Prof. Bryan Caplan

Econ 345

Fall, 1998


1. From the class webpage, download the new Eviews workfile hw4.bin. It contains data for both the U.S. and Italy for the 1881-1994 period. There are 8 variables for each country; their definitions are as follows:

Year - ranges from 1881-1994; note that you can create a Trend variable by subtracting 1880 from Year.

Inflation - inflation rate in given year

Real GDP growth - real GDP growth in a given year

Nominal GDP growth - nominal GDP growth in a given year

War - =1 if a country was at war in a given year, and 0 otherwise. (There was a minimum casualty rate needed to count; this is why e.g. the Gulf War is coded as a 0).

Foreign Soil - =1 if a country was fighting exclusively on foreign soil, and 0 otherwise. Thus, if War=1 and Foreign Soil=0, a country was fighting a war on domestic soil.

Months at War - This gives the number of months in a given year the country was at war. A fraction indicates a fractional number of months.

Casualty - The casualty rate of the country, equal to 1000* battle deaths/population.

2. Define a Trend variable, equal to Year-1880.

3. Regress U.S. inflation and real gdp growth on: {war}; {war, months}; {war, casualty}; {war, months, casualty.} Discuss your results.

4. Regress U.S. inflation on {war*soil, war*(1-soil)}. You should get some sort of error message. Why? (Hint: Have there been any wars on U.S. soil since 1881?)

5. Regress Italian inflation, real gdp growth, and nominal GDP growth on {war*soil, war*(1-soil)}. Do you see a difference for foreign vs. domestic wars?

6. Now re-do the regression in #5, adding the Trend variable as an additional regressor. How does this change your results?