Seminar: Untangling Fixed Effects and Constant Regressors

Fixed effects (FE) in panel data models overlap each other and prohibit the identification of the impacts of “constant” regressors. Think of regressors that are constant across countries in a country-time panel with time FE. We introduce “untangling normalization”, which orthogonalizes the FE to each
other and constant regressors. Untangled FE do not overlap, which eases interpretation. Moreover, exploiting the constant regressors shrinks the FE. If they are zero, the true values of the impacts of the constant regressors are identified, even though that is commonly considered impossible. In a gravity model for bilateral exports to the US the constant regressors US GDP, world GDP, and US effective exchange rate explain 98% of the time FE, making the FE redundant. We thus achieve identification.