« Summing up the AEAs | Main | Assorted links »
Rationality is a Property of Equilibrium
Some thoughts on rationality and economics, perhaps for a future paper, motivated by the financial panic:
Rationality is a property of equilibrium. By this I mean that rationality is habitual and experience-based and it becomes effective as it becomes embedded in the rules of thumb and collective wisdom of market participants. Rules of thumb approximate rational decision rules as market participants become familiar with an economic environment. Individuals per se are not very rational; shift the equilibrium enough so that the old rules of thumb no longer apply and we are likely to see bubbles, manias, panics and crashes. Significant innovation is thus almost always going to come accompanied with a wave of irrationality. When we shift to a significant, new equilibrium rationality itself is not strong enough to tie down behavior and unmoored by either reason or experience individuals flail about liked naked apes - this is the realm of behavioral economics. Given time, however, new rules of thumb evolve and rationality once again rules but only until the next big innovation arrives.
Posted by Alex Tabarrok on January 13, 2009 at 07:20 AM | Permalink
Comments
Hi;
This is rather evident (I am certainly wrong) but it looks to me like cafe arguments.... I hope to read the paper soon.
Rgds
Posted by: toto at Jan 13, 2009 7:58:06 AM
I think I see your point, but are you sure that it makes sense to call an actor 'irrational' if it acts to its best knowledge, but based on limited information and necessarily flawed rules?
I would say that 'irrational' is a useful word for an actor who prefers A over B, B over C and C over A, or A over A+B, but not for an actor who prefers A over B but would prefer B over A given enough experience.
Posted by: Zamfir at Jan 13, 2009 8:02:01 AM
Could the rational/irrational be recast in terms of risk? It might provide a firmer foundation, as rational/irrational definitions get slippery fast: when is it rational to act irrationally? Meanwhile, a la Atwood, keep clapping if you believe in banks!
Posted by: TD at Jan 13, 2009 8:03:52 AM
"...rationality is habitual and experience-based..."
I don't think rationality is the word you're looking for here. Something like "emotionality" would be better. People aren't generally creatures of reason, but rather, creatures of habit. Habit is an emotional response. We do what we did yesterday because it "feels" right. But other than that, I like the concept.
Posted by: Randy at Jan 13, 2009 8:06:39 AM
Given time, however, new rules of thumb evolve...
This rather implies that the ideally-rational outcome is also the evolutionarily stable outcome, something which is far from obviously true.
Posted by: david at Jan 13, 2009 8:09:43 AM
It's about time. Don't you think?
Individuals seem to get rational eventually.
It takes time for questions like "what is the most popular smartphone platform that provides the best networks?" to percolate through the economy.
"Do home prices ever stop going up? Do we ever run out of marginal mortgage recipients? Is this the same question?" The people who could have affected the situation were too busy raking in money to ask themselves these questions.
Is there a segment of economics that deals with time?
Posted by: Andrew at Jan 13, 2009 8:21:59 AM
I like Randy's "emotionality", it makes a distinction that might need to be clarified: do you (Alex Tabarrok) mean that far frome equilibrium, people become unavoidably worse at making judgements, or that on top of the unavoidable part they also start to make "emotional" errors they could have forseen, such as selling in a panic?
Posted by: Zamfir at Jan 13, 2009 8:22:21 AM
Then, the relevant question is when and how often is an economy in a state of equilibrium...
Posted by: arcop at Jan 13, 2009 8:27:18 AM
I think David made an important point. Is there any a priori reason to believe that repeated, rule-of-thumb based habitus will eventually approximate to what we call rational behavior?
Posted by: elmar at Jan 13, 2009 8:36:22 AM
I think it's a very interesting idea, but a few questions come to mind:
There is a language difficulty here. On one level, an equilibrium is defined by the actions of everybody aggregating to demand and supply in any given instant, so we are always in an equilibrium by definition. On another level, an equilibrium is a deeper, fundamental attractor that (at least in the short run) exists independently of people’s choices. In what follows, I will call the first “where we are” and the second “the attractor”.
Why would agents use rules of thumb instead of making decisions on a fully-rational basis? Is it just because they aren’t entirely rational people (not very satisfying) or are there constraints that induce a fully rational individual to use rules of thumb?
Under what market mechanisms do the individually sub-rational agents aggregate to collectively rational decision-making when we are close to the attractor and - potentially - to collectively irrational decision-making when we are far away from the attractor?
What form of decision rules do the sub-rational (rule of thumb) agents use? Could we say that agents use taylor-series approximations around the point they believe to be the attractor, with the exact location of the attractor being uncertain? If so, would it be interesting to imagine that simple agents use first-order (i.e. linear) approximations and sophisticated agents use second-order (quadratic) approximations?
What is the source of uncertainty? With my example in the previous paragraph, why doesn’t everybody instantly know the new location of the attractor and adjust their rules of thumb accordingly?
How do agents learn? Could we bypass this question by proposing that agents update their understanding of where the attractor is in a manner analogous to firms setting prices in the Calvo pricing (i.e. a fixed percentage of agents discover the truth in any given period)?
Posted by: John Barrdear at Jan 13, 2009 8:43:16 AM
Hi Alex:
First, as Zamfir's comment suggests, you need to define rationality. On a minimalist definition of rationality as transitivity at a single point in time, then a lot of apparently irrational behavior would be rational.
Second, your arguments are somewhat reminiscent of Doug North and Art Denzau's claim that rationality is a social, rather than individual, concept (see their paper on shared mental models).
On the other hand, it is almost certainly true that individual rationality can be inculcated, at least to some extent, by raising the costs of irrational behavior. The implication of this is that institutional (including market) design might affect individuals' responses to innovations, exogenous shocks (discontinuities), and shifting equilibria.
Finally, you seem to suggest that behavioral economics is concerned only with irrational behavior. I don't think that's true. It's concerned with all human behavior, including the perfectly rational, the imperfectly rational, and the irrational (however we might define those terms).
Best.
Dan
Posted by: dan cole at Jan 13, 2009 9:00:17 AM
I don't think it is rationality that calms things, it is fear. And once calm, greed again.
Posted by: odograph at Jan 13, 2009 9:02:52 AM
Alex,
Interesting idea to work on. However, do note that Paul Heyne has long argued this point in his presentation of the principles of the economic way of thinking --- see his discussion of rules of thumb and the question of cost plus mark-up pricing. I also think you can find similar discussions in Hayek, and his critique of "rational economic man".
Personally, I think this is also what Vernon Smith argued in his discussion of 'ecological rationality' -- at least that is how I have understood his argument. I think the major implication for economic theory is a switch from a focus on behavioral/cognitive assumptions to institutional context and the filter process that discipline decisions.
Let me be clear, I don't think these previous works have settled the issue, I am just saying that you will have some important thinkers to build on who have laid out some of the groundwork for you.
Pete
P.S.: BTW, John List's work might be added to the list.
Posted by: Peter Boettke at Jan 13, 2009 9:11:03 AM
there is probably a neurological analogue/cause for this. take a look at S.W.Porges, The Polyvagal Perspective, esp. section 3. It's available at http://web.mit.edu/autism/polyvagal%20perspective.pdf.
The basic idea is that the human nervous system is "tiered" and elements that are more "collaborative" (in their language, "spontaneous social engagement behaviors") function only during times when the organism considers itself "safe". when not safe, the organism exhibits fight or flight type behavior. The vagal "circuit" is what "switches" between the two by inhibiting fight or flight behaviors.
Posted by: babar at Jan 13, 2009 9:22:20 AM
This is why IHMO rules should be changed very slowly and carefully. Deregulation is mostly good but big changes in regulation can cause chaos in the short and medium time. It takes time for people to build systems to protect themselves. In the current financial crisis the market has finally lashed out and destroyed the more corrupt banks and with it much of the motivation to corruption; the worst thing to do now IMO is to make new regulations.
IMO this is also why interest rates should be changed slowly. A sharp drop can cause some chaos as we have seen.
I am a conservative libertarian.
Posted by: floccina at Jan 13, 2009 9:35:31 AM
Quick and dirty reasoning:
Rationality typically means "playing a best response". From an evolutionary game theory perspective, playing a best response maximizes a sub-population's "fitness" (or if you like, profit). Ergo over time, that sub-population will increase in size relative to other sub-populations (there's a variety of transfer mechanisms some memetic, some simple market survival) so that the aggregate population will be increasingly made up of the best-response playing subpopulation even if most of the population did not think it through beyond, "Hey, X made money than I did. I'll copy what X did." (For real-life examples of this evolution in action, see the proliferation of the West Coast offense and the Cover-2 defense.)
There's all sorts of caveats and cavils that can be made for specific cases (e.g., what's fitness maximizing can be dependent on the mix of strategies in the population), but in real brief that's why if the rules of the game aren't changing (i.e., no innovation) a market dynamic leads to "rationality" even if no one is thinking it through.
Hidden in the background of this quicky explanation is the implicit assumption that there is a strictly dominant strategy (e.g., West Coast > Run & Shoot for all possible defenses). But this is not a necessary condition. (For those more into the field, consider potential games and evolutionary games - a frequent topic of Sandholm).
Posted by: Jody at Jan 13, 2009 9:38:01 AM
(1) I agree that the word "rationality" here is a bit touchy, but I understand what you're getting at. People act AS IF they are rational when using rules of thumb. Throw the system out of equilibrium and they can no longer behave AS IF they were rational, because the rules of thumb no longer work.
(2) I agree with Pete that several people have worked on this area before. I would add to Pete's list some heterodox economists, such as my father (Edward Nell) and also agent based economists, in particular some econophysicists who have looked at financial markets, speculation, bubbles and crashes in this way.
(3) While you and my father seem to think that innovation can cause the upset, I personally think that some kind of regulation is more likely. When an innovation occurs you may have some herding effects, but I doubt you'd get a real upset. With innovation, you're not likely to lose all your rules of thumb - just some specific categories of "whats good." At first glance bubbles look like this kind of herding, with everyone jumping on the dot.com bandwagon, and maybe some are. But it also seems like you'd need more than an alluring investment opportunity to lose all rules of thumb -- instead a new law or regulation or lending environment would seem to throw people off more.
Posted by: liberty at Jan 13, 2009 9:48:01 AM
In a game theoretic setting, one can represent boundedness as some limit in the distance a person can move along their best response function.
Posted by: Keith at Jan 13, 2009 9:48:09 AM
I'm sorry, but that's really an odd idea - that there was something irrational about the behavior that led up to this market crash. The reverse is true - it would have been totally IRrational for a home buyer/hedge fund manager/mortgage broker etc NOT to get on the band wagon. Imagine it's 2004/5/6 - your neighbor/co-worker has just scored a good deal in a house or made a ton of money using complicated debt instruments - and the President is touting home ownership while the Fed Chairman is publicly totally confident in the ability of markets to regulate themselves .... you would be NUTS to turn away from doing what is so profitable to everyone else! (And if you were in 'the business', you would probably lose your job.)
The problem isn't irrationality - it's information, and timing. And although it's unfortunate, there is only one kind of social organization that can help here (ie has the interest of the whole at heart, and can act on its information with proper timing) - and that's government. I should add that the reason it's unfortunate is that a significant portion of the electorate sneers at the whole 'idea' of government - which is bad for Dems and Reps alike. My answer: don't demand LESS of government, demand MORE.
Analogy (yes, I know, analogies are in disfavor these days - but they work as long as they're not tortured - hopefully this one isn't): imagine the improvement in delivery times and lessened personal inconvenience if there were no automobile speed limits. It would be totally rational to go faster than you do today - in fact, because it is generally a bad idea to be going much slower than the average speed when traveling on a heavily used highway, it would be rational to speed up, like everyone else. But 'we' have made a judgment that the benefits aren't worth the toll. Etc.
What's irrational in my view, is to see what's coming but do nothing about it. Here's an excellent example: many many economists have been saying, for years, that the US is 'living beyond its means' - borrowing from China, not saving, running huge trade deficits, using gas guzzlers etc etc. So in many respects the current problems have been well foreseen. Yes, there have been some efforts to adjust but I would say they were at the margins. And granted, what was usually foreseen was a crash in the value of the dollar - which hasn't happened. And maybe the current crisis will solve that problem too, by increasing the savings rate and lowering imports and driving the auto makers toward more efficient vehicles, etc. But the fact is, the existence of a huge problem has NOT been a secret.
MORE competence in government, not less! Let the parties compete on the basis of SMARTS, not .... what they currently do.
Oops .... that's not too rational, is it?
Posted by: Skip at Jan 13, 2009 9:50:58 AM
Prices are set on the margin. In free open markets you may not need that many "rational" actors to create a rational market. I think Gary Becker in his Noble speech talked about individuals often being irrational but markets are rational.
I think you are wrong to assume that significant innovation leads to irrational actions. Unless you are assuming that this increases the cost of information or that the dumb and foolish love to jump into risky behavior.
I do think that the 401K and 403B markets are inefficient, but that is because sellers of these products are so good at lobbying Congress to keep purchasers in the dark. Almost no investor in a 401K understands the true cost structure of their plan, but that is more about government rules that aid such secrecy and restrict competition.
A quick way to resolve some of our current problems would be to create a secondary market (sell in the future markets) for financial derivatives such as CDOs to increase liquidity in these markets and fairly price them. Such an innovation may create significant movement in the market but would it be irrational? I think it would simply, and rather quickly, uncover the true value of these instruments, remove uncertainty from the market, and aid recovery.
Posted by: DanC at Jan 13, 2009 9:51:17 AM
I'm sorry, but that's really an odd idea - that there was something irrational about the behavior that led up to this market crash. The reverse is true - it would have been totally IRrational for a home buyer/hedge fund manager/mortgage broker etc NOT to get on the band wagon. Imagine it's 2004/5/6 - your neighbor/co-worker has just scored a good deal in a house or made a ton of money using complicated debt instruments - and the President is touting home ownership while the Fed Chairman is publicly totally confident in the ability of markets to regulate themselves .... you would be NUTS to turn away from doing what is so profitable to everyone else! (And if you were in 'the business', you would probably lose your job.)
The problem isn't irrationality - it's information, and timing. And although it's unfortunate, there is only one kind of social organization that can help here (ie has the interest of the whole at heart, and can act on its information with proper timing) - and that's government. I should add that the reason it's unfortunate is that a significant portion of the electorate sneers at the whole 'idea' of government - which is bad for Dems and Reps alike. My answer: don't demand LESS of government, demand MORE.
Analogy (yes, I know, analogies are in disfavor these days - but they work as long as they're not tortured - hopefully this one isn't): imagine the improvement in delivery times and lessened personal inconvenience if there were no automobile speed limits. It would be totally rational to go faster than you do today - in fact, because it is generally a bad idea to be going much slower than the average speed when traveling on a heavily used highway, it would be rational to speed up, like everyone else. But 'we' have made a judgment that the benefits aren't worth the toll. Etc.
What's irrational in my view, is to see what's coming but do nothing about it. Here's an excellent example: many many economists have been saying, for years, that the US is 'living beyond its means' - borrowing from China, not saving, running huge trade deficits, using gas guzzlers etc etc. So in many respects the current problems have been well foreseen. Yes, there have been some efforts to adjust but I would say they were at the margins. And granted, what was usually foreseen was a crash in the value of the dollar - which hasn't happened. And maybe the current crisis will solve that problem too, by increasing the savings rate and lowering imports and driving the auto makers toward more efficient vehicles, etc. But the fact is, the existence of a huge problem has NOT been a secret.
MORE competence in government, not less! Let the parties compete on the basis of SMARTS, not .... what they currently do.
Oops .... that's not too rational, is it?
Posted by: Skip at Jan 13, 2009 9:51:28 AM
Are we talking about the systematic element of learning costs?
Posted by: David Heigham at Jan 13, 2009 10:20:47 AM
ISn't "foresight" more the word we are looking for than "rational"? In the sense that in a static environment, actors behave as if they have close to perfect foresight, at least in a probabilistic "risk" sense, while in fact they only use heuristic rules with limited applicability.
This can have two different implications: from a modelling point of view, it means that perfect foresight is a good approximation for the actors in a static environment but less so for dynamic environments, and from a behaviouristic POV it might mean that during a change, people will believe they still have good foresight and behave too risky.
Posted by: Zamfir at Jan 13, 2009 10:25:50 AM
I think you are wrong to assume that significant innovation leads to irrational actions.
He doesn't assume but deduces it.
Posted by: JSK at Jan 13, 2009 10:28:19 AM
FYI - I very definitely did not post my comment twice. I previewed twice - posted once. The software wants looking at.
Posted by: Skip at Jan 13, 2009 10:29:34 AM
