent and occupation of a child

1. Fathers matter!
The role of parents in gender occupational segregation
Magdalena Smyk
GRAPE
10th July 2017
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 1 / 1

2. Gender dierences in occupations
Why is it interesting?
common and prevailing phenomenon
(Goldin, 2013, 2014; Blau et al. 2013)
contributes to explain gender wage dierences
(World Development Report, 2012)
\external" barriers and changes fail to explain it
(Bush and Holst 2011; Blau et al. 1998, 2013)
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 2 / 1

3. Motivation
Gender dierences in the choice of occupation:
preferences
scientically unsatisfactory explanation
(Becker 1996, Argwal 2000)
traits (risk aversion, competitiveness, altruism, etc.)
(Holt and Laury 2005, Niederle and Vesterlund 2011, Andreoni 1989)
small explanatory power
natureornurture?
(Finucane et al. 2000, Gneezy et al. 2009)
still unresolved
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 3 / 1

4. Contribution
The role of parents:
inheritance of a profession
(Laband and Lentz 1990, Korupp et al. 2002)
transmission of ules" related to choice of occupation
- literature gap
Hypothesis
Parents' choice of occupation (its gender context - norms) inuence
their children's choices of occupation.
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 4 / 1

5. Theoretical background
Roy's choice of occupation model (1951)
Gender identity (Akerlof and Kranton 2000, 2010)
Choice of occupation (Humlum et al. 2012)
Uj=Uj(wj(j; "j)); "j;Ij(j;cj; "j;P));
wherewj(j; "j) - wage in occupation"j,j- individual characteristics,cj-
social category (gender),P- gender rules related to choice of occupation
How to operationalizeP- gender rules jand parents (p)?
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 5 / 1

6. Theoretical background - extension
How the gender norms are shaped?
Pj=Pj(aj;aj;cj; j;Pj;p);
whereaj- actions of individualj,aj- actions of other people, cj- gender,j-
individual characteristics,Pj;p- parents' gender rules
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 6 / 1

7. Measure of gender intensity in an occupation
FemOcc Index
FemOcci;t=
share of womeni;t
share of woment
;
where i {narrowoccupational group, t { year
representation measure)data requirements
American Community Survey provides the largest consistently coded sample
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 7 / 1

8. Measure of gender intensity of an occupation
FemOcc Index
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 8 / 1

9. Measure of gender intensity of an occupation
FemOcc Index
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 9 / 1

10. Measure of gender intensity of an occupation
FemOcc Index
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 10 / 1

11. Inactive parents
Being inactive (especially for mothers) is an important occupation from
the perspective of gender rules:
FemOccinactive;t=
share of inactive woment
share of women in working age populationt
;
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 11 / 1

12. Transmission of norms related to the choice of occupation: the empirical
study
Do parents transmitgender rules related to choice of
occupationto their children?
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 12 / 1

13. Strategy
FemOcci=+1Father
0
s FemOcc+2Mother
0
s FemOcci+Xi+i
Variables
Dependent variable:
individuali(FemOcci)
Independent variables:
Father
0
s FemOcc- gender intensity ofi's father occupation
Mother
0
s FemOcc- gender intensity ofi's mother occupation
Control variables (Xi): gender, education, birth year, siblings order,
family xed eects
Data:Panel Study of Income Dynamics1968-2013 merged with
measures ofFemOccifrom ACS
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 13 / 1

14. Transmission of rules related to choice of occupation
Panel Study of Income Dynamics Data 1968-2013
Multiple occupations for parents and children:
specication of parents occupations:
observed by the child between 13 and 18 years old
child occupation: the top occupation - (highest code) Inactive parents
Method
OLS regression with:
standard errors clustered within family
andfamily xed eects
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 14 / 1

15. Transmission of rules related to choice of occupation
What do we expect?
Correlation between FemOcc indexes of:
fatherandson-positive motherandson-negative(but possibly weaker) motheranddaughter-positive fatheranddaughter-negative(but possibly weaker)
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 15 / 1

16. Transmission of rules related to choice of occupation
TotalWithoutOnlysamplethe samesiblingsoccupation
FATHER - SON
0.085**0.080* 0.080**(0.042)(0.043) (0.041)
MOTHER - SON
0.002-0.05 -0.06(0.038)(0.041) (0.039)
MOTHER - DAUGHTER
-0.008-0.011 -0.011(0.047)(0.049) (0.047)
FATHER - DAUGHTER
0.000*0.024 0.024(0.046)(0.047) (0.045)
Observations
576756265193
R
2
0.770.770.75
Standard errors in parentheses. *** -pvalue<0.01, ** -pvalue<0.05, * -pvalue<0.1
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 16 / 1

17. Is the eect stable in time?
No.
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 17 / 1

18. Transmission of rules related to choice of occupation
Children born before 1970
Total Without Only
sample the same siblings
occupation
FATHER - SON 0.057 0.060 0.060
(0.063) (0.065) (0.063)
MOTHER - SON 0.006 -0.001 -0.001
(0.061) (0.065) (0.063)
MOTHER - DAUGHTER -0.052 -0.064 -0.064
(0.071) (0.074) (0.074)
FATHER - DAUGHTER -0.051 -0.018 -0.018
(0.062) (0.065) (0.063)
Observations 3008 2835 2662
R
2
0.75 0.76 0.75
Standard errors in parentheses. *** -pvalue<0.01, ** -pvalue<0.05, * -pvalue<0.1
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 18 / 1

19. Transmission of rules related to choice of occupation
Children born after 1970
Total Without Only
sample the same siblings
occupation
FATHER - SON 0.128* 0.136* 0.136**
(0.071) (0.070) (0.067)
MOTHER - SON -0.020 -0.018 -0.018
(0.061) (0.064) (0.061)
MOTHER - DAUGHTER 0.046 0.037 0.037
(0.080) (0.083 ) (0.079)
FATHER - DAUGHTER -0.048** -0.033** -0.033**
(0.0208) (0.079) (0.008)
Observations 2662 2680 2426
R
2
0.83 0.83 0.81
Standard errors in parentheses. *** -pvalue<0.01, ** -pvalue<0.05, * -pvalue<0.1
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 19 / 1

20. What about mothers?
Working mother - working child?
Total Born Born
sample before 1970 after 1970
MOTHER - SON 0.026 -0.065 0.154
(0.107) (0.149) (0.157)
MOTHER - DAUGHTER 0.240* 0.342* 0.088
(0.135) (0.190) (0.198)
Observations 7605 3953 3632
R
2
0.12 0.16 0.15
Standard errors in parentheses. *** -pvalue<0.01, ** -pvalue<0.05, * -pvalue<0.1
Working mother - working daughter!
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 20 / 1

21. What about mothers?
Other channels:
Division of housework between parents - checked and rejected Dierent than linear relationship
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 21 / 1

22. Conclusions
Robustness checks:
Placebo tests - done
Eect robust only for specic (current) period
Robustness check - indexes sensitive for within state variation
Do parents transmitgender ules" related to choice of occupationto
their children?
Results:
Transmission from mother - rejected for both daughters and sons
Transmission from father to son - conrmed Transmission from father to daughter - only for children born after
1970
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 22 / 1

23. Thank you for your attention
Magdalena Smyk
GRAPE | FAME
[email protected]
Magdalena Smyk (GRAPE) Fathers matter! 10th July 2017 23 / 1