Time series models for projected sales Suppose we are given various reports on projected sales of Product $X$ for $5$ years starting in $2013$. All of the reports just give the results and not the methodology as to how they arrived at the predictions. Suppose we are interested in the projected sales of Product $X$ for $10$ years. So we already have the projected prices from $2013-2018$ and are interested in the projected salesfrom $2019-2023$. Can we use a time series model for these years?
 A: You have a time series and can certainly use time series techniques. The question is: will you get anything reasonable? Personally, I think your results will be terribly misleading and inaccurate.
First, you evidently only have five years of data, which isn't much. By one rule of thumb, five years of monthly data is about a bare minimum. Five years of yearly-data (i.e. five data points) doesn't seem sufficient to do much of anything.
Second, it sounds like the 2013-2018 data are point values, with no indication of uncertainty. That's not a very useful forecast to begin with, and yet it's the foundation that you want to build on.
Third, it sounds like you had some data handed to you and don't really know enough about your data to make good choices when you apply time series techniques. Theoretically, those who made the forecast had quite a bit of data (perhaps decades of data, and perhaps dozens of other variables), and all you have now is a short set of one variable.
So, in summary, you can take your 2013-2018 data, plug it into an ARIMA procedure or something like that, and you'll get a numbers for 2019-2023. But I can't see how those numbers will be justified.
(I'm not a professional statistician, and I don't play one on TV and didn't sleep at a Holiday Inn Express last night.)
A: I would not fit a time series to the 2013-2018 projections.  But if you have historical data up to the present you could use that to fit a model.  If you fit a model such as for example an ARIMA model you can project it as far into the future as you would like.  What you lose as you go further out is accuracy.
