While previous studies to test Wagner’s hypothesis for Nigeria usedtotal government expenditure, this paper in addition to total government expenditure used adisaggregated government expenditure data from 1961 - 2007, specifically; expenditure ongeneral administration and that of community and social services to determine the specificgovernment expenditure that economic growth may have significant impact on. Economicconditions and policies change implying that it is not only economic growth that can affectgovernment expenditure hence the inclusion of other fiscal policy variable and politicalfreedom to augment the functional form of Wagner’s law. All the variables used were found tobe I(1) and long run relationship exist between the dependent and the independent variablesexcept in the case where only GDP was used as the independent variable. Wagner’s hypothesisdoes not hold in all the estimations rather Keynesian hypothesis was validated in all theestimation. Elasticity estimates and Granger causality results are in agreement.