The purpose of this study is to analyse the relationship between government expenditure and economic growth in Turkey. The study tests the validity of Wagner's law by applying autoregressive distributed lag (ARDL) cointegration technique using annual data over 1950-2010 period. In order to find out the possible impact of omitted variables, we first tested the standard bivariate versions of Wagner's law. In the next step by including a third variable - inflation ratio - the analysis extended on a trivariate system. The findings of each testing procedure indicate that omitted variables matter. Since, while there exists no long-run relationship between the variables in the first step of the testing procedure, a long-run correlation is found in the second step. The differences of this paper from the earlier studies testing the Law for Turkey are that, the causal link is examined within a trivariate framework and non-interest government expenditure is considered instead of total government expenditures.