Question

You have been asked to referee a paper that has been submitted to a journal. The paper looks at how war affects trade between two countries. The observations are indexed with “i” and “j” for the two countries that are trading with each other and with a “t” for year. The authors have data for 200 countries for 1990 through 1997.

The authors estimate the following equation using a fixed effects model:

          ln⁡(export)ijt=α0+β1Warijt+β2neutralijk+α1lnYit+α2lnYjt+α3lnDij+εijt

Where

Export ijt is exports from country i to country j in year t

Warijt is a dummy variable indicating that there is a war between countries i and j in period t

Neutralijt is a dummy variable indicating that either country i or j is involved in a war against a third country, in which the other country is neutral.

Yit is GDP in country i in year t

Yjt is GDP in country j in year t

Dij is distance from country i to country j

The results are:

	Log of Exports
Observations	251,902
War ijt	-1.87
	(0.22)
Neutral ijt	-0.13
	(0.02)
ln(Yit)	0.42
	(0.04)
ln(Yjt)	0.32
	(0.05)
ln (Dij)	-1.38
	(0.03)
R-Squared	0.14

Standard errors are in parentheses

1. Calculate t-statistics for each of the coefficients in the table. Which of these coefficients are statistically significant?

2. Given your calculations, add stars to the table above (***, **, * indicating significance at 1%, 5% and 10% levels respectively)

3. Look at the coefficients on GDP in country i and country j. Based on the point estimates, how much does GDP in country i and country j affect trade between country i and country j? Are the coefficients in the typical range for gravity models?

4. You notice that the coefficient on country i’s GDP is larger than the coefficient on country j’s GDP. Is this what you would expect? Explain how you would interpret this result.

5. Look at the coefficient on distance. Based on the point estimates, how much does distance between country i and country j affect trade between the two countries? Is the coefficient in the typical range for a gravity model?

6. Look at the coefficient on the dummy indicating that the two countries are at war. Does the coefficient have the sign you would expect? Explain why or why not?

7. How large is the effect of war on trade?

8. Look at the coefficient on the dummy indicating that one country is involved in a war in which the other is neutral. Does the coefficient have the sign you would expect? Explain why or why not.

1. How large is the effect of neutrality on trade?

10. Do you think the war and neutrality dummies are exogenous? Based on your answer, is the model the authors estimate appropriate? Explain why or why not.

11. The other referee suggests that the authors should add a dummy indicating that the countries share a common border. Explain how this might affect the coefficient on the dummy indicating the two countries are at war.

Answer 1