This paper explores the evaluation of forecast accuracy in economics, particularly when comparing a parsimonious null model to a larger nested model. The authors propose adjusting the Mean Squared Prediction Error (MSPE) to account for the noise introduced by the additional parameters in the larger model. They recommend using standard procedures to compute the standard error of the adjusted MSPE difference and apply these procedures to test for equal predictive accuracy. The paper discusses the asymptotic properties of the adjusted MSPE statistic and provides simulation evidence to support the recommended procedure. The results show that the adjusted MSPE statistic performs better in terms of size and power compared to other commonly used statistics, such as the raw MSPE difference and the Chao-Corradi-Swanson (CCS) statistic. The authors conclude that the adjusted MSPE statistic, with standard normal critical values, is a reliable method for comparing forecast accuracy in nested models.This paper explores the evaluation of forecast accuracy in economics, particularly when comparing a parsimonious null model to a larger nested model. The authors propose adjusting the Mean Squared Prediction Error (MSPE) to account for the noise introduced by the additional parameters in the larger model. They recommend using standard procedures to compute the standard error of the adjusted MSPE difference and apply these procedures to test for equal predictive accuracy. The paper discusses the asymptotic properties of the adjusted MSPE statistic and provides simulation evidence to support the recommended procedure. The results show that the adjusted MSPE statistic performs better in terms of size and power compared to other commonly used statistics, such as the raw MSPE difference and the Chao-Corradi-Swanson (CCS) statistic. The authors conclude that the adjusted MSPE statistic, with standard normal critical values, is a reliable method for comparing forecast accuracy in nested models.