Hyperbolic Discounting Isn’t About Willpower: Why Gym Members Quit Long-Term Goals

A recent study out of Harvard has found that willpower may not be the reason people quit on long-term goals. The problem is how they perceive the future. In fact, the research highlights the link between hyperbolic discounting and gym members who often struggle to stay committed over time.

Benjamin Enke asked a few hundred people to put a price on $50 they’d receive in twelve months. Standard stuff, the kind of intertemporal choice task that behavioural economists at Harvard have been running for decades. People wanted less money now rather than wait for more later, discounting future payments along the familiar hyperbolic curve: steep for near-term rewards, flatter for distant ones. Fifty years of research had a tidy explanation for why: people are impatient, self-control is limited, and the pull of right now always wins.

Then Enke and his colleagues Thomas Graeber and Ryan Oprea broke the explanation. They built what they called “atemporal mirrors,” problems with identical compound maths but no actual waiting. Value $50 that has been shrunk twelve times, each time by four per cent, paid in full right now. Nobody waited for anything, and willpower had nothing to do with the task. The participants discounted the money in exactly the same hyperbolic pattern.

The implication landed quickly outside economics. If discounting isn’t about impatience, the entire fitness industry’s answer to member dropout, motivate them harder, is aimed at the wrong problem.

What “Atemporal Mirrors” Revealed

The paper appeared in the Journal of the European Economic Association in October 2025. For every standard time-delay question, the authors built an “atemporal mirror” with the same compound arithmetic and no waiting. The logic was simple: if impatience causes the discounting, remove the delay and the hyperbolic pattern should vanish. It didn’t.

Participants undervalued rewards that required many compounding steps regardless of whether those steps represented months of waiting or rounds of multiplication applied to money they were about to receive. Simpler calculations got higher valuations and complex ones got lower ones. The curve was the same one researchers had spent fifty years calling a failure of self-control, and it showed up in a task where self-control was irrelevant.

Why the Complexity Explanation Holds

The evidence against the willpower explanation came from three directions.

People who discounted most heavily in the time-delay tasks were also the most inconsistent across every other task, including problems that involved no delay at all. Impatience would show up selectively in tasks where waiting was required, but the inconsistency was general. It looked like noise from overloaded computation, not a preference for now.

When the researchers made the atemporal calculations harder to evaluate mentally, the hyperbolic pattern got steeper. Nothing about the time delay changed. The sums just got more difficult, and people discounted more.

The people who discounted the future most steeply also reported the least confidence in their own answers, what the researchers called “cognitive uncertainty.” They weren’t choosing impatiently. What looked like impatience was confusion, because the calculation was too hard to do properly.

The Non-Monetary Penalty

Enke’s team studied money, where the units are at least familiar. A separate meta-analysis tells a worse story for anything that isn’t cash. Stephen Cheung, Agnieszka Tymula, and Xueting Wang aggregated 86 studies of quasi-hyperbolic discounting and found that the present-bias parameter for monetary rewards is 0.938 after correcting for selective reporting. For non-monetary rewards, that same parameter falls to 0.750.

In practical terms, a β of 0.938 means someone discounts money by about six per cent the moment a delay appears. A β of 0.750 means the same person, facing the same delay, discounts health outcomes, food, and real effort by twenty-five per cent. Four times the penalty, for the crime of not being denominated in dollars.

Fitness sits on the worst possible side of that divide. The reward for twelve months of consistent training is a shift in body composition, energy, joint health, and daily capacity that nobody can quantify in advance, predict with any confidence, or weigh against the concrete cost of dragging themselves to the gym at 6am on a Tuesday in February.

What This Means in a Gym

If the problem were motivation, the solution would look like most gyms already look: inspirational quotes on the walls, transformation photos in the lobby, a Spotify playlist engineered to make people feel something at 6am, and coaches trained to deliver pep talks when attendance drops. The complexity research says every one of those interventions is aimed at the wrong target.

A member standing at the front desk with a twelve-month contract in front of them is not being asked to resist temptation. They are being asked to run a calculation: what is the compounding value of something I can’t measure, arriving on a schedule I can’t predict, in units I can’t compare to the money and hours on this price sheet? That calculation would be difficult for an economist with a spreadsheet. The research says that when it gets hard enough, people just default to steep discounting, and no amount of motivation changes that.

Reducing the computational complexity of the decision does more than any motivational intervention. A twelve-month contract forces a member to evaluate twelve compounding steps of uncertain value. A four-week trial gives them one step. The trial doesn’t build discipline; it shrinks the calculation, and smaller calculations produce less discounting.

Progress tracking does something similar. When a member can see that they squatted 60kg four weeks ago and 72.5kg today, the fuzzy promise of “fitness in twelve months” becomes a concrete number that moved in four weeks. You don’t need to compute the value of a twelve-month outcome if a four-week result is already sitting on a whiteboard in kilograms.

Shorter feedback loops, concrete metrics, and tighter decision windows work because they change the structure of the problem, not because they motivate anyone. The brain handles compound value badly in general, and health outcomes make it worse than almost anything else.

The Question

If your members aren’t quitting because they lack willpower but because the maths of a long-term payoff is too complex to evaluate, what would your gym look like if every decision you asked them to make was one step instead of twelve?

People Also Ask

Why do gym members struggle with long-term fitness goals?

Research by Enke, Graeber, and Oprea (2025) suggests gym members struggle with long-term goals not because of weak willpower, but because the human brain handles multi-step compound calculations poorly. When a fitness outcome requires mentally evaluating months of uncertain, non-monetary payoffs, the computational difficulty causes people to heavily discount the future reward, making the immediate costs (time, effort, money) feel disproportionately large.

What is hyperbolic discounting in fitness?

Hyperbolic discounting in fitness is the tendency for gym members to heavily prefer immediate comfort over long-term health gains, with the preference being strongest for near-future decisions. A member who commits to training consistently “starting next month” but cancels tomorrow’s session is showing a classic hyperbolic pattern. Enke et al.’s research shows this pattern is driven by the difficulty of evaluating compound future value, not by laziness.

How can gym owners help members stick to fitness goals?

Rather than relying on motivation, gym owners can reduce the computational complexity of fitness decisions. This means shorter commitment windows (four weeks instead of twelve months), concrete and measurable progress metrics (weight lifted, distance run) rather than vague outcomes (“get fit”), and frequent feedback that converts uncertain future value into observable present value. The research suggests that simpler decisions produce less steep discounting and better long-term adherence.

The reason your members undervalue what you’re selling isn’t weakness. It’s that the calculation you’re asking them to do, estimating the compound return on something that isn’t money over a timeline they can’t see, is one of the hardest things a consumer brain gets asked to process.

References

1. Enke, B., Graeber, T., & Oprea, R. (2025). Complexity and Time. Journal of the European Economic Association, 23(5), 1838-1878. https://doi.org/10.1093/jeea/jvae055

2. Cheung, S. L., Tymula, A., & Wang, X. (2025). A meta-analysis of quasi-hyperbolic discounting. SSRN Working Paper. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5222564

If this changes how you think about your members’ commitment, we should talk.