Home › What the research says

Loss Aversion and Why a Funded Envelope Feels Different

The original claim

Daniel Kahneman and Amos Tversky published prospect theory in Econometrica in 1979 as a descriptive alternative to expected utility theory. The part that matters here is not the probability weighting but the value function: value is assigned to gains and losses relative to a reference point rather than to final wealth, and the function is generally steeper for losses than for gains. That is where loss aversion comes from. The practical consequence is that the same dollar is not one thing. Whether it feels like a gain or a loss depends entirely on what you were treating as the starting point when the dollar moved, and the starting point is often something you set yourself.

A field test where the framing mattered

Roland Fryer, Steven Levitt, John List, and Sally Sadoff ran a field experiment on teacher incentives, reported in a 2012 NBER working paper. Financial incentives for teachers had generally been ineffective. They tested a loss-framed version in which teachers were paid in advance and asked to give the money back if their students did not improve sufficiently. That arm increased math test scores between 0.201 and 0.398 standard deviations, which they describe as equivalent to raising teacher quality by more than one standard deviation. A second arm, identical in money but implemented as a standard end-of-period bonus, produced smaller results that were not statistically significant. Same payment, different reference point, different outcome.

The same trick failed somewhere else

Lamar Pierce, Alex Rees-Jones, and Charlotte Blank randomized whether sales bonuses were prepaid or postpaid across 294 car dealerships, published in the American Economic Journal: Economic Policy in 2025. Their analysis, which they describe as somewhat statistically imprecise, gives strong indications that loss framing had quantitatively important negative effects. They also document a mechanism: loss framing increased incentives for gaming behaviors, which is what happens when people who stand to lose money find ways other than performance to avoid losing it. They explicitly reassess the common wisdom that loss framing is desirable. Two field experiments, opposite signs, so the honest summary is that loss framing is context-dependent, not a reliable lever.

The critique of loss aversion itself

David Gal and Derek Rucker published a review in the Journal of Consumer Psychology in 2018 arguing that current evidence does not support the claim that losses, on balance, tend to be any more impactful than gains. They spend much of the article on why belief in loss aversion as a general principle has persisted among researchers despite this, and they argue for a more contextualized view of when losses matter more. You do not have to take a side to use the practical takeaway. Loss aversion is not a law you can count on to do your budgeting for you. It is a pattern that shows up in some settings, reverses in others, and depends on the reference point people are actually holding.

The envelope change to make

The durable, uncontroversial part of this literature is reference dependence, not the size of the asymmetry. A funded envelope sets a reference point in advance, so a purchase reads as depleting something you already have rather than as an abstract subtraction from an account balance you never look at. That is a real change in how the decision is framed, and it costs nothing to arrange. So fund envelopes at the start of the period rather than letting them accumulate spending against an open-ended limit, and check the remaining balance before the purchase, not after. If you only ever see the number afterward, there was no reference point at the moment that mattered.

Sources

Every source below was retrieved and checked. Findings are reported as the source states them.

• Prospect Theory: An Analysis of Decision under Risk — Econometrica, 1979
  Value is assigned to gains and losses relative to a reference point rather than to final assets, and the value function is generally steeper for losses than for gains.
• Enhancing the Efficacy of Teacher Incentives through Loss Aversion: A Field Experiment — NBER Working Paper No. 18237, 2012
  Paying teachers in advance and asking them to return the money if students did not improve sufficiently increased math test scores between 0.201 and 0.398 standard deviations; a second arm identical except implemented in the standard fashion yielded smaller and statistically insignificant results.
• The Negative Consequences of Loss-Framed Performance Incentives — American Economic Journal: Economic Policy, 2025
  Randomizing pre- or postpayment of sales bonuses at 294 car dealerships gave strong indications that loss framing had quantitatively important negative effects, partly through increased incentives for gaming behaviors.
• The Loss of Loss Aversion: Will It Loom Larger Than Its Gain? — Journal of Consumer Psychology, 2018
  A review concluding that current evidence does not support that losses, on balance, tend to be any more impactful than gains, and arguing for a more contextualized perspective on the relative impact of losses versus gains.

Common questions

Is loss aversion real or has it been debunked?

Neither, exactly. Gal and Rucker argue in a 2018 review that the evidence does not support losses being generally more impactful than gains, and they call for a more contextualized account. Others disagree. What is not in dispute is the underlying structure from prospect theory: value gets assigned relative to a reference point rather than to final wealth. That part is what a funded envelope uses. Whether the loss side is exactly twice as heavy is an open argument and not something your budget depends on.

Does that mean I should punish myself for overspending?

The evidence does not support that. The teacher study worked by moving money up front and making the target concrete, not by adding punishment, and the car dealership study found the same loss framing produced negative effects and gaming behavior. When the downside is severe enough, people route around the measurement instead of improving the outcome. Applied to a budget, that looks like not logging the purchase. A budget you stop recording is worse than one you overspend honestly.

Why does a card balance not create the same effect?

Because there is no reference point set in advance. An account balance is a single number that covers everything you might do, so no individual purchase depletes anything identifiable, and reference dependence needs an identifiable starting point. An envelope funded to a specific amount for a specific purpose creates that starting point before the decision is made. That is a structural difference, not a claim that the app makes you spend less.

Related