Pith. sign in

REVIEW 1 major objections 5 minor 243 references

In large random markets, Top Trading Cycles loses its priority edge and matches Random Serial Dictatorship on justified envy.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-14 09:00 UTC pith:F5COVIA5

load-bearing objection Clean asymptotic proof that TTC's priority advantage over RSD vanishes in the standard large balanced market; the Markov forest characterization is the reusable piece. the 1 major comments →

arxiv 2607.10819 v1 pith:F5COVIA5 submitted 2026-07-12 econ.TH

Top Trading Cycles in Large Markets: The Asymptotic Irrelevance of Priorities

Yeon-Koo Che , Olivier Tercieux This is my paper
classification econ.TH
keywords Top Trading CyclesRandom Serial Dictatorshipjustified envylarge marketspriority matchingMarkov chainschool choicestrategy-proofness
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Top Trading Cycles is the standard Pareto-efficient, strategy-proof assignment rule that uses institutional priorities. This paper asks how much fairness those priorities actually buy once the market is large. Under a canonical model with independent uniform preferences and priorities, the answer is almost none. The share of agents assigned through short cycles (the only assignments that enforce priorities) vanishes, so the joint distribution of preference ranks and realized priority ranks under TTC converges to that of Random Serial Dictatorship, a mechanism that ignores priorities entirely. Consequently the usual fairness statistics—incidence of justified envy, fraction of agents who justifiably envy someone, and normalized blocking pairs—become asymptotically identical under the two rules. The result quantifies a long-standing policy worry: TTC’s theoretical respect for priorities may not survive scale.

Core claim

In a balanced one-to-one market with iid uniform preferences and priorities, the profile of normalized priority and preference ranks produced by TTC converges in distribution, as market size goes to infinity, to the corresponding profile under Random Serial Dictatorship (iid uniform priorities independent of the common agent-side rank distribution). All standard justified-envy statistics therefore become asymptotically identical under the two mechanisms.

What carries the argument

A novel Markov-chain characterization of the number of agents and objects remaining after each round of TTC. The chain depends only on current market size; it implies that TTC finishes in a sublinear number of rounds while the expected number of short-cycle assignments per round stays bounded by 2, so the fraction of short-cycle assignments vanishes and priorities become irrelevant.

Load-bearing premise

Preferences and priorities are drawn independently and uniformly at random; if they become strongly correlated the short-cycle share need not vanish and priorities can remain relevant.

What would settle it

Simulate balanced markets of increasing size under the same iid uniform draws and check whether the incidence ratio of justified envy under TTC converges to one-half (the exact RSD value) and whether the fraction of short-cycle assignments goes to zero.

Watch this falsifier — get emailed when new claim-graph text bears on it.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

1 major / 5 minor

Summary. The paper analyzes prioritized Top Trading Cycles (TTC) in large one-to-one matching markets under the canonical model of iid uniform preferences and priorities. It establishes that the numbers of remaining agents and objects form a Markov chain with explicit transition probabilities (Theorem 1), that TTC terminates in a sublinear number of rounds (Proposition 1), and that the expected number of short-cycle assignments per round is at most 2 (Proposition C.1). Combining these with a symmetry argument for long-cycle ranks, the authors prove that the joint profile of normalized preference and priority ranks under TTC converges (in the growing-vector sense of Definition 1) to the corresponding profile under priority-blind Random Serial Dictatorship (Theorem 2). Consequently, standard justified-envy metrics—the incidence ratio, the fraction of agents with justified envy, and the normalized number of blocking pairs—become asymptotically identical under the two mechanisms (Proposition 2 and Corollary 1). Section 6 discusses robustness under partial correlation and many-to-one environments.

Significance. If the asymptotic irrelevance result holds, it substantially revises the practical case for TTC as a mechanism that balances Pareto efficiency, strategy-proofness, and priority respect in large markets such as school choice. The paper strengthens the classical finite-market equivalence of TTC and RSD (which concerns only agent-side ranks) to the joint distribution of preference and priority ranks, and it supplies a novel Markov characterization of TTC via random spanning forests that is of independent technical interest. The proofs are self-contained, rely on explicit counting and deferred-decision arguments rather than fitted parameters, and carefully delimit the scope of the iid benchmark. These features make the contribution both theoretically sharp and policy-relevant.

major comments (1)
  1. The central asymptotic claims (Theorem 2 and Proposition 2) are fully proved under the stated iid uniform model, and the chain from the Markov projection through the short-cycle bound and long-cycle rank uniformity is closed. No load-bearing gap appears in the derivation. The modeling assumption itself is the natural weakest point, but the paper already flags its scope in Section 6 and supplies simulations for partial correlation and many-to-one settings; those extensions are not claimed as theorems and do not undermine the main result.
minor comments (5)
  1. Definition 1 of growing-vector convergence is nonstandard; a brief comparison with the usual notions of finite-dimensional convergence plus tightness (or with empirical-measure convergence) would help readers unfamiliar with the device.
  2. In the proof of Proposition 1 (Appendix B), the constant c = 2 / sqrt(pi delta) is introduced without an immediate reference to the asymptotic mean clearance rate derived in Online Appendix OA.2; a forward pointer would improve readability.
  3. Figure 1 caption states averages over 100 draws; it would be useful to report standard errors or interquartile ranges, especially for the incidence-ratio panel at large n, so that the visual convergence can be assessed more precisely.
  4. The discussion of correlated priorities (Section 6.2) asserts within-tier uniformity but does not state a formal proposition; even a short corollary would make the claim easier to cite.
  5. Minor typographical inconsistencies appear in the arXiv header date (July 14, 2026) and in a few cross-references to Online Appendix sections; these should be cleaned before publication.

Circularity Check

0 steps flagged

No significant circularity: asymptotic irrelevance is derived from the iid uniform primitive via a self-contained Markov/forest argument, not by construction or fitted inputs.

full rationale

The paper's central claim (Theorem 2 and Proposition 2) is that, under the balanced one-to-one market with iid uniform preferences and priorities, the joint distribution of normalized priority and preference ranks under TTC converges to that under RSD, so justified-envy statistics become asymptotically identical. The derivation chain is: (i) random spanning forests induced by TTC form a Markov chain whose projection onto remaining market sizes (n,o) has closed-form transitions (Theorem 1, Appendix A); (ii) expected short-cycle assignments per round are at most 2 for any forest (Proposition C.1); (iii) completion time is sublinear (Proposition 1); hence the short-cycle fraction vanishes and almost all objects are assigned via long cycles with R*_o = o_p(n); (iv) long-cycle symmetry (Lemma D.1 / Proposition D.1) forces the realized priority rank of such an object to be uniform on {R*_o+1,...,n}, independent of agent ranks, yielding the growing-vector convergence of Definition 1. None of these steps is definitional of the target, none fits a free parameter to data and re-labels it a prediction, and none imports a uniqueness theorem that forbids alternatives. Self-citations (Pathak–Sethuraman 2011 / Carroll 2014 for agent-side rank equivalence; Che–Tercieux 2018 for the log^{1+ε}n rank bound under Pareto efficiency; Abdulkadiroğlu et al. 2020 for envy-minimality) supply independent intermediate facts used as black boxes; they do not reduce the target to an input by construction. The modeling assumption (iid uniform) is explicit and scoped (Section 6), not circular. Score 1 reflects only ordinary, non-load-bearing self-citation of intermediate lemmas.

Axiom & Free-Parameter Ledger

0 free parameters · 4 axioms · 1 invented entities

The paper is pure theory under a standard random-market model. No numerical parameters are fitted. The only non-standard modeling choices are the uniform iid draws and the one-to-one balanced asymptotics; both are explicitly flagged as the canonical benchmark and are relaxed in the discussion section.

axioms (4)
  • domain assumption Preferences and priorities are drawn independently and uniformly at random (iid uniform model).
    Section 3; all limit theorems are proved under this measure. The authors note that perfect correlation restores priority respect, so the assumption is load-bearing.
  • domain assumption Market is balanced one-to-one (|I|=|O|=n) with n o∞.
    Section 5; the many-to-one and unbalanced cases are treated only by simulation or by separate remarks.
  • domain assumption All agents find all objects acceptable (and vice versa).
    Section 3; the authors claim the result extends when the number of acceptable partners grows linearly, but the proofs are written for the full-support case.
  • standard math Standard combinatorial facts on random bipartite mappings and spanning forests (Jaworski 1985, Jin–Liu 2004, Lovász).
    Used to count cyclic vertices and forests in the Online Appendix; these are classical enumeration results.
invented entities (1)
  • Random spanning-forest process induced by TTC rounds independent evidence
    purpose: To obtain a Markov chain on the macroscopic state (n_t,o_t) after projecting away microscopic history dependence.
    The process is a natural encoding of TTC’s residual graph; it is not postulated ad hoc but derived from the algorithm’s definition. Independent evidence is the closed-form transition probability that matches direct simulation.

pith-pipeline@v1.1.0-grok45 · 43036 in / 2660 out tokens · 36741 ms · 2026-07-14T09:00:19.947829+00:00 · methodology

0 comments
read the original abstract

Top Trading Cycles (TTC) is Pareto efficient and strategy-proof and explicitly uses agents' priorities. Although TTC favors higher-priority agents in each round, we show that this priority advantage vanishes as the market grows large under a canonical random model of preferences and priorities. In the limit, TTC produces assignments with virtually the same incidence of justified envy as Random Serial Dictatorship (RSD) -- a mechanism entirely blind to priorities. This stark asymptotic equivalence implies that TTC effectively fails to satisfy standard fairness criteria in large markets, casting significant doubt on its practical appeal for balancing efficiency and fairness.

Figures

Figures reproduced from arXiv: 2607.10819 by Olivier Tercieux, Yeon-Koo Che.

Figure 1
Figure 1. Figure 1: Difference between TTC and RSD. 3 Model We consider a market with a set of agents (I) and a set of objects (O). The market may be unbalanced, meaning the number of agents and objects need not be equal. A matching is a map µ : I → O ∪ {ø} such that i ̸= i ′ implies that either µ(i) ̸= µ(i ′ ) or µ(i) = µ(i ′ ) = ø, where ø denotes being unmatched. In Section 5, we will focus on large markets, assuming 9 [P… view at source ↗
Figure 2
Figure 2. Figure 2: Markov dynamics of TTC in a balanced market ( [PITH_FULL_IMAGE:figures/full_fig_p013_2.png] view at source ↗
Figure 4
Figure 4. Figure 4: Incidences of justified envy among incidences of envy. [PITH_FULL_IMAGE:figures/full_fig_p022_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Difference between TTC and RSD Interestingly, convergence is restored if we evaluate performance using an alternative metric: the ratio of agents who experience justified envy out of all agents who experience any form of envy.23 Panel (b) of [PITH_FULL_IMAGE:figures/full_fig_p023_5.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

243 extracted references

  1. [1]

    2019 , eprint=

    Monotone Comparative Statics without Lattices , author=. 2019 , eprint=

    2019

  2. [2]

    American Economic Review , volume=

    Obviously strategy-proof mechanisms , author=. American Economic Review , volume=. 2017 , publisher=

    2017

  3. [3]

    2023 , publisher=

    Strategyproofness-Exposing Descriptions of Matching Mechanisms , author=. 2023 , publisher=

    2023

  4. [4]

    John William Hatfield and Fuhito Kojima , title =

  5. [5]

    American Economic Journal: Microeconomics , Volume =

    Dogan, Battal and Ehlers, Lars , Title =. American Economic Journal: Microeconomics , Volume =. 2022 , Month =

    2022

  6. [6]

    Gabriel Carroll , title =

  7. [7]

    Sophie Bade , title =

  8. [8]

    Combe, Julien and Dur, Umut and Tercieux, Olivier and Terrier, Camille and Unver, Utku , title =

  9. [9]

    The Quarterly Journal of Economics , volume=

    The dynamics of reorganization in matching markets: A laboratory experiment motivated by a natural experiment , author=. The Quarterly Journal of Economics , volume=. 2000 , publisher=

    2000

  10. [10]

    2010 , institution=

    Contract design and stability in matching markets , author=. 2010 , institution=

    2010

  11. [11]

    2023 , institution=

    Minimalist Market Design: A Framework for Economists with Policy Aspirations , author=. 2023 , institution=

    2023

  12. [12]

    International Economic Review , volume=

    Course Bidding at Business Schools , author=. International Economic Review , volume=. 2010 , publisher=

    2010

  13. [13]

    Yan Chen and Onur Kesten , title =

  14. [14]

    Theoretical Economics , volume =

    Yeon-Koo Che and Olivier Tercieux , title =. Theoretical Economics , volume =

  15. [15]

    2014 , author =

    Matching mechanisms and matching quality: Evidence from a top university in China , journal =. 2014 , author =

    2014

  16. [16]

    Strategy-Proofness Made Simpler , journal =

    Katu. Strategy-Proofness Made Simpler , journal =

  17. [17]

    Pathak and Alvin E

    Atila Abdulkadiroglu and Parag A. Pathak and Alvin E. Roth and Tayfun S\". Changing the Boston School Choice Mechanism , note =

  18. [18]

    Gabrielle Fack and Julien Grenet and Yinghua He , title =

  19. [19]

    Hoffman, AJ and Kruskal, JB , journal=

  20. [20]

    2009 , note =

    A Field Study on Matching with Network Externalities , author=. 2009 , note =

    2009

  21. [21]

    1974 , publisher=

    Topology and normed spaces , author=. 1974 , publisher=

    1974

  22. [22]

    Lov. Ann. Discrete Math , year=

  23. [23]

    2011 , publisher=

    Real analysis with economic applications , author=. 2011 , publisher=

    2011

  24. [24]

    American Economic Journal: Microeconomics , volume=

    Assignment messages and exchanges , author=. American Economic Journal: Microeconomics , volume=. 2009 , publisher=

    2009

  25. [25]

    College Admission with Affirmative Action , journal = IJGT, volume = 33, pages =

    Atila Abdulkadiro. College Admission with Affirmative Action , journal = IJGT, volume = 33, pages =

  26. [26]

    Pathak and Alvin E

    Atila Abdulkadiroglu and Parag A. Pathak and Alvin E. Roth and Tayfun Sonmez , title =

  27. [27]

    and Moldovanu, B

    Jehiel, P. and Moldovanu, B. , title =. The Proceedings of the 9th World Congress of the Econometric Society , publisher =

  28. [28]

    Pathak and Alvin E

    Atila Abdulkadiroglu and Parag A. Pathak and Alvin E. Roth , title =

  29. [29]

    Management Science , volume=

    The fair and efficient division of the Winsor family silver , author=. Management Science , volume=. 1990 , publisher=

    1990

  30. [30]

    Clayton Featherstone , title =

  31. [31]

    Random Serial Dictatorship and the Core from Random Endowments in House Allocation Problems , journal = EMA, volume = 66, pages =

    Atila Abdulkadiroglu and Tayfun S\". Random Serial Dictatorship and the Core from Random Endowments in House Allocation Problems , journal = EMA, volume = 66, pages =

  32. [32]

    House Allocation with Existing Tenants , journal = JET, volume = 88, pages =

    Atila Abdulkadiro. House Allocation with Existing Tenants , journal = JET, volume = 88, pages =

  33. [33]

    Ordinal Efficiency and Dominated Sets of Assignments , journal = JET, volume = 112, pages =

    Atila Abdulkadiro. Ordinal Efficiency and Dominated Sets of Assignments , journal = JET, volume = 112, pages =

  34. [34]

    A supply and demand framework for two-sided matching markets , author=

  35. [35]

    Unpublished mimeo, University of Chicago and University of Pennsylvania , year=

    Strategy-proofness in the Large , author=. Unpublished mimeo, University of Chicago and University of Pennsylvania , year=

  36. [36]

    2012 , note =

    Complementarity and Multidimensional Heterogeneity in Matching Markets , author=. 2012 , note =

    2012

  37. [37]

    Azevedo, Eduardo M and Weyl, E Glen and White Alexander , title =

  38. [38]

    Atila Abdulkadiroglu and Tayfun Sonmez , title =

  39. [39]

    Atila Abdulkadiroglu and Yeon-Koo Che and Yosuke Yasuda , title =

  40. [40]

    Jose Alcade and Salvador Barbera , title =

  41. [41]

    Implementation of stable solutions to the marriage problem , journal = JET, volume = 69, pages =

    Jos\'. Implementation of stable solutions to the marriage problem , journal = JET, volume = 69, pages =

  42. [42]

    2006 , publisher=

    Infinite Dimensional Analysis: A Hitchhiker's Guide , author=. 2006 , publisher=

    2006

  43. [43]

    Ahmet Alkan and David Gale , title =

  44. [44]

    Georgy Artemov , title =

  45. [45]

    1977 , publisher=

    Real Analysis and Probability , author=. 1977 , publisher=

    1977

  46. [46]

    Susan Athey , title =

  47. [47]

    Frontiers of Theoretical Economics , volume = 1, pages =

    Larry Ausubel and Paul Milgrom , title =. Frontiers of Theoretical Economics , volume = 1, pages =

  48. [48]

    The Importance of Irrelevance of Rejected Contracts in Matching under Weakened Substitutes Conditions , note =

    Orhan Ayg\". The Importance of Irrelevance of Rejected Contracts in Matching under Weakened Substitutes Conditions , note =

  49. [49]

    Atila Abdulkadiroglu and Yeon-Koo Che and Parag Pathak and Alvin Roth and Olivier Tercieux , title =

  50. [50]

    Marketing Science , volume = 15, pages =

    The Value of Purchase History Data in Target Marketing , author=. Marketing Science , volume = 15, pages =

  51. [51]

    Many-to-many matching: stable polyandrous polygamy (or polygomous polyandry) , journal = DAM, volume = 101, pages =

    Mourad Ba\". Many-to-many matching: stable polyandrous polygamy (or polygomous polyandry) , journal = DAM, volume = 101, pages =

  52. [52]

    A tale of two mechanisms: student placement , journal = JET, volume = 84, pages =

    Michel Balinski and Tayfun S\". A tale of two mechanisms: student placement , journal = JET, volume = 84, pages =

  53. [53]

    Dirk Bergemann and Stephen Morris , title =

  54. [54]

    Information Acquisition and Efficient Mechanism Design , journal = EMA, volume = 70, pages =

    Dirk Bergemann and Juuso V\". Information Acquisition and Efficient Mechanism Design , journal = EMA, volume = 70, pages =

  55. [55]

    Allocating Indivisible Goods , journal =

    Ivona Bez\'. Allocating Indivisible Goods , journal =

  56. [56]

    Sushil Bikhchandani , title =

  57. [57]

    Garrett Birkhoff , title =. Revi. Univ. Nac. Tucuman, ser A , volume = 5, pages =

  58. [58]

    Blum and Alvin Roth and Uriel Rothblum , title =

    Y. Blum and Alvin Roth and Uriel Rothblum , title =

  59. [59]

    Random Matching Under Dichotomous Preferences , journal = EMA, volume = 72, pages =

    Anna Bogolomania and Herv\'. Random Matching Under Dichotomous Preferences , journal = EMA, volume = 72, pages =

  60. [60]

    A New Solution to the Random Assignment Problem , journal = JET, volume = 100, pages =

    Anna Bogomolnaia and Herv\'. A New Solution to the Random Assignment Problem , journal = JET, volume = 100, pages =

  61. [61]

    A Simple Random Assignment Problem with a Unique Solution , journal = ET, volume = 19, pages =

    Anna Bogomolnaia and Herv\'. A Simple Random Assignment Problem with a Unique Solution , journal = ET, volume = 19, pages =

  62. [62]

    Bela Bollobas , title =

  63. [63]

    , journal=

    Ghouila-Houri, A. , journal=. Caract

  64. [64]

    Border , title =

    Kim C. Border , title =

  65. [65]

    Brams and Alan D

    Steven J. Brams and Alan D. Taylor , title =

  66. [66]

    Eric Budish , title =

  67. [67]

    Eric Budish and Estelle Cantillon , title =

  68. [68]

    Eric Budish and Yeon-Koo Che and Fuhito Kojima and Paul Milgrom , title =

  69. [69]

    Jeremy Bulow and Jonathan Levin , title =

  70. [70]

    The Handbook of Experimental Economics , publisher =

    Colin Camerer , title =. The Handbook of Experimental Economics , publisher =

  71. [71]

    The American Economic Review , pages=

    Stability in supply chain networks , author=. The American Economic Review , pages=. 2008 , publisher=

    2008

  72. [72]

    Archishman Chakraborty and Alessandro Citanna and Michael Ostrovsky , title =

  73. [73]

    Yeon-Koo Che and Ian Gale , title =

  74. [74]

    Yeon-Koo Che and Daniele Condorelli and Jinwoo Kim , title =

  75. [75]

    Yeon-Koo Che and Ian Gale and Jinwoo Kim , title =

  76. [76]

    Yeon-Koo Che and Youngwoo Koh , title =

  77. [77]

    and Levin, Henry M

    Corcoran, Sean P. and Levin, Henry M. , title =. Education Reform in New York City: Ambitious Change in the Nation's Most Complex School System , publisher =

  78. [78]

    Yeon-Koo Che and Fuhito Kojima , title =

  79. [79]

    Yeon-Koo Che and Jinwoo Kim and Fuhito Kojima , title =

  80. [80]

    Journal of Political Economy , volume=

    Efficiency and stability in large matching markets , author=. Journal of Political Economy , volume=. 2019 , publisher=

    2019

Showing first 80 references.