Choi (2009) offers a useful demonstration that Li and Resnick's (2003) analysis of the effect of regime type on foreign direct investment (FDI) inflows is vulnerable to the impact of outliers. However, his first remedial method of controlling for outliers with a dummy variable leads to qualitatively the same findings of Li and Resnick. His second remedial method, a robust regression estimator, drops 80 observations and its result on democracy contradicts the finding from median regression. Choi claims that the origin of the outlier problem in Li and Resnick is the inappropriate operationalization of the concept: FDI inflows. Analysts should choose FDI/GDP over net FDI inflows. This research demonstrates how scholars often conflate the two measures conceptually and empirically. Fundamentally, FDI/GDP reflects a country's openness to or reliance on foreign capital whereas net FDI inflows indicate the amount of investment. Based on FDI/GDP, scholars cannot draw correct inferences about the impact of democracy on the amount of foreign investment. Hence, conclusions in several published studies have to be revised and corrected. One promising solution for outliers in FDI data is to log-transform net FDI inflows. This research has important implications not only for resolving the democracy-FDI controversy but also for studies of other causes of FDI flows.
The Effect of Outliers on Regression Analysis: Regime Type and Foreign Direct Investment , Quarterly Journal of Political Science, Volume 4, Issue 2 10.1561/100.00008021
Understanding Outliers in Foreign Direct Investment Data Analysis , Quarterly Journal of Political Science, Volume 4, Issue 2 10.1561/100.00009035