Hello AYDot -
The way I look at it is as follows: the forecasted values should be derived by the data which you have at hand - which could mean that you are creating negative forecast (if you are facing a negative trend for example).
If you have additional knowledge about the business question at hand (for example: when forecasting sales non-negative numbers might not make a lot of sense) you can go ahead and modify the output of the statistical forecasting engine in a post-process step.
As such you might want to consider the follow approach:
Start with creating forecasted values using SAS/ETS. If the prediction is negative simply set the values (prediction and confidence interval) to 0 using a data step for example.
Hope this makes sense.
Regards,
Udo
Catch up on SAS Innovate 2026
Nearly 200 sessions are now available on demand in the Innovate Hub.