I don't think you are correct.

The miser knows each time he will not get the reward, and that he will save on food and drink. That is the real reward, and the rest is a kabuki play he puts on for less-important impulses, to temporarily allow him to restrain them in service of his larger goal. The end pleasure of savings will provide strong positive reinforcement.

This could probably be empirically tested, to see if it is true and would work as a technique. I can imagine a test where someone is promised candy, and anticipates it while acting to fulfill a task, and then is rewarded instead with a dollar. Do they learn disappointment, or does the greater pleasure of money outweigh the candy? This is predicated on the idea that they would prefer the money, of course - you would need to tinker with amounts before the experiment might give useful results.