Great idea, but we can only assume that Subaru doesn't think the sales volume would be enough to amortize the costs of development.
Just to provide an illustration:
Lets say it costs them $150,000 to update the ordering system, catalog, promotional materials, website configurator, assembly line, parts suppliers, safety certs, EPA tests, and pay the engineers to actually do it, then they need to make it back. If this sells 200 cars, then it's a $750 option before they even profit. If they goof and build 500, that's 300 cars they need to discount, probably by $1k or so to clear them at the end of the year. Whoops, that means the costs just went from $150k to $450k in the blink of an eye. It should have been a $2,500+ option. They'll think of this and hedge that option price to say... $2,000.
I suspect that the real cost is significantly higher. But even at $2,000 extra would you still go for it? Bundled with other goodies to make a special edition for an extra $5k? At that point I think the car is well north of $40k and still isn't likely to appeal to the wider performance crowd.
Remember, the LGT Spec B special edition only sold 500 units the first year, had an even higher price and it was a sedan, so the maxima/a4/Acura TL crowd would notice.