WorksEconomics, Econometrics and Finance
An Enforcement Foundation for Constrained Equal Awards
Abstract
Standard claims problems take claims as given. We ask how rule selection changes when a certified claim can be augmented by a privately known component that is costly to verify. In a binary environment, equal treatment of equals, claim boundedness, and efficiency impose a sharp lower bound on the gain from exaggeration and hence on the expected auditing needed for truthful reporting to be weakly dominant. The constrained equal awards (CEA) rule attains this bound at every report profile through a correlated audit policy. Conversely, every profilewise audit-minimal rule is CEA. The pointwise result also makes CEA the unique admissible rule that minimizes expected verification cost under every prior over true claims.
Introduction
Claims problems divide an insufficient estate among competing claims. The classical model treats claims as established before allocation [9, 11]. When a documented claim can be augmented by an increment requiring further evidence, the allocation rule also affects verification costs.1
The canonical claims problem is a triple , where and . A rule chooses awards satisfying and . The constrained equal awards rule (CEA) assigns
Thus all unsatisfied claimants receive the same award. Dagan [4] characterizes CEA by equal treatment of equals (equal claims receive equal awards), composition up (for , allocating and then to the residual claims using the same rule yields the allocation for the larger estate ), and claims-truncation invariance (the part of a claim that exceeds the estate does not matter). By contrast, the proportional rule assigns : it preserves claim ratios and equalizes satisfaction ratios for positive claims, rather than award levels.
This raises a rule-selection question: among claim-bounded and efficient rules that treat equals equally (henceforth admissible), which is cheapest to enforce when eligibility for the increment is privately known? A larger payoff gain from a high report strengthens the incentive to exaggerate and therefore requires more auditing. Verification burden is thus an endogenous property of the rule.
We study a binary environment. Every claimant has the same certified component and a total claim of either or ; eligibility for the increment is privately known, and an audit reveals perfectly whether an asserted increment is genuine. The estate covers all certified components but not all high claims. “Certified” describes evidence, not a guaranteed award.
Our main result characterizes CEA. An admissible rule is profilewise audit-minimal if no admissible rule can make truthful reporting weakly dominant with fewer expected audits at any report profile. Admissibility imposes a lower bound on the payoff premium from a high report. CEA reaches the resulting audit bound at every profile and is the only rule that does so. The sanction changes the amount of auditing, but not the selected rule.
Recent work derives CEA from restrictions on maximal reasonable claims (Flores-Szwagrzak and Treibich [6]) or immunity to merging manipulations (Yin et al. [13]). Strategic approaches instead endogenize claims or bargaining (Ashlagi et al. [1]; Li and Ju [7]). For rationing with single-peaked preferences, Sprumont [10] characterizes the uniform rule by strategy-proofness, efficiency, and anonymity. Under excess demand it gives the CEA allocation when peaks are interpreted as claims. Under this rule, overstating a peak cannot increase the award of an agent rationed below it; an agent already at her peak does not prefer more. Here claims are entitlements, while preferences increase with the monetary award, including amounts above the true claim. Verification is therefore needed to induce truthful CEA.2 Whereas the costly-verification literature designs allocation mechanisms subject to verification constraints (Townsend [12]; Border and Sobel [3]; Ben-Porath et al. [2]; Mylovanov and Zapechelnyuk [8]), we compare admissible claims rules by the minimum auditing needed to implement each one.
For intuition, let be the estate remaining after reserving for every claimant. At a scarce profile with high reports, claim boundedness caps each low reporter at , so efficiency and equal treatment of equals force each high reporter to receive at least . A low claimant facing the same high-reporting opponents can therefore gain at least by reporting high. CEA makes both bounds tight.
Model
Let , , be the set of claimants. Fix claim levels , an estate with , a per-audit cost , and a sanction for a detected false high report. Each claimant has a true claim and submits a report ; and denote the true and report profiles, both in . Let be the possible increment and
the estate remaining after hypothetically covering every certified component. For a profile , let be its set of high reporters, and write for the profile obtained by replacing ’s report with .
A rule assigns an award vector to every report profile (the estate is fixed and suppressed). It is admissible if it is claim bounded ( for every ), efficient (), and satisfies equal treatment of equals (ETE): implies .3 Efficiency exhausts the compensation budget whenever claim bounds permit. If , claims are paid in full and the authority retains the surplus. This binary rationing problem restricts to the usual claims domain when total claims are at least .
At a profile with high reports, ETE gives high reporters a common award () and low reporters a common award (). The profile is nonscarce if and scarce otherwise. Claim boundedness and efficiency fully compensate every reporter at nonscarce profiles. At a scarce profile,
where the first term is absent when . On this domain CEA depends only on :
Note that is nonincreasing in , so whenever .
We now turn to verification. An audit policy assigns to every report profile a probability distribution over subsets . Only high reports are audited, because only the asserted increment requires verification. After observing , the authority draws and audits its members simultaneously, so cannot depend on audit outcomes or on . Each audit costs and reveals the audited claimant’s true claim. Audited false high reports are corrected to before is applied, and each detected liar pays ; the sanction does not augment , and truthful audited claimants are not sanctioned. Claimants are risk neutral and value their award net of any sanction.
Write for the probability that is audited. The number of audits at is the random variable , where , and its expected value is
Expected audits depend only on marginals; incentives depend on the joint corrections, which determine the profile to which is applied.
An audit policy makes dominant-strategy incentive compatible (DSIC) if, for every , true profile , and reports of the others, reporting weakly maximizes ’s expected payoff. Let denote these policies. For every admissible , auditing every high report is DSIC: exaggerations are corrected and penalized, while truthful high types receive more than (see the proof of Lemma 1) and low reports yield at most . Hence is nonempty. It is closed in a finite product of simplices, so the minimum
exists. The fixed sanction is suppressed in and subsequent cost notation.
Definition 1 (Profilewise audit-minimality). An admissible rule is profilewise audit-minimal if for every admissible rule and every .
Fix with high reports and . Define
as the reporting premium, the gain from reporting high without audits. Suppose and , so others’ reports are truthful and never corrected. Truth pays ; exaggeration pays without an audit and otherwise. DSIC requires
Equivalently, . Solving at equality for a positive premium gives the threshold
which is zero for a nonpositive premium, since no audit is needed to deter this deviation. Every thus satisfies . This condition is necessary for DSIC but not sufficient, since it considers only truthful opponents. Lemma 1 establishes strictly positive premiums. CEA’s premium with high reports is from (2), giving the marginal audit threshold
Results
Lemma 1 (Premium floor). For every admissible , every with high reports, and every , , with equality for . Proof. If , both and are nonscarce, so , , and . If , claim boundedness gives when , and (1) yields ; when , (1) gives directly. Since also , we obtain . Equality for CEA follows from (2).
Lemma 2 (Audit floor). For every admissible and every with high reports,
Proof. Every satisfies ; minimizing over gives the first inequality. The second follows from Lemma 1, since is strictly increasing on .
Theorem 1 (Audit-minimality characterization of CEA). The audit policy defined by if and, if has high reports,
makes CEA DSIC and attains at every profile with high reports, with at the all-low profile. Moreover, for an admissible rule , the following are equivalent:
(a) ;
(b) for every with high reports and every ;
(c) is profilewise audit-minimal.
Proof. Fix and the others’ true claims and reports, and let count high reports when reports . Under , each high reporter is audited with probability , so . Reporting always yields under CEA.
Let . Reporting yields if nobody is audited. Otherwise is detected, corrected, and receives . The expected payoff is , so truthful reporting is weakly optimal.
Let . Reporting yields if nobody is audited. Otherwise is confirmed while false high reports of others are corrected, leaving high reports with ; receives because is nonincreasing. Reporting is strictly optimal. Hence , and with Lemma 2, ; at the all-low profile, .
is the equality case of Lemma 1.
. Nonscarce profiles are fixed by admissibility. Let be scarce with high reports and . The proof of Lemma 1 gives
By (b), , so both inequalities bind and . If , (1) then gives . Every scarce profile’s awards thus coincide with (2), so .
follows from and Lemma 2.
. If is profilewise audit-minimal, then for every with high reports,
Since each by Lemma 1 and the monotonicity of , each equals . Strict monotonicity of then gives .
The joint audit distribution matters for attaining the bound. Let , , , , and , so . Fix a claimant, suppose both true claims are low, and let the opponent report . Reporting pays 1. If both report and each is audited independently with probability , the four equally likely outcomes (neither audited, only the opponent, only the claimant, both) give a claimant net payoffs 2, 3, 0, 0: catching only the opponent raises the unaudited liar’s award to 3. Exaggeration therefore pays . Under , both are audited or neither is, and exaggeration pays . Correlation thus allows the minimum marginal audit probabilities to deter exaggeration even when others lie. Independent audits at higher rates, including full auditing, can still provide DSIC.
The proportional benchmark also illustrates the strictness of the bound. At a scarce mixed profile (one with high reports), let . Proportional awards are and . By efficiency, ; since , it follows that for every . Proportional therefore requires strictly more auditing at that profile. If no mixed profile is scarce, its full-compensation extension and CEA coincide on the binary domain.
The sanction changes required auditing, while CEA remains uniquely optimal for every finite , including . Such penalties are feasible in the model because net payoffs may be negative. Since as , institutional restrictions justify taking as fixed or bounded [3, 8]. If the authority chooses without a cost of setting the sanction, expands with , so choosing minimizes audits. A cost incurred only when a sanction is imposed is zero along the truthful equilibrium used to evaluate costs. An ex ante cost of establishing a sanction may favor a smaller . If this cost depends only on , replacing any rule by CEA at the same positive sanction weakly lowers total expected cost. Thus any attained joint optimum can be achieved with CEA.4
We close with expected verification cost. For a prior over true profiles , the minimum expected verification cost of implementing truthfully is
Costs are evaluated at truthful reports . Under , low types are indifferent between both reports, so truthfulness need not be the unique equilibrium. The prior may assign all its weight to nonscarce profiles, where every admissible rule pays claims in full.
Corollary 1 (Prior-free verification cost). For every prior and every admissible rule , . Moreover, an admissible rule that minimizes expected verification cost for every prior is CEA.
Proof. For every and every , by the definition of and Theorem 1. At the all-low profile nobody reports high, so every policy audits nobody and all these terms are zero. Taking expectations under and minimizing over gives . For uniqueness, let minimize cost for every prior and take degenerate on an arbitrary profile . Then , and for every admissible by Theorem 1. As is arbitrary, is profilewise audit-minimal, and Theorem 1 gives . □
Conclusion
Costly verification changes the comparison of claims rules because a rule determines not only how the estate is divided but also how profitable it is to exaggerate. In the binary environment, admissibility imposes a lower bound on this reporting gain. CEA uniquely attains the bound simultaneously at every profile and, together with a correlated audit policy, makes truthful reporting weakly dominant with the fewest expected audits. The contrast with proportional division identifies the source of the result: protecting each certified component leaves the smallest possible premium for a high report. Because the bound is pointwise, the same argument delivers verification-cost optimality under every prior.
The characterization is deliberately tied to perfect, one-sided, simultaneous verification. Richer claim spaces, heterogeneous evidence or sanctions, and adaptive or imperfect audits may select different rules.
Declaration of generative AI and AI-assisted technologies in manuscript preparation
ChatGPT (OpenAI) was used to assist with exposition and checks of arguments, notation, and references. The author reviewed and revised the resulting material and takes responsibility for the article.
References
- [1]Itai Ashlagi, Emin Karagözoğlu, and Bettina Klaus. A non-cooperative support for equal division in estate division problems. Mathematical Social Sciences, 63(3):228–233, 2012. doi: 10.1016/j.mathsocsci.2012.01.004.
- [2]Elchanan Ben-Porath, Eddie Dekel, and Barton L. Lipman. Optimal allocation with costly verification. American Economic Review, 104(12):3779–3813, 2014. doi: 10.1257/aer.104.12.3779.
- [3]Kim C. Border and Joel Sobel. Samurai accountant: A theory of auditing and plunder. Review of Economic Studies, 54(4):525–540, 1987. doi: 10.2307/2297481.DOI
- [4]Nir Dagan. New characterizations of old bankruptcy rules. Social Choice and Welfare, 13(1):51–59, 1996. doi: 10.1007/BF00179098.DOI
- [5]Lloyd Dixon, Geoffrey McGovern, and Amy Coombe. Asbestos bankruptcy trusts: An overview of trust structure and activity with detailed reports on the largest trusts. Technical Report TR-872-ICJ, RAND Institute for Civil Justice, Santa Monica, CA, 2010. URL https://www.rand.org/pubs/technical_reports/TR872.html.DOI
- [6]Karol Flores-Szwagrzak and Rafael Treibich. How much can you claim? Economics Letters, 246:112068, 2025. doi: 10.1016/j.econlet.2024.112068.
- [7]Jiawen Li and Yuan Ju. Divide and choose: An informationally robust strategic approach to bankruptcy problems. Journal of Mathematical Economics, 107:102862, 2023. doi: 10.1016/j.jmateco.2023.102862.DOI
- [8]Tymofiy Mylovanov and Andriy Zapechelnyuk. Optimal allocation with ex post verification and limited penalties. American Economic Review, 107(9):2666–2694, 2017. doi: 10.1257/aer.20140494.DOI
- [9]Barry O’Neill. A problem of rights arbitration from the Talmud. Mathematical Social Sciences, 2(4):345–371, 1982. doi: 10.1016/0165-4896(82)90029-4.DOI
- [10]Yves Sprumont. The division problem with single-peaked preferences: A characterization of the uniform allocation rule. Econometrica, 59(2):509–519, 1991. doi: 10.2307/2938268.DOI
- [11]William Thomson. How to Divide When There Isn’t Enough: From Aristotle, the Talmud, and Maimonides to the Axiomatics of Resource Allocation. Cambridge University Press, Cambridge, 2019. doi: 10.1017/9781108161107.DOI
- [12]Robert M. Townsend. Optimal contracts and competitive markets with costly state verification. Journal of Economic Theory, 21(2):265–293, 1979. doi: 10.1016/0022-0531(79)90031-0.DOI
- [13]Xiuxia Yin, Pedro Calleja, and Josep M. Izquierdo. Claims manipulations discriminate between egalitarian rules. Theory and Decision, 100(4):1041–1059, 2026. doi: 10.1007/s11238-025-10096-2.