Dynamic Pooling and Regional Participation in Deceased-Donor Organ Allocation

arXiv:2609.18147 · econ.TH, cs.GT · Submitted 2026-09-16 · 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: I'm Ruth.

Warren: And I'm Warren, here with Iris.

Ruth: We're looking at the paper: "Dynamic Pooling and Regional Participation in Deceased-Donor Organ Allocation".

Warren: Here's the gist of the paper. Moving from geographically fragmented to pooled waiting lists in deceased-donor organ transplantation can improve efficiency, but it raises concerns about regional fairness and participation incentives The paper studies Pareto gains from such transitions in a multiclass queueing model with impatient agents, perishable items, and a tractable homogeneous compatibility friction Our analysis identifies dynamic risk pooling as the main source of Pareto improvements for regions Model Formulation The problem is formulated as a multi-class queueing system with abandonment, in the spirit of standard models (e.g., Ansari et al.

Ruth: (2025)), but with multiple regions, a pooled organ-arrival process, and a compatibility friction The state space is S = Z K+, where K = RΘ, consisting of vectors q = (qi)i, where qi denotes the number of waiting patients of class i Organs arrive from region r according to a Poisson process with rate µ r > 0, yielding a total arrival rate µ = P r∈R µ r To incorporate incompatibility while preserving tractability, we assume an ex-ante homogeneous random compatibility structure Optimality of Policy Structure The paper establishes that for any set of nonnegative welfare weights, the optimal allocation rule retains a simple form even under homogeneous random incompatibility: a static index policy This result is the building block for the Pareto frontier characterization in the next section The optimal policy for maximizing weighted social welfare is a static index policy, which assigns a fixed priority score to each patient class and allocates each arriving organ to the highest-index compatible waiting class Achievable Utility Set and Constraints The achievable set U is characterized as the set of nonnegative utility vectors dominated by some convex combination of utility vectors generated by static index policies The Pareto frontier of regional utilities (u r) is formed by the convex hull of regional utility vectors corresponding to different priority orderings The benchmark for regional participation is the autarky utility, which defines the participation constraint u r ≥ u r sep Constrained Optimization Algorithm The Constrained Pareto Frontier Search (CPFS) algorithm efficiently solves the high-dimensional constrained problem by exploiting the optimality of index policies established in Section 4 CPFS computes an approximately optimal allocation policy with arbitrarily small participation-constraint violations

Title and authors: Ruth: Let's talk about who wrote this: Genta Okada is one of the researchers here, and Warren is the quantitative guy running things as a trader. They’re looking at how these allocation systems work under these pooled conditions.

Warren: I’m interested in what they call "dynamic risk pooling" because that sounds like a strategy you'd use in trading to manage volatility, but here it applies to organ supply and demand uncertainty.

Iris: From an econometric standpoint, I’m curious about the model setup they use—they are using a multiclass queueing system with impatient agents and perishable items, which is a lot of complexity to handle when you're trying to get real answers about fairness.

Ruth: That complexity is there because real life isn't simple; people get impatient waiting, organs expire, and compatibility isn't perfect across regions. The paper sets up this model with an ex-ante homogeneous random compatibility structure just to keep things manageable while still capturing that friction.

Warren: So the authors are trying to build a framework that captures these messy real-world factors—like the stochastic supply and demand fluctuations—within a mathematical structure so they can test policy changes.

Iris: That tractability is key, but it raises questions about what's being ignored; specifically, if the compatibility friction is truly random and homogeneous, does that mean we’re missing important regional differences in how organs actually match up?

The paper's summary: Ruth: Moving on to what the paper actually summarizes—it establishes that for any set of welfare weights you choose, there is a specific kind of allocation rule that is optimal, even with this random compatibility friction.

Warren: That sounds like they’ve managed to simplify the decision-making process down to something very structural, which is good for building a strategy. What's the actual policy they land on?

Iris: They find that the optimal policy for maximizing weighted social welfare is a static index policy, meaning you assign a fixed priority score to each patient class and then allocate any incoming organ to the highest-index class that is compatible.

Ruth: So instead of constantly recalculating based on every single arrival, it boils down to setting up these fixed scores for each type of patient and sticking to them. This structure is what allows them to define the achievable utility set as a convex polytope made up of these index policies.

Warren: That simplifies the search space considerably, which is helpful when you’re trying to find a good allocation rule under those dynamic conditions they described in "Dynamic Pooling and Regional Participation in Deceased-Donor Organ Allocation."

Iris: But this structure means the achievable set itself is defined by these static index policies. It shows that even though we have dynamic risks, the optimal solution is fundamentally rooted in a fixed priority assignment.

The paper's improvements: Ruth: The real improvement they highlight isn't just the model itself, but how it suggests we can actually use this structure to solve the problem practically. They propose the Constrained Pareto Frontier Search, or CPFS algorithm.

Warren: I’m interested in that because if you have a high-dimensional problem like this one—with many regions and patient classes—you don't want to check every possible priority permutation manually.

Iris: The CPFS algorithm is presented as a polynomial-time procedure that finds an allocation policy that respects the regional participation constraints, even when we’re dealing with these complex utility vectors. It gets you very close to the best possible outcome without getting stuck in intractable calculations.

Ruth: So they’ve connected the theoretical optimality of static index policies with a concrete algorithm that can actually compute an approximately optimal allocation policy for these constrained problems.

Warren: That means we can potentially use this index policy idea to bypass having to enumerate all those factorial priority permutations when designing a system for deceased-donor organ allocation.

Iris: And there's also this dual approach paired with a subgradient oracle, which lets them find the maximum over an affine function by solving an inner maximization problem that is induced by those modified welfare weights. It’s a sophisticated way to find the right policy without brute force.

Conclusion: Ruth: To wrap up, they show that dynamic risk pooling optimization allows AI systems to maximize global efficiency while making sure local participation incentives stay intact by adjusting those priority indices within the bounds of what is called the "gains from pooled supply, demand, and compatibility risk."

Warren: It seems the main structural finding is that dynamic risk pooling is actually the core source of Pareto improvements for regions when they aren't perfectly balanced.

Iris: I think what this means for us as researchers is that we can use the index policy structure to solve these high-dimensional constrained problems by relying on the optimality of static index policies, which saves us from having to check every single priority permutation.

Ruth: And finally, they suggest a stationary randomized policy or a time-sharing approach that lets AI systems implement allocation rules based on those indices for every patient class.

Warren: That time-sharing idea sounds like a deterministic way to switch between the index policies over time, ensuring that the long-run average utility during any block assigned to one policy converges exactly to its optimal value as time goes on.

Iris: It’s interesting because they also pointed out a limitation: while they can compute an epsilon-optimal allocation subject to constraints, they aren't guaranteeing perfect Pareto improvement over autarky when regions are asymmetric, which is a fair caveat to keep in mind.

Ruth: So, the paper "Dynamic Pooling and Regional Participation in Deceased-Donor Organ Allocation" gives us a way to design allocation rules that are efficient while respecting regional fairness constraints by leveraging static index policies and dynamic risk pooling.

Graduate School of Economics, University of Tokyo

econ.TH, cs.GT

Submitted: 2026-09-16

Updated: 2026-11-12

Comments: 47 pages, 3 figures

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 90/100

The gist: Moving from geographically fragmented to pooled waiting lists in deceased-donor organ transplantation can improve efficiency…

Key concepts

Multi-class queueing model
This is a mathematical framework used to model the organ allocation system. It treats waiting patients as different classes, each with specific needs, and models how organs arrive from various sources over time. This helps predict patient flow and waiting times under different conditions.
Static index policy
This is an optimal allocation rule where every patient class is assigned a fixed priority score. When an organ arrives, it goes to the compatible waiting class with the highest index score. The paper proves this simple policy structure is optimal for maximizing social welfare in this system.
Dynamic risk pooling
This refers to adjusting the priority indices over time based on current conditions like supply and demand. By dynamically changing which classes get higher priority, the system can manage compatibility risks and ensure that benefits are shared fairly across different regions.
Participation constraint
This is a rule ensuring that every region receives a minimum level of service or utility. In this context, it sets a lower bound on the utility received by each region to prevent any single region from being completely ignored in the allocation process.

Terminology

Summary

Moving from geographically fragmented to pooled waiting lists in deceased-donor organ transplantation can improve efficiency, but it raises concerns about regional fairness and participation incentives The paper studies Pareto gains from such transitions in a multiclass queueing model with impatient agents, perishable items, and a tractable homogeneous compatibility friction Our analysis identifies dynamic risk pooling as the main source of Pareto improvements for regions: pooling reduces organ waste due to stochastic supply-demand fluctuations, temporal mismatches, and random compatibility outcomes To resolve this conflict, we show that the set of achievable utilities can be described by mixtures of static priority index policies Building on this structure, we propose the Constrained Pareto Frontier Search (CPFS), a polynomial-time procedure that computes an approximately optimal allocation policy with arbitrarily small participation-constraint violations. p. 0

Where the paper says it

This paper studies Pareto gains from such transitions in a multiclass queueing model with impatient agents, perishable items, and a tractable homogeneous compatibility friction. Unlike standard approaches that focus on static match quality or ignore regional incentives in dynamic settings, our analysis identifies dynamic risk pooling as the main source of Pareto improvements for regions: pooling reduces organ waste due to stochastic supply-demand fluctuations, temporal mismatches, and random compatibility outcomes.

Model Formulation and Framework

The problem is formulated as a multi-class queueing system with abandonment, in the spirit of standard models (e.g., Ansari et al. p. 2

Where the paper says it

Comparable optimality results appear in Down et al. (2011) for two-class systems and in Ansari et al.

(2025)), but with multiple regions, a pooled organ-arrival process, and a compatibility friction The state space is S = Z K+, where K = RΘ, consisting of vectors q = (qi)i, where qi denotes the number of waiting patients of class i. p. 3
Where the paper says it

, Ansari et al. (2025)), but with multiple regions, a pooled organ-arrival process, and a compatibility friction.

A stationary policy σ specifies allocation probabilities after both the queue state and the compatibility realization are observed. p. 3
Where the paper says it

Because patient-level compatibilities are independent, the random set of compatible classes is a subset A ⊆ J(q) with probability pq(A) = Y i∈A gρ(qi) Y j∈J(q)\A 1 − gρ(qj ) . A stationary policy σ specifies allocation probabilities after both the queue state and the compatibility realization are observed.

The transition rates νσ(q, q′) from state q to state q′ under policy σ are defined by λi if q′ = q + ei, diqi + µσ¯i(q) if q′ = q − ei, and-Qσ(q) if q′ = q. p. 4
Where the paper says it

across organ arrivals and can be integrated out conditional on q. The transition rates ν σ (q, q′ ) from state q to state q ′ under policy σ are ν σ (q, q′ ) =    λi if q ′ = q + ei , diqi + µσ¯i(q) if q ′ = q − ei , −Qσ (q) if q ′ = q, 0 otherwise, where Q σ (q) = X i λi + X i diqi + µ X i σ¯i(q) is the total rate of state-changing transitions out of state q.

Optimality of Allocation Rules

The optimal allocation rule for maximizing weighted social welfare is a static index policy This result shows that for any set of nonnegative welfare weights, the optimal allocation rule retains a simple form even under homogeneous random incompatibility: a static index policy. The utility vector u(σ) is characterized by mixtures of utility vectors generated by static index policies. p. 7

Where the paper says it

Proposition 1. U is the set of nonnegative utility vectors dominated by some convex combination of utility vectors generated by static index policies.

Under uniform abandonment, the average utility for class k is given by uk = vk / λk (Tρ(Λk, µ) − Tρ(Λk−1, µ)). p. 8
Where the paper says it

, K in decreasing order of priority, and let Λk = Pk ℓ=1 λℓ with Λ0 = 0. Then the average utility for class k is uk = vk λk (T ρ (Λk, µ) − T ρ (Λk−1, µ)).

Achievable Utility Set and Constraints

The set of achievable utility vectors U is the set of nonnegative utility vectors dominated by some convex combination of utility vectors generated by static index policies. The Pareto frontier of regional utilities (u r) is formed by the convex hull of regional utility vectors corresponding to different priority orderings. p. 8

Where the paper says it

The aggregate regional utility u r is the weighted average of utilities of all patient types within region r: u r = 1 Λr X θ∈Θ λ(r,θ)u(r,θ) , where Λr = P θ λ(r,θ) . The Pareto frontier of regional utilities (u r )r∈R is formed by the convex hull of regional utility vectors corresponding to different priority orderings.

The autarky benchmark, denoted by sep, is constructed by evaluating each region as a separate queue with patient-arrival rate λ r and organ-arrival rate µ r. p. 9
Where the paper says it

1 Utilitarian Policies in a Multi-Region Setting The utilitarian benchmark maximizes the total utility rate of all patients and weights each arriving patient equally. This sets welfare weights equal to arrival rates, wi = λi , so the priority index reduces to bi = wivi λi = vi .

Constrained Optimization Algorithm

The Constrained Pareto Frontier Search (CPFS) algorithm efficiently solves the high-dimensional constrained problem by exploiting the optimality of index policies established in Section 4 It uses a Lagrangian dual approach paired with an efficient subgradient oracle to compute an approximately optimal allocation policy with arbitrarily small participation-constraint violations. The final output is a mixture of static index policies, denoted by C, with utility vector uC. p. 5

Where the paper says it

Actual transplantation satisfies none of them exactly; compatibility varies with blood type, sensitization level, organ quality and donor characteristics, and impatience varies with diagnosis and comorbidity. The first two make organs interchangeable once the compatibility draw is realized, which leaves the queue-length vector a controlled Markov process with static priority rankings among its optimal policies; the third establishes the optimality of index policies for all welfare weights, enabling our characterization of achievable utility vectors.

Empirical Calibration and Results

The model is coarsely calibrated to 2023 U.S. kidney-transplant data to show how participation constraints can bind in practice The CPFS policy (C) preserves the same aggregate throughput while satisfying every constraint. p. 0

Where the paper says it

To resolve this conflict, we show that the set of achievable utilities can be described by mixtures of static priority index policies. Building on this structure, we propose the Constrained Pareto Frontier Search (CPFS), a polynomial-time procedure that computes an approximately optimal allocation policy with arbitrarily small participation-constraint violations.

Table 3 reports the aggregate performance in the U. p. 4
Where the paper says it

Furthermore, by the ergodic theorem for Markov processes, the long-run time average of any function f(q) that has a finite expectation converges almost surely to its expected value under the stationary distribution π σ : lim T→∞ 1 T Z T 0 f(q(t)) dt = Eπσ [f(q)] a. This property allows us to evaluate performance metrics, such as long-run average costs or utilities, by calculating expectations with respect to π σ .

S. kidney-transplant calibration, showing that CPFS matches the brute-force objective in every case to machine precision.

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. Dynamic risk pooling optimization allows AI systems to maximize global efficiency while preserving local participation incentives by adjusting priority indices to keep redistribution within the gains from pooled supply, demand, and compatibility risk.

  2. The Constrained Pareto Frontier Search (CPFS) algorithm provides a polynomial-time procedure that computes an approximately optimal allocation policy with arbitrarily small participation-constraint violations.

  3. AI systems can utilize the index policy structure to solve high-dimensional constrained problems by exploiting the optimality of static index policies established in Section 4, bypassing the need to enumerate all factorial priority permutations.

  4. The dual approach paired with a subgradient oracle allows for the computation of optimal policies by finding a maximum over an affine function, where the inner maximization is solved by the index policy induced by the modified welfare weights.

  5. A stationary randomized policy can be constructed that attains the mixed utility vector, allowing AI systems to implement allocation rules based on P sigma i (q, C) = P sigma in S a s pi (q) sigma i (q, C) for every class.

  6. The time-sharing policy offers a deterministic alternative where the system deterministically switches between the active index policies sigma in S over time, ensuring that "the long-run average utility during the block assigned to policy sigma converges exactly to u(sigma) as T -> infinity."

Abstract

Moving from geographically fragmented to pooled waiting lists in deceased-donor organ transplantation can improve efficiency, but it raises concerns about regional fairness and participation incentives. This paper studies Pareto gains from such transitions in a multi-class queueing model with impatient agents, perishable items, and a tractable homogeneous compatibility friction. In contrast to standard approaches that emphasize static match quality or overlook regional incentives in dynamic contexts, this analysis identifies dynamic risk pooling as the primary source of Pareto improvements for regions. Pooling reduces organ waste resulting from stochastic supply-demand fluctuations, temporal mismatches, and random compatibility outcomes. Although this dynamic surplus theoretically enables Pareto improvements, a purely utilitarian policy that treats all patients equally across regions often fails to achieve these improvements when regions are asymmetric. In these situations, utilitarian redistribution removes an excessive number of organs from high-supply regions and violates their participation constraints. To address this conflict, it is demonstrated that the set of achievable utilities can be represented by mixtures of static priority index policies. Building on this structure, the Constrained Pareto Frontier Search (CPFS) is proposed as a polynomial-time procedure that computes an approximately optimal allocation policy with arbitrarily small participation-constraint violations. The results indicate that, by adjusting priority indices to ensure redistribution remains within the gains from pooled supply, demand, and compatibility risk, policymakers can achieve global efficiency while maintaining local participation incentives.

Related papers