A mini-debate has ensued, starting with Michael Mandel weighing in on the chief methodological problem with game theory — its presupposition of the adoption of perfectly mutually-adjusted strategies by actors. Tyler Cowen ignores this critique by taking the somewhat dialectical position of saying that if it improves our ability to think through economic problems, then it’s useful. Russell Roberts seems to agree, but notes that the argument cuts both ways: “The problem is that game theory can organize your thinking the wrong way because it tends to cause its users to underestimate the power of competition. I assume this is a result of taking payoffs as given when in fact they are often endogenous and affected by market forces.”
All of this brings to mind the fact that long before I took a single course in economics, I had taken a course in Control Systems theory, using this book. It could be argued, as Cowen seems to suggest, that the mathematics of transfer functions could also be fruitfully applied in enhancing understanding of economic phenomena. In some ways, I think my knowledge of transfer functions has improved my understanding of economic phenomena in ways that other mathematical constructs cannot, since a fundamental property of transfer functions is that their response to a given input varies over time.
But, like virtually-useless indifference curves, I have to wonder about the opportunity cost of learning and expressing economic thoughts in game theory or any other sort of mathematics. Roberts’s anecdote certainly shows that there are some graduate student economists that have not absorbed the basic lessons that the vast literature of economics offers.
If Cowen’s defense of game theory were valid, then it would be reasonable for me to say “the strongest argument for including the mathematics of transfer functions (and the attendant mathematics of differential equations, LaPlace transforms, and Heaviside’s operational calculus) into an education in economics is simply to chat with someone who doesn’t know any”. Bad logic. Bad economist, bad!
Oh, for you non-engineers — when I say transfer functions, I mean this kind, not the econometric kind.
{ 10 comments }
Heck, the problem I have with “game” theory is that it has little room for win-win or positive sum games. You’d think these people had never played D&D!
Curt, I don’t think that’s quite true. I believe the prisoner’s dilemma, for example, can be recast such that the traditional utility values represent opportunity costs instead. So the function U(P1,P2)=(-2,-2) could mean something like each agent gains two fewer utility units than they would have if both players had cooperated. In this sense — and I’m just talking off the top of my head, so please correct me if I’m way off base — game theory could be brought closer to Austrian economics by changing the traditionally cardinal utility values to ordinal ones. So perhaps (-2,-2) should be, instead, (3rd, 3rd).
Lowell R. – Most of the “payoffs” I see in game theory are actually utilities and are by necessity ranked ordinally already.
Curt,
I’m with you. Having read some years ago a book titled “The Evolution of Cooperation” (can’t remember the author and can’t locate the book on my shelves), what became clear to me is that much of game theory is based on “The Prisoners Dilemma,” which obviously has nothing to do with the voluntary cooperation upon which a free and open society is based.
When people can communicate with each other without interference (i.e., without the state), there is no dilemma at all — no “game” to be played other than arriving at the mutually agreed-upon terms for a market (i.e., social) transaction to take place.
The rest, to me, is but the strategizing — the gamesmanship, if you will — of firms competing to better cooperate with their existing and would-be customers.
If you have sometime, check out this article regarding real-world business applications of ‘game theory’ — You Got Game Theory!
Game theory application in economics is a perfect example of MISapplied mathematics. While the math itself can be perfectly good there is absolutely zero evidence that the behaviour of actual humans can in any way be described by transfer functions or other analythically tractable models, primarily because humans have memory and engage in information exchange for decidedly extra-economical reasons.
David White, unless you are using that as a definition of a “free and open society” – in which case it is a utopian ideal and not what we actually have – it just isn’t true that you don’t get prisoners’ dilemmas and suchlike.
The thing is, there is a paradox: just as Adam Smith noticed, if left free people tend to rig things, which makes things less free. For instance, the merchant fee structure that credit card companies prefer actually creates a Tragedy of the Commons (which is pretty much a Prisoner’s Dilemma).
That’s not a big deal, biut it shows how things work out. It establishes the principle. But there is a really serious modern example: the labour market. That’s not such a clear example, since these days state-implemented and tax funded (or similar) social security creates a Tragedy of the Commons. However, that is just substituting for “vagrancy costs”, the cost of having unassisted unemployed people around and raising crime and so on.
So states have muddied the waters on favouring unemployment, but there is still a very real and naturally occurring games theory problem going on there. I have quite a bit of material on it at http://member.netlink.com.au/~peterl/publicns.html.
P.M. Lawrence,
I’m unfamiliar with your reference to Adam Smith and would appreciate your pointing me to it, as I have a hard time squaring it with his famous “invisible hand.”
In any case, “by free and open society,” I mean stateless society, just as I said. Whether it is utopian remains to be seen (since the state won’t allow us to try it), my point being that because parties would be free to interact directly with each other, rather than being separated in a prison environment, there would be no “dilemma” about what to do. Everyone would be known by their prior actions — i.e., their reputations — and decisions about whether to “defect” or not would be made on these grounds.
The Adam Smith reference is to that part where he says that merchants rarely get together, even for the most innocent purpose, but that they generally end up thinking up ways to make a cartel or some such (my paraphrase, from memory). To be sure, he also points out that such schemes have a natural tendency to fall apart – but the important thing to notice is that the steady state doesn’t involve all interactions being “free”.
As for your other remarks, they do show a serious utopian inclination. The thing is, starting from here, you find that a simple change doesn’t go straight to a steady state. Everybody gets locked in with whichever assets they had on day -1, and it takes a generations long transition for trickle down and carrying costs to sort that out.
In the first instance, you get a whole load of people with inadequate resources, unable to price themselves into work since they need a living wage, which creates a pool of people causing vagrancy costs that fall on all those that have material levels of assets (hence the carrying costs). After other historical periods of disruption they were called things like “broken men”, if you want to google on that for an example.
It’s not that there’s any flaw in your theoretical scheme, just that there’s no easy way to get there from here within even the medium term. Of course, there are incremental methods that do have improvement every step of the way, but they rather rely on being so slow that earlier generations leave the stage without any individuals having a fall in living standards.
“It’s not that there’s any flaw in your theoretical scheme, just that there’s no easy way to get there from here within even the medium term.”
Granted, as the massive dislocations from centuries of state oppression have so unbalanced the world’s population that some three billion law-wage workers have, with the collapse of the communist bloc and the opening of the Indian subcontinent, been dumped on the global labor market.
Result? Increasingly rapid capital flight, as corporations abandon their domestic work forces in order in search of cheap labor abroad. It’s a prescription for disaster, however, since those abandoned live increasingly off of debt and will sooner or later (probably sooner) cease to be the consumers of last resort.
THAT is the REAL Prisoner’s Dilemma, since we are all held captive by what Sean Corrigan has written so superbly about in this morning’s Daily Article — even if the vast majority of us are either oblivious of this fact.
Comments on this entry are closed.