Interesting, but there seem to be a lot of assumptions baked into the default "used car market" example, that don't necessarily hold up. For example, the idea that the owner knows whether they hold a "peach" or a "lemon". Clearly he/she has <i>some</i> idea, but they definitely do not have perfect knowledge of this. In fact, this very lack of knowledge ("how many more miles will I get out of this car before the transmission falls out?") may be the reason the owner has chosen to manage risk by buying a newer car. But the actual answer to that question could be anywhere from "1 more mile" to "another 250,000+ miles".<p>Likewise, this statement:<p><i>Given the fixed price at which buyers will buy, sellers will sell only when they hold "lemons" (since plemon < pavg) and they will leave the market when they hold "peaches"</i><p>is iffy. As the owner of a used car, I may choose to buy a newer car for any number of reasons, including just wanting something to impress my peers, or impress members of the opposite sex, whatever. Say I buy a nice, sexy, new sports-car. I don't need my old car anymore, regardless of whether it is a peach or a lemon. But my decision to sell it may not involve its status as "peach" or "lemon". It may simply be that I don't want to pay property taxes and registration fees on a car I'm not using. Maybe I just don't have anywhere to park a second car. Or maybe I'm just feeling generous and want to help somebody else out by selling them my no-longer-needed car.<p>All of that said, I'm sure the math is valid and this theory has uses... but this used car thing feels a little dodgy.