Key - Problem Set 6 - 2011

Key - Problem Set 6 - 2011 - Key - Problem Set 6...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
Key - Problem Set 6 Econometrics I Resource Economics 702 1. The following model was specified to estimate the effects of human capital on U.S. wages. The data are from the 2004 Current Population Survey and are available on the course website under Problem Sets: 2 01 2 i 3 . ii i i i i i i i ff u p h wage yrsed xp xp fe xpfe unioncov pension hinspd u ee e   where: wage is measured in $/hour; YRSED is number of years of education; Exp is number of years of potential experience; fe is a binary variable (0 if male, 1 if female); Expfe is an interaction variable and the remaining three dummy variables indicate the individual is covered by a union contract, has a pension program and has paid health insurance. a. Review the results. Which variables are statistically significant? Would you suggest dropping some of the variables from the model? The estimated coefficients for yrsed , exp, and expsq are all statistically different from zero. We would text one-tail tests in the right tail for yrsed and exp , while we would conduct a one-tail test in the left tail for expsq (to get a “hill shaped” quadratic effect). It does not appear that the variables f and fexp have statistically significant effects. Thus, from these t-tests, we might conclude we could drop f and fexp without concern. But, looking at the VIFs that are included leads us to suspect that we may make a Type II Error. Although the VIFs suggest that the variances for exp and expsq are inflated, we should proceed with caution in drawing inferences from t-tests when multicollinearity may exist. We should check the joint test below before deciding. The REG Procedure Model: MODEL1 Dependent Variable: wage Analysis of Variance Sum of Mean Source DF Squares Square F Value Pr > F Model 8 11058 1382.30008 17.81 <.0001 Error 748 58046 77.60109 Corrected Total 756 69104 Root MSE 8.80915 R-Square 0.1600 Dependent Mean 16.64101 Adj R-Sq 0.1510 Coeff Var 52.93638 Parameter Estimates Parameter Standard Variance Variable Label DF Estimate Error t Value Pr > |t| Inflation Intercept Intercept 1 -6.54604 2.19902 -2.98 0.0030 0 yrsed yrsed 1 1.36201 0.15076 9.03 <.0001 1.09229 exp 1 0.44915 0.09027 4.98 <.0001 11.79023 expsq 1 -0.00699 0.00181 -3.85 0.0001 10.78093 f 1 -1.01175 1.29626 -0.78 0.4353 4.09725 fexp 1 -0.04465 0.05291 -0.84 0.3990 5.25282 unioncov 1 1.07338 3.63720 0.30 0.7680 1.01475 pension 1 0.59840 0.69662 0.86 0.3906 1.11731 hinspd 1 1.63787 0.69916 2.34 0.0194 1.13821
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
b. Test the null hypothesis implied by the following statement. “There are no gender differences in wages; all differences can be explained by human capital differences.” Revisit your conclusions from part a. The implied test is that the two parameters, δ and γ , are jointly zero: :0 ; : . OA H H either or or both are not zero   The calculated F-statistic of 5.00 (P-value = 0.007) provided by SAS suggests that we reject this hypothesis. Thus, there are significant differences between male and female wages. Test 1 Results for Dependent Variable wage Mean Source DF Square F Value Pr > F Numerator 2 387.69720 5.00 0.0070 Denominator 748 77.60109 c. Interpret the estimated effect of experience on wage for the model specified.
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 12/08/2011 for the course ECON 702 taught by Professor Staff during the Spring '08 term at UMass (Amherst).

Page1 / 10

Key - Problem Set 6 - 2011 - Key - Problem Set 6...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online