This preview shows pages 1–3. Sign up to view the full content.
University of California, Santa Barbara
Olivier Deschenes
Department of Economics
Winter 2011
Economics 140B
Individual Assignment 3: Answer Key
Due in class 02/15/11
Question 1:
The standard model for estimating the return to education is the loglinear wage equation:
i
i
i
i
i
u
EXP
EXP
ED
y
2
2
1
1
0
)
1
(
Where
y
i
is the log hourly wage of worker
i
,
ED
i
is years of education, and
EXP
i
is the labor
market experience (measured in years). This equation is often called the “Mincer equation” in
honor of the labor economist Jacob Mincer, who pioneered its usage. Since actual labor market
experience is rarely measured in data sets, labor economists proxy it with “years of potential
labor market experience” (defined as ageeducation6). While potential experience is an error
ridden measure of true labor market experience, we will ignore this issue in this exercise.
The parameter of central interest is
β
1
, which is interpreted as the “return to education” (since it
measures the percent increase in hourly wages associated with an additional year of
education). Note that we typically include other variables in the Mincer equation (e.g. sex, race,
marital status, etc.). A major concern in the estimation of equation (1) is the presence of omitted
variables that are correlated with both education and wages. Labor economists often refer to
such omitted variables as “ability”, and the resulting bias in the OLS estimates as the “ability
bias”
In this exercise, you will investigate different approaches to resolve the ability bias problem,
using two different data sets. The description and questions related to the first data set follow
immediately in “Part I”, while the description and questions related to the second data set are in
“Part II” at the end.
PART I
The first approach is to include “proxy” measures of ability directly in the wage equation and to
make use of panel data on wages and education of workers, and use fixed effect and random
effect models to account for unobserved ability.
The file “nlsy.dta” contains data on 1410 young white workers during their first five years in the
labor market (these are not the same years for every worker). The data were compiled from the
National Longitudinal Survey of Youth (NLSY) that we discussed in class. There are 5
observations on each worker (denoted by the variable “id”) observed over the years 19801991.
The key variables for the analysis are: log hourly wages (lnwage), years of education (educ),
age (age), marital status (married), and union coverage (ucov). In addition, the NLSY contains a
measure of “ability”, derived from the worker’s score on the Armed Forces Qualifying Test
(AFQT). The AFQT is the average of four scores from the Armed Forces Vocational Aptitude
Battery (AFVAB), which is designed to summarize aptitude and ability along various
dimensions. It is not a measure of IQ.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document (A) Estimate the wage regression using OLS, and including potential experience, potential
This is the end of the preview. Sign up
to
access the rest of the document.
This note was uploaded on 09/04/2011 for the course ECON 140b taught by Professor Staff during the Winter '08 term at UCSB.
 Winter '08
 Staff
 Economics

Click to edit the document details