I’ve never seen an Austrian opinion on Game Theory - at least like it’s done by neoclassical economists. I myself tend to think it’s a collection of tautologies with a lot of implicit assumptions about the ends chosen by the agents.
Take for an example the prisioner’s dillemma. It pressuposes that the expected reward got on defaulting is bigger than the expected reward on not defaulting for both agents. But it is not self evident why it should be - hence, the theory could not be general.
But then, what is the value of a theory that says something like: “if both agents believe that they will be better off acting in such a way, they will do it.” That’s just the action axiom rewritten…
If there’s already a thread on it, plz send the link and let this one go down,
Check this out, Peter Leeson also used game theory in his JPE paper, as do a few other self identified “Austrians”. Personally I don’t see any problems with game theory.
Game theory is one of the best branches of economics out there. Most logical, most testable, most ingrained with realistic human action.
The prisoner’s dilemma is one of many situations in game theory. -Two actors acting in their own self-interest leaving a worse outcome for both. This is certainly a possible situation in the real world. This seemingly paradoxical situation is solved completely by the freedom of contract, and contractual enforcement as well as increased transparency.
Game theory is very compatible with Austrian Economics.
Thanks for the link, that was an interesting paper. Not sure Foss accurately represents Mises in parts of his account but nevertheless helped clarify a few points and provided some food for thought.
I’m not entirely sure if it’s “compatible” with AE. In the prisoner’s dilemma, the actors take turns choosing either C or D. In real life, most people are game changers. The rules change. They do something unexpected, etc…
I think one of the main points about AE is that it’s very difficult/impossible to model economic situations and to get meaningful use from them.
The point isn’t whether or not humans are “game changers”, whatever that means, the important point is that in some cases the circumstances of real life will very closely approximate those of the prisoners dilemma. In certain cases, such as public goods, it may well be that the payoffs and other aspects of the game are very close to those of the prisoners dilemma. The problem comes when a model is applied carelessly and without much thought, as Ostrom has shown, there are cases where humans aren’t trapped by their rationality and stuck in a prisoners dilemma, even where it would otherwise appear they are. Rules and institutions evolve to allow for the provision of public goods, but these rules aren’t always the result of Reason, as your post seems to imply.
Stated as it is, the Prisoner’s Dilemma would appear to be statist. But when that very silly assumption of “one-time games” is dropped, than it develops into a purely “my best choice is our bets choice” scenario, totally consistent with a free-market. In the real world there are no such things as one-time games. That would amount to a world without consequences. The NAP would hardly be observed in such a world.
Game theory is a branch of pure math. So that if the permises are valid in a given situation, the conclusions MUST follow, for the same reason as they must in praxeology or geometry.
So that the only question is “Are there real world situations where the premises apply?” They either do or they don’t, in a given situation. And wether they do or not is usually pretty obvious.
This is unlike the usual anti AE stance on using Math in economics. There the argument is that mainstream economics is using premises that quite clearly are false with regard to the real world.
My impression was that game theory was invented in a flash of brilliance in the 50’s, gave some interesting initial results, showed how it has application to a great variety of VERY SIMPLE real life situations, and then laid down and died. It has not developed to the stage where it is really useful for tackling serious real world questions. No scientist in any field tackles a real world problem by saying “OK, lemme see what game theory has to say here.”
If someone knows I’m wrong, and/or has a link, I’d be glad to hear about it.
I believe mechanism design has some real world applications outside the classroom, there was even some articles on it and some fanfare on this site after Hurwicz won the nobel prize for it and cited Mises and Hayek as 2 figures that have inspired his work(he was in Mises’ geneva graduate seminar I believe). See the following:
I even remember getting taught about Hurwicz criterion in an applied maths course in the first year of my degree (before I was Austrian, and when I was still a positivist), leaving me with the impression that “decision maths” was a bunch of baloney nonsense that provided mental masturbation for mathematicians and helped them feel self important but with not much practical application. Maybe I was wrong… not convinced yet.
As for game theory, I’m sitting on the fence on that one. Anything else would be dishonest until I learn more.
I’ll try to develop more what I meant (still had no time to read that Foss paper):
Let’s take the Prisioneer’s Dillemma again. Game theory says that every agent will do the best for himself taking on account that the partner will do the same. That’s a big tautology, every agent tries to remove the most uneaseness possible with his action.
Ok. As I already said, the standard “equilibrium” for this game has a implicit assumption: the agents have an expected reward for defaulting bigger than that expected for not defaulting. Besides that, it is implicit that both agents know they’re playing a one shot game, and it is assumed that both agents have the right knowledge about the other’s value scales.
At this point I say: the “theory” is just a big bunch of tautologies developed under a lot of unrealistic assumptions. If two people believe they’re playing a one-shot PD, prefer the default reward and believe that the other will value more the default reward than others, it is obvious that they will both default.
But let’s contrast this vision with a more Austrian one. Agents are not free floating in the “theoretical nirvana”. Agents know a particular set of informations and interpret this info anyway they want it. So, there’s no pre-set theoretical equilibrium for any kind of game in the Austrian vision. Agents will try to understand (in the Misesian sense) his enviroment and possible previous actions by the other(s) to form a judgment about possible future scenarios. In some cases there could even be ex-ante plan coordination (optimality), or any other game result.
Sure. It studies human action [the subject of prax], using deductive reasoning [=math], the methods of prax. The sub field is described in wikipedia:
“Game theory attempts to mathematically capture behavior in strategic situations, in which an individual’s success in making choices depends on the choices of others.”