Labour (actually useful) Empirics term 1 Flashcards
Wages on sleep
Biddle and Hamermesh (1990):
w up 20% implies sleep down 34 minutes / week, despite no clear substitutes.
Phasing out of benefits UK
EMTR of roughly 80%! This is the unemployment trap
NIT
Friedman, 1962
NIT empirics
US Gov 1970s (Seattle). Single mothers enter at extensive margin. Husbands / wives drawn into benefits system.
NIT can be welfare max despite drawing in to benefits system effect (Fortin et al 1993)
Manipulate budget constraint
Theoretically impossible to devise a budget constraint (fixed expenditure) which will unambiguously lead to Ls up. (Moffitt, 2003)
Household based system of tax / benefits
Will have effect on incentives of family structure (Alm et al 1999)
Evidenced by Moffitt (2003) on support for single mothers
Work requirement
Blundell and Shephard (2010): Min 24 to 16 UK. Effects both ways (draw diagram). ‘Employability’ must be chosen and tagged by gov (Akerlof 1978). Subjective, continuous, perverse incentives
EITC USA
(since 1975)
Eissa and Liebman 1996: Phase in: DiD of 2.4% on participation rate of single mothers.
Bunching at kinks of a progressive tax system
Saez (2010)
Effect of childcare of child development
Ambiguous: Neg on behaviour? (Baker et al, 2008) / pos on performance in Scandinavia (Black et al, 2011)
Against pooling?
Justifies what models
Redistribution from father to mother: Lundberg, Pollack, Wales 1997.
Weighting with bargaining power (Chiappori et al, 1988)
Economic incentives on fertility
Gans et al (2009): Baby boom following $3,000 bonus if born after July 1, 2004 in Australia
Skill complementarity hypothesis
Low skilled workers are sub for capital, but high skill are complements.
Skill biased tech change?
Acemoglu, 1998
Card, 1990
Mariel boat lift: Influx of 125,000 Cubans to Miami 1980. Suggest caused 1% Ue fall for blacks
Vs Card 1990
Angrist Krueger 1999. Comparison cities are not well chosen. 1994 boat lift should not have impacted Miami and yet the DiD suggests a 6.3% Ue rise for blacks.
UK gov estimates on grad earnings
Roughly 20% higher
Proximity as instrument for schooling
Card, 1995: IV> OLS by 30% due to attenuation bias.
Exclusion violated? (People sort naturally into nicer areas?)
Quarter of birth as an instrument for schooling?
Angrist and Krueger (1991) - Drop out laws at 16 thus if born earlier in year, lower school pre drop out. Also IV>OLS. Issue is that exclusion is violated as wealthy families plan births away from the winter (Buckles and Hungerman, 2013)
Min schooling age as an IV
Harmon and Walker, 1995: Issue is that this happened 1947, meaning it has massive ID ussues.
Twins to find return to educ
Again find OLS as underestimates (on this small subset) (Ashenfelter and Krueger 1994)
Perry Pre School
2 years from age 3.
Initial improvement in IQ with eventual convergence.
Pre determined variation
Black, Devereux, Salvanes (2007): Birth weight corr with earnings and IQ.
Factors: Eg substance abuse / pollution (Currie and Walker, 2011): EZ Pass via DiD
Feed me better
Belot and James, 2001:
Positive effect on outcomes (but not random and media attention?!)
Adoptees: Twin studies
If we assume
- Randomly assigned and treated identically to bio kids.
Sacerdote (2004): Genetics > nurture for educ, < for drinking behaviour
Intergenerational mobility measure
Intergenerational income elasticity. Or: use percentile rankings.
Mobility = 1 - Beta1
0.071 Denmark / 0.517 USA
Issues with IIE
Life cycle bias: Biased down as those with higher life incomes have steeper income profiles.
Attenuation bias: Y1 is a noisy measure of Ynpv.
N up leads to H?
H up (Imbens et al, 2001) using lottery winners