negatives in a time series

Hi, I needed to extend my independent variables to the forecasting period (eg. Dec'12) to help in the forecast of the dependent variable (Price, which i sonly up to De'11). But some of the variables because I used proc forecast in SAS become negatives.. for example if the real data I have is on a downward trend.. they become negative if gone too far.. How do I deal with negatives in a time series model.? Thanks
Transform your dependent variable, i.e. in for instance spss-forecasting module you can choose to transform your DV; Square root, or Log. By doing that, you will not get values smaller than zero (if i'm right). I suggest you try this.