Behavioural Economics (Critique of Maximising)

Why the standard model is challenged

The standard model assumes consumers maximise utility and firms maximise profit, with perfect information and unlimited ability to process it. Behavioural economics does not discard this; it asks where the assumption systematically fails and what follows for policy.

The word systematically matters. If people simply made random errors, those errors would cancel out and the model would still predict well on average. What behavioural research finds is that people deviate in predictable, patterned ways, which means the deviations can be anticipated, and deliberately used to improve outcomes.

It is worth being precise about what is and is not being claimed. The standard model remains an excellent approximation in many markets, particularly where stakes are high, choices are repeated, and feedback is quick. It performs worst where decisions are rare, complex, emotionally loaded, or where the costs arrive far in the future, pensions, insurance, health, energy tariffs. Those are exactly the markets where governments intervene most.

The three bounds

Bounded rationality

People have limited information, limited time and limited ability to calculate. Rather than optimising, they satisfice, they search until they find an option that is good enough, and stop.

▶

They also rely on heuristics, mental shortcuts that work adequately most of the time: choosing a familiar brand, following what others do, using price as a proxy for quality.

1. Choosing between thirty broadly similar insurance policies, a consumer cannot compute the expected value of each
2. they pick a recognised brand, or the one recommended
3. the choice is reasonable given the constraints, but is not utility-maximising.

Two named patterns worth having ready, because they explain a great deal of consumer behaviour:

• Anchoring: an initial figure shapes later judgements even when it carries no information. The "was £200, now £120" tag works on this.
• Availability: people judge how likely something is by how easily an example comes to mind, which is why vivid, well-publicised risks are overestimated and common, undramatic ones underestimated.

Choice overload follows directly: more options can reduce welfare rather than raise it, because the cost of evaluating them exceeds the gain from a better match. That is a genuine challenge to the standard assumption that more choice is always weakly better.

Bounded self-control

Even knowing what is best for them, people often fail to act on it. Present satisfaction is weighted more heavily than future benefit, so people under-save for retirement, over-eat, procrastinate and postpone exercise.

The gap here is between intention and action, not between knowledge and ignorance, which is why information campaigns alone frequently fail to change behaviour.

This is also the strongest economic argument for treating some goods as demerit goods: if consumers systematically over-weight immediate pleasure against future harm, then over-consumption of tobacco, alcohol or sugar is not simply a preference to be respected. It links this topic directly to intervention in 2.7 and 2.8.

Bounded selfishness

People are not purely self-interested. They give to charity anonymously, tip in restaurants they will never revisit, volunteer, and reject offers they consider unfair even at a cost to themselves. Fairness is a genuine motivation, and models assuming pure self-interest mispredict behaviour in these settings.

The standard demonstration is the ultimatum game: one person proposes how to split a sum, the other can accept, both get the split, or reject, in which case both get nothing. Pure self-interest predicts that any positive offer is accepted, since something beats nothing. In practice, offers seen as unfair are routinely rejected, and proposers anticipate this and offer far more than the minimum. People pay real money to punish unfairness.

Loss aversion

A fourth finding, and one the syllabus's examples repeatedly rely on: losses feel larger than equivalent gains. Losing £50 hurts more than gaining £50 pleases.

Two consequences follow that are worth naming:

• The endowment effect: people demand more to give up something they own than they would have paid to acquire it, which is why free trials and return guarantees are so effective.
• The sunk cost fallacy: people continue a failing course of action because of what they have already spent, even though sunk costs should be irrelevant to a marginal decision. This directly contradicts the marginal analysis in Unit 1.

Loss aversion also explains why framing works: presenting the same outcome as a loss avoided rather than a gain achieved changes the decision.

Choice architecture

Choice architecture is the design of the environment in which choices are made. There is no neutral design, options must be presented in some order, with some default, so the question is not whether to influence choices but how.

Defaults are the most powerful tool, precisely because bounded rationality and bounded self-control mean people rarely change them. Countries with opt-out organ donation have dramatically higher registration than opt-in countries with similar attitudes, the difference is the default, not the values.

Firms use the same tools, which is worth saying because it cuts against the assumption that choice architecture is something governments do to people. Anchoring in sale pricing, defaults in pre-ticked add-on boxes, framing in "only 3 left at this price", and complex tariff structures that exploit choice overload are all commercial applications of exactly this research. That is itself an argument for regulation: if the architecture is going to be designed by someone; it may as well be designed in the consumer's interest.

Why this matters for policy

Nudges are attractive because they are typically cheap and preserve freedom of choice, nobody is forced. Compared with taxes, bans or subsidies; they are politically easier and less distorting.

But there are genuine objections, and a strong answer raises them:

• Effect sizes are often small, and may fade once novelty passes.
• Paternalism. Someone must decide which behaviour to nudge people towards, which is a value judgement about what is good for them.
• Transparency and manipulation. A nudge that works because people do not notice it sits uncomfortably with informed consent.
• They may not address the underlying cause. Nudging people to save more does not help someone whose income is too low to save at all.
• Evidence quality varies, and several well-known findings have proved hard to replicate.

The balanced conclusion is that behavioural tools are a useful complement to conventional policy, not a replacement for it.

Worked example

A government wants to raise retirement saving.

The conventional approach is a tax incentive: make pension contributions tax-deductible so the return to saving rises. This assumes people are rational maximisers who will respond to the improved incentive.

Why it under-performs. Bounded rationality means many people never work out how much the incentive is worth or how much they should save. Bounded self-control means that even those who intend to save postpone acting. The incentive also disproportionately benefits higher earners, who pay more tax and would often have saved anyway.

The behavioural approach is automatic enrolment: employees are enrolled by default, and must actively opt out to stop contributing.

1. The default switches from "not saving" to "saving"
2. inertia now works in favour of the desired outcome rather than against it
3. participation rises sharply
4. and because opting out remains available, freedom of choice is preserved.

Evaluation. Automatic enrolment is cheap relative to tax relief and has produced large participation increases where used. But the default contribution rate may be too low to fund an adequate pension, so people may enrol and still under-save. It does nothing for those whose incomes leave no surplus. And it is paternalistic, the government has decided that saving more is better for you. The strongest policy probably combines both: automatic enrolment to overcome inertia, plus incentives and adequate default rates.

Second worked example: reducing sugar consumption

Compare three instruments for the same objective.

• A sugar tax raises price and reduces quantity, and raises revenue, but its effect depends on PED, and demand for sugary drinks is fairly inelastic, so consumption may fall less than hoped. It is also regressive.
• Information campaigns address bounded rationality but not bounded self-control, which is why they consistently underperform: people already know sugar is harmful.
• Choice architecture: reformulation targets, smaller default portion sizes, checkout placement rules, works with bounded self-control rather than against it, is cheap, and preserves the option to buy sugar anyway.

The judgement. No single instrument dominates. The tax changes relative prices and funds something; the nudge changes the path of least resistance; neither reaches someone determined to buy sugar, and nor should they. Saying which bound each instrument addresses, and therefore where each fails, is the analysis a 15-mark question is built around.

Common exam mistakes

• Claiming behavioural economics shows people are "irrational". It shows rationality is bounded and that deviations are systematic.
• Confusing the three bounds, rationality is about capacity, self-control about acting on intentions, selfishness about caring for others.
• Describing a nudge that removes choice. A nudge preserves the option; if choice is removed it is regulation.
• Treating nudges as costless and universally effective.
• Presenting choice architecture as something only governments do.
• Ignoring the paternalism objection, which most questions expect.

Exam technique

Name the specific bound or tool and apply it to the scenario, "this is a default effect, working because bounded self-control means few people actively opt out" is worth far more than a definition.

Where a question asks about a policy, contrast the conventional prediction (what a rational maximiser would do) with the behavioural prediction (what people actually do). That contrast is the analysis.

When comparing instruments, say which bound each one addresses. An information campaign targets bounded rationality; a default targets bounded self-control; a tax assumes the standard model works. Matching instrument to bound is the sharpest evaluative move available in this topic.

For evaluation: effect size and durability, paternalism and transparency, whether the root cause is addressed, and comparison with conventional instruments such as taxes and regulation.

Quick revision

• Behavioural economics identifies systematic, predictable deviations from maximising.
• Bounded rationality: limited information and computation → satisficing, heuristics, anchoring, availability, choice overload.
• Bounded self-control: present bias, so intentions fail to become actions, the case for demerit-good intervention.
• Bounded selfishness: fairness and altruism genuinely motivate people; the ultimatum game is the evidence.
• Loss aversion: losses outweigh equivalent gains, giving the endowment effect and the sunk cost fallacy.
• Choice architecture: defaults, framing, anchoring, nudges, restricted and mandated choice.
• Defaults are the most powerful tool, because inertia is strong.
• Firms use the same tools commercially, which is itself an argument for regulation.
• Nudges are cheap and preserve choice, but are paternalistic and often small in effect.

Not the topic you were looking for? Describe what you are stuck on in your own words and we will take you to the notes that answer it.