Imagine taking a walk in your neighborhood. The sun is shining, the people around you are going about their business, and life is good. But then you pass someone on the sidewalk who makes you pause. He has a stack of $100 bills. He peels a bill from the top of the stack and sets it on fire with his lighter. Together you watch it burn away to nothing. He peels another bill from the stack and sets it on fire just the same way. And then another. The question: is what this man is doing rational?

To an economist, the answer is: maybe! Rationality is the conception of human behavior that underpins all of economics, and yet our money burner has not necessarily done anything that we can call irrational. How can this be? Why is it so crucial for economics to invoke this concept called “rationality”, even as it seems so far away from what the word means in normal language?

The Naming of Parts

Economics is about the allocation of resources: how can our society make the best use of our abilities and of the gifts that the world gives us? And what does “best” mean, anyway? One reason that these questions are tricky is that there are people involved. We have to think about how people think and behave in order to understand and influence how resources are used. Fortunately, we have a way to think about human behavior that is so powerful and so flexible that to this day, we follow this way almost exclusively.

Unfortunately, we got the name wrong. Of all the missteps that the discipline of economics has ever made, the worst may be that we gave our central concept of human behavior a name—rationality—that connotes something quite different to the layperson from what the modern economist means by the word.

But this word choice was no accident. When economics was first emerging from the generalist melting pot of Enlightenment intellectualism, it really was a discipline concerned with enrichment and wealth both for the individual and for society. It asked, essentially, how would a cold, calculating person achieve the best financial return in a given situation? But times change. Economics came to realize that the concept of rationality, with its whiff of moralizing, was in fact capable of so much more. As we shall see, this has unlocked many nuanced ways to understand and view the concept.

Yet confusion over the definition persists. Sensible people have claimed that perfectly reasonable things like altruism, spite, generosity, and gift-giving are irrational in the economic sense. But only for the most narrow-minded definition of rationality could this be true. These, and behaviors far more outlandish, are well within the purview of rationality. Any behavior is. Here is what rationality really means: when a person faces a choice, she will pick the thing that she likes best.

Stories About the World

Like any human endeavor, economics tries to tell stories about the world. Typically we call these stories models. More often than not they are written in math, but they are stories nevertheless. They say: here is how I think things work. If this bill passes, what do I think will change? If we are hit by an unexpected event, what do I think will happen?

The least you can ask of me as an economist is that my story must hang together. It must be internally consistent, so that the ending of the story follows sensibly from the setting and the characters. So my characters must be well-motivated. What are their wants, their needs, their desires, their temptations? What obstacles do they face? Are they rich or poor? What are their responsibilities? Who is their antagonist?

Rationality provides internal consistency for the characters, and so it closes the model: we may have a situation and a character, but we have no prediction until we decide what our character wants. Our characters can contain multitudes, but they are never, ever crazy. When the story calls for them to make a choice, they always make a choice that follows from how we have motivated them.

Notice that this makes no claim about what the motivations of the character actually are. Our rational protagonist could be humble, monomaniacal, greedy, reckless—so long as her choice follows, we call it rational. Rationality is the model within the model. It’s a story of how people choose, which we can then nest inside the larger story we are trying to tell.

The Heart Wants What it Wants

Since there are arbitrarily many motivations for us to choose from when building our character, there are arbitrarily many ways for someone to be rational, and arbitrarily many ways to close an economic model. This makes sense: when presented with exactly the same situation, you might behave differently than me. To assume that a person is rational doesn’t imply anything about what her choice will be until we decide what she likes. A good model, then, involves some artistry and judgment. What motivations do we think are plausible for our character to hold?

Another implication of the rationality concept is that economic models are underdetermined in a non-trivial way. If you try to predict my choice using the concept of rationality, how should you feel if what I do isn’t what you predicted? There are two possibilities. Either I’m not rational—I don’t always choose the thing I like best from the available options—or your guess about what I like was wrong. Your assumption about my preferences is an inescapable auxiliary hypothesis of your model. You cannot separate it. You cannot observe my choice and determine one way or the other whether I am rational.

This is why the money burner that we considered earlier is maybe rational and maybe not. Maybe he likes to burn money, or is doing performance art, or enjoys causing a scene—or hates Ben Franklin!—and he is making a perfectly sensible choice. Or maybe he likes money and is choosing irrationally.

And so rationality creates underdetermination. This makes economic models tricky to falsify. A person can certainly be irrational in the economic sense, but we can’t prove it. If you expected to see the man not burn his money, your model didn’t work. We can, at best, say that either your understanding of the situation was wrong, or your assumption of what the man liked was wrong. Conversely, a person can certainly be rational in the economic sense, but again we can’t prove it. We are in a perpetual state of Humean skepticism over cause-and-effect. As Milton Friedman puts it in “The Methodology of Positive Economics,” “[i]f there is one hypothesis that is consistent with the available evidence, there are always an infinite number that are.”

In economics, our alternative hypotheses are staring us in the face: there are a multitude of ways to rationalize a choice. A famous exploitation of underdetermination is Bobby Henderson’s Church of the Flying Spaghetti Monster, whose central tenet is the unfalsifiable claim that an invisible and undetectable being created every aspect of the universe and controls it to this day. It’s outlandish, but completely consistent with the evidence. Likewise, in economics, rationality gives us the flexibility to consider many explanations that are consistent with what we see. Our job is to evaluate these explanations for their plausibility and predictive generalizability.

Its flexibility makes rationality a surprisingly optimistic and empathetic concept. Despite the popular conception of “economic man” being a callous sociopath, drained of emotion and compassion, economic rationality requires no such thing. You’re the boss: whatever preferences you truly hold are quite acceptable, and rationality will not wag its finger at you to chastise you for your foolishness.

When we see a person do or say something that we struggle to understand, it is easy to call them crazy. But this is boring. It precludes understanding and creates an “other” that did not exist before. John Stuart Mill and other early economists laid a foundation for the discipline that assumed—controversially and optimistically—that the position and wellbeing of all the people of the world deserved equal consideration in our thought. The economist will never call you crazy. Instead she will seek to understand. Why did you choose this way? What motivation could you have had that led you to choose something that I didn’t expect?

Rationality All the Way Down

Economists certainly don’t believe that rationality is the way people actually choose in reality. Instead, when we do turn to observed choices, we consider the implications of what we see as if the person making the choice had been rational. To drive home the point that anything could be rationalized in this way required some provocative arguments. Most famously, Gary Becker, Nobel Prize Winner in Economics, insisted that the “economic approach”, including its reliance on rational choice, could be successfully applied to questions that are quite outside the realm of material goods and cash. Criminality, addiction, infidelity, suicide—even the most seemingly self-destructive behaviors can be rationalized. This might seem horrific. But by refusing to treat the chooser as an “other,” we can empathize. By imagining their choices as rational, it might just be that we learn something about the choices and, if we want, how to discourage them.

Rationality begins to seem like a tautology. We can write down an internally consistent story in which a person chooses just about anything we “need” her to do for the purposes of the message we want to convey. As a last resort against the tautology, we can look for rationality to give us a consistency condition: I will try to predict your next choice from what I know about your previous choice. But the universe never offers the same situation twice. We can continue to rationalize anything with a sufficiently rich set of rationales. If we need to, we can explicitly model your multitudes: the angel on your shoulder can rationalize one choice, but the devil on the other shoulder can rationalize another. There can be no smoking gun to “prove” that a person has made an irrational choice, short of being a passenger inside their head.

But this is a feature of economic rationality, not a bug. It reveals economic models to be the abstractions that they are. It insists that we have to begin a dialog with reality. What do people’s motivations actually look like? What choices do we actually observe? We need stories to organize our thoughts and to understand a complicated world, but any good story must have some kind of relationship to reality.

The good news is that economists and others who study human behavior know more than ever about how people actually choose and what shapes their choices, thanks to sophisticated experimental design and empirical analysis. We can create characters whose motivations we trust. We understand how nuanced human behavior really is, and we can make informed judgments about how to design the models of humans who will play in our models.

The meaning of rationality, though, is still debated. We have those who are hardline tautologists; we have those who say that emotionless rationality is a prescription for how a person “should” behave; we have those who say that rationality is predictive since the irrational will be weeded out; we have those who take rationality to mean choosing the best available given the time and energy you devote to making the decision. And yet these debates play around the edges. The spirit of rationality is supreme and immovable. It is too valuable to our work to give up—too seductive to resist.

Economists Are People Too

Popular discourse has fallen behind. It’s fair to say that economists are not terribly well-liked by the average layperson. Some of this has to do with the conflation of economics with business and finance. This is unfortunate, and I understand it. Yet some of this dislike has to do with precisely what we’ve discussed here. In popular articles about economics, misuse of the word “rationality” makes us economists seem like rather horrible people:

In 1982, some economists came up with a little game to study negotiating strategies. The results showed that rationality is subservient to more powerful drives—and demonstrated why human beings so easily conclude they are being wronged. The idea of the “ultimatum game” is simple. Player A is given 20 $1 bills and told that, in order to keep any of the money, A must share it with Player B. If B accepts A’s offer, they both pocket whatever they’ve agreed to. If B rejects the offer, they both get nothing. Economists naturally expected the players to do the rational thing: A would offer the lowest possible amount—$1; and B, knowing $1 was more than zero, would accept. Ha!

—Emily Yoffe, Slate

Now that we’ve learned what rationality is, should we call this rational? Maybe! It depends on what the player cares about. Player B may quite justifiably prefer to receive no money but punish A for an insultingly low offer rather than receive a mere dollar. The example shows only that either the punisher is irrational, or that they care about more than money.

It turns out that the second player often rejects a 20 percent offer [in the ultimatum game], which means that both players walk away empty-handed.

Many economists cannot understand why they’d do such a thing. To an economist, an offer of even 1 percent would be worth accepting since it is free money, and because for the second player it is ultimately irrelevant how much money the first player takes home.

But most people do not think like economists. When offered 10 percent or 20 percent or even 30 percent of the total, they are disgusted by the inequity—and willing to pay the price for that disgust by rejecting the offer.

—Stephen Dubner, Freakonomics.com

In the classroom of life, Dubner’s “economist” appears to have skipped the lecture on human emotion. His economist told a story of a rational player who cares only about money. But I’m an economist who told a story of a rational player who also cares about punishing people who mistreat her. Economists think both ways and more.

Understanding what economists mean by “rationality” is not just useful to understand how we build our stories of the world. It might also help to rehabilitate the battered image of the economist as a person who can’t comprehend how people truly think.

The Moral of the Story

All models are wrong. This is inescapable and by design. All stories are wrong. All narratives are wrong. To reason about the world is to live a Catch-22: the only perfect model is reality, but this is precisely what we need models to understand. But there is nothing wrong with being wrong. We do not ask for models that are true, or models that are realistic, just models that are useful. And so it is with rationality. It is the ultimate useful model. It is not designed to be true.

In the end, do economic models relate the real world? Some more than others—by adding more and more detail and by engaging in the conversation with evidence, we might get a story that describes reality quite remarkably well.

But once upon a time there was a tortoise and a hare. The fable is neither true nor realistic, yet it contains the essence of a truth. We learn something interesting about persistence and hubris by reflecting on the story. And so it is in economics: modeling and theorizing is story-telling. We try to create fables that have meaning. In our fables we have people, and the people come in all shapes and sizes. The concept of rationality can speak to them all. Rationality buys us flexibility in our fables, not rigidity. All that we have to do is use it.


Further Reading:


Image Credit:  Simon Cunningham from Flickr

About The Author

James Campbell
Academic Correspondent, Economics

James Campbell is an economist specializing in applying methods from game theory, microeconomic theory, and graph theory to topics including targeting in social networks, privacy regulation, coordination problems, and open access policies. He received a B.A. in Economics and Management from Oxford University in 2005 and a Ph.D. in Economics from Brown University in 2010. Since 2010 he has been appointed as an Assistant Professor at the University of Toronto, and is currently appointed as a Visiting Assistant Professor at Brown University. His research has been published in the Review of Network Economics, Quantitative Marketing and Economics, and the Journal of Economics and Management Strategy.

  • jon

    Do predictive models presuppose some kind of ontological bet that other ways of making sense of the world wouldn’t allow for? If, for existence, we all see the world differently and we can’t agree on the basic facts that would allow for assessment, wouldn’t models asking for predictive validation make no sense? Would models that allowed for useful results be so complicated (temporally-specific institutional variables, interactions of humans with completely different motivations) as to render them incomprehensible? Do aggregate models need individual microfoundations? Anyway, I think “rationality” is less problematic for those wanting to take on traditional economics but instead we should concentrate critical fire on the idea of stable equilibria.

    • James D. Campbell

      All models are contingent, so in a way looking for validation for a model is a red herring. But I agree that we can talk past each other in this world. One of the virtues of theorizing is that we can in principle agree on the crux of an issue before we get the data. There’s some research out there on when and if it’s possible that we “agree to disagree” in the Bayesian sense which is probably relevant to what you’re getting at.

      On your second point there are certainly scale and scope tradeoffs. The more intricate and detailed your model becomes, the less tractable they are, but also the less generalizable they are. For that reason I see no fundamental need for microfoundations if the story can hang together without.

      I’ve always been a bit confused by the insistence that we figure out if we’re “in equilibrium” and if equilibria are “stable”, to be honest. Maybe “equilibrium” is another label that attaches more mystique to its concept than it deserves.

      • jon

        I think equilibrium analysis is kind of fundamental to neoclassical economics. You have marginalist models (perhaps add in some “frictions” like sticky prices or asymmetric information) and come out with a point prediction. Well, that’s one kind of model. Then there are models like Wynn Godley’s sectoral balances approach (note, called an approach rather than a “model” interestingly) which doesn’t make point predictions but enables explanation and more of a narrative-based story for how the economy is as it is or might need to change. It doesn’t, however, wish away institutions (where are the institutions in a Phillips curve? And yet people generate a prediction for the inflation/unemployment trade-off anyway), it doesn’t create intricate models which leaves out the role of credit-creating banks etc etc. It seems to me that “modeling” means something very different in the neoclassical world and the heterodox world and it’s worth spelling out the differences between them. One could argue that the difference comes down to ontology- modeling approaches look very different if you subscribe to philosophical critical realism compared to Popperian neo-positivism. But perhaps the prove between the two research paradigms is how progressive or degenerative they are: the heterodox economists like Dean Baker, Thomas Palley and others predicted a financial crash, neoclassical economists? Not so much.

  • Pingback: What an Economist Means by “Rationality” | Hippo Reads | Public Philosophy Journal()