Academic research suggests that the size of the marketability discount is in the neighborhood of 13.5c/c and more recent work has suggested it may he as low-as 7.23%. The regression models of Silber, and Hertzel and Smith have provided both the intellectual and empirical basis for these conclusions. These models were initially developed to study the determinants of the marketability discount. It has been suggested that they should now be used as a basis for forecasting the marketability discount. This paper demonstrates that these models should not be used for this purpose because the forecast errors are likely to be large. Moreover, based on the structure of these models and their prediction errors, it is not possible to state with any certainty that a 13.5% marketability discount is statistically different than a discount of say 25%.