Abstract Using disaggregated data extracted from the UN’s COFOG functional classification, this study seeks to verify the Keynesian view versus Wagner’s law on the relationship between government expenditure and output, considering Turkey as a case study. To do so, the study attempts to capture the nature of the causal relationship between the two variables by applying linear and nonlinear Granger-causality tests. Rather than Wagner law, the empirical findings of the study provide supportive evidence for the Keynesian view, subject to the economically meaningful components of government expenditure considered. Overall, the empirical findings show that government expenditure on defense, economic affairs, education, health, housing and community amenities, and social protection positively affect output that comes out through fiscal multiplier and investment-accelerator mechanisms proposed by John Maynard Keynes.