Do formal intellectual property rights hinder the free flow of scientific knowledge? An empirical test of the anti-commons hypothesis