Econometric analysis of health care expenditures: can positive analysis help answer normative questions?

The size of national health care expenditure is an important research and policy issue. This paper reviews theoretical and empirical analyses of an implied optimal size for a health sector. Various economic theories are explicitly or implicitly invoked, but none is fully satisfactory. Theory provides, at best, a loose justification for empirical specifications of health sector behaviour. Nevertheless, this has a large and growing empirical research industry. The complexity of the issues provides an excuse for reliance on empirical analyses using ad hoc models. The paper analyses aggregate time-series data, using the cointegration approach, on health, health care expenditures and national income. Only one national model met both statistical criteria and showed a significant relationship: between potential life years lost and health care expenditure in the UK. The case for any general relationships remains unproven. There is no objective scientific method to determine optimal health expenditure, nor should we expect one. However, positive analyses can help with normative questions. A better understanding of health expenditure determination would arise from better specification of the relationships, perhaps by analysis at a lower level of aggregation.