The assumption that people behave rationally does a lot of work in economics, but perhaps its most important function is to allow economists to assume that mutually beneficial deals always get done. If a seller places a value of a $5 on a good, and a buyer a value of $10, the assumption goes, the seller and buyer will agree on a price somewhere between $5 and $10, and trade will take place, simply because the exchange is mutually beneficial.
Economists and their detractors have spent at least half a century tearing apart the assumption that good deals always get done, first through the lens of transaction costs, and later through behavioral economics. Transaction costs dealt only a glancing blow to the assumption, however, because additional costs don’t really undermine it. One can certainly accept that some deals do not get done because the cost of negotiating them–the legal fees, the time required to induce the other party to accept a particular share of the benefits generated by the deal, and so on–are too high, without giving up on the notion that good deals, defined to be those that are mutually beneficial after transaction costs are taken into account, still always do get done.
Behavioral economics has turned out to be harder to dismiss because it suggests that neither party to a transaction may actually want to execute mutually beneficial trades. If the seller doesn’t place the right value on his good, thinking it is worth $20 to him when instead it is worth $5, and the buyer thinks the good is worth $5 to him when instead it is worth $10, then the two will not be able to agree on a price, and so a mutually beneficial trade will not take place. But objections based on behavioral economics are not what interest me about the assumption that good deals always get done.
What is really interesting about the struggle over whether good deals get done is that economics has always needed the fact that some good deals do not get done to create the tension that gives economic inquiry its meaning. An economics in which good deals always get done is an utterly uninteresting, unrealistic, and indeed solipsistic undertaking. And economics has always understood that. Long before transaction cost economics and behavioral economics, economists were careful to build into their models discrete loci of irrationality in order to give the models meaning. Without these areas of irrationality, the models would lack what a creative writing teacher would tell you is the essential element of any story: conflict.
But if a novelist were to try to introduce tension into a plot this way, by asking the main character to treat similarly situated supporting characters differently for arbitrary and unexplained reasons, the novel would be panned.
Consider, for example, as doctrinaire and orthodox a model as the general equilibrium model of Arrow and Debreu. If these men had really taken the assumption that all good deals get done seriously, they would have started with a bunch of households and firms, written down their utility and production functions, and then: bam! The model would have been done. For the assumption that all good deals get done would then have ensured that all trades that, according to the utility functions and production functions they had written down, are mutually beneficial, would then immediately be carried out.
To give the story the conflict it needs to be of interest, Arrow and Debreu added another assumption: that prices are uniform in all markets. (This assumption does not of course originate with them, but their model represents a sort of apotheosis of orthodox economics, making it useful to frame the discussion around it.) Uniform pricing creates tension because when prices are uniform a seller can make more money by intentionally refusing to sell to certain buyers, even when those sales would be mutually beneficial. This is the classic problem of the inefficiency of the uniformly-pricing monopolist.
Consider a seller who places $5 of value on a good and has two prospective buyers, one who places $100 of value on the good and the other who places $10 of value on the good. Without the uniform pricing restriction, the seller would always sell to both buyers, because whatever profits he happened to generate from his sale to the first buyer he could always increase by selling to the second buyer as well.
That changes with uniform pricing, because then the price the seller charges the first buyer must be the same as the price the seller charges the second buyer. If the seller is able to negotiate a price of $95 with the first buyer (a price the first buyer will, under the all-good-deals-get-done assumption, accept because he places a value of $100 on the good, and so would enjoy a net gain of $5 from the deal), then the seller will not sell at all to the second buyer, who is only willing to pay up to $10 for the good and therefore won’t buy at a price of $95. So a deal with the second buyer becomes impossible, even though, if a lower price could be charged to the second buyer, a deal would be mutually beneficial. If the seller and the second buyer could agree on a price of $7, for example, the seller would earn an additional $2 of profit.
But that price is impossible under uniform pricing, because to charge the second buyer $7 would require that the seller charge the first buyer $7 as well, eliminating $83 of profit from the deal with the first buyer relative to the $90 earned at a price of $95, in exchange for a paltry gain of only $2 in profit on the second deal. The seller could still ensure that all good deals get done, by charging that $7 price, or any price between $5 and $10, but it is not in the interest of the seller to do that.
Now the Arrow and Debreu model has the opportunity to become interesting, by giving the conditions under which all mutually beneficial deals will still get done, in spite of the uniform pricing restriction and therefore in spite of the failure of the assumption that all good deals get done as a general matter. In particular, the Arrow and Debreu model makes clear that perfect competition, or some other mechanism that leads to competitive prices, is required for all good deals to get done when prices are uniform. Competition ensures that if one seller tries to charge $95, the $90 in profits generated thereby will induce other sellers to enter the market and steal the buyer’s business by charging a slightly lower price, and as competition intensifies that price will be bid down to the $5 of value that sellers place on the product, ensuring that the second buyer is able to purchase the product as well. All good deals get done after all. By circumscribing the assumption that good deals always get done using a restriction that is realistic–many goods are sold at uniform prices–the model poses a problem that has a certain verisimilitude–how to ensure that all good deals get done when prices are uniform–and then gives the conditions sufficient to solve the problem (e.g., competitive markets).
All economic models follow the same playbook: all economic models create tension and practical interest by limiting the general economic assumption that all good deals get done in some way (usually, but not always by assuming that prices are uniform), and then trying to show what legal rules or policy interventions might be needed to ensure that all good deals do get done anyway. (Another example is the assumption of risk aversion in insurance economics.)
What is so peculiar about this rhetorical posture of economics is that the baseline assumption is always that good deals do always get done, and the model is then built around the introduction of some discrete deviation from that assumption. The model never starts from the assumption that good deals never get done.
Which gives all economic models an internally discordant character.
Why, for example, should I assume that when the monopolist charges $95 to the first buyer, that buyer will magically trade at that price, simply because trade is mutually beneficial, but at the same time I should also accept that the seller won’t try to charge a lower price in order to be able to engage in mutually-beneficial trade with the second buyer? Yes, the seller generates more profit by charging the higher price and selling only to one buyer. But by the same token, the first buyer could enjoy a greater net gain from the transaction by insisting on paying no more than $80 for the good, as opposed to the $95 price that I am asked to assume that the buyer will accept. The buyer does better insisting on a lower price, and if the seller insists on a higher price, then the two might never reach a deal, as Robert Cooter so insightfully pointed out years ago. I am therefore asked to accept that the profit motive is not the be-all-and-end-all for the seller and the first buyer, otherwise I could not assume that the good will sell at $95, and yet I am asked to accept that the profit motive is the be-all-and-end-all for the seller in relation to the second buyer, which is why the seller won’t think twice about pricing the second buyer out of the market and missing an opportunity for mutually beneficial trade with the second buyer. Why ever would that be the case?
Of course, it is in the nature of the introduction of a deviation from the assumption that good deals always get done to have such dissonance. But that just begs the question: does it make sense to rely upon inconsistent behavioral assumptions within the same model?
Keep in mind that in order for uniform pricing to give rise to tension in the Arrow and Debreu model, the same individual seller must be willing to compromise profits for the sake of completing mutually-beneficial transactions with buyers who are willing to pay high prices–inframarginal buyers, they’re called–but not be willing to compromise profits for the sake of completing mutually-beneficial transactions with buyers who are able to pay only lower prices–marginal buyers, these are called. There seems to be no basis for assuming that sellers are socially oriented with respect to inframarginal buyers but rapaciously-profit-driven with respect to marginal buyers, other than the rhetorical need of model builders to introduce tension into the stories they are telling about economic activity.
But if a novelist were to try to introduce tension into a plot this way, by asking the main character to treat similarly situated supporting characters differently for arbitrary and unexplained reasons, the novel would be panned. The trouble for economists is that if they start adding content to the personalities of economic actors, they end up falling down the behavioral economics rabbit hole. There are too many different personality types from which to select , and the mathematics required to build models in any case becomes intractable. But if economists stick with the basic assumption that all good deals get done, then they paint a Panglossian portrait of economic activity that leaves them unable to identify economic problems or solve them. The result is an economic theory built on arbitrary and self-contradictory assumptions about when deals get done.
A more tenable theoretical approach would be to accept that good deals don’t always get done, all the time, in all circumstances. That means that even in competitive markets, sellers will fail to sell to buyers at the market price. That also means that in monopoly markets, sellers may fail to sell to inframarginal buyers at the monopoly price. Laying off absolute assumptions regarding whether deals always get done should also release economics from going to the opposite extreme: assuming that when good deals do not always get done good deals must therefore never get done. Which means that we should not be surprised to come upon monopolists that charge competitive prices.
Jettisoning absolute assumptions about whether good deals get done would prevent economics from making grand claims, such as the claim of the Arrow and Debreu model that competitive markets are always efficient. But it would not make economic theory useless. Economic theory could still tell us plenty about potentials: such as the amount of gain that would be created were policymakers to encourage buyers and sellers to strive to make mutually beneficial deals whenever possible. (Guido Calabresi makes a similar point when he argues that economics should focus less on how to expand the production possibilities curve and more on how to get the economy to that curve.) It would also help explain economic institutions that economics has so far been unable to penetrate.
Like advertising. The classic economic explanation for advertising is that it provides consumers with useful product information, something that is almost impossible to believe in the information age, if it ever was credible. But in a world in which good deals don’t always get done, there is another potential economic justification for advertising: that it seeks to overcome whatever cognitive or bargaining failures otherwise prevent good deals from getting done. In a world in which mutually beneficial transactions don’t always happen, because consumers are irrational, one would expect to find sellers spending large amounts of money trying to cajole buyers into buying, even in situations in which the deals on offer are good for buyers and so in theory they should embrace them without needing to be persuaded to do so. (That would go some ways toward undermining my own argument here that persuasive advertising must be bad for consumers, because absent advertising rational consumers always purchase the products that are best for them.)
There’s nothing wrong with the use of simplifying assumptions in economics, or in thought of any kind. But the use of inconsistent assumptions about behavior in the same model–often in relation to the same economic actors in the model–is a different story.
And all of economic theory is based upon doing just that.