An outlier-robust extreme bounds analysis of the determinants of health-care expenditure growth

Hartwig (2008) has presented empirical evidence that the difference between real wage growth and productivity growth at the macroeconomic level is a robust explanatory variable for deflated health-care expenditure growth in OECD countries. In this paper, we test whether this finding is robust to the inclusion of additional covariates, applying different versions of Extreme Bounds Analysis (EBA) to data for 33 OECD countries over the period 1970-2010. As far as it is statistically feasible, all macroeconomic and institutional determinants of health-care expenditure growth that have been suggested in the literature are included in the EBA. Furthermore, we analyse to what extent outliers in the data influence the results using an outlier-robust MM estimator. Our results confirm Hartwig's earlier finding. A number of additional both covariate- and outlier-robust determinants are also identified.