Ch. 1: Open thread
Aaron Swartz
Discussed in this chapter:
farming, externalities, mechanisms, tragedy of the commons,
overfishing, Pareto frontier, game theory, evolution, Nash equilibria,
dominance, cooperative games, common interest games, assurance games,
division games, prisoners' dilemma, stag hunt, property rights,
invisible hand, recursion, asymmetry, farming (reprise)
Random thoughts:
- I found the term cooperative / noncooperative games confusing at
first. The way I decided to remember it is that in cooperative games,
the parties are _forced_ to cooperate. (If you're selling a house,
it's tough to accuse you of being uncooperative. If you're slacking
off at work, though, you can be so accused.)
- I don't buy that "Many of the differences among scholars and policy
makers grappling with questions of institutional design can be traced
to whether they believe that the ills of society are the result of
common interest problems like traffic jams or of conflicting interest
problems like the division of a fixed pie."
Next week: chapter two.
Aaron Swartz
_Economics: A Very Short Introduction_, which covers why some
countries are rich and some are poor, institutions, game theory, and
all that jazz in a clear and brief way:
http://rapidshare.com/files/270933361/Economics_-_A_Very_Short_Introduction.pdf
Chris Mealy
enforceable/non-enforceable for it to make any sense.
I think Bowles was being polite with the "Many of the differences
among scholars ..." part. Some scholars really are just apologists
for power, selfishness, and greed.
Chris Mealy
chapter 1, but in the interest of drumming up a little traffic I
thought I'd share some of the points I liked enough to underline:
* As he promised in the prologue, economics is a lot harder when you
admit that people are strategic. That means we all have to learn game
theory, which is too bad, because I'm having a really hard time
keeping the game theory terms straight (Nash equilibrium, best
response, dominant strategy, etc all sound the same to me.)
* Bowles's emphasis on cooperation/coordination is a huge contrast to
the purely positive econ 101 view of competition.
* The constitutional conundrum -- "how can social interactions be
structured so that people are free to choose their own actions while
avoiding outcomes that none would have chosen?" (p. 25)
* The expression on page 39 for the degree of common interest versus
conflict is really cool.
* I'm not sure if I completely get this. It sounds really important.
I'd love to hear what everybody else thinks about it:
"The key difference between prisoners’ dilemmas and Assurance
Games is that in the former the undesirable outcome is the only Nash
equilibrium, so the only way that any of the other outcomes can be
supported is by a permanent intervention to change the payoffs or the
rules of the game. In the assurance game, by contrast, a desirable out-
come (mutual early planting, for example) is an equilibrium, so the
challenge to governance is limited to the less challenging how to get
there problem rather than also having to solve the more demanding
how to stay there problem. In debates on the appropriate type (and
duration) of government interventions in the economy, key differences
among economists and others concern whether one believes that the
underlying problem resembles a Prisoners’ Dilemma Game or an Assur-
ance Game. Interventions may be called for in both cases, but Assur-
ance Game problems may sometimes be reasonably well addressed by
one-time rather than permanent interventions. It is partly for this
reason
that a common approach to averting coordination failures is to devise
policies or constitutions that transform the payoff matrix so as to con-
vert a prisoners’ dilemma into an Assurance Game by making the mu-
tual cooperate outcome a Nash equilibrium." (p. 44)
* The difference between risk dominant equilibrium and payoff dominant
equilibrium. (p. 46)
* I have a feeling there will be plenty more definitions for
institution. Might as well start keeping track of them now:
"When a particular set of
mutual best responses is virtually universal in a population over an ex-
tended period of time, it constitutes one or more institutions." (p. 48)
* More on the theme of the persistence of coordination failures:
"The exaggerated emphasis on
two-person games (due in part to their pedagogical value) that are ame-
nable to solution in a repeated game framework may have contributed
to the view that coordination failures are exceptional rather than ge-
neric aspects of social interactions." (p. 51)
* I think Bowles is casting ideology/politics in terms of game theory.
I like it:
"The overlapping character of games is also impor-
tant because the structure of one game teaches the players lessons and
imparts direction to cultural evolution, affecting not only how they
play
the game in subsequent periods but how they play the other games they
are involved in." (p. 53)
"The changes in the rules of the game necessary
to avert a particular coordination failure may be resisted due to the
open endedness of institutions and the losses some players might as
result fear due to the effect of institutional changes on some other
game." (p. 54)
Aaron Swartz
> love to hear what everybody else thinks about it:
>
> "The key difference between prisoners’ dilemmas and Assurance
> Games is that in the former the undesirable outcome is the only Nash
> equilibrium, so the only way that any of the other outcomes can be
> supported is by a permanent intervention to change the payoffs or the
> rules of the game. In the assurance game, by contrast, a desirable out-
> come (mutual early planting, for example) is an equilibrium, so the
> challenge to governance is limited to the less challenging how to get
> there problem rather than also having to solve the more demanding
> how to stay there problem. In debates on the appropriate type (and
> duration) of government interventions in the economy, key differences
> among economists and others concern whether one believes that the
> underlying problem resembles a Prisoners’ Dilemma Game or an Assur-
> ance Game. Interventions may be called for in both cases, but Assur-
> ance Game problems may sometimes be reasonably well addressed by
> one-time rather than permanent interventions. It is partly for this reason
> that a common approach to averting coordination failures is to devise
> policies or constitutions that transform the payoff matrix so as to con-
> vert a prisoners’ dilemma into an Assurance Game by making the mu-
> tual cooperate outcome a Nash equilibrium." (p. 44)
My translation: Imagine two games: planting and hunting. In planting,
everyone plants either early or late. If everyone plants late,
everyone does better, but if everyone is planting early and you plant
late, your crops get eaten. In hunting, everyone either helps take
down a stag or goes off and catches a rabbit. If everyone catches the
stag, everyone does better. But there's always the inclination to go
off and grab a rabbit while the rest of the group catches the stag and
eat the rabbit now and the stag later. And, obviously, if everyone
does that, you're back to rabbits and you're worse off.
In planting, to fix things, all you need to do is get everyone to
switch to planting late once. Then they'll stick to it, since if
anyone defects their crops get eaten. But in hunting, you need to make
sure no one runs off and catches rabbits every single time, since the
temptation is ever-present.
In the real world, we make rules to switch things from hunting-type
games to planting-type games -- instead of the birds coming to eat
your crops, the government comes and fines you for helping catch stag.
That way everyone stays on the straight and narrow.
Andrey Fedorov
"Assurance game" means that once everyone cooperates, there's no benefit to defecting.
"Prisoner's Dilemma" means that while everyone cooperating is nice, an individual wins biggest when he defects and the other parties cooperate. Hence, if everyone is cooperating, the temptation of defection always looms.
Cheers,
Andrey
Aaron Swartz
an assurance game.
Benj. Mako Hill
> * As he promised in the prologue, economics is a lot harder when you
> admit that people are strategic. That means we all have to learn game
> theory, which is too bad, because I'm having a really hard time keeping
> the game theory terms straight (Nash equilibrium, best response, dominant
> strategy, etc all sound the same to me.)
Quite a few of these distinctions between similar sounding terms are
much more clearly expressed in mathematical formalism which I've looked
up in the cases that Bowles didn't provide it. Bowles seems to be trying
hard to reduce the number of equations --- especially so early in the
book --- in order to focus on the core underlying concepts. I think
that's probably the right thing to do but I've also found the result to
be a bit disorientating. It sounds like I'm not alone.
Bowles must realize this too. The first question in the problem set for
the book a the back asks for a number of these definitions.
Perhaps I'll put together a little glossary on a wiki page.
> * I have a feeling there will be plenty more definitions for
> institution. Might as well start keeping track of them now:
>
> "When a particular set of mutual best responses is virtually universal
> in a population over an ex- tended period of time, it constitutes one
> or more institutions." (p. 48)
I was struck by just how different this is from the definitions floated
around in sociology (Philip Selznick famously defined an institution as
an organization "infused with value"). I'm ready to write this off as
one of those weird idiosyncratic economics definitions of words you
thought you knew. Don't get me started on "technology." :)
Regards,
Mako
--
Benjamin Mako Hill
ma...@atdot.cc
http://mako.cc/
Creativity can be a social contribution, but only in so far
as society is free to use the results. --GNU Manifesto
Andrey Fedorov
- Andrey
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.9 (GNU/Linux)
iEYEARECAAYFAkqnQH0ACgkQic1LIWB1WeZe9gCg6ToFyXs1zLJUZHQXUzoTKv4B
rfUAoNMVLqsETwR59S0ab0RjKvYs6RCc
=Zd9R
-----END PGP SIGNATURE-----