Externality of Defense Expenditure in the United States: A New Analytical Technique to Overcome Multicollinearity

Abstract

This study estimates a three-sector Feder–Ram model using US annual data for 1965–2014 to confirm the externality of defense expenditure in the United States. Although the model is often used in the literature to scrutinize whether this effect exists, a flaw intrinsic to this model is the appearance of multicollinearity. In this study, I introduced novel techniques, namely: the standardization and estimation of a simple slope, to estimate the model. The results are as follows. First, I prove that the multicollinearity problem can be resolved by standardization. Second, externality, which is judged to conventionally exist, is not found. Third, increases in defense expenditure bring about positive but limited economic growth when the ratio of private to defense expenditure in the previous year ranges from 5.09 to 6.82%. By re-estimating the model, this study contributes to developing the Feder–Ram model within the related literature.