The purpose of this study is to analyse the effect of not treating Seasonal Level Shift (SLS) outliers on the Model estimate, Forecast accuracy, point forecasts and prediction intervals from SARIMA models with the time of occurrence of SLS through simulation experiment. We propose a strategy which first estimates the model parameters and outlier effects using the procedure of Kaiser and Marvell (2001) to reduce the bias in the parameter estimates, and then used this data for further analysis. Along with the SARIMA models performance in the presence of SLS, study also investigate the performance of Seasonal Adjustment in the presence of SLS through X-13-ARIMA. We demonstrate that SLS significantly increases bias in the SARIMA estimate, increases inaccuracy of the SARIMA models, and significantly affect the prediction intervals. However, after detection and adjustment of SLS, SARIMA estimate become less bias in a result of less bias, forecast accuracy measure, and prediction interval significantly improve. Furthermore we establish that most of the time X-13-ARIMA reject seasonal adjustment when SLS were detect through automatic procedure. However, rejection frequency of seasonal adjustment become small when SLS were deal through manual intervention. Three real examples are employed to illustrate the issues discussed. Supervisor:- Dr. Amena Urooj