Break-even effect
Sitting on a loss, people reach for risky bets that offer a shot at getting back to even.
What it means
The break-even effect is the increased appetite for risk that arises when people are behind and a gamble offers a chance to return to their reference point, even at the cost of a larger loss. After prior losses, options that could erase the deficit become disproportionately attractive because reaching break-even has special psychological value, consistent with the convex loss limb of the value function near the reference point. It is the counterpart to the house money effect: gains loosen risk-taking through one channel, while losses provoke desperate risk-taking through another. The pattern appears among gamblers chasing losses late in a session and among traders doubling down on losing positions. It illustrates how a salient reference point — here, 'even' — reshapes risk attitudes dynamically.
Why break-even is special
Thaler and Johnson framed prior outcomes through mental accounting: a person integrates a new gamble with the running balance of the current 'account' rather than judging each bet in isolation. While behind, the reference point that separates gains from losses stays anchored to the session's starting wealth, so any option that reaches zero converts a certain loss into no loss at all — a discrete, categorical jump that dominates smaller reductions of the deficit. The convex loss limb of prospect theory's value function already makes risk attractive across the loss region, but break-even adds a kink: the reference point itself carries value, so gambles that just touch it are prized beyond what the curve alone predicts. Thaler and Johnson called this pattern of combining outcomes 'quasi-hedonic editing.'
What the evidence shows
Break-even is one of behavioral economics' better-known dynamic effects, but its empirical record is uneven. The original demonstrations were hypothetical two-stage choices; field data are messier. Analyzing real online-gambling accounts, Chen and colleagues found chasing is multifaceted — players raised stakes more after wins yet played faster after losses — so no single break-even signature appears across every behavior. A real-money laboratory task by Juergensen and colleagues found no break-even effect at all: the apparent rise in risk as bankrolls shrank was a mathematical artifact of measuring the proportion wagered, and a mood-maintenance account fit the data better. Reviews of loss-chasing conclude the phenomenon is genuine but heterogeneous, shaped by individual differences and by whether a loss is framed as recoverable.
Where it shows up
Outside the casino the same reach for the reference point recurs. In equity markets it is a cousin of the disposition effect: Odean showed investors hold losers far longer than winners, reluctant to realize a loss and often adding to a sinking position in hope of returning to the purchase price. Sales forces facing a quota, gamblers late in a session, and bettors near closing time all show the pattern, because a deadline sharpens the reference point and narrows attention to the single outcome that would erase the gap. Product and subscription design can trigger it too, when a 'you're almost back to your streak' prompt or a partial-refund frame makes a risky top-up feel like the route back to whole.
Limits and caveats
Two conditions strongly moderate the effect. First, whether the loss is realized: people take less risk once a loss is booked and the account is psychologically closed, and more while it still feels open and reversible. Second, whether the gamble can actually reach the reference point — options that merely shrink the deficit, without a real path back to even, lose their special pull. The effect also assumes a stable reference point, which real decision-makers update; a bettor who resets 'even' to their current, lower balance stops chasing. Because break-even coexists with the opposite house-money effect and with plain mood repair, predicting behavior from prior outcomes alone is unreliable. Treat it as one force among several, strongest when a fixed target, a deadline, and an unrealized loss line up together.
Examples
Down $100 at the end of the night, a bettor takes a long-shot wager that could lose $200 but might just get them back to zero.
A salesperson far short of quota with a week left stops working the safe small deals and stakes everything on one improbable large account that would close the whole gap at once.
A player who has lost a long win streak spends real money on a long-shot loot box, chasing the position they held an hour ago rather than simply starting again.
A contractor who has already overrun a fixed-price job's budget sinks the remaining funds into one unproven shortcut that could finish on time and recoup the overrun, risking a defect that sinks the whole contract rather than booking the smaller, certain loss.
A negotiator who has already conceded on price rejects a reasonable middle offer and holds out for an all-or-nothing clause that would recover everything given up, risking walking away with no deal at all.
First described in Thaler & Johnson (1990).
Key references
- Chen, Z., Doekemeijer, R. A., Noel, X., & Verbruggen, F. (2022). Winning and losing in online gambling: Effects on within-session chasing. PLoS ONE, 17(8), e0273359. doi.org/10.1371/journal.pone.0273359
- Zhang, K., & Clark, L. (2020). Loss-chasing in gambling behaviour: Neurocognitive and behavioural economic perspectives. Current Opinion in Behavioral Sciences, 31, 1-7. doi.org/10.1016/j.cobeha.2019.10.006
- Juergensen, J., Weaver, J. S., May, C. N., & Demaree, H. A. (2018). More Than Money: Experienced Positive Affect Reduces Risk-Taking Behavior on a Real-World Gambling Task. Frontiers in Psychology, 9, 2116. doi.org/10.3389/fpsyg.2018.02116
- Odean, T. (1998). Are Investors Reluctant to Realize Their Losses? The Journal of Finance, 53(5), 1775-1798. doi.org/10.1111/0022-1082.00072
- Thaler, R. H., & Johnson, E. J. (1990). Gambling with the House Money and Trying to Break Even: The Effects of Prior Outcomes on Risky Choice. Management Science, 36(6), 643-660. doi.org/10.1287/mnsc.36.6.643