This paper has two objectives. First, it explores intertemporal relations between between public revenues and expenditures in the case of Turkey for the period 1955 to 1995.Utilizing the Johansen-Juselius multivariete cointegration procedure and Granger causality tests, we find evidence that changes in public expenditures lead to changes in public revenues in Turkey. Second, this paper attempts to investigate Keynesian public expenditure / income hypothesis by employing cointegration technique and error correction models. The Keynesian proposition of government expenditure as a policy instrument to encourage and lead growth in the economy is not supported by the data for Turkey. This result is consistent with the findings of some recent cross-country studies emphasizing that public expenditure is not conducive to growth in countries with large public sectors and weak institutions.