Health Care Spending and Economic Growth: Armey-Rahn Curve in a Panel of European Economies

The paper examines the Armey-Rahn hypothesis of the inverted U-shaped relationship between public health care expenditure and economic growth in the European economies over the 1995–2018 period. The study’s research objective is to establish the optimal level of healthcare spending, defined as the maximum economic growth rate associated with a particular level of spending. This allows determining whether health care is under-, over-, or optimally provided and whether healthcare spending has positive, negative, or neutral effects on economic growth and output level. To this end, the aggregate production function (in levels or logarithms) augmented by spending and economic openness terms is estimated. The fixed-effects panel regression, the panel quantile regression with fixed effects, and the panel ARDL models are used for empirical analysis. The focus on European economies sets a specific context for the results: the optimal levels of spending may differ across the economies of Europe, the efforts to curb overspending may require concerted effort of national and supra-national governments, and the convergence in healthcare spending levels may be protracted and may be related to broader economic convergence. The paper unequivocally indicates the existence of the Armey-Rahn curve and the negative effects of health care spending on the output (per capita) beyond the optimal spending level. Irrespective of the functional form of the model or the definition of dependent or independent variables, the optimal level was estimated to be smaller than the actual average health care spending level for the period (5.99% of GDP), indicating the over-provision of public health care. Under-provision of public health care was documented for the transition economies in Eastern Europe (that were characterised by comparatively small size of GDP, low per capita output and higher optimal spending levels, economic transition challenges, and lagging health care spending, in addition to indivisibilities of the public health investment). The study is subject to a number of limitations, such as short time-series dimension of the panel, the use of single model and metric, potential endogeneity problem, and abstraction from a broader complexity of spending problem. The study outlines avenues for future research, such as derivation of multi-dimensional optimal spending metrics, political-economic and qualitative analysis, and consideration of systemic aspects of overspending.