Spurious regression problems in the determinants of health care expenditure: a comment on Hitiris (1997)

In a recent article in this journal Hitiris used panel data for ten EC countries to analyse the determinants of aggregate health care expenditure. This comment shows that the model was plagued by a spurious regression problem. The data are reexamined using standard unit root and cointegration testing procedures, as well as new tests for unit roots in panels and for long-run relationships when the orders of integration of the underlying regressors are not known. There is overwhelming evidence for non-stationarity of the variables, and no conclusive evidence regarding the existence of equilibrium relationships. The apparent significance of the dependency rate and the rate of inflation in the Hitiris model were simply due to the influence of common stochastic trends. New results confirm the overriding importance of income in determining aggregate health care spending, but suggest a shortrun income elasticity significantly less than one. The results also uncover a number of problems in modelling the determinants of health care expenditure and warn against drawing firm conclusions from aggregate level models in an area where theory provides little guidance.