The chooser the model forgot
Two neighboring countries, alike in almost every way — language, wealth, religion, the shape of daily life. In one, about twelve percent of people are registered organ donors. In the other, nearly everyone is. The gap isn't culture, or generosity, or how much people value saving a life. It's a box on a form.
In the first country, the form asks you to check a box to opt in. In the second, you're already in unless you check a box to opt out. Same people, same stakes, one different default — and a life-and-death decision swings from rare to nearly universal.
For twenty-four chapters this book has trusted a quiet assumption: that people know what they want and choose it. If that were simply true, a checkbox couldn't move a decision this large. So either people don't care about organ donation as much as they say — or the choices we make are telling us something less reliable than we thought.
To see what behavioral economics breaks, first name the thing it breaks.
Almost every argument in this course has leaned, quietly, on one character: a rational chooser. Demand curves assume people buy more as price falls because they're weighing value against cost. Gains from trade assume both sides judge their own benefit correctly. Consumer surplus assumes willingness to pay tracks how much a thing is actually worth to you. The whole apparatus runs on revealed preference — the move that lets an economist say: if you chose it, it must have been best for you, so your choice reveals your welfare.
That assumption is not stupid. Most of the time, for most decisions, people are reasonable judges of their own interest, and treating them as such respects their freedom and usually gets the prediction right. The assumption has earned its keep across twenty-four chapters. But "most of the time" is not "always," and the gap between them is where this chapter lives. Behavioral economics asks a simple, destabilizing question: what if the chooser is predictable in the ways they get it wrong?
If our errors were random, they'd cancel out. The trouble is that they lean.
Here is the move that matters. A rational-choice defender can absorb the idea that people make mistakes — if the mistakes are random. Random errors scatter in all directions and wash out across millions of buyers, leaving the market's averages intact. What behavioral economics found is that human errors are not random. They are systematic: they point the same direction, person after person, so they don't cancel — they accumulate.
Loss aversion makes a $20 loss sting more than a $20 gain pleases, which is why people hold losing investments too long and demand more to give up a thing than they'd have paid for it (the endowment effect). Framing means a surgery described as "90% survive" gets chosen far more often than the identical "10% die." Anchoring lets a meaningless number — a list price, a first offer — drag your sense of fair value toward it. And present bias makes us plan to save, study, and start the diet tomorrow, then choose now every time tomorrow arrives. None of these is occasional. Each one leans the same way in nearly everyone.
In a classic endowment-effect study, students randomly given a mug demanded about twice as much to sell it as other students were willing to pay to buy the very same mug. Ownership of a few minutes changed its "value" — something a pure rational chooser, weighing only the mug itself, would not do.
Present bias has a famous fix. The Save More Tomorrow program let workers commit in advance to raising their savings rate out of future raises — so the cost always fell on a later self. Enrollment and saving rates rose sharply. The lesson cuts both ways: the bias is real, and good design can work with it.
Random errors cancel out across many people, so market averages survive them and revealed preference still roughly works. Systematic errors all lean the same way, so they don't cancel — they add up, and the gap between "what people choose" and "what's good for them" becomes real and predictable, not noise.
The deepest crack isn't that we misjudge the world. It's that we misjudge ourselves.
The biases so far concern how we read the world: prices, odds, options. But there's a more unsettling one, and it aims inward. Affective forecasting is your ability to predict how a future event will make you feel — and the research is clear that people do it badly, almost always in the same direction. We overestimate how much the raise, the car, the new phone, the move to the coast, or the dream job will lift us, and how long the lift will last. Psychologists call it the impact bias: the imagined future is more intense and more permanent than the real one turns out to be.
Sit with why this is so much worse for the model than ordinary bias. The whole defense of revealed preference is that you are the best judge of your own good. Affective forecasting says that even you — sincere, informed, trying hard — systematically misjudge what will make you happy. When you buy the thing because you expect it to make you happier, your expectation is itself biased. The market then performs flawlessly: it delivers exactly the thing you chose. The disappointment that follows isn't a market failure at all. The market did its job. The forecast was wrong.
That is the thought the final chapter picks up. If we can't reliably predict what will satisfy us — and if, even when we get exactly what we chose, the glow fades faster than we expected — then "give people what they're willing to pay for" stops being a complete account of doing right by them.
If people are predictably biased, someone will use it. The only question is who, and for whom.
Discovering systematic bias is like discovering a lever. Levers get pulled. The question behavioral economics forces is who gets to pull this one, and in whose interest.
One answer is the market itself. A firm that knows about defaults sets the costly auto-renewal as the default and buries the cancel button. A firm that knows about present bias offers the free trial that quietly converts. A firm that knows about anchoring shows the crossed-out "original" price first. These are sometimes called dark patterns, and they are simply biases turned into business models — the consumer-sovereignty story running in reverse, the market exploiting the chooser instead of serving them.
The other answer is policy. If a default can trap people, a default can also protect them: auto-enroll workers into retirement savings (they can opt out), set organ donation to opt-out, make the healthy or cheaper option the easy one. This is the nudge — steering without forbidding, sometimes called libertarian paternalism. It promises to help people get what they themselves say they want, while leaving every option open.
But here is the warning light, and it is the chapter's most important sentence. Behavioral economics does not hand the win to the planner. The same bounded rationality that clouds the shopper clouds the expert designing the nudge — and the nudger has interests of their own. Who decides which default is "for your own good"? What stops a nudge from sliding into manipulation, or from serving the nudger rather than the nudged? The honest conclusion cuts both ways: systematic bias is a reason to distrust the simple claim that consumers always know best, and equally a reason to distrust the claim that experts know better. It indicts the easy faith in the market and the easy faith in the planner at the same time.
It proves no such thing. People are boundedly rational, not helpless — and the experts are boundedly rational too, with their own biases and incentives. The finding undercuts uncritical trust in markets and uncritical trust in central planners. It argues for humility and good design, not for handing your decisions to someone who is also, predictably, human.
A firm can set a costly default (auto-renewal) to trap inattentive customers; a policymaker can set a protective default (auto-enrolled savings) to help them. The mechanism is identical — the default sticks. It doesn't settle policy because the designer of the protective default is also boundedly rational and self-interested, so we still have to ask who designs it, and for whom.
The moves you can now make
Honest heads-up: behavioral economics is a small slice of the AP exam. These checks are conceptual — name the bias, explain why it strains revealed preference, weigh a nudge.
Identify
A streaming service offers a free month, then charges automatically unless you cancel. Most subscribers who mean to cancel never do. Name two behavioral forces at work.
Present bias (canceling is a small cost now, so it's always put off) and status-quo / default bias (staying subscribed is the default, and defaults stick). The design turns both into revenue.
Reasoning
Explain why systematic biases, unlike random mistakes, undermine the claim that "if people bought it, it was good for them."
Random mistakes cancel across many people, so revealed preference survives on average. Systematic biases all lean the same way, so they accumulate rather than cancel — meaning choices can predictably diverge from welfare, and "they chose it" no longer reliably means "it served them."
Evaluation
A government auto-enrolls workers into retirement savings, with an opt-out. Give one argument for the policy and one principled objection.
For: it counters present bias and helps people reach the savings they themselves say they want, while leaving the choice open (they can opt out). Objection: it relies on a designer who is also boundedly rational and may choose the default to serve goals other than the worker's — and steering, even gentle, can shade into manipulation.
Defaults, bias, and welfare
A firm makes its subscription auto-renew by default and hides the cancellation link. A nearby city sets its public retirement plan to auto-enroll, with an easy opt-out. Both rely on the same behavioral force.
Real people depart from the rational chooser in systematic, predictable ways — so the choices a market faithfully delivers are not always the choices that serve us, and that should make us slower to trust the consumer and the planner alike.
they would stop treating every purchase as proof it was good for the buyer, and start asking, case by case, when the choice belonged to the person and when it belonged to the way the choice was framed.
You can't fully trust your gut forecast of what will make you happy, and present-you will undercut future-you. So build the defaults yourself: automate the saving, add friction to the things you'll regret, and decide once instead of every day.
If you design choices, you're already nudging — there's no neutral default. The same lever that protects a customer can trap them. Decide on purpose which one you're building, because "they clicked agree" won't carry the weight you want it to.