ECON 497 Assignment 4 Page 1 of 2

Metropolitan State University

ECON 497: Research and Forecasting

Bellas, Spring 2011

Assignment #4 Due 4/12

1. Which came first, the chicken or the egg?

2. What is the null hypothesis in the Durbin-Watson test for serial correlation?

3. Who doesn’t love panel data? Seriously, it’s sort of the ultimate in data.

A. How does panel data differ from cross sectional data?

B. What sorts of techniques should you use if you have access to panel data?

In SPSS

4. Use the Indy 500 data available on the web page to estimate the equations:

St = b0 + b1Yt + et and St = b0 + b1St-1 + et

where

St = Indy 500 pole sitter's qualifying speed from year t

Yt = year

A. Use the estimated coefficients to generate predictions from each model for the pole sitter's qualifying speed in the next race. Do the predictions differ very much?

B. Which model would you use in making a prediction? Why?

C. Based on analysis of the residuals, is there evidence of serial correlation in each of the models? Explain.

D. Conduct a Durbin-Watson test for serial correlation in each of the models. Are the results consistent with what you saw in the residuals?

E. What is the technical problem with conducting a Durbin-Watson test on the second model?

5. Use the industrial data set on the web site to estimate first a pooled and then a fixed effects model of logged output as a function of logged employment and logged overtime. Discuss how your estimates of the employment and overtime elasticities of output differ between the two types of models.

6. Use the seatbelt dataset on the web site to estimate a fixed effects model of fatalities by state. Comment briefly on the results. Here are the variables:

Fatality rate fatalities per million traffic miles

Sb_useage seat belt useage rate

Speed65 65 mph speed imit dummy

Speed 70 70 mph or higher speed limit dummy

Ba08 dummy variable for blood alcohol limit <= 0.8%

Drinkage21 dummy variable for a 21 year old drinking age

Income per capita income

Age mean age

Primary dummy variable for primar enforcement of seat belt laws

Secondary dummy variable for seconday enforcement of seat belt laws

Vmt millions of traffic miles per year

State which state the data are from

Year which year the data were observed

Fips state ID code