Indeed, and that will hold in both directions. We can't perform an ideal experiment here.

However, we do know funding work with copyright-backed models has been widely successful for a sustained period. We also know that alternative funding models aren't precluded by copyright and that several have been tried with varying but almost always much lower degrees of success.

As 'JoshTriplett mentioned in another post, one confounding possibility is that something about how copyright works today does in fact inhibit what would otherwise be a more successful alternative model. However, so far I've yet to see any compelling examples of that, either in this discussion or elsewhere, and even then you'd need that confounding factor to cancel out multiple orders of magnitude of benefit from some other model for the most successful alternatives I've seen so far to catch up with the most successful copyright-supported work.

That is certainly possible, but until there's evidence to the contrary, it seems highly unlikely.