I am extremely stuck on trying to think of an instrument for the location of FDI within a country. The instrument must be correlated with foreign firms, but not correlated with domestically-owned firms.

Note: there is no Government policy I can exploit.

Any suggestions, papers, anything at all would be greatly appreciated.

My dependent variable is the log of house price, and my regressors of interest are 1) the log of distance to FDI 2) the log of job creation at the particular firm.

Isolating the impact of FDI on house prices will be difficult, given the presence of domestically-owned investments in close proximity.

  • $\begingroup$ This question is unclear. Can you please write down your regression model, the reasons for endogeneity and the question you are trying to answer? $\endgroup$
    – ChinG
    Dec 13, 2015 at 17:18

1 Answer 1


Without much information on your regression model it's hard to make a good suggestion. But you may have a look at the Frankel and Romer AER's paper. They construct an instrument for trade, but it could be used for FDI, based on geographic characteristics. I assume that geographic distance between two countries is a good predictor of FDI flows and that ceteris paribus a Canadian firm based in Vancouver has a higher probability to set up an affiliate in Seattle than in Dallas.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.