ECON 497 Final Exam Page 1 of 3

ECON 497: Economic Research and Forecasting Name:______

Spring 2008 Bellas

Final Exam

Return this exam to me by 4:00 on Wednesday, April 23. It may be e-mailed to me. It may be delivered to my office in Minneapolis or faxed to me before 4:00 on the 25th at 612-659-7268. You could even drop it by my house if you like, and I’d be pleased to introduce you to my wife and my kids. You can also send it to my office through the regular post or to my home via regular post, but it should be post-marked by the 23rd.

You may consult any written source you like regarding the answers to these questions but you may not ask any person other than me any questions about this test. Questions to me must be sent via e-mail and responses will be sent to the entire class.

Answer all questions, and explain your answers. Fifty points total, points per part indicated in parentheses.

1. It will probably come as no surprise to you that economists love a good fight. Graduate school was a long series of scuffles, brawls and flat out donnybrooks. Happily, we always noted who was involved and, more importantly, who won. We estimated a binomial logit model of the fight outcomes and came up with the following:

lnPi1-Pi=-1.9+0.4Fi+0.2MIi+0.01Ai

Where Fi is a female dummy variable, MIi is a microeconomist dummy and Ai is the age of the participant in question.

A. Without knowing any of the characteristics of his opponent (an admitted weakness of the model) calculate the probability that a 40 year old male microeconomist will win a fight. (3)

B. This is a bit trickier. Based on this model and assuming that no econ fight ever ended in a tie, discuss how you might arrive at the probability a 25 year old male macroeconomist would win a fight against a 50 year old female microeconomist. Remember, and this is the trick, that there are no ties and that the probability of all possible outcomes must sum to one. (2)

2. Use the coffee data from assignment #2 to estimate the attendance elasticity of coffee demand (basically dCtCtdAtAt or, if you prefer, %∆C%∆A ). You should also include the other explanatory variables in your model.

A. Present your results. (2)

B. Explain the implications of this elasticity being either greater than one or less than one. (3)

3. Use the coffee data from assignment #2 to estimate a linear model of coffee sales (C) on temperature (T), attendance (A) and the night game dummy (N).

A. Present your results. (2)

B. Do a Park test for heteroskedasticity and discuss the results of this test. (3)

4. The best way to determine if there is multicollinearity in your model is to calculate (or ask a software package to calculate) VIFs for the explanatory variables. Explain carefully where these VIF numbers come from. (3)

5. There is data on real personal consumption spending (in billions of year 2000 dollars) and population for the U.S. available on the course web site. The data are from Microeconomics: Principle and Policy, 10th edition, by William J. Baumol and Alan S. Blinder. Use this data to do the following.

A. Estimate a linear model of real consumption spending as a function of population and year. Briefly discuss your results. (2)

B. Is there evidence of serial correlation in your model in part A? Offer three different pieces of supporting evidence. (3)

C. Estimate a semi-log model in which the dependent variable is the natural log of per capita consumption spending and the explanatory variable is year. Present your results. (2)

D. What is the interpretation of the estimated coefficient on year in the model in part C? (2)

E. Estimate a generalized least squares (GLS) model to correct for the serial correlation in your model from part C. Show clearly how you estimate the GLS model and present your results. (2)

6. Use the Age-Wisdom data that is available on the web site to do some stuff.

A. Estimate a pooled model in which wisdom is the dependent variable and age is the explanatory variable. Present your results. (3)

B. Briefly discuss your results from part A. (3)

C. Estimate a fixed-effects model in which wisdom is the dependent variable. Present your results. (3)

D. Explain the important difference between the results of the pooled and the fixed effects models. Discuss what these results mean in a couple of plain old, English language sentences that don’t have any numbers in them. (3)

7. Use the Metropolitan State University library resources to access the article “Do Students Go to Class? Should They?” by David Romer which appeared in The Journal of Economic Perspectives, Vol. 7, No. 3, (Summer, 1993), pp. 167-174. Answer the following questions.

A. In which types of economics courses is the rate of student attendance higher? (3)

B. According to the paper, does attending class more often help students do better, or is it simply the case that better students generally attend class more often and attending doesn’t really seem to matter to a student’s grade given her pre-existing level of talent? Support your answer with material from the paper. (3)

8. What three bits of advice would you offer to students who take this course in the future? They should be three distinct things, please.

A. (1)

B. (1)

C. (1)