> From a management perspective, you're forgetting about the obvious: the less than productive people. The benefit of giving employees no- or few-strings attached time to work on whatever is clear. Things like GMail, Apple 1, etc. The "cost" is that people are doing things that don't necessarily contribute to the bottom line -- for every GMail, there are 1,000 low-impact ideas.
Actually, there are other reason for non-rigid 20% time (that is, the 20% is a target with considerable variation in the short term); it means that resources across the company and in any team aren't fully committed on critical tasks routinely, so surging due to an emergent need doesn't mean dropping the ball somewhere else. When routine utilization gets too high, this rapidly becomes a very significant effect.
Its also a way that people gain experience and understanding and avoid developing tunnel vision around the way things are done on their exisiting primary products, and so become more efficient.