Determinants of U.S. health expenditure: Evidence from autoregressive distributed lag (ARDL) approach to cointegration

Using 1960–2012 annual time-series data for modelling, we apply the Autoregressive Distributed Lag Cointegration (ARDL) approach, to identify some major drivers of per capita real U.S. health spending. The ARDL Bounds testing procedure (Pesaran et al., 1999; 2001) has several econometric advantages compared with the standard Johansen and Juselius cointegration method. One distinguishing feature of this ARDL Bounds testing procedure is its ability to estimate the long-run economic relationship without requiring pre-testing the time-series for the presence of unit roots in the data generating process incorporated in the cointegration model. The empirical findings in this study indicate that per capita real income (INCOME), the population percent above 65 years (AGE) and the level of health care technology (HRD), measured as the level of Research & Development expenditure in health care are cointegrated. INCOME, AGE and HRD exert positive effects on U.S. health expenditure per capita. Unlike prior studies, this paper presents new empirical evidence indicating that the U.S. health care is a necessity, with an income elasticity estimate of around 0.92. We also find that medical technology advances play a major role in the long run rise of the U.S. health expenditure. We discuss implications of these findings.