Keeping in view the importance of economic growth in a country’s development, this study intended to examine the relationship[ between the government size and other determinants on economic growth using a time series data over the period 1973-2012. To specify the growth equation, we have followed the Barro (1990) model of endogenous growth. The exogenous variables in the model consisted of, government size, employment, inflation, capital and trade openness. To examine the impact of the 9/11 incident, the earth quake in 2005 and financial crises, we have introduced three dummies in our growth equation. Keeping in view the nature of variables and possible endogenity in the model, we have used the VAR methodology which is believed to overcome the possible endogenity. The estimation strategy comprised of two steps. In the first step, we have estimated the long run growth equation using the Johansen co-integration technique. In the second step, we have estimated the ECM model to arrive at the short run growth elasticities with respect to the variables concerned. Moreover, to estimate the optimum size of the government, we have employed the Scully methodology which is based on the assumption of a balance budget. The long run results indicated that almost all the variables have found out to be significant with their expected signs except for trade openness which carried negative coefficient. The negative and significant coefficient of the government size suggested that large government size negatively affect economic growth of Pakistan. On the other hand, the positive and significant coefficient of capital indicated that increase in capital holdings enhances economic growth. The positive and significant long run coefficients of inflation and employment highlight that economic growth increase along with increase in inflation and employment. The trade openness variable was found to be significant with positive sign which is the only significant variable in the ECM model except the dummies. The ECM term in the error correction model has carried out a significant coefficient with negative sign and plausible magnitude which highlights the stability of the model.. The findings have suggested a reduction in total government spending since the estimated size of the government was found to be 17 percent which is less than the current size of the government which is 18 percent. Supervisor:- Dr. Abdul Qayyum