Commitment Device¶
Core Idea¶
A commitment device is the structural pattern in which an agent, anticipating that a later self (or a later state of the world) will be tempted to deviate from a currently preferred course, deliberately alters the future choice set now — removing options, raising the cost of defection, or delegating the choice away — so that the tempting action becomes impossible or unattractive. The essential commitment is strategic self-limitation against time-inconsistent preferences: voluntarily shrinking one's own future freedom to make the desired behavior the path of least resistance, a logic Strotz (1955) first formalized in his analysis of how a rational agent who foresees the inconsistency of his future preferences will "precommit" by constraining the actions available to his later self. [1] The pattern presupposes a divided agent — a present self who knows what it wants and a future self who, under the pull of temptation, immediate reward, or changed incentives, will want something different — and resolves the conflict in the present self's favor not by strengthening resolve but by foreclosing the option to renege, a move Schelling (1960) recognized as the strategic essence of binding oneself to gain credibility or self-control. [2]
How would you explain it like I'm…
Tying your future hands
Locking yourself in early
Binding your future choices
Structural Signature¶
A commitment device encodes a structural pattern: anticipated future temptation → present-tense constriction of the choice set → altered downstream incentives → desired behavior secured by infeasibility rather than will. It separates two temporal positions of a single agent (the foresighted present self and the tempted future self) and names the work the present self does to bind the future self, an asymmetry Elster (1979) analyzed under the figure of Ulysses ordering his crew to lash him to the mast and stop his ears so that he could hear the Sirens without steering toward them. [3]
Equivalent framings:
- Voluntary constraint imposed by a present self on a future self
- Removing or raising the cost of a tempting option in advance
- Strategic self-limitation against time-inconsistent preferences
- Tying one's own hands to gain credibility or self-control
- Making defection infeasible rather than merely undesirable
- Delegating a choice away so it cannot be reversed under temptation
- Paying a real cost now to foreclose a worse choice later
The structural insight is robust and counterintuitive: a dieter, a central bank, a recovering gambler, a general burning bridges behind his army, and a constitutional convention all execute the same move — they make the future cheaper to navigate by making it narrower. Lowering one's own degrees of freedom, ordinarily a loss, becomes a gain precisely because the foreclosed options are the ones a weaker future self would have chosen. Laibson (1997) showed that under hyperbolic discounting this is not irrational at all but the optimal response of a sophisticated agent who correctly predicts his own future impatience and prices illiquidity as a feature. [4]
What It Is Not¶
A commitment device is not willpower or self-control exercised in the moment. Willpower re-fights the temptation each time it arises, drawing on a depletable resource of resolve; a commitment device fights the battle once, in advance, by changing the situation so that the temptation either cannot be acted on or is no longer attractive. The whole point is to avoid relying on in-the-moment resolve, which the agent expects to fail. A person who throws away the cigarettes is not being more disciplined than one who keeps them and resists; they are being more strategic — they have removed the need for discipline. [5]
Nor is it a mere preference, intention, or plan. An intention to save more, a New Year's resolution, or a stated plan to exercise carries no structural force; the future self can simply revise it. A commitment device is distinguished precisely by the fact that it bites — it imposes a real cost, removes a real option, or delegates a real decision, so that reneging is no longer free. The signature is the binding mechanism, not the declaration of intent.
A commitment device also does not require that the binding be perfect or irreversible. Many real devices are merely costly to escape rather than impossible — a savings account with a withdrawal penalty, a deposit forfeited if a goal is missed, a public promise whose breach damages reputation. What matters structurally is that the cost of defection has been raised enough, in advance, to shift the future self's choice. The device works by re-pricing options, and the price need only be high enough to deter, not infinite.
Finally, a commitment device makes no claim that constraining the future self is always wise. The present self can misjudge — binding against a "temptation" that turns out to be valuable new information, or locking in a course that should have been abandoned. The prime names the mechanism (present-tense self-binding against anticipated future deviation), not a verdict on whether any particular act of self-binding improves welfare. A commitment device can trap an agent in a course that conditions have rendered obsolete just as surely as it can protect a goal worth protecting.
Broad Use¶
Behavioral economics: Locking savings in accounts with withdrawal penalties; pre-paying a gym membership; "commitment savings" products where depositors voluntarily forfeit liquidity until a self-set goal is reached; apps that automatically donate money to a disliked cause ("anti-charity") if a pledged goal is missed. Ashraf, Karlan, and Yin (2006) demonstrated in a Philippine field experiment that offering a voluntary commitment savings account — one that restricted withdrawals until a goal date or amount — raised savings balances by over 80 percent relative to controls, and that take-up was concentrated among exactly the women who exhibited hyperbolic, present-biased preferences. [6]
Game theory and strategy (non-obvious): "Burning the bridges," dismantling one's own line of retreat, or publicly and irrevocably committing so that an opponent knows backing down is impossible — credibility achieved by destroying one's own options. The deterrent value of a threat or the binding value of a promise depends on having removed the option to renege; a threat one can costlessly abandon is not believed. Schelling (1960) treated this as the paradox at the heart of bargaining: the power to constrain an adversary often derives from the power to constrain oneself first. [2]
Monetary and macroeconomic policy: Central-bank independence, rules-based inflation targets, currency boards, and statutory mandates bind future policymakers against the temptation to inflate for short-run output gains. Kydland and Prescott (1977) gave this its canonical formal treatment, showing that a discretionary policymaker who can always re-optimize produces worse outcomes than one bound by a fixed rule, because economic actors anticipate the discretion and price in the expected reneging — the "time-inconsistency" problem that a credible commitment technology resolves. [7]
Clinical psychology and addiction: Disulfiram (which makes drinking acutely unpleasant), self-exclusion lists that legally bar problem gamblers from casinos, and the pre-emptive removal of substances or paraphernalia from the home — all pre-remove or poison the means of relapse before craving arrives. The structure is the foresighted self protecting the goal against a future self it expects to be overpowered by craving.
Constitutional and institutional design: Entrenched rights, supermajority amendment thresholds, and fixed terms bind future legislatures and transient majorities against impulses the framing generation wished to foreclose. Holmes (1995) analyzed constitutions as collective precommitments — a polity binding its own future deliberations — and argued that such self-imposed constraints can be enabling rather than merely restrictive, structuring future action so as to make stable self-government possible. [8]
Personal productivity: Website and app blockers, scheduled-send and "do not disturb" defaults, accountability partners, and deposit contracts (such as StickK-style pledges that forfeit money on failure) make backsliding costly enough to deter it. Each converts a repeated in-the-moment willpower problem into a single up-front design decision.
Clarity¶
Naming the commitment device makes visible a counterintuitive move that ordinary "more options are better" reasoning obscures: that reducing one's own choice set can increase welfare or bargaining power. It separates the rational, forward-looking act of self-binding from both mere willpower (which it deliberately bypasses) and from passive constraint (which is imposed from outside rather than chosen). Above all it exposes the precondition that licenses the whole move — a predicted conflict between present and future selves, whether driven by hyperbolic discounting, by temptation, or by the strategic need for credibility. O'Donoghue and Rabin (1999) sharpened this by distinguishing "sophisticated" agents, who foresee their own future present-bias and therefore demand commitment, from "naïve" agents, who do not foresee it and so never reach for the device at all — making the perception of the conflict the hinge on which commitment turns. [9]
The concept also clarifies why a constraint can be valuable to its own bearer. Outside the lens of time-inconsistent preferences, voluntarily forfeiting liquidity, flexibility, or the option to change one's mind looks simply irrational. The commitment-device frame reveals that the forfeited options are precisely the ones a weaker or differently-motivated future self would exploit, so giving them up now is not a loss of value but a transfer of control from the future self back to the present self.
Manages Complexity¶
A commitment device collapses a hard, repeated, in-the-moment self-control problem into a single up-front design decision. Rather than re-litigating temptation at every future juncture — each meal, each market open, each policy meeting, each idle browser tab — the agent makes one binding choice and lets the altered choice architecture carry the rest, freeing cognitive and emotional resources that would otherwise be consumed by continuous vigilance. [10] Thaler and Benartzi (2004) exploited exactly this collapse in their "Save More Tomorrow" program, in which employees commit in advance to allocating future salary raises to retirement savings; because the costly decision is made once, and at a moment when the future raise feels abstract rather than tempting, participation and savings rates rose dramatically without any further act of self-control. [11]
The complexity-management payoff is structural, not motivational. The device does not make the agent want the good outcome more strongly; it relocates the decision from a high-temptation context (where the future self is weak) to a low-temptation context (where the present self is strong), and then makes the present self's choice binding. Complexity is managed by changing when and under what conditions the choice gets made, not by demanding more of the chooser at the hard moment.
Abstract Reasoning¶
Recognizing the pattern licenses a family of inferences that are otherwise hard to reach. It grounds reasoning about credibility: a threat or a promise is believable to others only to the extent that the option to renege has been visibly removed, so that the analysis of deterrence, of contract, and of bargaining all reduce in part to the analysis of which commitment technologies are available. It licenses reasoning about the value of deliberately foreclosing flexibility — inverting the default presumption, ubiquitous in decision theory, that a larger choice set is weakly better, because here some options are liabilities that a future self would misuse. And it explains why agents will pay real costs for constraints, an observation that looks like a market anomaly until the time-inconsistency structure is supplied. [12] Bénabou and Tirole (2004) extended this reasoning to the internal economy of self-image and willpower, modeling personal rules and self-imposed bright lines as commitment devices through which a present self manages an imperfectly-known future self by staking reputation-with-oneself on not defecting.
Knowledge Transfer¶
Schelling's strategic insight that throwing away the steering wheel can win a game of chicken transfers directly across substrates that share no surface vocabulary. In monetary policy, a central bank "ties its hands" via an independence statute or a currency board to gain anti-inflation credibility — the same move as the driver who ostentatiously rips out the wheel so the oncoming car must swerve. In personal finance, a locked savings account is a contract a present self signs against its own future self, the household-scale analogue of the central bank's rule against its own future discretion. [2] The clinical pre-removal-of-means logic — emptying the house of alcohol before craving arrives — transfers to organizational and constitutional design as irreversible governance precommitments and entrenched supermajority thresholds, where a founding body removes from its successors options it expects them to be tempted to take. Bryan, Karlan, and Nelson (2010), surveying the empirical literature, document that the same underlying mechanism — voluntary, costly constriction of one's own future choice set — recurs across savings, health, addiction, and productivity interventions, and that its effectiveness depends consistently on whether the agent is sophisticated enough to recognize the future conflict in advance. [13]
Examples¶
Formal/abstract¶
Game of chicken (Schelling): Two drivers speed toward each other; whoever swerves loses face, but if neither swerves both crash. The standard analysis treats more freedom of action as advantageous. Yet the player who visibly disables his own steering — throwing the wheel out the window where the opponent can see it — wins, because he has converted his swerve from a choice into an impossibility. The opponent, now facing a driver who cannot yield, must yield himself. The commitment is achieved by destroying an option, and its strategic force comes entirely from being irreversible and observable. Mapped back: The structure is the prime in its purest form: a present-tense, costly, observable constriction of one's own choice set that alters the downstream game by making one's preferred outcome follow from infeasibility rather than from continued resolve. The disabled steering is to the driver what the withdrawal penalty is to the saver and the independence statute is to the central bank.
Time-inconsistency in policy (Kydland–Prescott): A central bank announces low inflation; firms and workers set wages and prices on that basis. Once expectations are locked in, the bank faces a tempting one-time gain from surprise inflation (lower real wages, higher output). But rational actors foresee this temptation, disbelieve the announcement, and build expected inflation into their behavior in advance — so the discretionary bank ends up with high inflation and no output gain, a worse equilibrium than if it had been unable to inflate. A binding rule, or an independent central bank with an inflation mandate, removes the bank's option to surprise-inflate and thereby makes the low-inflation promise credible. Mapped back: This is the commitment device operating between a present self (the rule-bound institution) and a future self (the discretionary policymaker who would be tempted ex post). The welfare gain comes not from wanting low inflation more, but from foreclosing the future option to deviate — exactly the dieter's logic of removing the cookies, scaled to a national economy.
Applied/industry¶
Commitment savings accounts (behavioral finance): A bank offers two products: an ordinary savings account and a "commitment" account that pays the same interest but blocks withdrawals until the depositor reaches a self-chosen date or balance. Rationally, no one should prefer the strictly-less-flexible product — yet present-biased savers choose it, and they save substantially more as a result. The illiquidity is the feature: it removes the future self's option to raid the account on a whim, so the savings goal is protected by infeasibility rather than by monthly willpower. Mapped back: The depositor is a present self correctly predicting that a future self, under the pull of immediate consumption, would withdraw the funds; the locked account is the present self's instrument for binding that future self. The forfeited liquidity is not a cost mistakenly accepted but a deliberately purchased constraint — the household analogue of burning the bridges and of the central bank's rule.
Deposit-contract productivity platforms (consumer technology): A user pledges a sum of money to a goal — write daily, quit smoking, run three times a week — and authorizes the platform to forfeit the stake (often to a cause the user despises) if an independent referee reports failure. Nothing about the user's intrinsic motivation changes; what changes is the price of defection, which has been raised in advance from zero to a real and aversive amount, and placed beyond the user's reach to cancel in the moment of temptation. Take-up and success rates rise with the size of the stake and with the unpleasantness of the anti-charity, precisely as the re-pricing logic predicts. Mapped back: The platform converts a repeated in-the-moment willpower problem into one up-front design decision, exactly as Save More Tomorrow does for retirement saving and as the self-exclusion list does for the problem gambler. The structural identity across these cases — a foresighted present self raising the cost of its future self's defection — is what licenses treating savings products, anti-charity pledges, and casino bans as instances of a single prime rather than as unrelated tricks.
Structural Tensions¶
T1: The present self that binds is not obviously more authoritative than the future self that is bound. A commitment device privileges the preferences of the self who sets the device over the self who must live under it, but there is no neutral ground from which to declare the "cold," foresighted self the true self and the "hot," tempted self an impostor. The dieter who locks the pantry may be protecting a considered goal — or may be imposing an ascetic ideal on a future self who would have been right to enjoy the meal. The device executes a transfer of control across time, and which self deserves control is a genuinely contested question the mechanism cannot settle.
T2: A device strong enough to bind reliably is a device that cannot adapt to new information. The value of foreclosing the future self's options collapses if conditions change in ways the present self did not anticipate. A central bank rigidly bound to a fixed rule cannot respond to a genuine emergency; a constitution entrenched against amendment cannot correct a founding error; a saver who has locked away funds cannot meet a true emergency. The same irreversibility that gives the device its binding force also strips the agent of the flexibility to revise when revision is warranted. Strength of commitment and capacity to learn trade off directly.
T3: Commitment must be visible to others to provide credibility, but visibility invites circumvention and signaling games. Schelling-style strategic commitments work only if the adversary can observe that retreat has been foreclosed; a secretly burned bridge deters no one. But once commitment is a public signal, parties have incentives to fake irreversibility, to manufacture the appearance of bound hands while retaining a hidden escape, and counterparties learn to discount the display. The device's power as a credibility technology is in tension with the arms race of feigned commitment it sets off.
T4: Self-binding can be indistinguishable from external coercion, and the difference matters morally even when the structure is identical. A self-exclusion list a gambler signs voluntarily and a casino ban imposed by a court have the same structural effect on the future choice set, yet one is autonomy-enhancing and the other paternalistic. Because the prime is defined by the mechanism rather than by who initiates it, devices that look like enabling precommitments can shade into devices that simply remove a person's freedom "for their own good," and the boundary between protecting an agent's deeper goals and overriding their judgment is rarely sharp.
T5: The device assumes the agent correctly predicts the future conflict, but mispredicting it in either direction is costly. A sophisticated agent who foresees future temptation reaches for commitment; a naïve one does not. But sophistication can also overshoot: an agent who over-predicts future weakness will over-constrain, paying for binding that was never needed and forfeiting flexibility for no gain. Conversely, partial sophistication can be worse than full naïveté, because a partially-aware agent may adopt a weak device that licenses the future self to defect ("I have the gym membership, so one skipped day is fine") while feeling protected. Calibration of the predicted conflict is itself a hard problem the device presupposes rather than solves.
T6: A commitment device protects a goal by making the agent's future irreversible, but the goal it protects may itself be wrong. The mechanism is indifferent to the merit of what is locked in. The same constriction that secures a worthy savings plan or a credible peace can entrench a destructive course: a regime that burns its own diplomatic bridges to look resolute may foreclose the only exit from a ruinous war; a founder who entrenches governance rules may bind successors to a flawed design they cannot fix. Because the device converts a revisable choice into a fixed one, an error committed at the moment of binding propagates forward with the device's full protective force behind it.
Structural–Framed Character¶
Commitment Device is a framed prime on the structural–framed spectrum: it names the pattern in which an agent, anticipating that a later self will be tempted to deviate, deliberately alters the future choice set now — removing options, raising the cost of defection, or delegating the choice away — so the tempting action becomes impossible or unattractive. Its essence is strategic self-limitation against time-inconsistent preferences.
The concept arose in behavioral economics and game theory, and that origin shows: its whole vocabulary of choice sets, defection, and time-inconsistent preferences travels with it, and it cannot be stated without presupposing agents who have preferences, anticipate their own future temptation, and act on them — an irreducibly practice-laden setup. There is some evaluative tinge in casting self-binding as a remedy, and naming a structure a commitment device imports that strategic-choice perspective rather than merely spotting a pattern already there, as when Ulysses lashes himself to the mast or a saver locks funds in an account with a withdrawal penalty. On nearly every diagnostic, it reads framed.
Substrate Independence¶
Commitment Device is a moderately substrate-independent prime — composite 3 / 5 on the substrate-independence scale. The structure is clean and even slightly counterintuitive: an agent deliberately shrinks its own future choice set now to defend against time-inconsistent preferences later. It shows up vividly across behavioral economics, game theory, monetary policy, and clinical psychology. But notice that these are all agent- and decision-flavored substrates; there is no genuine physical, biological, or formal-systems instance, because the whole idea presupposes a chooser who could otherwise defect. The transfer is real but stays within the social-cognitive decision family, which keeps it in the middle tier.
- Composite substrate independence — 3 / 5
- Domain breadth — 3 / 5
- Structural abstraction — 4 / 5
- Transfer evidence — 3 / 5
Relationships to Other Primes¶
Parents (2) — more general patterns this builds on
-
Commitment Device presupposes Constraint
A commitment device works by deliberately altering the future choice set so that tempting deviations become impossible or costly — voluntarily imposing a binding restriction on one's later admissible actions. Without constraint's machinery of binding restriction on admissible configurations, there would be no structural difference between announcing an intention and locking in a course: the device's whole logic is to convert a soft preference into a hard constraint. The parent prime supplies the binding-restriction structure that the commitment device deploys against time-inconsistent preferences.
-
Commitment Device presupposes Temporal Inconsistency and Preference Reversals
A commitment device is a deliberate restriction of one's future choice set imposed by the present self to bind a later self against tempting deviation. The strategy is only intelligible when later preferences are expected to differ from current ones in a predictable way — when the agent anticipates preference reversal as the future becomes present. Temporal inconsistency names exactly this pattern: the same information and goals produce reversed preferences as horizon shortens. Commitment devices presuppose this dynamic as the threat they are designed to neutralize through preemptive self-binding.
Path to root: Commitment Device → Constraint
Neighborhood in Abstraction Space¶
Commitment Device sits among the more crowded primes in the catalog (24th percentile for distinctiveness): several abstractions describe nearly the same structure, so a description that fits it will tend to fit its neighbors too — transporting it usually means disambiguating within this family rather than landing on it exactly.
Family — Cooperation, Trust & Institutional Bonds (19 primes)
Nearest neighbors
- Foresight — 0.84
- Self Control — 0.84
- Institution — 0.80
- Decision — 0.80
- Reputation — 0.80
Computed from structural-signature embeddings · 2026-05-29
Not to Be Confused With¶
A commitment device must be distinguished from Escalation of Commitment, with which it shares the word "commitment" but which is nearly its temporal mirror image. Escalation of commitment is the irrational continuation of a failing course of action because of resources already sunk into it — the manager who pours more money into a doomed project to justify the original investment, the investor who holds a losing stock to avoid realizing the loss. Its causal arrow points backward: past expenditure drives present persistence, against the agent's own forward-looking interest. A commitment device points the other way. It is a deliberate, forward-looking act in which a present self binds a future self against an anticipated deviation, and it is adopted precisely because the present self is reasoning correctly about the future rather than being captured by the past. Escalation is a bias to be overcome; a commitment device is a tool to be wielded. The two can even interact perversely — a poorly chosen commitment device can lock an agent into a course that later escalation-style reasoning then refuses to abandon — but their core logics are opposite: one is the seduction of sunk costs pulling an agent forward into a bad continuation, the other is foresight reaching backward in time, as it were, to constrain a self that has not yet been tempted.
Nor is a commitment device the same as Sunk Cost (and irreversible commitment understood as a psychological trap). Sunk cost describes an unintended and typically irrational barrier: resources already spent that ought to be ignored in forward-looking decisions but that instead distort them, generating the reluctance to abandon a course simply because much has been invested in it. The sunk cost is a byproduct, an accident of accounting and psychology, and the rational prescription is to disregard it. A commitment device is the opposite in intent and in valence: it is an intentionally engineered constraint, created on purpose for its binding effect, and the cost it imposes on future defection is not a regrettable residue but the very thing the agent is buying. Where the sunk-cost fallacy is a failure of reasoning that traps an agent, a commitment device is a success of reasoning that protects one. The confusion arises because both involve "being committed" to a course in a way that resists change; the distinction is whether the commitment was designed (device) or merely accumulated and then irrationally honored (sunk cost).
Finally, a commitment device must be separated from Lock-In and Path Dependence, which also produce constrained, hard-to-reverse futures but through an entirely different causal route. Lock-in and path dependence describe how a sequence of locally reasonable choices, network effects, switching costs, and self-reinforcing dynamics can leave a system trapped in a particular configuration — the QWERTY keyboard, a dominant but technically inferior standard, an organization shaped by its accumulated history — without anyone having chosen the trap as a binding constraint. The constraint is emergent: it arises as an unintended consequence of prior choices and feedback, and it is frequently regretted. A commitment device, by contrast, is a chosen self-constraint adopted precisely for its binding effect; the agent wants to be bound and engineers the binding deliberately. The structural family resemblance — a present that narrows the future, a state that is costly to escape — is real, which is why the three are neighbors, but the diagnostic question is intentionality and direction. Path dependence happens to an agent through the accretion of past choices; a commitment device is something an agent does to its own future on purpose. An agent can, of course, exploit lock-in dynamics as a commitment technology — choosing a standard precisely because its switching costs will later prevent backsliding — but at that point the lock-in has been recruited into a commitment device, which only underscores that the prime is defined by the deliberate, forward-looking self-binding, not by the constrainedness of the resulting state.
Solution Archetypes¶
No catalogued solution archetypes reference this prime yet.
Notes¶
The commitment device sits inside the broader theory of time-inconsistent preferences, and its analysis is most powerful when paired with the distinction between sophisticated and naïve agents. Only an agent who foresees its own future conflict will reach for a device; a fully naïve agent never perceives the need, and a partially sophisticated agent may adopt devices that are too weak to bind, gaining a false sense of protection. Diagnosing where an agent sits on this spectrum is often the first practical step in deciding whether a commitment device is the right intervention at all.
Commitment devices vary along at least three structural dimensions that are worth keeping separate: the mechanism of binding (removing an option entirely, raising the cost of defection, or delegating the choice to a third party), the degree of binding (perfectly irreversible versus merely costly to escape), and the source of the binding (self-imposed versus externally imposed but voluntarily accepted). Many disputes about whether something "really is" a commitment device dissolve once these dimensions are named, because devices that look different on the surface often share the same structural core while differing on one of these axes.
The prime is morally and prudentially neutral. The same mechanism that protects a savings goal or makes a peace credible can entrench a bad policy, lock a regime into a ruinous war, or impose one self's ascetic ideal on another self who would have been right to deviate. Critical reasoning about whether the goal being locked in deserves protection must always accompany the technical reasoning about how to lock it in — a caution that parallels the way activation-energy logic can be applied perversely to mobilize harmful movements.
A recurring practical subtlety is that visibility cuts both ways. For strategic commitment (deterrence, bargaining), the binding must be observable to the counterparty or it provides no credibility; for self-control commitment, observability is often irrelevant or even counterproductive. Conflating these two uses leads to errors — designing a private self-control device as though it needed to signal resolve to others, or designing a strategic commitment that is binding but invisible and therefore strategically worthless.
References¶
[1] Strotz, R. H. (1955). Myopia and inconsistency in dynamic utility maximization. The Review of Economic Studies, 23(3), 165–180. Foundational formalization of dynamic inconsistency: an agent's optimal plan over future consumption is not in general the plan that the agent's later self will choose to follow, generating preference reversals as the decision horizon approaches. ↩
[2] Schelling, T. C. (1960). The Strategy of Conflict. Harvard University Press. Introduces strategic pre-commitment and commitment devices as deliberate self-binding mechanisms; the contrast with inadvertent lock-in is structural — both produce future-self constraint, but commitment devices are sought while lock-ins emerge as side effects of locally reasonable choices. ↩
[3] Elster, J. (1979). Ulysses and the Sirens: Studies in Rationality and Irrationality. Cambridge University Press. Philosophical analysis of pre-commitment and weakness of will: distinguishes akrasia (intention-action gap) from the underlying preference reversal that generates it, motivating Ulysses-style binding strategies. ↩
[4] Laibson, D. (1997). Golden eggs and hyperbolic discounting. The Quarterly Journal of Economics, 112(2), 443–477. Introduces the quasi-hyperbolic (beta-delta) discount function as a tractable model of distance-dependent valuation; shows how preferences expressed at temporal distance reverse at temporal proximity. ↩
[5] Baumeister, R. F., & Vohs, K. D. (2007). Self-regulation, ego depletion, and motivation. Social and Personality Psychology Compass, 1(1), 115–128. Develops the limited-resource (strength) model of self-control, supporting the distinction between depletable in-the-moment willpower and situational self-binding that bypasses it. ↩
[6] Ashraf, N., Karlan, D., & Yin, W. (2006). Tying Odysseus to the mast: Evidence from a commitment savings product in the Philippines. The Quarterly Journal of Economics, 121(2), 635–672. Field experiment showing that a voluntary withdrawal-restricting savings account raised balances substantially, with take-up concentrated among present-biased (hyperbolic) savers. ↩
[7] Kydland, F. E., & Prescott, E. C. (1977). Rules rather than discretion: The inconsistency of optimal plans. Journal of Political Economy, 85(3), 473–491. Time-consistency problem in macroeconomic policy: optimal plans formulated at distance are abandoned at the moment of execution; institutional analogue of individual preference reversal in long-horizon corporate commitments. ↩
[8] Holmes, S. (1995). Passions and Constraint: On the Theory of Liberal Democracy. University of Chicago Press. Analyzes constitutions as collective precommitments — a polity binding its own future deliberation — and argues such self-imposed constraints can be enabling rather than merely restrictive. ↩
[9] O'Donoghue, T., & Rabin, M. (1999). Doing it now or later. American Economic Review, 89(1), 103–124. Formal analysis of present-biased preferences: reframes procrastination and self-control failure from moral weakness to a structural feature of distance-dependent discount functions, motivating commitment-device interventions. ↩
[10] Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving Decisions about Health, Wealth, and Happiness. Yale University Press. Develops choice architecture and friction reduction as policy-level activation-energy lowering: defaults, simplification, and removal of small barriers transform thermodynamically favorable but kinetically blocked behaviors. ↩
[11] Thaler, R. H., & Benartzi, S. (2004). Save More Tomorrow: Using behavioral economics to increase employee saving. Journal of Political Economy, 112(S1), S164–S187. The SMarT program in which employees commit in advance to allocating future salary raises to retirement savings, raising participation and saving rates without further in-the-moment self-control. ↩
[12] Bénabou, R., & Tirole, J. (2004). Willpower and personal rules. Journal of Political Economy, 112(4), 848–886. Models personal rules and self-imposed bright lines as internal commitment devices, with a present self managing an imperfectly-known future self by staking self-reputation on not defecting. ↩
[13] Bryan, G., Karlan, D., & Nelson, S. (2010). Commitment devices. Annual Review of Economics, 2, 671–698. Survey of the empirical literature showing the same voluntary, costly constriction of one's own future choice set recurs across savings, health, addiction, and productivity, with effectiveness depending on the agent's sophistication. ↩