Brilliant.

I think you need to set an m&m quota for a day. He can get them all if he goes x times a day, but going more than x times doesn't lead to extra m&ms. Also, you should institute a minimimum pee amount before the m&m reward is triggered. I think all this is called "regulation".

There's always the free-market approach, which would allow him to eat as many M&Ms as he wants until he pukes, which would presumably count as a major disincentive to such behavior.

Only after that would come the bailout, in the form of bread and other foodstuffs to replace the purged items. Suppose we could let him starve, but where's the utility in that?

That's the problem with regulation based on positive rather than negative incentives --you can end up with the regulation working *too well*.

I don't approve of the Matthew solution, which will only lead to dampened effectivenes of the M&M instrument and negative externalities in the form of both puke and pee going everywhere. This would be a definite case of the intervention being worse than the problem it was designed to correct.

The Laffont and Tirole approved, market-friendly positive incentive approach is to provide OSR with a stream of M&Ms at a rate below that which induces vomiting or long-term obesity, but make clear to him that the stream will stop for a pre-defined period if he pees elsewhere than in the potty.

To work, this approach is crucially dependent on the regulatory authority having the ability to carry through on the threat to withdraw the M&M flow. This ability could be compromised if the regulator becomes 'captured' by the regulated entity (perhaps due to nepotistic relationships between the two). So the safest course is to give the M&Ms for me to administer as an independent third party.

Another reason that puke is not an effective solution is the systemic problems it might stimulate, i.e., mass family puking. This is why we intervene: we privatize candy and socialize the clean-up.

I doubt, however, that appointing an independent M&M distributor is particularly ideal. Chances are said regulator will do nothing but complain about how he never has enough M&Ms to properly discharge his duties.

Actually I favour Matty's free market approach in the hope that the regurgitative effect of too many m&ms will eventually lower the perceived utility of m&ms, which does seem an overly plentiful commodity, or in simpler terms, he'll get sick of them, which is what happened to me and bananas when i lived in jamaica when i was 4. is that a giffen good? i can't remember.

of course your big problem if this goes too far is that he might begin to associate going pee pee in potty with being sick from eating bad chocolate, and go back to peeing anywhere but there, on computers for instance, in shoes etc., which would have no prima facie negative consequences for him.

(P=M squared) with apologies to Einstein is the theory of potty training