The post earlier about hair splitting is right on.
There just isn't a huge difference in viscosity between a typical 5W20 and a typical 5W30... a fact that could be used by both sides of the argument, I guess. I think that's why so many manufacturers didn't have any trouble changing the recommendations on engines that were originally designed/spec'ed for a 30. If you look through the UOAs, however, you see that the gap narrows further as many 30s lose viscosity over time and miles. Overall, 5W20s seem to hold their grade better, so your 30 may be a 20 by the time you're 2/3 of the way to oil change time but your 20 has stayed pretty near it starting viscosity. That may be another factor used by the OEs in their recommendations.