According to Chang et al. (2013), social welfare under licensing is lower than under no-licensing if the R&D efficiency is high. We show that this result is not correct, as when R&D efficiency is high social welfare is higher under licensing than under no-licensing.