Regression of annual series on a monthly associated series (e.g.
annual GDP and monthly industrial production) is probably the best
way to go, saving the predicted values. To accomplish this convert
the annual series to a monthly series that is observed, say, each
December, with missing values for the other months. -predict- will
fill in all months' values.
