Nick Rowe explains the overlapping generations model.
Imagine an infinite line of people, each holding one beer. One equilibrium is where each person drinks one beer. But there is a second equilibrium, where each person gives his beer to the person in front. The person at the front of the line drinks two beers, and everyone else drinks one. The second equilibrium is Pareto Superior to the first, because the person at the front of the line drinks more beer, and everyone else is the same. You can imagine the first person in line giving the person second in line a bit of paper, in exchange for the beer. That bit of paper (money) travels down the line in exchange for the beers traveling up the line.
I think that the OLG model is perhaps the stupidest idea in monetary economics. But if you have to learn it, at least go to Nick’s post in order to understand it.
The OLG model is probably a reasonably good way to think about Social Security. I think that this was probably the original intent. I think that the profession would be better off if it had never been viewed as a way to think about money.
[update: Roger Farmer has an excellent overview, and he points out that Samuelson did originally intend it as a model of money.]
Thanks Arnold!
” I think that the profession would be better off if it had never been viewed as a way to think about money.”
I tend to agree, except: if we interpret “generation” very very metaphorically, it might be a good way to think about money. Think about a Wicksellian triangle, where A produces apples and wants bananas, B produces bananas and wants carrots, and C produces carrots and wants apples. Now expand the triangle into a circle, or break the circle and stretch it out to infinity. You can only trade pairwise, with the people to right and left of you. It’s impossible (or very hard) to get everybody together in one place at one time to do an n-way trade of everything for everything.
The PSST problem would be impossibly hard without a medium of exchange.
I think OLG models are also a good way to think about whether government debt imposes a burden on future generations. After all, unless you have at least 2 generations, you can’t even think about that question.
If there is no last person there is also no first person.
So if there is no last positive integer there is no one?
If the last guy gets two beers the first guy gets no beers.
The mental model, if based on an infinite number of beer drinkers, has to run infinitely in both directions (and assume everyone likes beer).
And assume pop-top beer or assume a can opener for old style cans. Jk.
In other words, you can’t have it both ways. You can’t say there is no first guy who gets no beer in order to claim the last guy gets two beers.
Otoh, it may be onto something if finance is included. For a split second, depending on accounting conventions, everyone gets to put two beers on their balance sheets and borrow to finance a third beer.
As an academic monetary economist who was trying to earn tenure in the late 80s and 90s, I can attest to the stranglehold that OLG-types, mainly out of Minnesota, had on the field. Obsessed with their ignorant way of modeling money, which they defended on the grounds that it was less “ad hoc” than other approaches (as if any formal apparatus for motivating a demand for “money,” no matter how ludicrously at odds with reality, deserved to be considered an improvement over informal but at least sensible alternatives*), they did their best to keep other approaches from making it into print. And since the Minnesota-types turned out dozens of PhD students every year, their blocking power was considerable. Thankfully one of those students–Randy Wright–succeeded in convincing even his own mentors of the error of their ways. By then, however, the damage done to the discipline was considerable.
For anyone looking for more good dirt on OLG models of money to supplement Nick’s nice review, I highly recommend the chapter on the topic in Kevin Hoover’s The New Classical Macroeconomics.
*My personal favorite candidate for a foolishly ad hoc claim–and a particular favorite of the OLG-types–is “It takes a model to beat a model.” If you find this persuasive I daresay it’s because you have a stupid model you’d rather not see subjected to criticism.