The study seeks to find out the impact of pension reform on government recurrent expenditure in Nigeria. The Augmented Dickey Fuller and Phillips-Perron unit root tests conducted showed the variables for the research to be integrated at different orders of I(0) and I(1), necessitating the adoption of Autoregressive Distributive Lag (ARDL) for the analysis. The optimal lag structure was 1 which was determined through the Akaike Information criteria. The null hypothesis of no co-integration was rejected based on the ARDL bound test. The error correction term indicates that the speed of adjustment emanating from any short run disequilibrium in the variables is about twenty-one (21) times in the long run. Granger causality showed unidirectional causality running from recurrent expenditure to national pension fund, bidirectional causality running from recurrent expenditure to government revenue and vice versa and unidirectional causality running from national pension fund to government revenue. Diagnostic test also indicated the model to be appropriate, it has no serial correlation, no heteroskedasticity and the model was properly specified. The study recommends increase participation in the contributory pension scheme among public sector employees, increased awareness campaign on the benefits of the contributory pension especially in states and local government, contributor's confidence should be taken into consideration among other recommendations.