We’ve had various versions of the case for and against the use of (micro)economic rational actor models in the social sciences lately, so I thought I would weigh in with my version of the case against. It has three main elements
First, most rational actor models assume that “rationality” can be represented as “maximization of self-interest”. This assumption is either false or vacuous. Those committed to egoistic rationality tend, when challenged, to oscillate between falsehood and vacuity, in much the manner of the function sin (1/x) as x approaches zero.
A very good example arises with voting. In this 1987 paper pointing out the empirical weakness of rational actor models, I observed out that, while egoistic rationalists will not vote (or rather will vote only in very small numbers), quite a limited amount of altruism suffices to making voting rational. Andrew Gelman made the same point more recently.
Second, the fact that egoistic rationality assumptions work well in a lot of microeconomic applications proves little or nothing about their viability in other contexts; these include macroeconomics as well as sociology, political science and other targets of microeconomic imperialism. Abstractions that work well in one context don’t work well in another. In particular, deviations from given assumptions that roughly cancel out locally may nonetheless be significant in the aggregate. I conjecture that something like this underlies the notion of emergent phenomena, beloved of critics of reductionism in general.
Finally, game theory is much more problematic than is commonly realised. To derive a Nash equilibrium, it is necessary to define the strategy space. In real games this is not a problem. In social, economic and political operations, however, it requires that the participants have shared understandings of the problem, accessible to the modeller. In practice this is hardly ever true, and game theoretic analyses typically proceed with an essentially arbitrary assignment of strategies to players.
{ 37 comments }
Walt 07.02.09 at 12:28 pm
John, did you hit post too soon? The last paragraph seems to end abruptly.
tom s. 07.02.09 at 12:31 pm
If John did hit post what appeared to be too soon, I’m sure he had a perfectly good reason for it.
david 07.02.09 at 12:39 pm
He didn’t hit post too soon. His HTML got messed up, is all.
This is what got eaten:
Daniel 07.02.09 at 1:06 pm
I disagree that the HTML is messed up. A proper and rigorous formulation of John’s preference function would reveal that the amount of this post which was concealed between malformed <a> tags was in fact both locally and globally optimal.
The proof is quite simple – if John had spotted the mistake, he wouldn’t have made it, therefore it isn’t a mistake.
It is even robust to relaxation of one of the key assumptions; even if John had made a mistake, Henry would have anticipated this and given me the necessary access level to correct it. He didn’t, so it isn’t a mistake. The fact that you are still seeing the malformed tags therefore proves that they were intentional, however “counterintuitive” this might seem to you bunch of luddites and humanities graduates.
Daniel 07.02.09 at 1:09 pm
#4 apparently demonstrates that the results are so scientifically valid that they can be reproduced. I wanted to write
“I disagree that the HTML is messed up. A proper and rigorous formulation of John’s preference function would reveal that the amount of this post which was concealed between malformed anchor tags was both locally and globally optimal
The proof is quite simple – if John had spotted the mistake, he wouldn’t have made it, therefore it isn’t a mistake.
It is even robust to relaxation of one of the key assumptions; even if John had made a mistake, Henry would have anticipated this and given me the necessary access level to correct it. He didn’t, so it isn’t a mistake. The fact that you are still seeing the malformed tags therefore proves that they were intentional, however “counterintuitive†this might seem to you bunch of luddites and humanities graduates.”
in fact, I wanted to write it so much that I intentionally made a mistake so I could write it twice. It didn’t seem so at the time – I must confess I felt almost as if I’d made a mistake. But I have always been a slightly heterodox economist and often have urges toward sociologism.
If John has noticed the error by now and rendered this post and the preceding one an even bigger waste of time, I clearly intended that too.
Michael Harris 07.02.09 at 1:28 pm
There may be a time-inconsistency issue. The point at which John hit “post” may turn out to not be the point at which he wished he hit “post” when he looks at it again later on.
He really ought to be anticipating this possibility and constraining himself in advance to not hit “post” prematurely.
tom s. 07.02.09 at 1:54 pm
The discussion is vacuous. Clearly someone must have paid him.
Henry 07.02.09 at 1:55 pm
I’ve noticed the mistake ex-post and corrected it – dunno what that does to the rationality assumptions …
Daniel 07.02.09 at 2:04 pm
I expected that
dr 07.02.09 at 2:09 pm
Nerds. Every last one of you.
(keep it up)
peter ramus 07.02.09 at 2:25 pm
Can weigh in on “way in?”
JoB 07.02.09 at 2:26 pm
Daniel was expecting to be expecting that. And this.
Henry 07.02.09 at 3:55 pm
And did you expect me to tell you that you should actually already have the requisite level of permissions to have corrected John’s post yourself? (I sense a failure of common knowledge here – but then I am a mere foolish political scientist).
dsquared 07.02.09 at 3:57 pm
Revealed preference demonstates that I didn’t check, because there is no possible world in which I don’t prefer making facetious comments to doing the hard working of maintaining the site.
lemuel pitkin 07.02.09 at 5:04 pm
I observed out that, while egoistic rationalists will not vote (or rather will vote only in very small numbers), quite a limited amount of altruism suffices to making voting rational.
This is one way of looking at it. Another is that the relevant actors are not always individuals. As a matter of fact, voting on strictly individual preferences would be irrational, and people generally don’t do so (and are generally regarded as foolish or confused when they do.) But for groups that are large with respect to the electorate, voting is completely rational; when people vote, or participate in political activity in general, it’s precisely because they are experiencing some collective identity, rather than a strictly personal one. As a practical matter much of political organizing, rhetoric, etc., is about fostering collective identities.
StevenAttewell 07.02.09 at 7:26 pm
Lemuel:
Collective identities…like political parties? Or ideologies? Whoo hoo, they exist again! Gah, I hate rational choice even more when it seeps into political science.
peter 07.02.09 at 7:52 pm
JQ wrote: “Finally, game theory is much more problematic than is commonly realised.”
As you probably know, some of the other problems are: Nash equilibria may not exist, or there may be multiple such equilibria, or they may not be computable, or not in polynomial time, or their properties (including existence) may depend on common knowledge of and/or infinite processing by and/or infinite memory resources of and/or perfect, noise-free communications between the participants, or they may not be robust in the face of small changes to the underlying assumptions. For example, while a Nash equilibrium is the best-response to a best-reponse, it is not necessarily the best-response to an almost-best-response, as for example, when human participants try their best to play Nash, but do so only imperfectly.
IMO, the best demonstration of the impoverishment and fallibility of game theory was the 1994 FCC auctions of PCS (cell-phone) spectrum in the USA, which is usually touted as a success of the theory. However, these auctions required (a) experiments with students, (b) computer simulations, AND (c) test runs with potential auction entrants before the auction mechanism rules were finally decided. One has to wonder just how great a theory is when the practical application of it to a (very narrowly-defined) real-world problem still required supplementation with practical know-how and trial-and-error test runs.
robertdfeinman 07.02.09 at 10:43 pm
What is lacking in these days of bashing conventional economics is any discussion of ethics. Bringing in other disciplines like psychology or anthropology may help explain what people do, but not what they should do.
The BBC just broadcast a series of four half-hour lectures by political philosopher Michael Sandel. You seem to be limited to listening via their own web page, but it’s worth it. I especially recommend the first “Morals and Markets”. The link:
http://www.bbc.co.uk/programmes/b00729d9
It is generally taken as obvious that capitalism (disguised as a market economy) is the best of all possible worlds and that maximizing “efficiency” should be the goal. In the coming era of resource shortages these assumptions will no longer be valid. Who gets how much of limited resources can only be treated as a moral question.
Michael Harris 07.03.09 at 3:44 am
Pffft. If Henry hadn’t heavy-handedly “corrected” John’s “error”, the free market would have sorted it out, without the need for force.
Chris 07.03.09 at 1:23 pm
@18: But isn’t economics often defined as the study of the allocation of limited resources? How will an age of less abundance (assuming for the sake of argument that we are *inevitably* headed for one) make economics any less important?
And the reason discussions of ethics are rare is that ethicists have still not found a way over that huge stumbling block, the arbitrariness of first principles. It makes it difficult to discuss any ethical system except with those who already believe in it.
Jim Johnson 07.03.09 at 4:25 pm
John,
I think your last point is the most crucial. That said, I am not sure of the point of all this. Are you saying that game theorists don’t admit the problems you delineate? That is not defensible – see, for instance, Rubinstein’s 1991 Econometrica paper “Comments on the Interpretation of Game Theory” or an earlier paper by Myerson in Rationality & Society “On the Uses of Game Theory in Social Science.” And the point surely cannot be that we just chuck the enterprise altogether. It seems to me that your complaint boils down to the observation that a plausible use of game theory in any application needs considerable ballast in the form of reasonable knowledge about the situation the theorist is trying to model, hence mitigating the arbitrariness you talk about. That seems like common sense to me.
P.S.: I am not claiming that most game theoretic work in, say, political science actually displays such knowledge. This is more about reform strategies.
Tracy W 07.03.09 at 4:36 pm
robertdfeinman : In the coming era of resource shortages these assumptions will no longer be valid.
Two points:
1. Market-allocated prices are the most efficient way we know of of solving the knowledge problem about how to allocate resources in a world where all humans are very limited in how much we know. For example, take food. There are a vast range of possible ways of getting food from the ground to people’s stomachs, with very complicated production possibilities. Different geographies support different crops, this can vary in a very detailed way. One place may have more fertile soil than another, but be further from major population centres. Some geographies can grow many different crops, others can only grow a far more limited range. Some locations can be altered with some capital expenditure, eg building a brick wall to shelter frost-vulnerable crops. Different crops also require different inputs of labour at different times of year, there are various ways that labour could be supplied (eg by hand, by machine), crops respond differently to treatment, eg some fruits bruise more easily than others, different crops have different vulnerabilities to disease, people have different reactions to different foods, eg coeliac sufferers have to keep away from gluten, some people can’t digest lactose, etc. People have different likes and dislikes about foods too, ranging from pure taste to different opinions about nutrition. Once the crops have been grown there are different ways of transporting them to the would-be eaters. Market-determined prices provide a guide to making those decisions – the price we pay for food at a market includes the relevant costs of making it and transporting it to the market. If a food can only be sold for less than the costs of making it and transporting it then the people in the chain will eventually switch to doing something else. If you slightly prefer food A to food B, but food B is far cheaper to produce, then you will have an incentive to buy more of food B. There are many market imperfections, yes, but I have not come across another way of solving this fundamental knowledge problem. This is why environmental economists come back to Pigovian taxes or cap-and-trade systems as a way of dealing with externalities – rather than people having massive arguments about food miles and the carbon costs of different farming methods, they see the carbon price and everyone adjusts to their alternatives. This is the argument for market-determined prices. (It’s not an argument against regulation per se, but it argues for certain sorts of regulations, ones that allow for the price discovery process to continue to operate.)
2. Assuming that you have a goal of maximising efficiency, why would an increased resource shortage make this goal invalid? I thought that a basic result of philosophy is that you can’t get from “is” statements about the world to “ought” without introducing some more or less arbitrary first principle. Has someone figured out a way of doing this? What other goals do you think could be invalidated by increased resource shortages?
lemuel pitkin 07.03.09 at 5:08 pm
Chris and Tracy are correct, I think: It’s abundance, not shortage that (will) make economics, and markets, obsolete.
That said, exhaustion of certain resources doesn’t change the fact that an ever-increasing proportion of genuine goods are becoming abundant, as evidenced e.g. by discussions here of IP.
robertdfeinman 07.03.09 at 5:27 pm
The problem with “efficiency” arguments is that the word itself is open to various interpretations. If I switch from digging up the earth to get the gold out and instead pour in cyanide, I’ve lowered my cost of production and improved my “efficiency”.
Where is the cost of the environment damage factored in? Where is the fact that the gold is never going to be available from this source again factored in? The two biggest criticisms of the standard efficiency argument are demonstrated in this example. The externalities of depletion and pollution are not included in the calculations.
As for capitalism itself, it is based upon a model where the “value” of resources in taken as zero. The only cost factored in is that of extraction. Oil is priced at the cost of pumping it from the ground. The fact that it is non-renewable is not considered.
The second defect with capitalism is that it is highly wasteful. In the US 80% of all new firms fail within 5 years. All the resources used to build these firms are lost. The idea that the best man wins still doesn’t compensate for the waste. The flip side is the me-too effect. Company A makes widgets which are a hit. Before you know it companies B, C, and D are making similar widgets. The market becomes saturated and the glut causes the excess to be dumped and firms to fail. The waste is called “creative destruction” since, presumably, the best won out. However we have seen time and again that it is not the “best” that may win, but the firm which uses its marketing or monopoly power or just flat out breaks the law to achieve its aims.
Capitalism was created during a period when the world was mostly empty. There was plenty of land and resources for the size of the population. If one place became depleted or polluted people could just move on. If the trees were all cut down for fuel then coal was substituted. When coal became uneconomic (as in the UK) then oil was favored. Such substitution is standard economic theory, but doesn’t allow for the case where no substitutes exist.
Viewing economics as having its primary focus on the allocating of scarce resources is inadequate in an age when the population will increase by 3 billion over the next 40 years. Allocation has a moral component whether traditional economists want to acknowledge it or not. If they don’t then let’s bring in the ethicists to make the choices apparent.
engels 07.03.09 at 5:34 pm
Well, the UK did in fact suspend the market in food the last time there were severe, widespread shortages, in WW2, with rationing. I have no idea what the Economics 101 judgment of that policy would be: presumably there’s no logical reason why one shouldn’t have a free market even if it would mean lots of people starving to death–maybe it’s still pareto-efficicient–but there might nevertheless be a, ahem, political reason that would rule it out. It doesn’t seem to me inconceivable as a response to possible future environmental events.
peter 07.03.09 at 7:03 pm
robertdfeinman @ # 24:
“The second defect with capitalism is that it is highly wasteful. In the US 80% of all new firms fail within 5 years. All the resources used to build these firms are lost.”
Hardly. The assets of the failed firms (both physical and intangible assets) don’t just disappear in a puff of smoke. Usually, they are purchased by successful firms and redeployed. I doubt any programmer who worked on the development of CP/M, for example, spent much time in his or her career subsequently unemployed.
“The idea that the best man wins still doesn’t compensate for the waste. The flip side is the me-too effect. Company A makes widgets which are a hit. Before you know it companies B, C, and D are making similar widgets. The market becomes saturated and the glut causes the excess to be dumped and firms to fail.”
Your analysis is incomplete, since you ignored the demand side. Potential buyers of widgets get loads of choice (products from all four companies A, B, C and D), and presumably are able to purchase their widgets at reduced prices if there is a glut. It is also extremely rare for companies to exactly duplicate their competitors’ offerings, since doing so would indicate a marketing department lacking in ideas (and self-esteem). Even when there is over-supply of very similar products, there may therefore be differentiation and innovation in (eg) design, advertising, production processes, distribution methods, market segmentati0n, targeting, customer financing arrangements, after-sales service, etc. These innovations, likewise, do not disappear in a puff of smoke when a company fails, but are instead adopted by the other companies remaining or by companies in other industries. IBM’s entire business strategy used to be based on adopting such innovations pioneered by others, as captured in the slogan, “Second, but better”.
True, some of the investors in one or more of the failing companies may lose their investments (although this is not necessarily the case), but they should have known that was possible when they made their investments or joined the firm. To the extent that investors or employees of a company have influence over its policies and actions, then they must bear responsibility for entering a crowded market and/or its failure to survive a shakeout there. In other words, if the investors and employees of company D were that wonderful, why did their product not beat those of A, B and C?
Even allowing for the losses suffered by the investors and employees of the failed companies, it is not at all obvious to me that the net society-wide costs of product duplication always outweigh the net society-wide benefits. Indeed, I don’t believe I’ve ever seen a case where they do.
Jock Bowden 07.03.09 at 11:43 pm
First, most rational actor models assume that “rationality†can be represented as “maximization of self-interestâ€.
JQ, how many is “most”? Which requires you to answer:
1. How many different ‘rational actor models’ are there?
2. Can you connect q.1 to the claim “‘rationality’ can be represented as ‘maximisation of self-interest?
WHO makes these alleged claims that “‘rationality’ can be represented as ‘maximisation of self-interest?” I am younger than you, and even I know that since the 1960s, the world’s foremost economists and philosophers have explored the epistemological and political opportuinities opened by “the theory of the second best”.
This assumption is either false or vacuous.
Well of course it is ‘false’ as it is no more than a discourse constructed by you for other battles. Oh and it is simultaneously vacuous, as the historical record attests.
Those committed to egoistic rationality
John are you compartentalsing social scientists on the form or rationality they use – egoistic bor, er, er….?
Jock Bowden 07.04.09 at 2:20 pm
My post is in moderation. Have I used a banned word?
You’re on automoderation, and are likely to remain so. I’ve posted your last, but I can’t say I thought it was a particularly useful contribution.
Maynard Handley 07.04.09 at 9:47 pm
“Hardly. The assets of the failed firms (both physical and intangible assets) don’t just disappear in a puff of smoke. Usually, they are purchased by successful firms and redeployed. I doubt any programmer who worked on the development of CP/M, for example, spent much time in his or her career subsequently unemployed.”
This assumes that those assets are valuable and can *be* redeployed.
There was recently this glut of overbuilding of homes in the US, you may have heard of it. We now have a number of excess houses in places where people apparently don’t want to live. How do you propose to redeploy those houses?
Henri Vieuxtemps 07.04.09 at 10:06 pm
Without a doubt, many millions of man-hours (as well as tangible resources, of course) were completely wasted in the dot-com bubble, for example. Even in ordinary times every bankruptcy or merger or spin-off is a huge waste. Sure, one could argue that it’s a necessary evil, but denying it is ludicrous.
lemuel pitkin 07.04.09 at 10:26 pm
We did get a lot of fiber-optic cable, tho….
Min 07.05.09 at 5:17 am
Maynard Handley: “There was recently this glut of overbuilding of homes in the US, you may have heard of it. We now have a number of excess houses in places where people apparently don’t want to live. How do you propose to redeploy those houses?”
Follow the Nancy Reagan plan, put the homeless in them. “Just get a house.” ;)
Actually, something like that happened in Honolulu in the 1970s. Overdevelopment led to the phenomenon of the welfare condominium. :)
Michael Harris 07.05.09 at 11:45 am
Robert @ 24
The externalities of depletion and pollution are not included in the calculations.
and
Oil is priced at the cost of pumping it from the ground. The fact that it is non-renewable is not considered.
These may be problems with how capitalism operates, but “economics”, the discipline, recognises these outcomes as not being efficient. The external costs of the pollution (in the first instance) and the (possible, depending on market structure) lack of recognition of the scarcity rents associated with a non-renewable resource are fundamental issues confronted in any treatment of resource and environmental economics.
Of course allocation questions have an ethical component. (You said “moral”, but then you said to bring in the ethicists. I would like to keep the moralists out of it too as much as we can, but sure, bring on the ethics.) Geoffrey Brennan once decribed normative economics as being a branch of applied ethics, which sounds about right to me.
I’ve been reading over these The Problem With Economics threads on CT the last week or so to see if I can gather up any useful new insights, and it’s quite disheartening to see how easily the discussion gets derailed by people giving the distinct impression they do not quite know what they’re talking about.
oh why not 07.05.09 at 3:14 pm
“Oil is priced at the cost of pumping it from the ground.”
No. Just No.
Call it irrational self-interest.
http://forums.somethingawful.com/showthread.php?threadid=3159732&pagenumber=1
“While the global supply of oil will eventually dry up, the short-term flow has actually been increasing. In the six months before prices spiked, according to the U.S. Energy Information Administration, the world oil supply rose from 85.24 million barrels a day to 85.72 million. Over the same period, world oil demand dropped from 86.82 million barrels a day to 86.07 million. Not only was the short-term supply of oil rising, the demand for it was falling – which, in classic economic terms, should have brought prices at the pump down.
So what caused the huge spike in oil prices? Take a wild guess. Obviously Goldman had help – there were other players in the physical-commodities market – but the root cause had almost everything to do with the behavior of a few powerful actors determined to turn the once-solid market into a speculative casino. Goldman did it by persuading pension funds and other large institutional investors to invest in oil futures – agreeing to buy oil at a certain price on a fixed date. The push transformed oil from a physical commodity, rigidly subject to supply and demand, into something to bet on, like a stock. Between 2003 and 2008, the amount of speculative money in commodities grew from $13 billion to $317 billion, an increase of 2,300 percent. By 2008, a barrel of oil was traded 27 times, on average, before it was actually delivered and consumed.”
Robert 07.06.09 at 9:36 am
Michael, I think regular readers and commentators have tired of commenting on broad generalizations about mainstream economics since a year or two. Some threads on topic reached over 200 comments, if I recall correctly. I don’t think any minds were changed, though.
I’m sure many will agree that we’ve seen “people giving the distinct impression they do not quite know what they’re talking about.” Getting more specific – that is, specifying which people – might produce more heat, though. I sometimes see critiques which I think could be improved if the critic wasted more time studying mainstream economics.
tps12 07.06.09 at 2:20 pm
If your argument against capitalism’s waste is that loads of otherwise indistinguishable products nonetheless differentiate themselves thanks to the variety of their branding strategies, then I think you’ve basically conceded the point.
quanticle 07.07.09 at 9:19 pm
The first externality, pollution, can be internalized by mandating that polluters clean up after themselves or use technologies that do not pollute in the first place. Indeed, this is exactly what is done in most locales. If the cost of extraction is now higher than the value of the gold in the ground, the company is free to walk away and dig somewhere else.
The second externality isn’t an externality at all. Its up to the company to try to decide whether it is best to pull the gold out of the ground now, or to leave it in the ground and wait for higher prices. If the company decides to pull the gold out of the ground now, and prices rise in the future, then its value (as measured by its market capitalization) will go down when its investors realize that the management has wasted the corporation’s most valuable resource. If the corporation decides the opposite, and prices decrease, the same will occur as investors will realize that management has just wasted an opportunity to build up the firm’s profits.
Au contraire. Capitalism (with its attendant property rights) tends to exist in situations of maximum scarcity, not minimum scarcity. Indeed, some of our most socialistic cultural norms relate to intellectual property, which is completely unbound from the limits of scarcity. The costs of copying a CD are trivial, so it doesn’t seem morally wrong to do so even if many would argue that it violates the intellectual rights of the original creator.
Comments on this entry are closed.