My opinion is it depends on the grafting.
(a) is there enough excess bone from the jaw surgery to use in a graph?
(b) does the surgeon have skills in doing that graft?
(c) does the surgeon allow proper healing time of the graft (e.g. some will do the graph, then tell you to chew as soon as possible)
It's amazing the amount of incompetence out there just on those three issues. If there is excess bone, the graft is done right, and there is proper healing time, it should be stable. The one complication out of anyone's control is that the body might reject the graft. This is more unlikely if it's your own bone, which is why (a) is important. The surgeons who report stable results and argue with the logic that the graph is unstable probably know how to handle the three points above. There's a lot of literature about the topic, and you just need to use Google or the "education" section here.