Doing X for specially crafted incentive Y for the intrinsic value of X is a form of corruption. It's not always possible that all decisions are made for the intrinsic value but if you have a political enviroment where there a lot pressure to do Y's.

Especially if you can't get any political power without Y, you won't have many people who persue political goals for their intrinsic value in your political system.

This seems like possibly quite a useful bit of abstraction and offer the potential of arguing the merits of a single principle that appears in many manifestations, in politics, corporations, volunteer organizations, etc. But I'm just having trouble getting it clearly in my head. Two things might help.

1) One or 2 concrete examples where you flesh out "X" and "Y". I spent 2 years in a math Ph.D. program, which is long enough to know that to move forward with an abstraction, it is best to start with at least a couple of examples.

2) Consider the "agency problem" (or "principle-agent problem"), which to me seems the most promising abstraction for reasoning about corruption. See http://en.wikipedia.org/wiki/Agency_problem, and maybe very close to what you're aiming at.