Scribd
Upload
84 views6 pages

Honorsexam 2011 Econometrics PDF

The document summarizes regressions examining the relationship between height, gender, education and economic achievement using a dataset of over 18,000 American adults. It reports regression results with log income and years of education as the dependent variables. Independent variables include height, gender, age, race and education. The regressions are used to test hypotheses about the effects of height and gender on earnings and education. Questions are also asked about potential omitted variable bias from excluding intelligence, and whether additional data could help address this.

Uploaded by

ADITYA
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
84 views6 pages

Honorsexam 2011 Econometrics PDF

The document summarizes regressions examining the relationship between height, gender, education and economic achievement using a dataset of over 18,000 American adults. It reports regression results with log income and years of education as the dependent variables. Independent variables include height, gender, age, race and education. The regressions are used to test hypotheses about the effects of height and gender on earnings and education. Questions are also asked about potential omitted variable bias from excluding intelligence, and whether additional data could help address this.

Uploaded by

ADITYA
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 6

Econometrics (60 points)

Question 7 (60 points)


The regressions in Table 1 examine the relationship between height and economic achievement using
cross-sectional data on n = 18,028 randomly chosen adults (ages 30+) in the United States. The data are
from the U.S. National Health Interview Survey (NHIS) for 1994.
1) Consider the gender gap in earnings estimated in regressions (2) and (4).
a) (5 points) Use regression (4) to compute the predicted percentage difference in earnings for a
woman and a man who are each 5 foot 9 inches, and who have the same values of the other
variables in the regression.
b) (5 points) Compare the gender gap estimated in (a) to the gender gap estimated using regression
(2). Why is one of the estimated gaps greater than the other?
2) Use regression (7) to test the following hypotheses at the 5% significance level, holding constant the
other variables in the regression:
a) (5 points) For both men and women, the effect of height on years of education is zero.
b) (5 points) For women, the effect of height on years of education is zero.
c) (5 points) For men, the effect of height on years of education is zero.
3) Now focus on estimating the causal effect on log earnings of an additional year of school.
a) (5 points) Consider the omitted variable intelligence. Does the omission of intelligence from
regression (3) introduce omitted variable bias and, if so, is the estimated effect of education
arguably biased up (too large) or down? Explain.
b) (5 points) One determinant of height is early childhood nutrition, and it is known that good
childhood nutrition enhances learning, performance in school, and even IQ. A researcher
therefore suggests using Height and MaleHeight to control for intelligence and thereby
eliminate this source of omitted variable bias. In your judgment, does including Height and
MaleHeight in regression (5) adequately control for intelligence? Explain.
c) A second researcher proposes to reestimate regression (3) by two stage least squares using
Height and MaleHeight as instruments for Educ.
i) (5 points) Are the proposed instruments weak or strong? Explain.
ii) (5 points) In your judgment, are the proposed instruments exogenous? Explain.
d) A third researcher points out that these data are for the first of four consecutive years in which
the same individuals were surveyed, and suggests that by using panel data regression methods on
the full data set the omitted variable bias in question 4(a) could be eliminated.
i) (5 points) Suggest a panel data estimator that would eliminate the omitted variable bias and
explain why it would do so.
ii) (5 points) Suggest three different types of standard errors that could potentially be computed
for the panel data estimator in (i). Which would you use, and why? Be precise.
iii) (5 points) Would you suggest pursuing this strategy here (that is, using the panel data
estimator from (i) and your preferred standard errors from (ii))? Explain.

Table 1. Regression Results Using NHIS data on Adults Ages 30+


Dependent Variable
Regressor
Height
Male
Height
Male

(1)
0.020
(0.001)

Log of Income
(3)

(2)

0.215
(0.016)
0.090
(0.005)
-0.0011
(0.0001)
-0.289
(0.026)
-0.301
(0.021)
-0.091
(0.030)

Educ
(4)
0 .010
(0.003)
0.013
(0.004)
-0.820
(0.243)
0.090
(0.004)
-0.001
(0.0001)
-0.245
(0.021)
-0.340
(0.018)
-.078
(0.028)

0.210
(0.016)
Age
0.089
(0.005)
Age2
-0.0010
(0.0001)
Hispanic
-0.255
(0.026)
Black
-0.280
(0.022)
Race_Other
-0.089
(0.029)
Educ
0.082
(0.003)
Intercept
8.746
5.655
6.066
7.187
(0.094)
(0.202)
(0.212)
(0.191)
F-statistics testing for coefficients equaling zero on:
Height and
47.1
MaleHeight
Measures of Fit
SER
0.730
0.683
0.675
0.670
0.012
0.152
0.184
0.168

(5)
0.005
(0.003)
0.009
(0.003)
-0.563
(0.236)
0.083
(0.004)
-0.001
(0.0001)
-0.143
(0.020)
-0.313
(0.017)
-0.111
(0.027)
0.075
(0.002)
6.746
(0.187)

(6)
0.119
(0.007)

-0.618
(0.058)
0.131
(0.015)
-0.002
(0.0002)
-1.728
(0.104)
-0.710
(0.057)
0.529
(0.112)

(7)
0.093
(0.010)
0.055
(0.015)
-4.329
(1.023)
0.130
(0.016)
-0.002
(0.0002)
-1.712
(0.104)
-0.708
(0.057)
0.542
(0.112)

3.290
(0.578)

5.407
(0.710)

18.43

0.648
0.222

130.9

2.559
0.068

2.558
0.068

Note: Heteroskedasticity-robust standard errors appear in parentheses below estimated coefficients.

Variable Definitions
Variable
Male
Age
Educ
Height
Hispanic
Black_NH
White_NH
Race_other

Description
= 1 if male, = 0 if female
Age, in years
Years of education
Height without shoes (in inches)
Race/Ethnicity = Hispanic (0,1 indicator)
Race/Ethnicity = non-Hispanic Black (0,1 indicator)
Race/Ethnicity = non-Hispanic White (0,1 indicator)
Race/Ethnicity = other (0,1 indicator)

Selected Tables from Stock and Watson, Introduction to Econometrics

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy