Rating Bias?
I've always suspected there is some inherent bias in these ratings, in that people may expect more of a certain brand. Therefore they may report more things gone wrong. For instance, if I buy a certain brand, that is not known for high quality, I may overlook a small defect and not report it.
If this is true, maybe some got into the top end because people didn't report everything, 100%, as they might with a brand they expected much more from.
So, if this is the case, then staying on top would get increasingly difficult. Being on the bottom might mean there was even more wrong than was showing up in the ratings - people were just not actually reporting everything as much.
Not to say JD Powers initial quality should be dismissed outright - it does have value - but it also may be seriously skewed. I wonder how closely this correlates with ratings like Consumer Reports (but of course, they could suffer from the same bias somewhat)?