Fuck "Revealed Preference"

 

 

 Give 'Em What They Want

 

There is no greater thought-terminating cliche than "revealed preference". If you ever meet a person using this phrase, run.

For those not familiar with this phrase, it was coined by economist Paul Samuelson in 1938, in his paper, "A Note on the Pure Theory of Consumer's Behavior."  Samuelson was a self-described Keynesian economist who later wrote a popular economics textbook, "Economics: An Introductory Analysis."  (As Samuelson himself once remarked, "Let those who will write the nation's laws, if I can write its textbooks.")  He also wrote a popular newspaper column about economics.  Gradually, the term spread through the field of economics, and then into broader culture.

Samuelson's initial discussion of revealed preference was modest and technical, describing a way of measuring consumer behavior.  And understood that far, revealed preference is a perfectly reasonable standard by which to make measurements in economics - one among many.  Indeed, Samuelson is to be lauded for recommending that economists actually go out and collect data about consumers, rather than making unfounded assumptions about them.  But as the term has spread among heterodox economists and non-economists (some of whom may think they are economists), it accumulated all kinds of errors of interpretation and eventually developed into its own ideology.

WARP, SARP, and GARP

WARP, or the Weak Axiom of Revealed Preference, first proposed by Samuelson (though he did not use that term), states, in essence, that if a consumer prefers one bundle of commodities over another at a specific price point, they will continue to do so.  At every price point, the consumer will prefer one, and only one, bundle.  Thus economists can draw demand curves that intersect at the point where the cost will motivate them to switch preferences.

SARP, or the Strong Axiom of Revealed Preference, is similar, except with an additional assumption that transitivity is preserved: if a consumer prefers A to B, and B to C, then they will prefer A to C.

And GARP, or the General Axiom of Revealed Preference, allows for situations in which a consumer is indifferent, where they will be equally satisfied by two bundles of commodities (to use the lingo, it allows for non-convex demand curves).  There are situations in which economists need a more flexible model like this than to measure consumer preference.  (There are also other possible axioms that they can use, if they wish, like VARP, etc.)

Note that all of the above are axioms.  They are not, as is sometimes unconsciously inferred, anything that has been "proven" in economics - either through factual, statistical evidence or through a logical theorem.  They are more like assumptions that economists make, because to do so can make calculation easier.  And that's all that it is: an assumption.

In fact, a theorem has been proven about revealed preference: Afriat's Theorem, which shows that the results of revealed preference theory work out to be equivalent to the utility function theory that it was supposed to replace.  Samuelson took that as a vindication of his theory, but another economist, Stanley Wong, showed that if its results are equivalent to the theory it was supposed to replace, it has no explanatory power - as it turns out, Samuelson was unknowingly smuggling in all the assumptions of utility function theory but in a new mathematical vocabulary.  Wong went so far as to call revealed preference theory "a failed research program".

And it's not hard to see problems with the revealed preference argument.  For one thing, how do we know that preferences remain the same over time?  They may change, for irrational or unconscious reasons that have nothing to do with price.  Often people don't do what they actually prefer.  They do that which is convenient, or that which they are used to, or they conform to the preferences of their peers, or what their parents tell them to do, or their husbands, or their priests, etc., etc..  

But there's a deeper problem with revealed preference.  In the 1950s and 1960s, cognitive revolution

But by that time, that is, at the same time that the experts were coming around to recognizing that revealed preference theory was proving to be a failed research program, the concept of revealed preference had leaked into popular culture and was now used by all kinds of people making all kinds of arguments, while mangling the theory and mashing it to fit whatever agenda they happened to support.  In short, it had become the thought-terminating cliche we know today. 

Often, when people are confronted with something that doesn't fit into their worldview, they nonetheless remain convinced of their prior beliefs and indeed, having been challenged, dig in and become all the more adamant.  

Connotation of hypocrisy

To read revealed preference as normative is to mistake a method of collecting data for the result of that data collection. 

What helps enforce this moralistic reading of revealed preference theory is the very word "revealed".  It was revealed to me!  How can it be wrong?  Who are you, to question revelation? 

The revealed preference of 

AI slop 

According to the dogmatic logic of revealed preference, if Coke is the most popular drink, then it must be the case that Coke is the superior drink that most people prefer.  No need to perform a blind taste test.  No need for double-blind experiments.  No need for science.  The market has spoken.  All must heed its pronouncements.

Enough of this.  If economics is to be a science, then mere quantitative observation of what people are doing, "in the wild" so to speak, is not enough.  Economic experiments must be performed in a controlled laboratory setting.

The problem is that, if you push this logic of revealed preference theory far enough, you eventually come to bizarre conclusions, such as that the revealed preference of heroin addicts is doing heroin.  The revealed preference of serfs was serfdom.  The revealed preference of slaves was slavery.  The revealed preference of rape victims is... well, you get the idea.   

I can imagine the objection that someone is yelling at their computer screen now: "But that's not their free choice!  Those people were coerced!"  But that's exactly the question.  The doctrine of "revealed preference" is that that which consumers do is their free choice.  It deliberately ignores the possibility that they were somehow manipulated into this social arrangement.  But then it presents this unquestioned assumption as the very thing it sets out to prove. 

Everywhere, the empty phrase "revealed preference" is used to justify oppression and exploitation in a triumphalist gesture that amounts to little more than "that's the way it is" or "that's what people do" or "people do what they want to do" or "people desire what they in fact prefer" or "people get what they want," conveniently ignoring all of the ways this might not always be the case.  The use of this phrase signifies that the user is either incapable of, or refuses to engage in, critical thought.  Indeed, what the very phrase "revealed preference" means, ultimately, is that there is no need for criticism.  Instead, users of this phrase conform to the status quo, the powers that be, those who have conquered and won in "the marketplace of ideas".  As Debord put it, "The spectacle presents itself as something enormously positive, indisputable and inaccessible. It says nothing more than 'that which appears is good, that which is good appears.'"  (Then, when people are confronted with the fact that, in reality, the market is full of "inadequate equilibria," where, in stark contrast to their dogma, the market has not produced the optimal utopia, they are forced to come up with arcane, complex rationalizations to explain how come.)

The old saying goes, "The winners write the history books."  I would say that "The winners write the economics textbooks," but typically people who use the phrase "revealed preference" are not winners - they are sniveling toady losers trying desperately to suck up to the winners.