A Time-Series Investigation of the U.S. Real Health Expenditure: Evidence from Nonlinear Unit Root Tests

Abstract

This paper contributes to the literature on health expenditure studies by applying the non-linear unit root tests formulated recently by Kapetanios et al. (Journal of Econometrics 112(2):359–79, 2003) to empirically test whether the U.S. real health expenditure time-series are non-stationary or non-linear and globally stationary during a relatively long period from 1965 to 2009. For comparison purposes, it also reports the results of a battery of traditional linear unit root tests and the Lee and Strazicich’s minimum LM unit root (Review of Economics and Statistics 85(4):1082–1089, 2003) test for structural breaks. The empirical findings of the paper show that in the United States during the period under investigation, the real health expenditure time-series are non-stationary in levels. Policy implications of the empirical findings are discussed.