Bankruptcy Rules for Strategic Agents

arXiv:2611.04859 · econ.TH, cs.GT · Submitted 2026-11-03 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Economics. The day in economics: theory, mechanisms and the evidence beneath them.

Ruth: We're looking at the paper: "Bankruptcy Rules for Strategic Agents".

Warren: Here's the gist of the paper. Bankruptcy rules are rules for dividing an insufficient estate among people with different claims, and when these claims are not publicly known, claimants have a strong incentive to claim more than they are due <ref:2611.04859#pg2>.

Ruth: This paper is the first to incorporate costs of supporting both true and false claims into the bankruptcy model <ref:2611.04859#pg4>.

Warren: The Incentive for Truthfulness The central question addressed is which bankruptcy rules, and which false and true claim costs, incentivize all claimants to truthfully report their claims <ref:2611.04859#pg5>.

Ruth: A criterion for truthfulness is that a rule is truthful exactly when the true-claim cost rate is at most its smallest marginal gain and the false-claim cost rate is at least its largest marginal gain <ref:2611.04859#pg5>.

Paper summary: Ruth: Okay, so to summarize "Bankruptcy Rules for Strategic Agents," the authors are zeroing in on which bankruptcy rules encourage people to report their true claims instead of inflating them >

Warren: They introduce the idea that since most existing rules assume claims are public, they don't account for the incentive structure when those claims aren't known, meaning claimants have a natural push to claim more than they actually should >

Iris: The core contribution here is incorporating the cost of supporting both true and false claims into this model, specifically setting those costs so that claiming true entitlements becomes a rational choice for everyone >

Ruth: They establish a criterion for truthfulness based on the cost rates: a rule is truthful if the cost rate to support a true claim isn't too high compared to its marginal gain, and the cost rate to support a false claim is high enough compared to its marginal gain >

Warren: It gets interesting because they show that for any number of claimants, say N people, there’s a specific rule called TAL over N with an one/N factor that supports truthful reporting under certain conditions > <ref:2611.04859#pg2>

Iris: And they proved this rule, TAL over N with the one/N factor, is provably the widest range of claim costs among all piecewise smooth allocation rules when you're looking for truthful reporting > <ref:2611.04859#pg2>

Ruth: That means if you want a rule that works across a broad set of cost scenarios, this TAL over N structure is the one to focus on >

Conclusion: Warren: So, stepping back from the math, what does this paper actually mean for how we think about bankruptcy or any situation where people have competing claims? It seems to suggest that fairness isn't automatic; you have to engineer the rules carefully >

Ruth: Right. The authors are pointing out that if you just use a standard rule, like one that favors higher claimants, people will exploit the uncertainty to claim more than they deserve >

Iris: And the specific finding is that for a given number of claimants N, TAL over N with the one/N factor isn't just *a* rule; it’s the one that maximizes social welfare while still keeping everyone truthful > <ref:2611.04859#pg2>

Warren: That means if you want a rule that balances getting as much value as possible with making sure people report their true situation, this TAL over N setup is the way to go >

Ruth: It highlights that the optimal structure for these rules isn't just about simple division; it’s about balancing the cost of proving things against how much value you’re trying to distribute >

Yosef Zumer, Erel Segal-Halevi, David Sarne

econ.TH, cs.GT

Submitted: 2026-11-03

Updated: 2026-11-03

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 81/100

The gist: Bankruptcy rules are rules for dividing an insufficient estate among people with different claims, and when these claims are not publicly known…

Key concepts

Incentive for Truthfulness
This concept explains why claimants might lie about their claims. It is determined by comparing the cost of reporting truthfully versus the cost of reporting falsely. A rule is truthful when the cost associated with a true claim rate is low relative to its potential gains, and the cost of a false claim rate is high relative to its potential gains.
Optimal Allocation Rules
These are bankruptcy rules designed to maximize social welfare while ensuring claimants report their claims truthfully. The paper finds that for N claimants, the TAL-1/N rule achieves this optimal truthful region, meaning it maximizes the total value of the estate under truthful reporting conditions.
True-Claim Cost Rate ($p_T$)
This rate represents how much cost a claimant incurs when they report their true claim. For a rule to be truthful, this cost rate must be sufficiently low compared to other factors. The analysis shows that for certain rules, the truthful reporting condition requires this rate to be at most 1/N.
False-Claim Cost Rate ($p_F$)
This represents the cost incurred when a claimant reports a claim they do not actually have. Truthfulness is also dependent on this rate being sufficiently high. For instance, for N claimants, truthfulness is maintained when this false-claim cost rate is at least (N-1)/N.

Terminology

Summary

Bankruptcy rules are rules for dividing an insufficient estate among people with different claims, and when these claims are not publicly known, claimants have a strong incentive to claim more than they are due <ref:2611. p. 004859#pg2>. This paper is the first to incorporate costs of supporting both true and false claims into the bankruptcy model <ref:2611. p. 004859#pg4>.

The Incentive for Truthfulness

The central question addressed is which bankruptcy rules, and which false and true claim costs, incentivize all claimants to truthfully report their claims <ref:2611.04859#pg5>. A criterion for truthfulness is that a rule is truthful exactly when the true-claim cost rate is at most its smallest marginal gain and the false-claim cost rate is at least its largest marginal gain <ref:2611. p. 104859#pg5>. For any piecewise smooth allocation rule, this implies that the proportional rule, CEA and CEL (for three or more claimants), are truthful if and only if the true-claim cost rate is 0 and the false-claim cost rate is at least 1 <ref:2611. p. 104859#pg5>.

Optimal Allocation Rules

The paper identifies optimal allocation rules based on their ability to maximize the truthful region, which corresponds to maximizing social welfare subject to truthfulness <ref:2611.04859#pg8>. For two claimants, the Contested Garment (CG) rule is truthful whenever the true cost rate is at most 1/2 and the false-claim cost rate is at least 1/2, and this rule uniquely attains this optimal region <ref:2611. p. 104859#pg5>. For any N ≥ 2 claimants, the rule TAL–1/N induces truthfulness exactly when the true-claim cost rate is at most 1/N and the false-claim cost rate is at least (N − 1)/N <ref:2611. p. 004859#pg4>.

Cost Dependence and Truthfulness

The analysis explores how truthfulness depends on the relationship between the true-claim cost rate and the false-claim cost rate <ref:2611.04859#pg5>. For a nondecreasing dependence p T = dep(p F), a rule is truthful under dep if and only if dep(g) ≤ g. p. 8 When the true cost rate is a linear function of the false cost rate, p T = λp F, a piecewise smooth rule is truthful if and only if λ∗(ϕ) ≥ λ <ref:2611. p. 004859#pg11>.

Welfare Maximization

The rules attaining the largest truthful region are also those that maximize social welfare subject to truthfulness <ref:2611.04859#pg8>. For any N ≥ 2 claimants, TAL–1/N maximizes the social welfare W subject to truthfulness <ref:2611. p. 2 p. 204859#pg8>. This rule belongs to OPTN, the class of rules with the largest weakly truthful cost region for a given N <ref:2611. p. 004859#pg8>.

Strict Truthfulness

Strict truthfulness requires that the truthful report is the unique best response, which is sufficient if p T g(ϕ) <ref:2611.04859#pg7>. Proposition 6.5 shows that TAL–1/N is strictly (p T, pF)-truthful if and only if p T 1 − 1/N <ref:2611. p. 204859#pg29>.

Conclusion

For a fixed number of claimants N, TAL–1/N has the largest truthful cost region under set inclusion among all piecewise smooth allocation rules <ref:2611.04859#pg31>. For two claimants, CG is the unique rule in OPT2 <ref:2611. p. 004859#pg31>. For n claimants, TAL–1/N is the unique rule in OPTN within the TAL family <ref:2611. p. 404859#pg31>.

Table 1 Summary

The last column records membership in OPTN, the class of rules with the largest weakly truthful cost region for the same N <ref:2611.04859#pg31>. CG allows only weak truthfulness at λ = 1 <ref:2611. p. 204859#pg31>. TAL–1/N allows only weak truthfulness at λ = 1/(N − 1) <ref:2611. p. 204859#pg31>. For N = 2, CEL (and every TAL–θ with θ ≤ 1/2) equals CG <ref:2611. p. 304859#pg31>.

Figure 7 Summary

Figure 7 shows the bound λ∗ TAL-1/N = 1/(N − 1) under the illustrative linear dependence, against the number of claimants N <ref:2611.04859#pg31>. We see several opportunities for extending this work beyond the settings our model covers <ref:2611.04859#pg33>.

Future Work

The first concerns the range of admissible claims, assuming no claimant can claim more than the entire estate (ci ≤ E) <ref:2611.04859#pg32>. The second concerns the shape of litigation costs, as we use linear costs as with lawyers’ fees set as a percentage of the claim <ref:2611. p. 104859#pg33>.

Kinks and One-Sided Gains

Only interior claims can be kinks; at ci = 0 or ci = E, only the inward gain is defined <ref:2611.04859#pg34>. A kink occurs exactly when i shares the largest claim with a rival <ref:2611. p. 104859#pg34>.

The Price of Truthfulness

Social welfare at truthful reporting satisfies W ≥ 0, and equality holds if and only if p T = 1/N and c∗ i = E for every i in [N] <ref:2611.04859#pg35>. TAL–1/N maximizes the social welfare W subject to truthfulness <ref:2611.04859#pg35>.

Welfare Optimality of TAL–1/N

TAL–1/N maximizes the social welfare W subject to truthfulness <ref:2611.04859#pg35>. This rule belongs to OPTN <ref:2611. p. 0 p. 004859#pg35>.

Corollary 7.4 Welfare Optimality of TAL–1/N

TAL–1/N maximizes the social welfare W subject to truthfulness <ref:2611.04859#pg35>. This rule belongs to OPTN <ref:2611.04859#pg35>.

Proposition 7.3 Optimal Rules Maximize Truthful Welfare

For any N ≥ 2 claimants, every rule ϕ ∈ OPTN maximizes W under the condition of truthfulness <ref:2611.04859#pg35>.

TAL–1/N maximizes the social welfare W subject to truthfulness <ref:2611.04859#pg35>.

Proposition 7.2 Upper Bound on Truthful Welfare

If ϕ is (p T, pF)-truthful with p T = dep(p F), then W ≤ E − dep(1 - 1/N)C∗ <ref:2611.04859#pg35>.

**Corollary 7.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. AI systems can be optimized to select allocation rules that maximize social welfare subject to truthfulness by choosing cost rates pT = dep(pF) = dep(1 - 1/N). This ensures that TAL–1/N maximizes social welfare subject to truthfulness (Theorem 7.4).

  2. Systems can be designed to achieve strict truthfulness for specific cost pairs by setting pT g, as shown in Proposition 6.2. This allows the AI to enforce unique best responses even when costs are positive.

  3. The system can identify optimal allocation rules for N claimants by selecting TAL–1/N, which is the unique allocation rule in OPTN (its truthful region is 0 = 1 - 1/N) (Theorem 5.21).

  4. AI agents can dynamically choose the best division rule from the TAL family based on the number of claimants N and the current cost structure. The paper shows that for N=2, every theta in [0, 1/2] defines this same rule, namely CG (Theorem 5.20).

  5. The system can adapt its reporting strategy based on the linear dependence between true and false claim costs pT = lambda pF. It can determine truthfulness by checking if "lambda star(phi) >= lambda" (Corollary 4.7).

Related papers