Ch. 6: Summary
Aaron Swartz
which Bowles calls "Utopian Capitalism".
Imagine you and I have a pie and a pizza. And let's imagine that
having pie or pizza makes us happy, and that happiness can be measured
in utiles. So how do we divide them up? Well, one minimal assumption
is that I make sure you get at least a couple utiles and that I get as
many as I can. (Do these assumptions sound ridiculous? They sound less
ridiculous when they're written as equations, I suppose!) But if I was
doing that, I would divide things up so that [equation I don't really
understand] which means that neither of us would be better off by
swapping a piece of pizza for pie. That makes sense right? If I could
do better and you wouldn't do any worse, obviously I'd swap the pizza
for pie. (The state where no one can be better off is called Pareto
optimality.)
But in the real world, this is totally impossible, since I have no
idea how many utiles a piece of pizza gives you and you have no
incentive to tell me the truth. (You'd just lie and say you didn't
like them that much so I'd give you more to make you even a little
happy.) And here comes the market! Imagine I have some pizza and you
have some pie. Well, you might want some pizza and I might want some
pie, so we'll trade. But how do we decide the terms of trade?
Classical economics followed the "Walrasian" model, assuming there was
a third-party Auctioneer who kept suggesting terms of trade until he
hit upon one where for every piece I agreed to sell at that price, you
agreed to buy. And, sure enough, if that happens, we reach Pareto
optimality. Amazing! (No, really -- be amazed.) This is known as the
First Fundamental Theorem of Welfare Economics, basically that if
everything can be traded, such an auction will lead to Pareto
optimality. (Of course Pareto optimality isn't a particularly
interesting goal. Me getting all the pizza and all the pie is Pareto
optimal.)
That brings us to the Second Theorem, which says if we assume there
are no increasing returns, any Pareto optimal allocation will be the
result of some initial distribution of resources. This is interesting
because it means if poor people want houses and you want to give poor
people houses, you don't actually have to give them houses -- you can
give them whatever you want and they can trade it for a house.
All this has led to a powerful ideology of the market -- namely that
markets can only make things better. If you don't like what some
people end up with, it's not the market's fault -- the problem is that
people didn't have enough to start out with. The market just takes
what people have and makes it a little better -- it may not make it a
lot better, but it doesn't make it worse.
Of course, that doesn't seem very likely, and indeed the First
Fundamental Theorem has some problems. 1: There is no Auctioneer in
real life. Now the standard story is that if I can't sell all the
pizza I want to sell, I lower my price, and so I end up at the price
the Auctioneer would have picked. But while that makes some sense with
only one buyer, why would I do that when there are millions? 2: Do we
ever stop trading? Do we get in odd loops where we keep trading pizza
back and forth? More technical work has shown that there are no
reasonable assumptions you can make that avoid this. 3: There are all
sorts of possible outcomes compatible with the Walrasian model and no
way to know which one you'll end up with. But most problematically, 4:
not everything can be traded. It used to be thought that things that
couldn't be traded were only special exceptions, but it turns out
almost everything has this problem.
And this is an even bigger deal when you introduce the General Theorem
of the Second-Best, which notes that if you have one bad thing, a
whole bunch of other bad things might actually turn out to be good. As
an example, if you have pollution, then monopolies might actually be a
good thing since they can choose not to pollute (whereas in a
competitive market they'd be forced to pollute). And since we do have
pollution, this is a pretty big problem for Econ 101.
There is some hope, though -- Duncan Foley (!) showed that if instead
of assuming a bunch of ideal people, we assume a bunch of kind-of
decent people, their mistakes all balance out in the end and you get
pretty close to the ideal (helping with 1 and 2). Of course, this is
only true on average -- there are still some people who get really
good deals and some people who get really bad deals. It's not clear
how big a deal this is, but it does mean that we can't just assume the
market never makes things worse.
Now the canonical solution to 4 is that the government steps in and
imposes taxes on pollution (e.g. cap-and-trade). That way people
actually have to pay the costs of doing the bad thing and then they
don't do it unless it's actually worth it. This was the solution
endorsed by prominent classical economists. But in 1960, Ronald Coase
showed that if we accept the normal assumption that there are no
transaction costs and everyone's a perfect bargainer, this isn't true.
In that case, everyone who gets hurt by pollution will band together
and pay polluters to stop, providing the exact same incentives as a
government tax. Woohoo, the market wins again! (Here Bowles quotes
economist Harold Demsetz who points out that this means we didn't need
to outlaw slavery -- if slaves really wanted to be free they would
have just pooled their money and bought themselves. Take that,
Lincoln!)
Of course, these are some ridiculous assumptions -- we showed in the
last chapter how hard bargaining is and it doesn't take an genius to
realize that transactions can be pretty expensive. Indeed, it's not
really clear how this is any improvement on the First Theorem, which
assumed everything was tradable. But you can think of it as a
contribution like Foley's: it shows that Walrasian auctioning isn't
the only way to get to Pareto optimality -- bargaining works too. And
that means the government can make things better by improving
bargaining, not just taxing.
Well, that's enough for utopia. Next week we'll start talking about
the real world.
Chris Mealy
from libertarians, as in "the Coase Theorem dictates that we will end
up with preference maximization".[1] Bowles's careful treatment
doesn't really resemble the glibertarian perspective (the chapter is
called "Utopian Capitalism" after all). The way to read the CT isn't
that that the free market is magic but that there are even more
requirements for market efficiency than you thought.
The paper that introduced the idea that came to be called the Coase
theorem (not by Coase, btw), "The Problem of Social Cost," doesn't
really have a solution to what to do about the typical case of costly
bargaining. There's a terrific paper about this called '"Coase v.
Pigou" Reexamined' by AWB Simpson.[2] Simpson points out that when
disputes wind up in court it's up to judges to choose the assignment
of rights that maximize productive capacity. There are some problems
with that. Are judges supposed to do cost-benefit analysis to compare
the social product of different assignments of rights? Are judges
really even capable of doing that? Does the law allow them to do that?
Wouldn't they wind up ruling inconsistently from case to case? And how
weird is that this implies we should have a command economy run by
judges?! Simpson's take is that Coase sees these problems and backs
off on making any sweeping recommendations. If you take the CT too
seriously this is where you wind up.
[1] http://economicsofcontempt.blogspot.com/2008/04/megan-mcardle-invokes-coast-theorem.html
[2] http://scholar.google.com/scholar?cluster=1141293931410627756&hl=en
Aaron Swartz
there are two things:
1. Transaction costs exist.
2. Equity concerns are real.
Right?
Chris Mealy
The second part of Coase's paper, "The Problem of Social Cost," is
about how in real world assigning rights -- regulation -- can improve
efficiency. Next time you hear the "Coase said we don't need
government" routine ask them if they've read the second half of "The
Problem of Social Cost."
(For kicks google "Teaching the Coase Theorem: Are We Getting It
Right?" The answer is no.)