Sensitivity of elasticity estimates for OECD health care spending: analysis of a dynamic heterogeneous data field

Current econometric research on the determinants of health care spending is characterized by three serious shortcomings: a lack of attention to dynamics, a failure to deal with heterogeneity and a neglect of sensitivity testing. The latter shortcoming is a particular problem in an area where there is little theoretical guidance with which to judge the validity of estimates. This paper deals with these three problems by adopting recently developed techniques for analysing dynamic heterogeneous data fields containing non-stationary variables. This method allows for parameter heterogeneity across countries and is preferred to pooling with country specific effects, since the latter does not yield consistent estimates in this context, despite its common use in studies of health care spending. Long-run and short-run elasticity estimates are found to be sensitive to both model specification and sample composition. An attempt is made to compensate for the weak theoretical base by increasing the emphasis on statistical and econometric criteria for model evaluation, and by subjecting the results to rigorous sensitivity testing.