Quasi-hyperbolic discounting is a model of intertemporal choice in which an individual gives immediate outcomes a special weight relative to all future outcomes. It combines an extra discount on future utility with a constant discount factor between successive future periods. Widely used in behavioral economics, it provides a simple representation of present bias and explains how someone can make a plan, then prefer to abandon it when the time to act arrives.
Mathematical formulation
For a sequence of consumption outcomes, a standard finite-horizon formulation is
where:
- is the value of the sequence as evaluated at time ;
- is consumption periods ahead;
- is the period utility function;
- , usually satisfying , is the long-run discount factor;
- , usually satisfying , measures the additional discount applied to everything outside the present.
The corresponding discount function is
An infinite-horizon version is also common, provided the discounted utility sum converges.
The parameters have distinct roles. A smaller means less weight on later utility throughout the future. A smaller means a stronger preference for immediacy. Thus, an individual can be relatively patient when comparing two distant outcomes yet strongly prefer an immediate outcome to a slightly delayed one.
When , the model reduces to exponential discounting. When , there is an additional present–future wedge:
The distinctive feature is therefore not that the discount factor continually changes with delay, but that the first step away from the present receives an extra discount.
Preference reversals and dynamic inconsistency
Quasi-hyperbolic discounting can produce dynamic inconsistency: a ranking of alternatives changes simply because time passes, even if the outcomes, available information, and underlying parameters remain unchanged.
Suppose utility is linear in money, , and . An individual chooses between receiving 100 units of money one period from now and 110 units two periods from now. The current valuations are
The individual prefers the later payment. One period later, however, the same alternatives become 100 immediately or 110 one period ahead:
The preference reverses toward the smaller, immediate payment.
The example isolates the model’s mechanism rather than describing all monetary choices. Actual choices can also depend on liquidity, payment reliability, borrowing opportunities, and the timing of consumption.
Present bias does not mean that every choice must reverse. If one option is sufficiently attractive, it may be preferred both before and after it becomes immediate. The model identifies conditions under which reversals are possible.
Relationship to hyperbolic discounting
Quasi-hyperbolic discounting is related to, but mathematically distinct from, hyperbolic discounting. A simple hyperbolic discount function is
Under this function, the proportional discount rate declines continuously with delay. Under the quasi-hyperbolic model, all strictly future periods are discounted exponentially relative to one another; the departure from exponential discounting is concentrated at the boundary between the present and the future.
The term “quasi-hyperbolic” reflects the model’s ability to capture a prominent implication of hyperbolic discounting—greater impatience over near-term trade-offs—without reproducing its full discount curve. The beta–delta specification is often easier to incorporate into models of saving, borrowing, and repeated decision-making.
Neither specification should be confused with ordinary financial discounting used to calculate present value. A market interest rate describes an intertemporal exchange opportunity; and describe preferences. Market opportunities and preferences jointly shape behavior.
Historical development
An important early antecedent is the discounting structure used by Edmund Phelps and Robert Pollak in their 1968 analysis of intergenerational saving. Their setting concerned how successive generations valued their own consumption relative to that of later generations, rather than an individual’s short-run self-control.
David Laibson’s 1997 paper “Golden Eggs and Hyperbolic Discounting” helped establish the beta–delta formulation in modern behavioral economics. It used present-biased preferences to examine saving and the role of assets that restrict immediate access to wealth.
Subsequent research developed the implications of different beliefs about future behavior, distinguishing individuals who anticipate their future preference reversals from those who do not. The formulation became a standard analytical tool because it separates long-run patience from the special appeal of immediate outcomes using only two discount parameters.
Sophistication, naïveté, and commitment
Preferences alone do not fully determine behavior when current and future choices may conflict. Predictions also depend on what an individual believes about future decisions.
Sophisticated individuals correctly anticipate their future present bias. They recognize that a plan preferred today may be abandoned tomorrow and take that prospect into account. Dynamic models commonly represent this as a strategic interaction among successive “selves,” each controlling decisions at a different date.
Naïve individuals fail to anticipate future preference reversals. In a standard formulation, they expect their future selves to follow the plan currently preferred. This can generate repeated postponement: a task is always scheduled for later, but the immediate cost becomes unattractive whenever the scheduled date arrives.
Partially naïve individuals anticipate some future present bias but underestimate its strength.
These distinctions matter for the demand for a commitment device—an arrangement that restricts future choices or makes deviations costly. Examples include withdrawal restrictions, deadlines with penalties, and contractual limits on spending. A sophisticated present-biased individual may value such arrangements because they alter future incentives. Commitment is not universally desirable, however: restrictions can be costly, and preserving flexibility may be valuable when future needs are uncertain.
Economic applications
Saving and borrowing
The model helps explain why an individual may plan to save but spend available resources when the decision becomes immediate. It is used to study saving, retirement wealth accumulation, and the coexistence of expensive debt with assets that are difficult to access.
Such patterns do not uniquely establish present bias. Differences in returns, transaction costs, uncertainty, and borrowing constraints can also explain them.
Effort and procrastination
Tasks with an immediate effort cost and delayed benefits are natural applications. A present-biased individual may prefer to complete a task early when evaluating it in advance, then postpone it once effort must be supplied immediately.
The reverse timing pattern also matters: activities with immediate rewards and delayed costs can become more attractive at the moment of choice than they appeared in advance. The model therefore concerns the timing of costs and benefits, not a general preference for inactivity.
Contracts and policy design
Quasi-hyperbolic preferences have been incorporated into contract theory and the analysis of consumer markets. Firms may design contracts around consumers’ beliefs about future usage, repayment, cancellation, or effort. Predictions depend substantially on whether consumers correctly anticipate their future behavior.
The model is also used in research on defaults, deadlines, and nudges. These applications require distinguishing a change in behavior from an improvement in welfare: an intervention can alter choices without establishing which preferences should govern evaluation.
Evidence and identification
Experiments often compare a choice involving an immediate outcome with an otherwise similar choice in which both outcomes are delayed. Greater impatience in the first comparison is consistent with present bias. Researchers also study preference reversals, demand for commitment, and behavior across repeated decisions.
Identifying and nevertheless requires assumptions. Important complications include:
- Payment credibility: immediate and delayed payments may differ in perceived reliability.
- Financial constraints: urgent cash needs can make immediate money especially valuable without implying present-biased preferences.
- Utility curvature: monetary choices reflect both discounting and marginal utility.
- Consumption timing: receiving money does not necessarily mean consuming it immediately.
- Uncertainty and changing circumstances: later choices may rationally differ because information or needs have changed.
The meaning of a “period” is another limitation. An estimated depends on how the present and future are operationalized. A discrete-period parameter cannot automatically be transferred unchanged between daily, monthly, and annual models.
Welfare and conceptual limitations
Dynamic inconsistency complicates welfare economics because different temporal perspectives may favor different outcomes. The plan preferred before a temptation arises can conflict with the choice preferred when it becomes immediate. Selecting one perspective as authoritative requires an additional normative criterion; it does not follow from the beta–delta equation itself.
The specification also compresses diverse behavior into a single present-bias parameter. Attention, uncertainty, changing tastes, and mistaken beliefs can produce behavior resembling present bias without sharing its mechanism. Its usefulness lies in providing a tractable account of a particular pattern of intertemporal choice, rather than a complete explanation of impatience or self-control.