The empirical findings on the relationship between gross domestic product (GDP) and health expenditure are diverse. The influence of income levels on this causal relationship is unclear. This study examines if the direction of causality and income elasticity of health expenditure varies with income level. It uses the 1995-2014 panel data of 161 countries divided into four income groups. Unit root, cointegration and causality tests were employed to examine the relationship between GDP and health expenditure. Impulse-response functions and forecast-error variance decomposition tests were conducted to measure the responsiveness of health expenditure to changes in GDP. Finally, the common correlated effects mean group method was used to examine the income elasticity of health expenditure. Findings show that no long-term cointegration exists, and the growth in health expenditure and GDP across income levels has a different causal relationship when cross-sectional dependence in the panel is accounted for. About 43% of the variation in global health expenditure growth can be explained by economic growth. Income shocks affect health expenditure of high-income countries more than lower-income countries. Lastly, the income elasticity of health expenditure is less than one for all income levels. Therefore, healthcare is a necessity. In comparison with markets, governments have greater obligation to provide essential health care services. Such results have noticeable policy implications, especially for low-income countries where GDP growth does not cause increased health expenditure.
Keywords: Causality analysis; Common correlated effects; Gross domestic product; Health expenditure; Impulse response function; Westerlund cointergration.