latest possible sample for forecasting

[nupogodi]

Could you please suggest how to determine the latest possible sample for out-of sample forecasting given an equation.
Namely, an equation may contain exogenous variables with different number of observations (end points) and different lag lengths.
Introduction of longer lags extends the possible sample size for out-of sample forecasting.

For example, the adjusted sample of an equation is 1998m1 to 2013m12, possible out-of sample range for forecasting is from date ... to date ...

[EViews Gareth]

To be honest, probably the easiest way to do it is to forecast from the equation until the end of the workfile, then see which periods have forecast values and which don't.