Chapter 31: Risk Policies

Type: Chapter 31

Category: Psychology

Difficulty: Intermediate

Reading time: 6 min

Core idea

A company’s executives each face a business decision with an 80% chance of a $2M gain and a 20% chance of a $500K loss. Each executive declines — the expected value is positive ($1.5M), but loss aversion makes the individual gamble unattractive. The CEO, surveying the situation across 100 such decisions made by 100 managers across the organization, says: “Take all of them.” At the portfolio level, the expected gain is enormous and the probability of any overall loss is vanishingly small.

The individual executives are guilty of narrow framing: evaluating each decision in isolation. The CEO is using broad framing: evaluating the aggregate policy. Loss aversion that is individually rational in isolation becomes collectively self-defeating when applied narrowly to many decisions that could be pooled.

Why it matters

The logic of aggregation

When a gamble with positive expected value is played many times, the law of large numbers guarantees that the aggregate outcome will be close to the positive expected value. Individual outcomes vary — but the aggregate converges. Loss aversion is calibrated to the individual-gamble level; it is not recalibrated when the same gamble is embedded in a long series.

This creates a systematic problem for organizations. Risk-averse executives will decline many positive-expected-value projects — not because those projects are bad individually, but because loss aversion makes each one feel too risky. The aggregate result is a portfolio that is far worse than the portfolio that would result from accepting all positive-expected-value projects.

The “risk policy” recommendation

Kahneman’s recommendation: establish explicit risk policies that define the standard of decision across many similar choices. Instead of evaluating “should I take this gamble?” frame it as “what is my policy for this class of gamble?” If the policy is “accept all gambles with >3:1 upside-to-downside ratio,” then each individual decision does not need to be re-agonized over. The policy does the work.

This is how insurance works in reverse: instead of paying a premium to offload risk to a larger entity, you become the larger entity yourself by pooling your own risk across many decisions.

When narrow framing is unavoidable

Narrow framing is most costly when decisions are independent and numerous. It is less problematic when decisions are few and large — where the expected value is genuinely at stake and cannot be diversified by repetition. Narrow framing is also more damaging when loss aversion is high relative to expected value — a small positive expected value will be rejected by highly loss-averse evaluators even though aggregation would justify acceptance.

Key takeaways

Key takeaways

• Narrow framing: evaluating each risky decision in isolation, applying loss aversion at the level of the single gamble rather than the portfolio.
• Broad framing: evaluating a class of similar decisions as a policy — accepting all decisions that meet a defined threshold across many repetitions.
• The aggregation argument: a gamble rejected individually on loss-aversion grounds would be accepted if played 100 times — the law of large numbers eliminates the tail risk that loss aversion is protecting against.
• Risk policy recommendation: instead of deciding 'should I take this gamble?', define a policy class and accept/reject all similar gambles by the policy standard.
• Organizational implication: executives using narrow framing systematically reject positive-expected-value projects; CEOs who frame decisions as policies unlock the aggregation benefit.
• Limits: broad framing assumes decisions are independent and numerous enough for aggregation to work; rare, large, or correlated bets do not aggregate away tail risk.

Mental model

click to zoom

Narrow vs. broad framing — how aggregating across many independent decisions changes the risk calculus.

Read it as: A single positive-expected-value gamble, evaluated narrowly, will be rejected by loss-averse decision-makers. The same gamble, evaluated as part of a policy for a class of similar decisions, will be accepted — because the aggregate expected value is high and the tail risk is diluted by repetition. The critical move is shifting from single-decision to policy-level framing.

Practical application

Writing explicit risk policies

To combat narrow framing in organizations, have decision-makers write their decision criteria before evaluating any specific case. “I will accept any project with positive expected value and a loss-to-gain ratio under 1:3” is a policy. Applied consistently, it produces better aggregate outcomes than case-by-case loss-averse evaluation.

Applications:

• Corporate R&D: individual product development decisions are typically narrow-framed; portfolio-level thinking is needed to realize the aggregation benefit. A research portfolio of 20 early-stage projects with positive expected value should be accepted as a group even if each individual project would be rejected on loss-aversion grounds.
• Sales negotiation authority: giving salespeople a risk policy (“accept any deal above $X margin and above Y% revenue retention”) replaces case-by-case escalations with consistent, policy-level decision-making that aggregates better.
• Personal finance: re-balancing and dollar-cost averaging are risk policies — they commit in advance to accept portfolio fluctuations rather than react to each market movement with narrow-framing loss aversion.

Example

A media company evaluates 50 content pitches per year. Each pitch is evaluated by a creative director who must decide to commission or pass. The director, applying narrow framing and loss aversion, rejects 35 of the 50. Many of those rejections were positive-expected-value projects.

If the company adopted a portfolio policy — “commission all pitches with expected audience above X and content quality above Y, regardless of individual project uncertainty” — it would accept more projects, some of which would fail, but the aggregate portfolio would generate higher revenue and a richer catalog. The individual evaluation frame creates systematic under-investment in content; the policy frame corrects it.

Related lessons

Related

Chapter 26: Prospect Theory The loss aversion that makes narrow framing costly — why individual decisions are evaluated below their expected value. Chapter 30: Rare Events Overweighting of low-probability bad outcomes — the same mechanism that makes narrow framing produce excessive risk aversion. Loss Aversion The asymmetric weighting of gains and losses that drives the narrow-framing problem.