Openness & instruments

[nilshg]

In Chapter 3 we claim that we replicate Chen's analysis using our data set, and boldly claim that we would instrument for openness.

Equation 12 and 13 both include the openness variable, while equation 11 does not.

We need to think about

1. Whether it makes sense to include openness at all and
2. How we could instrument for it if we include it

[nilshg]

With regards to the second point, it should be noted that Chen uses "bulkiness" of exports (weight of imports divided by value) and a gravity equation, which requires data on distance as well as market share of industries in the respective countries.

[nilshg]

I think unless we get bulkiness, we have to drop this altogether. I don't think this will be an issue given our regression specification; things like changes in oil prices, shared language, etc. will be picked up by first differences and should only bias the results to the effect that their time variation differs across industries, which is unlikely. Thoughts @rweldzius?