I don't think there is the symmetry between the two cases David
Groenewegen cites as he seems to imply. But before explaining why, we should not be surprised that any commercial venture presents information in a way that serves its interests. Surely anyone growing up in a market economy knows that. So we discount things that sales people tell us as a matter of course.
If subscription-based publications were simply adding articles to boost prices, then we would see that reflected in eroding impact factors. My experience is that traditional publishers more often have the opposite discussion, that is, whether to increase the rejection rate in order to improve impact factors. This is because market forces demand quality publications. This is the real virtue of the subscription model: it demands competition based on editorial quality, as libraries, with limited budgets, will cancel everything else.
The virtues of Gold OA are many, but this model is indeed subject to manipulation. This is precisely what Jeffrey Beall's work addresses. But we have seen even non-manipulative alterations of editorial regimes in OA publishing. Consider the development of PLoS ONE, which uses a different editorial strategy from the PLoS flagship journals. At PLoS the alteration of the editorial protocol changed the acceptance rate of papers, which has turned out to be highly profitable.
At this point the conversation breaks down into an argument (irresolvable, in my opinion) over whether the community-managed post-publication peer review of PLoS ONE is better than, inferior to, or simply different from the editorial policies of traditional publications. I subscribe to the "different from" school, but there is no end of argument over this.
I think it's a red herring to try to demonstrate equivalency between traditional and OA forms of publication. These are different systems; they have different characteristics. Different people will make different determinations, and these determinations will vary depending on context. I much prefer the NY Times to Reddit, but I dip into Reddit from time to time.
Joe Esposito