Professors who teach classes on incentives and pay-for-performance almost univerally point out that incentive systems usually work *exactly* as set up. Unfortunately, this doesn't mean that the system will reach the intended goal, since the mechanics of the incentives may or nay not be in line with the overall objective or they may be off in terms of the strength of the carrot on the stick. (Those of you computer programmers that insist that "the computer isn't doing what it's told" have a good idea of what I'm talking about here...like it or not, the computer is doing exactly as it was told- it can't help it if the instructions were erroneous!)
I illustrate via a personal example. When I was in kindergarten (full disclosure: I was probably a difficult child), my teacher tried to reward students for good work with a cute hand stamp of some sort. Now, I really don't like hand stamps- even now, I get very annoyed when I go to a club or a concert and the bouncer insists on the hand stmp policy...I think the skin on my hands is abnormally porous, and the stamps are a pain to wash off. Anyway, my mother had been very careful to teach me that when someone is violating my person in a way that makes me uncomfortable, I am to say "no" in my sternest voice possible. So of course I chose this moment to actually do what I was told. The other kids thought about my behavior for a bit and figured that there must be a reason that I didn't want the hand stamp, and a mutiny began. Now, my mother was not pleased when my teacher explained this to her, so she offered me a quarter for each time I came home with a hand stamp. Apparently I liked money more than I disliked hand stamps, since my response was to find a friend that had a stamp and make sure that she stamped my hand before I went home each day. I'm pretty sure my mom eventually caught on, but I think it illustrates my point nicely.
This cartoon also gets the point across...a picture really is worth a thousand words. Thanks to Jeff for the link. :)