10.5 Intertemporal choice and hyperbolic discounting

Intertemporal Choice in Economics

Intertemporal choice is about decisions where costs and benefits land at different points in time. Nearly every major financial decision fits this framework: saving for retirement, taking on student debt, or choosing between a job that pays well now versus one with better long-term prospects. Two concepts drive the analysis: time preferences (how much you value the present relative to the future) and discount rates (the rate at which you mathematically shrink future values to compare them with present values).

Key Components of Intertemporal Choice

• Present value calculates what a future sum of money is worth right now, given some rate of return. If someone offers you $110 a year from now and the discount rate is 10%, the present value is $PV = \frac{110}{1.10} = 100$.
• Future value works in the other direction: what a current amount grows to over time at an assumed rate.
• Opportunity cost captures what you give up by choosing one option. Spending $1,000 today means forgoing whatever that $1,000 could have earned if invested.
• Risk and uncertainty matter because future outcomes are never guaranteed, so riskier future payoffs get discounted more heavily.
• Time horizon shapes how much weight you place on short-term versus long-term consequences. A 25-year-old saving for retirement has a very different calculus than a 60-year-old.

In standard rational choice theory, a person maximizes discounted utility over time, weighing each period's utility by a discount factor. That's where the distinction between exponential and hyperbolic discounting becomes central.

Hyperbolic vs. Exponential Discounting

Exponential Discounting (The Standard Model)

Exponential discounting is the benchmark in traditional economics. It assumes a constant discount rate across all time periods. The discount function is:

$D(t) = \delta^t$

where $\delta$ is the per-period discount factor (between 0 and 1) and $t$ is the number of periods into the future. The discounted utility of a stream of payoffs is then $U = \sum_{t=0}^{T} \delta^t u(c_t)$, where $u(c_t)$ is the utility from consumption in period $t$.

The crucial property here is time consistency: your ranking of two options stays the same regardless of when you evaluate them. If you prefer $110 in 31 days over $100 in 30 days today, you'll still prefer it when day 30 actually arrives. Your preferences don't flip. Formally, this holds because the ratio of discount factors between any two adjacent periods is always $\delta$, no matter how far away those periods are.

Hyperbolic Discounting (The Behavioral Model)

Hyperbolic discounting captures something exponential discounting misses: people tend to be disproportionately impatient about the near future compared to the distant future. The discount rate isn't constant; it's high for short delays and drops off for longer delays. A common functional form is:

$D(t) = \frac{1}{1 + kt}$

where $k > 0$ controls how steeply the person discounts.

The discount function drops sharply at first, then flattens out. This shape produces preference reversals. Here's how that works:

1. You're asked today: would you rather have $100 in 30 days or $110 in 31 days? Most people pick $110. The one-day wait barely matters when both options are far away.
2. Now day 30 arrives. The choice becomes: $100 right now or $110 tomorrow. Suddenly many people grab the $100. The immediacy of the reward changes everything.

Under exponential discounting, this reversal can't happen because the discount rate between day 30 and day 31 is the same $\delta$ whether you evaluate it from day 0 or day 30. Under hyperbolic discounting, the discount rate between "right now" and "one day from now" is much steeper than the rate between "30 days from now" and "31 days from now." That asymmetry is what drives the reversal.

A widely used alternative that's more tractable in formal models is the quasi-hyperbolic (or $\beta$-$\delta$) model introduced by Laibson (1997):

$D(0) = 1, \quad D(t) = \beta \delta^t \text{ for } t \geq 1$

Here $\beta \in (0,1)$ captures the extra impatience for any delay versus the present, while $\delta$ handles the standard period-by-period discounting after that. When $\beta = 1$, you're back to the standard exponential model. This two-parameter setup generates preference reversals like the full hyperbolic model but is much easier to work with mathematically.

Side-by-Side Comparison

Empirical studies consistently find that hyperbolic and quasi-hyperbolic discounting fit actual human behavior better than the exponential model, which is why these models are a cornerstone of behavioral economics.

Implications of Hyperbolic Discounting

Financial Decision Making

Hyperbolic discounting explains a wide range of financial mistakes that the standard model struggles with:

• Undersaving for retirement. Future comfort feels abstract and distant, so people consistently overconsume in the present. Someone might genuinely plan to start saving "next month" but never does, because when next month arrives, the immediate cost of saving feels too steep. In the $\beta$-$\delta$ framework, the person's long-run self (reflecting $\delta$) wants to save, but the present-biased self (reflecting $\beta < 1$) keeps postponing.
• Difficulty sticking to financial plans. You set a budget on Sunday, then overspend on Tuesday. Your Sunday self and your Tuesday self effectively have different discount rates for the immediate moment.
• Preference for smaller-sooner rewards. Choosing a $500 bonus now over a $700 bonus in six months, even when the implied return on waiting far exceeds any available investment.

This is also why commitment devices work. Automatic payroll deductions into a 401(k), penalties for early withdrawal, or apps that lock your savings are all ways to bind your future self to the choices your current self knows are better. The person using these tools is essentially acknowledging their own time inconsistency.

A related distinction worth knowing: a sophisticated hyperbolic discounter is aware of their future self-control problems and seeks commitment devices. A naive hyperbolic discounter incorrectly believes their future self will follow through on current plans. Sophisticates do better on average, but both types deviate from the exponential benchmark.

Behavioral Consequences Beyond Finance

The same logic extends well past money:

• Procrastination is a textbook case. Studying has future benefits (good grade) but immediate costs (effort, boredom). Hyperbolic discounting makes the immediate cost loom large, so you delay.
• Addiction and health behaviors involve immediate gratification (the cigarette now) at the expense of future well-being (long-term health). The steep near-term discounting makes the immediate pleasure dominate.
• Cyclical self-control failures like yo-yo dieting follow the same pattern: you commit to a plan when the temptation is distant, then abandon it when the temptation is right in front of you.
• Career choices can be distorted too. A person might take a higher-paying but dead-end job over one with lower starting pay but a much better long-term trajectory.

Policy Implications of Intertemporal Choice

Retirement and Financial Planning

Because people predictably undersave, policy can work with hyperbolic discounting rather than against it:

• Automatic enrollment in retirement plans (with an opt-out rather than opt-in) dramatically increases participation. The default does the work that willpower can't.
• Escalation commitments like "Save More Tomorrow" programs let people commit now to increase their savings rate with future raises, sidestepping the pain of reducing current consumption.
• Matching contributions and tax incentives make the immediate reward of saving more visible, partially counteracting the pull of present consumption.
• Cooling-off periods for major purchases give the impulsive near-term self time to reconsider.

These aren't just theoretical suggestions. Research by Thaler and Benartzi on the Save More Tomorrow program showed that participation rates and savings levels increased substantially when the program design accounted for hyperbolic discounting. In one implementation, average savings rates rose from 3.5% to 13.6% over several years.

Environmental and Public Health Policies

The same discounting bias affects collective decisions:

• Climate change policy faces a massive discounting problem. The costs of action are immediate (higher energy prices, regulation), while the benefits are decades away. Framing policies around short-term co-benefits (cleaner air now, lower energy bills) can help overcome this.
• Immediate incentives for green behavior, like tax rebates for electric vehicles or energy-efficient appliances, work better than appeals to long-term environmental outcomes precisely because of hyperbolic discounting.
• Public health campaigns are more effective when they emphasize near-term benefits. Telling someone "exercise gives you more energy this week" tends to motivate better than "exercise reduces your risk of heart disease in 30 years."

The core insight across all these domains is the same: if people systematically overweight the present, effective policy meets them where they are rather than assuming they'll behave like perfectly patient, time-consistent optimizers.