Estimating the Best Model for Gross Domestic Product in Sudan Using Prediction Residual Error Sum of Squares Method (1982-2008)

Abstract

The importance of this study stems from discussing the problem of type of expenditure, and its impact on the GDP. The study cover all the period 1982 to 2008, this was the period during new policy were adopted by the government “Decentralization” and hence many change in type of expenditure happened. The study focus on the expenditure approach which is more important of the calculation of GDP. The study used the descriptive – analytic approach for the data was given by CBS /Sudan also it use the data from reports and publications from relevant sides like Central Bank of Sudan and Sudan household survey 2006. The findings of the study that the best model can be get by Prediction Residual Error Sum of Squares method, the prediction sum of squares is a useful statistic for comparing different models. We use PRESS to assess the model's predictive ability usually, the smaller the PRESS value, the better the model's predictive ability. PRESS is used to calculate the predicted R2. In this study we predict GDP for 2009 and 2010 and compare between the value obtained from the traditional method and PRESS method, we notice that PRESS method model bring estimated value for DGP more relevant to the actual value than traditional method which indicate that PRESS method is the best for estimating the future value. Based on these findings the study calls for more application of the criteria for selection the best model.