Money on the Line: Do Commitment Devices Work?

Odysseus did not trust the man he was about to become. Sailing toward the Sirens, he had the crew plug their ears with wax and tie him to the mast, under orders to ignore whatever he screamed. Present Odysseus struck a deal that future Odysseus could not undo. Behavioral economists call that arrangement a commitment device, and the Ulysses move is still the cleanest picture: you bind the weaker version of yourself before temptation starts singing.

The modern version skips the rope and goes for your wallet. Sign a contract, put real money down, and if you miss your goal the money is gone. Deadlines run on the same precommitment logic, and whether deadlines work turns out to be its own strange story. Cash raises the stakes from awkwardness to actual loss. The open question is whether paying for your own discipline buys any, and the research's answer should annoy fans and skeptics about equally.

A Savings Account That Bets Against Your Cigarettes

The flagship test ran in the Philippines. In a randomized trial with 2,000 smokers, economists Xavier Giné, Dean Karlan, and Jonathan Zinman offered a contract called CARES: open a savings account, deposit money into it for six months, then take a urine test for nicotine. Pass, and every peso comes back. Fail, and the balance is gone. Karlan is affiliated with stickK, where anyone can sign a contract like this today.

The average signer committed 550 pesos, about US$11, over six months: roughly 20 percent of a month's income for these smokers, and about six months of their baseline cigarette spending.

It worked, in the economist's precise sense. Smokers offered the contract were 3 percentage points more likely to pass a nicotine urine test at six months than the control group, and the effect persisted in surprise tests at twelve months, long after the account had closed. Three points reads as modest until you set it against control-group quit rates of roughly 10 to 18 percent; the twelve-month numbers work out to a 32 to 35 percent relative increase.

Then the part the success story leaves out.

Why Most People Refuse a Commitment Device

A second number in that trial matters more than the first. Only 11% of smokers offered the contract signed up: 83 people out of 781. The device worked, and nearly nine in ten of the people it was built for said no.

That refusal is the signature finding of the whole genre, and the sharpest version comes from a 2015 trial in the New England Journal of Medicine. Scott Halpern and colleagues randomized 2,538 smokers to usual care or one of four financial-incentive programs. Some programs paid pure rewards for staying smoke-free. Others required participants to put down a refundable $150 deposit of their own money, lost on failure.

The results read in two halves, and only together. Counting everyone assigned, rewards won: 15.7% versus 10.2% sustained abstinence at six months, largely because so few ever accepted the deposit requirement. But the same paper estimated outcomes among the 13.7% of smokers who would have accepted either program, and inside that group, deposit-based programs produced a 13.2-percentage-point higher six-month abstinence rate than reward-based ones. For the rare person willing to take the bet, risking their own money beat collecting someone else's.

When the Payments Stop

Rewards carry a quieter flaw, and it arrives on a schedule. In a field experiment with 1,000 employees at a Fortune-500 company, Heather Royer, Mark Stehr, and Justin Sydnor paid workers $10 per gym visit for a month. Attendance roughly doubled. Then the payments ended, and within about two months the incentive-only group had drifted back to the control group's level. The behavior was rented, and it went back when the lease ran out.

The second arm is the interesting one. After the paid month, some workers were offered a self-funded commitment contract: stake your own money on continuing to show up, lose it to charity if you stop. The familiar minority appeared. Only 12% took the contract, staking $58 on average; 63% of them kept their commitments, and 37% watched their money go to charity. But the group offered the contract was still attending about 25% more than the control group over the rest of the year.

One month of bribery bought one month of gym. A single signature, taken up by a few and lost by over a third of them, was the only intervention in the study still working years later.

Lowering the Stakes Until People Say Yes

If the deal-breaker is handing over cash, the stake can be softened. Janet Schwartz, Dan Ariely, and colleagues tested a gentler version with grocery shoppers in a large discount program. Households already earning a 25% discount on healthy food were offered a precommitment: pledge to hold your healthy purchases 5 percentage points above baseline every month; miss a month and you forfeit that month's discount. No cash leaves your pocket. A benefit stops arriving.

The acceptance problem eased immediately: 36% of households took the deal, well over double the deposit trial's 13.7%. And committers moved, averaging a 3.5-percentage-point increase in healthy purchases in each of the six months, while households that declined, and control households offered only a hypothetical commitment, showed no such increase.

Two footnotes belong here. The 3.5-point gain fell short of the 5-point target committers had pledged, so even volunteers under contract undershot their own bar. And the gain belongs to the self-selected 36% who opted in; the study cannot say what it would have done for those who refused. The trade still sits in plain sight: soften the stake, and more people let it bind them.

The Stake That Doesn't Cost a Dollar

Follow that trade all the way down and you reach a 2014 JAMA viewpoint. Todd Rogers, Katherine Milkman, and Kevin Volpp, the last of whom was also an author on the deposit trial, argued that commitment devices can only help the people who use them, and that uptake in research runs as low as a few percentage points. Their proposal was to widen what counts as a stake: penalties beyond losing money, such as disappointing a workout partner, would likely make commitment devices more appealing, and existing social networks may be a low-cost way to sustain healthier behavior.

It is an argument, not a trial, and should be graded as one. But notice how well it fits the data. The bottleneck in every money-stake study was never whether stakes bind; it was whether people will accept them. A promise to a person costs nothing to sign and something real to break. The moment a friend expects you at the gym at 7am, skipping acquires a price no bank holds. We have traced that force in what an accountability partner does and why sharing goals works. Future You is built around exactly that stake, people in your corner who notice when you go quiet, free on iOS and Android.

How to Choose Your Stake

Set the studies side by side and a decision guide falls out.

Answer the screening question first. Does imagining $150 of your own money on the line make you plan harder, or close the tab? Most people close the tab, and that is fine. But if the thought genuinely focuses you, you belong to the minority these contracts serve, where the deposit beat the reward.

Expect to lose sometimes, and price it in. Most CARES signers forfeited their deposits. So did 37% of the gym committers. A stake you might lose is the entire mechanism; a forfeited one is the invoice for the weeks it caught you. If losing the money would hurt worse than missing the goal, bet something else.

If cash repels you, walk down the ladder. A revocable perk, like Schwartz's discount, drew well over double the acceptance of a cash deposit. And a stake made of another person's expectations asks for no cash at all, which is why Rogers, Milkman, and Volpp expected it to reach the people money never will.

So the honest verdict on betting money on your goals: it works, for the minority who will take the bet, at real risk of losing the money, with staying power that plain rewards never showed. If that is you, make the bet. If not, skip the guilt; the research says the currency was wrong for you, nothing more.

Whatever stake you choose, choose it in calm water. That was the whole trick on Odysseus's ship: the deal was struck while the man making it could still think. Pick your stake this week, while quitting is still theoretical. By the time you want out, wanting out should already be the expensive option.