There are several problems intersecting simultaneously. There's a significant lack of capacity to actually move program, which is getting worse. There is a problem related to the sheer weight of process. When you combine those two, obviously that's a bad outcome. There's also a problem as it relates to the actual decision-making, typically at the front end, in order to get these things moving, so to speak.
There are other problems as well, but I think before we start throwing solutions around, we need to figure out which problem in what order we want to tackle, and recognize that the impact in one area could adversely affect another.
To me, those are the three key areas—capacity, process and decision-making. There are others, but those are the big ones that are hampering their ability to move forward.