On Sybil Proofness in Competitive Combinatorial Exchanges

We study Sybil manipulation in BRACE, a competitive equilibrium mechanism for combinatorial exchanges, by treating identity creation as a finite perturbation of the empirical distribution of reported types. Under standard regularity assumptions on the excess demand map and smoothness of principal utilities, we obtain explicit linear bounds on price and welfare deviations induced by bounded Sybil invasion. Using these bounds, we prove a sharp contrast: strategyproofness in the large holds if and only if each principal’s share of identities vanishes, whereas any principal with a persistent positive share can construct deviations yielding strictly positive limiting gains. We further show that the feasibility of BRACE fails in the event of an unbounded population of Sybils and provide a precise cost threshold that ensures disincentivization of such attacks in large markets.

💡 Research Summary

The paper investigates the vulnerability of the Budget‑Relaxed Approximate Competitive Equilibrium (BRACE) mechanism— a leading competitive‑equilibrium based solution for combinatorial exchanges— to Sybil attacks, where a single principal can create many fake identities at negligible cost. The authors model identity creation as a finite perturbation of the empirical distribution of reported types. Under standard regularity assumptions—namely that the excess‑demand correspondence is locally Lipschitz and that principals’ utilities are Lipschitz in prices—they derive explicit linear bounds on how prices and welfare deviate when a bounded number of Sybil identities are introduced.

First, the authors formalize the market: a finite set of identities N, a finite set of indivisible goods M with capacities, and each identity i having a feasible bundle set Ψ_i and ordinal preferences. BRACE introduces random budget relaxations (parameter δ) to smooth discrete demand, yielding a δ‑BRACE equilibrium (price vector p and random allocation ˜X) that satisfies optimal random demand, approximate market clearing, individual rationality, ordinal efficiency, and justified envy‑freeness up to one good (JEF1).

The core technical contribution is a “price‑stability” theorem (Theorem 4): if the empirical type distribution is perturbed by ε (measured in total variation), the resulting equilibrium price vector shifts by at most L·ε, where L is a Lipschitz constant derived from the excess‑demand map. Combining this with Lipschitz utilities gives Proposition 3, which states that each principal’s expected welfare loss is also O(ε). Thus, small‑scale Sybil attacks cause only bounded distortions.

The paper then shifts focus from identities to principals. Let s_{p,n}=|N_p|/|N| denote the share of identities controlled by principal p. Theorem 5 proves a sharp dichotomy: strategy‑proofness in the large (SP‑L) at the principal level holds iff max_p s_{p,n} → 0 as the market grows. If any principal retains a positive fraction of identities, its price impact does not vanish, and coordinated misreports can generate a strictly positive limiting gain, even though identity‑level SP‑L still decays at the classic O(n^{‑1/2+ε}) rate. This demonstrates that BRACE’s fairness guarantees do not aggregate across identities.

Further, Proposition 5 shows that when a principal’s identity share tends to one (i.e., an unbounded Sybil mass), the excess‑demand conditions required for equilibrium eventually fail, precluding any sequence of δ‑BRACE equilibria. In other words, BRACE cannot exist in markets overwhelmed by Sybils.

To address the “bad region” where regularity assumptions break down, Lemma 1 bounds expected utilities by the probability of entering that region; if this probability vanishes, Sybil sensitivity disappears. Building on the previous bounds, Proposition 6 derives an explicit cost threshold K* such that imposing a per‑identity cost K ≥ K* makes profitable Sybil strategies infeasible in sufficiently large economies. This provides a concrete design guideline: by calibrating budget‑relaxation δ and enforcing a minimal identity‑creation cost, a system can deter Sybil attacks while preserving BRACE’s desirable welfare properties.

Overall, the paper delivers a rigorous perturbation‑theoretic framework for analyzing Sybil manipulation in competitive combinatorial exchanges, establishes precise conditions under which BRACE remains strategy‑proof, identifies the exact point where it collapses, and offers actionable cost‑based deterrence mechanisms for practitioners designing decentralized markets.