Please enable JavaScript.
Coggle requires JavaScript to display documents.
Schooling Model (Estimation of MRR (Mincer Equation (Run a regression of…
Schooling Model
Estimation of MRR
Running regression of wages against schooling
- MRR = coefficient of schooling = dlnw(s)/ds
- We note that this model assumes constant MRR, however this is not possible as the wage-schooling locus is typically concave
Mincer Equation
Run a regression of wages against schooling, experience and squared experience. Idea is to find not just the impact of schooling, but also the impact of experience, which is the education you get at work.
Assumptions:
- Constant MRR
- Separability of experience and education
Education & Earnings
- The more education you have, the higher your earnings will be
- There is a college premium, or the earnings premium you get from completing college
Experience & Earnings
- Concave relationship
- Wages increase sharply from 0 to 20 years of experience
- After about 40 years of experience, wages start to fall
Why is the return to education smaller in the specification without experience?
- The more schooling you have for a given age, the less experience you have
- Since experience matters a lot for wages, we have underestimated the return on education
- We compared people with more schooling but less experience, which will have lowered their return on education
Ability Bias
In most cases, we expect that we will overestimate the returns to schooling, and both terms in the selection bias are positive.
3 different methods to estimate causal relationship between earnings and education:
- Find a proxy for ability
- IV
- Twin studies
Quality
Card & Krueger (1992)
- Showed negative effect of class size
- Showed importance of teacher quality
Other experimental studies showed:
- Infrastrcture matters only for low/mid-developed countries
- No evidence that interactive whiteboards etc matter
- No effect of selective, 'elite' school
-
-
-
-