10.1 Prospect theory and loss aversion

Prospect Theory and Loss Aversion

Prospect theory and loss aversion challenge traditional economic models by incorporating psychological factors into decision-making under uncertainty. These concepts explain why people often make choices that seem irrational under standard theory, like overvaluing losses compared to equivalent gains.

Understanding prospect theory helps you grasp real-world economic behaviors, from consumer choices to financial markets. It's also crucial for designing effective policies, marketing strategies, and incentives that account for how people actually make decisions, not just how rational-agent models say they should.

Prospect Theory vs. Expected Utility

Key Principles and Differences

Expected utility theory assumes people evaluate outcomes as final wealth states and weight probabilities linearly. Prospect theory, developed by Kahneman and Tversky (1979), departs from this in several important ways:

• Outcomes are evaluated relative to a reference point, not as final wealth states. A gain of $100 isn't just "$100 more wealth"; it's "$100 above where you started."
• The value function is concave for gains and convex for losses, reflecting diminishing sensitivity. The difference between gaining $10 and $20 feels larger than the difference between gaining $1,010 and $1,020.
• Decision weights replace raw probabilities. People tend to overweight low-probability events and underweight high-probability ones.
• Framing effects matter: how a choice is presented can change the decision. Expected utility theory, by contrast, satisfies the invariance axiom and treats logically equivalent framings identically.
• The certainty effect leads people to overvalue outcomes that are guaranteed relative to outcomes that are merely probable, even when expected values favor the probable option.

Value Function and Decision Weights

The value function $v(x)$ is the core of prospect theory. Its shape captures two key behavioral patterns:

• Risk aversion in gains: The concave region means you'd rather take a sure $50 than a 50/50 gamble for $100, even though the expected values are equal.
• Risk-seeking in losses: The convex region means you'd rather gamble on a 50/50 chance of losing $100 (vs. losing nothing) than accept a sure loss of $50. People take bigger risks to avoid locking in a loss.
• The function is steeper for losses than for gains, which is where loss aversion shows up mathematically. Formally, $v(-x) > v(x)$ in absolute value for all $x > 0$.

Kahneman and Tversky proposed a specific functional form:

$v(x) = x^{\alpha} \text{ for } x \geq 0$

$v(x) = -\lambda(-x)^{\beta} \text{ for } x < 0$

where $\alpha$ and $\beta$ capture diminishing sensitivity (estimated around 0.88) and $\lambda$ captures loss aversion (estimated around 2.25).

The probability weighting function $\pi(p)$ is typically inverse S-shaped:

• Small probabilities get overweighted ($\pi(p) > p$ for small $p$). This is why people buy lottery tickets (tiny chance of a huge gain feels more likely than it is) and why they buy insurance against rare catastrophic events.
• Large probabilities get underweighted ($\pi(p) < p$ for large $p$). A 95% chance of winning doesn't feel like a near-certainty; people still worry disproportionately about the 5% chance of losing.

The overall value of a prospect with outcomes $x$ and $y$ (where $x \leq 0 \leq y$) is:

$V = \pi(p) \cdot v(y) + \pi(q) \cdot v(x)$

This replaces the expected utility formula $EU = p \cdot u(W + y) + q \cdot u(W + x)$, where $W$ is total wealth.

Loss Aversion and Decision-Making

Psychological Impact and Behavioral Effects

Loss aversion means losses loom larger than equivalent gains. The standard estimate is that the psychological impact of a loss is roughly twice as strong as an equivalent gain ($\lambda \approx 2.25$ in Kahneman and Tversky's estimates, though the commonly cited round number is a 2:1 ratio). Losing $100 feels about as bad as gaining $200 feels good.

This asymmetry drives several well-documented behavioral effects:

• Endowment effect: Once you own something, you value it more than you would if you didn't own it. In classic experiments (Kahneman, Knetsch, and Thaler, 1990), people given a coffee mug demanded roughly twice as much to sell it as others were willing to pay to buy it. Ownership shifts the reference point so that giving up the mug registers as a loss.
• Disposition effect: Investors tend to sell winning stocks too quickly (to "lock in" gains) and hold losing stocks too long (to avoid realizing the loss). This is costly because it often means riding losers down further while cutting winners short.
• Status quo bias: People prefer the current state of affairs because any change involves potential losses, and those losses weigh more heavily than potential gains.

Implications for Marketing and Policy

Because losses feel worse than equivalent gains feel good, framing matters enormously:

• Marketing strategies often frame offers as avoiding a loss rather than acquiring a gain. "Don't miss out on this limited-time offer" triggers loss aversion more effectively than "Take advantage of this deal."
• Pricing strategies emphasize what you'll save (loss prevented) rather than what you'll spend.
• Policymakers need to anticipate that changes perceived as losses will face stronger resistance than equivalent changes perceived as foregone gains. Framing energy conservation as "avoiding $200 in waste" tends to be more motivating than "saving $200."
• Tax policy design reflects this too: people strongly prefer receiving a tax refund (a gain) over equivalent lower withholding throughout the year, even though lower withholding gives them use of the money sooner.

Reference Points and Gains vs. Losses

Influence on Decision-Making

Whether an outcome counts as a "gain" or a "loss" depends entirely on the reference point. This is what makes prospect theory context-dependent in a way expected utility theory is not. In expected utility theory, only final wealth levels matter, so there's no role for a reference point at all.

Reference points are often the status quo, but they can also be shaped by:

• Expectations: If you expected a 5% raise and got 3%, that 3% raise can feel like a loss, even though your salary increased. Kőszegi and Rabin (2006) formalized this idea by defining the reference point as rational expectations about outcomes.
• Aspirations: A student aiming for an A who gets a B+ experiences it differently than a student who expected a C.
• Social comparisons: Your salary feels different depending on what your peers earn.
• Adaptation over time: The hedonic treadmill shifts your reference point as you get used to new circumstances. A raise that felt great in January becomes your new baseline by June.

Framing and Manipulation

Reference points can be deliberately shifted through framing, which has practical consequences:

• Anchoring in negotiations: The first offer sets a reference point. If a seller lists a house at $500,000, a counteroffer of $450,000 feels like a $50,000 discount (gain). If the list price had been $400,000, that same $450,000 would feel like a $50,000 overpayment (loss).
• Segregating gains, integrating losses: Because the value function is concave for gains, $v(500) + v(500) > v(1000)$. Two separate $500 bonuses yield more total psychological value than one $1,000 bonus. Conversely, because the value function is convex for losses, $v(-1000) > v(-500) + v(-500)$ (less total pain), so people prefer a single $1,000 loss over two $500 losses. This is sometimes called the "silver lining" principle.
• Financial markets: Investors often anchor to purchase prices or historical highs, treating any price below those reference points as a loss regardless of the stock's fundamentals.

Prospect Theory in Economic Applications

Consumer Behavior and Financial Markets

Prospect theory provides explanations for several behaviors that expected utility theory struggles with:

• Sunk cost fallacy: You keep attending concerts from a season ticket package you're not enjoying because you've already paid. The purchase price acts as a reference point, and leaving early would mean "wasting" the money (realizing a loss), even though the money is gone regardless. Under standard theory, sunk costs should be irrelevant to forward-looking decisions.
• Insurance demand: People often pay premiums for low-deductible plans that have negative expected value. The overweighting of small-probability catastrophic losses, combined with loss aversion, makes the premium feel worthwhile even when the math says otherwise.
• Equity premium puzzle: Stocks have historically earned much higher returns than bonds (roughly 6% annually in excess returns). Benartzi and Thaler (1995) showed that myopic loss aversion can explain this. If investors evaluate their portfolios frequently (say, annually) and feel losses about twice as sharply as gains, they'll demand a large premium to hold volatile assets. The combination of loss aversion and short evaluation periods makes the observed equity premium consistent with prospect theory.

Labor Markets and Policy Design

• Wage rigidity: Workers resist nominal wage cuts far more than they push for equivalent raises. A 2% pay cut in a year with 0% inflation feels much worse than a 0% raise in a year with 2% inflation, even though the real wage change is similar. This asymmetry contributes to downward nominal wage rigidity, which has macroeconomic consequences for unemployment during recessions.
• Tax policy: People prefer receiving a lump-sum tax refund over equivalent reductions in withholding throughout the year. The refund feels like a discrete gain; lower withholding just feels like slightly less of a loss each paycheck.
• Incentive design: Framing bonuses as something employees could lose (e.g., "you've been awarded a $1,000 bonus that will be forfeited if targets aren't met") tends to be more motivating than framing them as something to earn, precisely because of loss aversion.
• Public health: Presenting health initiatives as "avoiding future illness" rather than "gaining better health" leverages loss aversion to increase compliance.