Consider a research design that is similar to the one studied by Greenstone, Hornbeck and

Moretti discussed in class. We observe firm productivities in a winning and a losing country.

The objective is to estimate the impact of opening a large new plant on incumbent firms'

productivity, i.e. estimate the magnitude of the agglomeration spill-overs. Let us consider

a plant that opened its operation in 2019. Suppose we use the average value of output in a

country to measure firms' productivities in a county.

a) We observe that the average value of output in the losing county was $280 thousand in

2018 and $300 thousand in 2020. Similarly we observe an average value of output of $350

thousand in the winning county in 2018 and $400 thousand in 2020. What is a reasonable

estimate of the impact of opening the plant on productivity that is consistent with the

approach taken by GHM.

b) Suppose you observe that the average value of output was $310 thousand in the losing

country and $390 in the winning country in 2021. What do you conclude from that?

c) Discuss two reasons why these estimates may be problematic.