Shortcuts, not failures
In the previous guide you met behavioral economics and its founding observation: real minds run on bounded rationality — limited attention, limited time, limited information. This guide opens the toolbox that bounded minds actually use. A heuristic is a fast, frugal rule of thumb that turns a hard question we cannot answer into an easy one we can. Asked "how likely is this?", we quietly swap in "how easily does an example come to mind?" — and answer that instead. Most of the time we never notice the substitution.
It is tempting to sneer at these shortcuts, but they are mostly a feature, not a bug. A creature that paused to compute the full expected value of every rustle in the grass would be eaten before it finished the sum. Heuristics trade a little accuracy for an enormous saving in effort, and across the ordinary business of living they usually land close enough. The interesting part — the part economics cares about — is that when they miss, they do not miss randomly. They miss in the same direction, again and again, for the same reasons. A repeated, directional error like that is what we call a cognitive bias.
Anchoring: the first number sticks
[[anchoring|Anchoring]] is the pull of an irrelevant starting number on a later judgement. When we have to estimate something uncertain, we grab whatever figure is nearby, treat it as a starting post, and then adjust away from it — but we adjust too little, so our final answer stays stuck near the anchor. The unsettling part is that the anchor need not carry any information at all. In one classic study, people spun a wheel rigged to land on 10 or 65, then guessed what share of African countries are in the UN. Those who saw 10 guessed about 25%; those who saw 65 guessed about 45%. A number from a roulette wheel — pure noise — moved their estimates of world geography.
Once you know the pattern you will see it everywhere money changes hands. A jacket marked "was $400, now $150" feels like a steal because $400 set the anchor, even if nobody ever paid it. A salary negotiation is largely a fight over who gets to name the first number. House listings, restaurant wine lists with one absurdly priced bottle, the "recommended" donation already filled in — all are anchors quietly steering your sense of what is reasonable. The price you would have paid in a calm vacuum barely matters; the price next to it does the work.
Availability: what comes to mind feels common
The [[availability-heuristic|availability heuristic]] judges how likely or how frequent something is by how easily examples spring to mind. It is a clever trick, because in a stable world common things really are easier to recall. But ease of recall is driven by more than frequency — it is inflated by anything vivid, recent, emotional, or heavily covered in the news. So our risk-meter quietly tracks memorability instead of probability, and the two come apart.
This is why a single plane crash, splashed across every screen for a week, can leave millions afraid to fly even though driving the same distance is far more dangerous per mile. The crash is unforgettable; the thousands who arrived safely make no headlines and so leave no trace in memory. The same machinery makes us overrate shark attacks and terrorism while underrating the slow, undramatic killers — poor diet, high blood pressure — that quietly do far more damage. Vividness, not frequency, is steering the wheel.
Availability has a quieter economic cousin worth naming: it inflates our confidence in our own contributions. Ask each member of a team what share of the work they did and the totals routinely add up to well past 100% — everyone recalls their own effort vividly and a teammate's only dimly. That slide from "easy to recall" to "therefore large" is one road into overconfidence, a bias we will return to, and a reliable source of busted budgets and blown deadlines.
Representativeness: when resemblance fools us
The third workhorse is representativeness: we judge how likely something belongs to a category by how much it resembles our mental picture of that category — and in doing so we ignore the plain background odds. Meet Linda: thirty-one, single, outspoken, a former philosophy student deeply concerned with justice and anti-nuclear causes. Which is more probable — that Linda is a bank teller, or that Linda is a bank teller and a feminist activist? Most people pick the second. But it cannot be more probable. Every feminist bank teller is, by definition, a bank teller, so the narrower group can never be larger than the group that contains it.
The story of Linda matches our stereotype of an activist so snugly that resemblance shouts down arithmetic. This is the conjunction fallacy — judging a specific, vivid combination as more likely than one of its own parts. Its close relative is base-rate neglect: ignoring how common something is to begin with. A worked example makes the cost vivid. Suppose a disease afflicts 1 in 1,000 people, and a test is 99% accurate. You test positive — how worried should you be?
Out of 10,000 people: truly sick = 10 -> ~10 test positive (true positives) truly well = 9,990 -> ~100 test positive (1% false positives) positives total = 110 chance you are actually sick = 10 / 110 ~= 9%
Most people, doctors included, guess the positive means a 90%-or-higher chance of illness. The true answer is roughly 9%, because among 10,000 people the handful who are genuinely sick is swamped by the larger crowd of healthy people who test positive by mistake. Representativeness whispers "the test looks sick, so you probably are"; the base rate — how rare the disease is — is what actually settles the question. The same blind spot makes us mistake a few good quarters for a brilliant fund manager, or a striking anecdote for a trend.
Why biases are predictable, not random
Here is the move that turns psychology into economics. If errors were random, they would scatter in all directions and wash out in any large group — and a model that assumes perfect rational choice would still describe the average person fine. But these errors are not random. They lean. Show a thousand strangers a high anchor and almost all of them estimate high. Make a risk vivid and almost everyone overrates it. Because the whole crowd tilts the same way, the mistakes pile up instead of cancelling out, and they show up in real prices, real markets, and real policy.
Predictable is the operative word, and it cuts two ways. Because the tilt is reliable, a researcher can forecast it: tell me the anchor and I can guess which way your estimate will bend. But anything predictable can also be priced, sold, and engineered. Marketers, casinos, and app designers have always exploited these levers by instinct. Behavioral economics simply names them, measures them, and asks how they aggregate. The flip side — using the same regularities to help rather than exploit — is the design of a nudge, the subject of a later guide.
Putting the toolkit to work
You cannot uninstall these shortcuts — they run beneath awareness, and even experts who study them fall for them. What you can do is build a few checkpoints for decisions that matter, where slowing down is worth the effort. The aim is not to distrust every snap judgement (most are fine and fast for a reason) but to recognise the handful of situations — big purchases, vivid scares, pattern-matching about people — where the shortcut is most likely to mislead.
- Spot the anchor. Ask: is some number — a list price, a first offer, last year's figure — quietly framing my estimate? Set your own figure before you look.
- Discount the vivid. When a story feels alarming or unforgettable, ask for the actual rate, not the headline. Memorable is not the same as common.
- Recall the base rate. Before trusting a resemblance — a diagnosis, a stereotype, a hot streak — ask how common the thing is to begin with.
- Consider the opposite. Force yourself to argue the other side for a moment; it is the cheapest, best-tested antidote to overconfidence.
These three heuristics are not the whole story — they are the foundation. The same machinery, applied to gains and losses rather than probabilities, produces the asymmetric way we feel about winning and losing, and the deeper map of choice under risk that economists call prospect theory. Closely related is the framing effect — how the very same option can feel different dressed as a gain or a loss. Those are the next steps on this rung. For now you hold the core insight: our errors are not noise, they are signal — patterned, predictable, and therefore something economics can study.