Ah, such a simple word, but such complexity hidden under its hood. Regulation can create rentier institutions like the taxi medallion system[1] alongside PG&E. So another way to phrase my question is: how can we distinguish regulations that result in PG&E from those that result in taxis? When do the benefits of regulating an established player outweigh the costs in reduced incentives for others to attempt to create future monopolies?
One potential resolution to the question is that you can't, that PG&E's failure to grow is just a temporary failure of imagination. Rentierism is a weed that, uh, finds a way, to quote Jeff Goldblum from Jurassic Park (https://www.youtube.com/watch?v=SkWeMvrNiOM). Under this worldview, a utility is merely a monopoly that nobody wants to compete with anymore -- for now. It has temporarily exhausted its growth/optimization curve, leading to predictable revenues. For a while the predictable revenues are divided between existing players (shareholders, govt, employees, etc.) in a stable equilibrium. But eventually, institutions that fail to grow (taxis) will run into hacks that do (Uber/Lyft).
An alternative solution would be that there is a distinction between good and bad regulations, perhaps that you have to weigh the size of the regulator against the size of the regulated industry. Taxi regulations were bad because they were suppressing a small industry. Regulation of large industries on the other hand, is more likely to do good than harm. Provided public attention can outlast private lobbying. That tension then becomes the new frontier for the eternal battle of public policy.
[1] More examples here: http://www.cloverfoodlab.com/blog/2008/09/05/licensing-and-o...; http://en.wikipedia.org/wiki/Licence_Raj