Cost overruns are common - don't forget the Mars Rovers went $100m over budget. Delays for things like MSL, Kepler (not entirely Kepler's fault), and Messenger also happen.

The "nuclear option" of cancelling a mission for blowing budgets and schedules almost never happens, especially after a lot of sunk costs, and it probably can't happen enough to create appropriate incentives. Transferring management to another entity is also impractical.

Requiring the managing entity to replace the top project manager, however, is a reasonable consequence, and that option should be part of every project contract. That could give managers the incentive to design realistic budgets and timelines and then stick to them.

If this is an already-existing legal option, then it should be excercised in practice, and not just in theory.