Behavioral Economics: The Errors Point the Same Way

Standard economics long assumed people make choices that best serve their own interests, given the information available. This assumption is not stupid — it is a simplification that works reasonably well in many settings and makes the mathematics tractable.

Behavioural economics began with a narrower observation: people depart from that model in consistent, predictable directions.

The consistency is the important part. If people simply made random errors, those errors would cancel out and the standard model would remain a good approximation. They do not cancel. They lean the same way, across many people and many situations, which means the departures can be measured, predicted, and designed around.

That is what makes the field practically useful rather than merely a list of ways humans are irrational.

The main departures

Several findings have held up across many replications, and they recur throughout the rest of this site because they explain so much.

Losses feel larger than equivalent gains. Roughly twice as large. This alone explains reluctance to sell losing investments, resistance to change, and why "you could lose X" outperforms "you could gain X" in persuasion. See loss aversion.

Reference points determine how outcomes feel. People judge changes from where they currently are, not final positions — so the identical outcome can register as a gain or a loss depending on framing.

Defaults have enormous influence. Making something the default rather than an option changes uptake dramatically, even when switching takes seconds. Partly effort, partly the implicit recommendation a default carries.

Present costs dominate future ones. People discount the future steeply and inconsistently, which produces plans they genuinely intend to follow and then don't.

Context changes preferences. Adding an irrelevant third option can change which of two options people prefer. This is difficult to reconcile with a model where preferences exist independently of how choices are presented.

Effort spent creates attachment. People value what they have worked for above its market value, which overlaps with cognitive dissonance.

What this changed in practice

The findings moved from academic curiosity to applied policy fairly quickly, because some produce large effects at almost no cost.

Automatic enrolment in savings and pension schemes is the clearest case. Switching from opt-in to opt-out raises participation substantially, without removing anyone's choice. The intervention costs essentially nothing and the effect is large.

Simplification turns out to matter more than incentives in many settings. Complicated forms reduce uptake of things people want, and removing steps often outperforms paying people.

Timing and framing of information. The same fact presented at the moment of decision, rather than in advance, changes behaviour considerably.

Commitment devices. Letting people bind their future selves — automatic transfers, pre-committed deadlines — works because it addresses the inconsistency directly rather than asking for more willpower.

There is a genuine caveat here that the field has had to confront. Several well-publicised findings have failed to replicate, and some effect sizes shrank substantially under scrutiny. The core results listed above have held up reasonably well; a number of more colourful ones have not. Anyone citing a single striking study should be read with the same caution the law of large numbers recommends.

The ethical question the field cannot avoid

If defaults and framing change behaviour, then whoever designs the choice environment shapes the decision — and someone always designs it. There is no neutral presentation.

That cuts both ways, and this is where the honest treatment differs from the enthusiastic one.

The same techniques that increase retirement saving also increase subscription sign-ups people did not intend and struggle to cancel. Difficult cancellation flows, pre-ticked boxes, and countdown timers are behavioural economics applied competently against the user's interest. The mechanism does not care who deploys it.

The usual defence of benign versions is that a nudge preserves choice — you can still opt out. That distinction is real and it is not sufficient, because the entire finding is that people frequently do not exercise available choices. If the effect depended on people easily reversing the default, it would not work.

A more defensible test asks two questions: is reversing genuinely easy in practice, not just in principle? And whose interest does the default serve? Those are questions about ethics and accountability rather than psychology, and the research cannot settle them.

What the field does establish is narrower and still valuable: people are predictably not the optimising agents the standard model assumes, the departures are systematic enough to design around, and anyone presenting a choice is influencing it whether or not they intend to. Pretending otherwise is not neutrality — it is just influence that nobody has taken responsibility for.