A Micro Approach to Poverty Analysis

1
Family Structure, Educational Attainment, and
Earnings Evidence from Peruvian Data
Gabriella Berloffa, Giuseppe Folloni, Francesca
Modena Università di Trento
2
Background Literature

• macro empirical studies (Barro1991 and 1993,
  Barro and Lee 2000, ecc.)
• Mincer approach, problems and extensions
  (Psacharopoulos and Patrinos 2002, Card 1999,
  Heckman et al. 2003)
• Intergenerational transfers of HC and other
  effects of family background on various outcomes
  (Ermisch, Francesconi and Pevalin, 2004 Ginther
  and Pollak, 2003 Heckman and Carneiro, 2003).

3
Mincer equation

• Implications
• log-earnings experience profiles are parallel
  across schooling levels
• log-earnings age profiles diverge with age across
  schooling levels
• Linearity of log-earnings in years of school

4
Figure 1
5
Figure 2
6
Figure 3
7
Figure 4
8
Figure 5
9
Empirical models

• a basic Mincer equation allowing for different
  returns to schooling at different schooling
  levels
• a model of the incidence of family background on
  school choices and grade attainment (Poisson)
• an enlarged Mincer equation with family
  background variables.

10
Table 4 Model 1 MINCER (Dep. Variable log
hourly wage)
11
Table 5 MODEL 2a (FB-gtEDUC) - Sample with
siblings data (Dep. Variable grade completed)
12
MODEL 2b (FB-gtEDUC) Enlarged sample (with
observations from LSMS 1985) Dep. Variable grade
completed
13
Table 6 MODEL 2c (FB-gtEDUC) Enlarged sample -
Separate estimation for indiv. born before 60
Dep. Variable grade completed Old cohorts
14
Table 6 MODEL 2c (FB-gtEDUC) Enlarged sample -
Separate estimation for indiv. born after 59
Dep. Variable grade completed
Individuals born after 1959
15
Table 7 MODEL 3 FB -gt WAGES Dep. Variable log
hourly wage Mincer model replicated on
subsample with information on parents
16
Table 8 MODEL 3 FB -gt WAGES Dep. Variable log
hourly wage adding family background variables
17
Conclusions (1)

• With respect to educational attainments
• positive and significant effect of parents
  education (father education more important for
  older individuals, the opposite for younger
  ones)
• negative effect of the mother not being married
  to her partner for younger individuals
• a gender gap in educational outcomes emerges only
  for older individuals.

18
Conclusions (2)

• With respect to wages
• Different rates of return for different levels of
  schooling, higher for higher educational levels
• Non-linear returns to experience (
  separability)
• Males (and self-employed) earn higher wages
• Strong regional and sectoral wage gaps
• Large, positive and significant direct effect of
  mother education on wages no evidence of a
  direct effect for other family background
  variables.