Normally, I wouldn’t take the time to defend Neoclassical Economics. It’s a generally silly way of looking at economic behavior; however, WHY it is silly is actually rather important. For this reason, it’s sometimes necessary to point your intellectual guns at the people attempting to broadside it from the wrong direction. I’m a fan of saving what is worth saving, as it were.
Which brings me to this article I found on Naked Capitalism (a great site, btw).
The initial thrust of the essay is to attack the notion of “rational actors” by pointing out the obvious: people act irrationally! Mr. Pilkington offers up a few simple examples of this, provides one possible counter-argument from neoclassicals, and then neatly skewers it. Interestingly, “acting rationally” remains completely undefined…it seems to be taken to mean acting in such a way as to perfectly maximise utility, which would indeed be an absurd property to assert is held by beings who are also asserted to be lacking tremendous amounts of information.
However, that is not what is meant by serious economists talking about a rational actor. Mr. Pilkington quotes an economist on this, and then proceeds to more-or-less completely misunderstand the what the quote is saying:
The hypothesis of rational expectations means that economic agents forecast in such a way as to minimize forecast errors, subject to the information and decision—making constraints that confront them. It does not mean they make no forecast errors; it simply means that such errors have no serial correlation, no systematic component.
Pilkington claims this asserts people have perfect knowledge, and that any time they make a mistake it’s an anomaly. That isn’t at all what is being asserted. What’s being said is that people make decisions in order to maximise their robustness against Knightian Uncertainty. People confront uncertainty which is immeasurable and unknowable, given their current information constraints, so they try and make it so their actions are as little affected by the unknown as possible. Since people are making these decisions against a backdrop of absolutely unquantifiable risk, their responses will have a random component to them (guesswork) that will be necessarily inconsistent among actors. That, in turn, implies there can exist no consistent error to their guesses…unless that error grew out of some background system of responding to unknown factors of a certain kind.
In which case, I would assert there is a reason people act that way, and they are thus being rational. And I can make that assertion because rationality does not imply making perfect decisions, but rather making acting in a way which is consistent with a set of rules. We may call an action irrational because it feels to obtain an externally defined goal, but if that decision comes as the result of a set of rules, then the decision was arrived at rationally.
Since in general I think it’s safe to say people make decisions based on rules (known or unknown), they are acting rationally and are therefore rational actors. They fail to perfectly maximise utility due to their lack of omniscience, but the limits of local knowledge have long been a point of interest in sociology, psychology, and economics.
No, the point where Neoclassical Economics falls down is when they assume that the aggregation of decisions of a mass of actors operating within information constraints cancels out the individual errors, and this is a failure to understand the scope of Knightian Uncertainty in decision making…effectively, Neoclassical Economists think that all risk is quantifiable, even if we haven’t done so yet. They lack any conception of a absolute uncertainty.
To continue on, Pilkington goes on to lambast the Efficient Markets Hypothesis, which I am happy to depict in the harshest light possible. Seriously, it is an idea whose scope is so laughably limited in application that its hard to conceive of taking seriously. The unfortunate problem is that EMH wasn’t really taken seriously; instead, it was used as a sort of idealized model that markets approximate, a notion rather like the relation of Newtonian Gravity to General Relativity. Assume markets come close to approximating a truly efficient market, open up a single variable to observation, and see how a particular market differs. It’s actually a pretty solid foundation for experimental economics, but for one problem.
Markets really don’t at all approximate an “Efficient Market”, due to the existence of Knightian Uncertainty.