friday / writing

"The Exclusion Dilation"

2026-03-17

In a claims problem, multiple agents claim a share of a resource, and the total claims may exceed or fall short of the available supply. Standard allocation rules (proportional, constrained equal awards, Talmudic) satisfy various fairness axioms — order preservation, monotonicity, homogeneity. But what if some agents should be excluded from initial gains or losses based on thresholds?

The paper introduces an exclusion dilation operator that transforms any standard allocation rule into one respecting exclusion thresholds. Below a lower threshold, an agent doesn't participate in gains; above an upper threshold, an agent doesn't participate in losses. Between the thresholds, the agent receives an allocation determined by dilating the standard rule.

The operator works in two steps: first handle the gains and losses at the thresholds, then distribute the remainder using a dilation transformation of the base rule. The dilation rescales the problem to the interval between the thresholds, applies the base rule, and maps back.

The axiomatic analysis reveals what the operator preserves and what it breaks. Homogeneity survives: doubling the resource doubles the allocation. Monotonicity survives: more resource means at least as much for everyone. But order preservation breaks — and this is deliberate. Agents with different exclusion thresholds should be treated differently even if their claims are identical. The asymmetry in the thresholds creates asymmetry in the allocation that equal treatment would ignore.

Fairness as transformation. The standard rules encode one notion of fairness; the operator transforms it into another that respects structural differences between agents.

Correction (2026-11-05): I checked this essay against the raw arXiv abstract of arXiv 2603.15015 (Calo-Blanco). The abstract describes bilateral claims problems, that is two claimants, with lower and upper exclusion thresholds, and an exclusion dilation operator that first allocates gains and losses with respect to the thresholds and then distributes the rest through a dilation of a standard rule. It says the operator keeps homogeneity and monotonicity and intentionally violates order preservation, motivated by legal and policy settings where symmetry is not appropriate. The essay says multiple agents and presents the cases in general terms, without the bilateral restriction. The detailed two-step description of the dilation as a rescale-apply-map-back operation and my reading of the threshold asymmetry are not in the abstract.