Home › What the research says

Present bias and hyperbolic discounting: what the research actually shows

The model, in plain terms

Standard economic models discount the future at a constant rate, which implies your preferences stay consistent over time. David Laibson's 1997 Quarterly Journal of Economics paper works instead with hyperbolic discount functions, which he notes induce dynamically inconsistent preferences and therefore give consumers a motive to constrain their own future choices. O'Donoghue and Rabin, in the American Economic Review in 1999, use the term present-biased preferences for the same idea and formalize it as beta-delta preferences: delta is ordinary long-run discounting, and beta is an extra discount applied to everything that is not right now. If beta equals one, you are the textbook agent. If beta is less than one, you are most people.

Two distinctions that matter

O'Donoghue and Rabin emphasize two splits. First, whether an activity has immediate costs and delayed rewards, which is the shape of saving or exercising, or immediate rewards and delayed costs, which is the shape of most spending. Second, whether the person is sophisticated, meaning they correctly anticipate their own future self-control problems, or naive, meaning they expect their future selves to behave. Their result is that naive people procrastinate on immediate-cost activities, while sophistication mitigates procrastination but makes the immediate-reward problem worse. They also report that with immediate costs a small present bias can severely harm only naive people, while with immediate rewards it can severely harm only sophisticated ones. The practical translation is that knowing you have a self-control problem is not automatically protective. It changes which mistakes you make.

The planner and the doer

Thaler and Shefrin, writing in the Journal of Political Economy in 1981, framed the same conflict differently. They modeled the individual as an organization containing both a farsighted planner and a myopic doer, and argued that the resulting conflict is fundamentally similar to the agency conflict between a firm's owners and its managers. Their point is that individuals and firms reach for the same class of solutions: rules, constraints and pre-commitment, rather than better intentions. This is a theoretical framework, not a measurement, and it is worth being clear about which of these papers are models and which are data.

The gym study, and what it actually found

DellaVigna and Malmendier, in the American Economic Review in 2006, analyzed contract choice and day-to-day attendance for 7,752 members of three US health clubs over three years. Members who chose a flat monthly fee of over $70 attended an average of 4.3 times a month, paying a price per expected visit of more than $17 when a 10-visit pass was available at $10 per visit. On average those users forwent $600 of savings over the course of their membership. The authors also found monthly members were 17 percent more likely to stay enrolled past a year than members who committed for a year.

The caveat the authors put on their own result

DellaVigna and Malmendier do not present overpayment as pure self-control failure. They write that the leading explanations for their findings are overconfidence about future self-control or about future efficiency: overconfident members overestimate how often they will attend, and also overestimate the probability that they will cancel an automatically renewing contract. That distinction matters. Beliefs about your future behavior are doing a large share of the work, not only your behavior itself. It also means the fix is partly informational: knowing your real attendance rate, or your real spending rate, changes the contract you would choose.

What this means for envelopes

If beta is less than one, the useful move is to shift decisions away from the moment of temptation, because that is the moment where the extra discount applies. Filling envelopes on payday is exactly that shift. The planner allocates $180 to dining out on a Friday morning, calmly, with the whole month visible. The doer standing in a store on the twenty-second does not get to re-run that decision from scratch; they get to read a balance. None of the cited research tested a budgeting app, and none of it shows that envelopes reduce annual spending. What it supports is narrower and still useful: deciding earlier is structurally different from deciding better.

Sources

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

• Golden Eggs and Hyperbolic Discounting — Quarterly Journal of Economics, 1997
  States that hyperbolic discount functions induce dynamically inconsistent preferences, implying a motive for consumers to constrain their own future choices, and analyzes a consumer using an illiquid asset as an imperfect commitment technology. Volume 112, issue 2, pages 443-478.
• Doing It Now or Later — American Economic Review, 1999
  Examines self-control problems modeled as time-inconsistent, present-biased preferences, distinguishing immediate-cost from immediate-reward activities and sophisticated from naive agents. Reports that naive people procrastinate on immediate-cost activities while sophistication mitigates procrastination but worsens the immediate-reward case.
• An Economic Theory of Self-Control — Journal of Political Economy, 1981
  Models the individual as an organization that is simultaneously a farsighted planner and a myopic doer, and argues the resulting conflict is fundamentally similar to the agency conflict between the owners and managers of a firm. Volume 89, number 2.
• Paying Not to Go to the Gym — American Economic Review, 2006
  Using contract and attendance data on 7,752 members of three US health clubs over three years, finds members on flat monthly fees over $70 attended 4.3 times per month and paid over $17 per expected visit versus $10 with a 10-visit pass, forgoing about $600 over their membership. The authors name overconfidence about future self-control or efficiency as the leading explanations.

Common questions

Is present bias the same as being impatient?

No. Ordinary impatience is captured by delta, a constant discount that applies evenly across time, and it does not cause preference reversals. Present bias is the extra weight, beta, applied to right now specifically. The signature is inconsistency: you prefer the larger reward in twelve months over the smaller one in eleven months, but flip when the smaller one becomes available today. A very impatient person with beta equal to one saves little but never contradicts themselves.

Does the gym study prove people are bad at self-control?

Not on its own. The authors say the leading explanations are overconfidence about future self-control or future efficiency, which is a claim about beliefs as much as willpower. Members overestimated how often they would go and how likely they were to cancel an auto-renewing contract. That is a forecasting error, and it means better information about your own past behavior could change your choice without any change in willpower at all.

Do envelopes fix present bias?

No study cited here tested that, and it would be dishonest to claim it. What the research supports is the general mechanism: present bias applies at the moment of consumption, so decisions moved earlier are made under different conditions. Whether a digital envelope on a phone recreates that effect strongly enough to change annual spending is an open empirical question. Treat it as a plausible design choice with a mechanism behind it, not a proven intervention.

Related