Skip to main content

Hyperbolic Discounting

TL;DR

Hyperbolic Discounting: People prefer $100 today over $110 tomorrow, but prefer $110 in 31 days over $100 in 30 days β€” even though both involve exactly the same one-day wait for $10. We discount the immediate future far more steeply than the distant future. This explains why we fail to save, exercise, or work on long-term projects even though we clearly want those outcomes.


What Is Hyperbolic Discounting?​

Rational economic theory assumes "exponential discounting" β€” the value of a future reward declines at a constant rate per unit of time. If you'd trade $100 now for $110 in a year, you should also trade $100 in 10 years for $110 in 11 years. The discount rate is constant.

Hyperbolic discounting, documented extensively by George Ainslie and others, shows that humans don't discount at a constant rate. Instead, we discount steeply between "now" and "soon," but relatively gently between "soon" and "later." The resulting hyperbolic discount curve creates preference reversals: a person who prefers $100 now over $110 next week will, once next week arrives, suddenly prefer $100 (now) over $110 (next week) β€” even though they previously said they'd prefer the $110.

This time-inconsistency is the source of self-control problems: we make plans for the future that our future selves don't follow through on.


How It Works​

Exponential discounting (rational): constant rate per time period
β€” $100 now = $110 in 1 year = $121 in 2 years (10% discount rate)

Hyperbolic discounting (human): steep near-term, shallow far-term
β€” $100 now > $110 tomorrow (steep near-term discount)
β€” $110 in 31 days > $100 in 30 days (shallow: same 1-day gap)
β€” This creates preference reversals as time passes

The present bias: people give disproportionate weight to "now"
vs. any future time, even the immediate future

Consequences:
β€” Procrastination: starting tomorrow costs little now but arrives never
β€” Under-saving: future self's retirement seems less important than
today's spending
β€” Health behaviour failures: diet starts Monday, exercise later
β€” Commitment devices: people seek external constraints on future selves
(automatic savings, gym contracts, deadlines)

Three Real-World Examples​

Retirement Savings Behaviour​

Despite knowing they need to save for retirement, most Americans save less than actuarially optimal amounts. A study by Thaler and Benartzi found that automatically enrolling employees in retirement savings plans (with opt-out) dramatically increased savings rates versus requiring active opt-in β€” even when the employer match was identical. The default removes the "decision now vs. saving later" trade-off that hyperbolic discounting wins. Opt-in requires an active present-biased decision; opt-out doesn't.

The "I'll Start Monday" Pattern​

Research on habit formation and dieting finds a strong "temporal landmark" effect: people are dramatically more likely to start a new healthy behaviour at the beginning of a week, month, or year than on an arbitrary date. This is hyperbolic discounting in action: "Monday" is in the future (cheap now), and the decision to start "tomorrow" is made repeatedly because when tomorrow arrives, it becomes "now" β€” and the discount rate on starting immediately vs. Tuesday spikes again.

Credit Card Debt and Payday Loans​

The interest rates on payday loans (400%+ APR) and credit card minimum-payment traps are economically rational only under hyperbolic discounting. A person who borrows $300 at 400% APR to get cash today, despite knowing they'll pay back $600, is exhibiting massive discounting of the future payment relative to the present need. Studies find that payday loan borrowers consistently underestimate how often they'll roll over loans β€” because their future self's situation seems distant and manageable, while the present need is urgent.


When to Recognise It​

🚨 Hyperbolic Discounting is likely operating when:

  • You repeatedly delay the same task to "tomorrow" or "Monday"
  • You choose immediate small rewards over larger delayed ones that you'd advise others to take
  • Your saving, exercise, or health behaviours are consistently below your stated goals
  • You make promises about future behaviour that you then break when the future arrives

βœ… Countermeasures:

  • Pre-commitment devices: automate savings, sign up for gym contracts, use deadline tools
  • Implementation intentions: specify when and how you'll do the future task, reducing the decision cost when the time arrives
  • Temporal reframing: "saving $10/day" vs "$3,650/year" β€” daily framing emphasises present cost less
  • Default design: make the desired future behaviour the default, requiring active choice to not do it
Pairs well withWhy
Pre-commitmentPre-commitment is the direct antidote to hyperbolic discounting
Status Quo BiasBoth lead to inaction; defaults exploit both biases constructively
Loss AversionLoss aversion amplifies present bias β€” losses now feel more urgent than future gains

Common Misuses and Limitations​

Conflating with simple impatience. Hyperbolic discounting specifically predicts preference reversals β€” not just that people prefer sooner to later, but that the preference reverses over time. Simple impatience (consistent preference for sooner) is not the fallacy. The distinguishing feature is time-inconsistency.

Assuming it's always irrational. Preferring present consumption over future consumption isn't always wrong. If the future is genuinely uncertain, discounting it is rational. The problem is specifically when the discount rate for the immediate future is inconsistent with the discount rate for more distant futures, creating preference reversals that undermine long-term plans.


ModelRelationship
Pre-commitmentPre-commitment devices are the direct antidote
Status Quo BiasBoth produce inaction; both can be addressed by smart defaults
Loss AversionLoss aversion amplifies present-biased decisions

Frequently Asked Questions​

Is hyperbolic discounting the same as procrastination?

Procrastination is one manifestation of hyperbolic discounting β€” the cost of starting a task now is discounted less steeply than the perceived cost of starting "soon," creating repeated deferral. But hyperbolic discounting is broader: it also explains under-saving, health behaviour failures, and any domain where present preferences consistently override future plans. Procrastination is the task-initiation specific application.

What are the most effective commitment devices?

Research by Ariely, Laibson, and others identifies several effective commitment devices: (1) automatic savings with pre-set contribution rates β€” removes the decision entirely; (2) the "Save More Tomorrow" programme β€” commits to saving future raises, avoiding present-tense sacrifice; (3) deadlines with penalties β€” external costs for non-completion that activate loss aversion; (4) temptation bundling β€” combining immediately gratifying activities with valuable long-term ones (audiobooks only while exercising). The common feature: they create a constraint that the present self enforces on the future self before the future arrives.

Does hyperbolic discounting vary across cultures?

Yes. Studies find higher discount rates (more present-biased) in lower-income populations β€” arguably rational, given higher present uncertainty β€” and lower discount rates (more patient) in higher-income populations with more stable futures. Cultural variation exists but is smaller than the individual variation driven by income, certainty, and personal history with resource scarcity. Discount rates are also domain-specific: the same person may be patient about money and impatient about food.


Further Reading​

  • Ainslie, G. (1975). "Specious Reward: A Behavioral Theory of Impulsiveness and Impulse Control." Psychological Bulletin
  • Thaler, R. & Benartzi, S. (2004). "Save More Tomorrow." Journal of Political Economy
  • O'Donoghue, T. & Rabin, M. (1999). "Doing It Now or Later." American Economic Review

Apply with AI​

πŸš€ Design commitment devices for your goals with MindMax β†’


This page is part of the MindMax Mental Models Knowledge Base.