pith. sign in

arxiv: 2312.04827 · v4 · pith:F3GHYQPUnew · submitted 2023-12-08 · 💰 econ.TH

A Separability Foundation for Random Coefficients Logit

Fedor Sandomirskiy , Po Hyun Sung , Omer Tamuz , Ben Wincelberg This is my paper

Pith reviewed 2026-05-24 05:29 UTC · model grok-4.3

classification 💰 econ.TH
keywords stochastic choicerandom coefficients logitseparabilitymonotonicitycontinuitydecision theorymenu choiceaxiomatic characterization
0
0 comments X

The pith

A consistency condition on choice in separable menus, together with monotonicity and continuity, characterizes the random coefficients logit family.

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

The paper establishes a separability consistency condition for stochastic choice: when a decision problem consists of two separable components, the choice probabilities must match those obtained by considering each component separately. Combined with monotonicity and continuity, this condition uniquely identifies the family of random coefficients logit rules. A reader would care because the result supplies an axiomatic basis for these models rather than treating them as an ad hoc functional form. The central object is the requirement that probabilities remain invariant under decomposition of menus into independent parts.

Core claim

We study stochastic choice across decision problems, each represented as a menu of action labels paired with observable outcome vectors. We propose a consistency condition for behavior in decision problems composed of two separable components: choice probabilities must agree with those obtained when each component is considered in isolation. Together with monotonicity and continuity, this separability requirement characterizes the family of random coefficients logit rules.

What carries the argument

The separability consistency condition requiring that choice probabilities in a menu of two separable components match the probabilities from each component considered in isolation.

If this is right

  • Random coefficients logit is the only stochastic choice rule that satisfies separability, monotonicity, and continuity.
  • Choice probabilities in any menu can be recovered from the probabilities in its separable sub-components.
  • Any rule outside the random coefficients logit family must violate the separability condition on some composed menu.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Researchers could test the model by presenting subjects with menus that combine known separable components and checking whether probabilities match the isolated cases.
  • Similar consistency conditions might be used to characterize other families of discrete choice models beyond logit.
  • In empirical applications the result supplies a direct way to check whether the maintained model is consistent with separability in the observed menus.

Load-bearing premise

That choice probabilities must stay the same when a menu is formed by combining two separable components versus examining each component alone.

What would settle it

Choice data from menus built from two separable components where the observed probabilities in the combined menu differ from those recorded when each component appears by itself, while monotonicity and continuity still hold.

read the original abstract

We study stochastic choice across decision problems, each represented as a menu of action labels paired with observable outcome vectors. We propose a consistency condition for behavior in decision problems composed of two separable components: choice probabilities must agree with those obtained when each component is considered in isolation. Together with monotonicity and continuity, this separability requirement characterizes the family of random coefficients logit rules.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

0 major / 3 minor

Summary. The manuscript studies stochastic choice over menus of action labels paired with observable outcome vectors. It introduces a separability consistency condition requiring that choice probabilities in a decision problem composed of two separable components coincide with those obtained when each component is evaluated in isolation. The central result is a characterization theorem: separability, together with monotonicity and continuity, is necessary and sufficient for the choice rule to belong to the family of random coefficients logit rules.

Significance. If the characterization holds, the paper supplies an axiomatic foundation for the random coefficients logit model that is widely used in empirical work on discrete choice with unobserved heterogeneity. The separability axiom is a natural consistency requirement that may be testable in data and clarifies the behavioral content of the model without relying on parametric assumptions beyond the logit form. The result is of interest to both decision theorists and applied econometricians working on identification of random-coefficient specifications.

minor comments (3)
  1. The notation for menus and outcome vectors is introduced in §2 but the precise definition of 'separable components' is only stated informally in the abstract; a formal definition should appear before the statement of the main theorem.
  2. Theorem 1 (or whichever number contains the characterization) would benefit from an explicit statement of the domain of menus and the codomain of the choice rule to make the continuity assumption fully precise.
  3. The paper should include a short discussion or example showing how the separability condition fails for a non-logit random-coefficients specification, to illustrate necessity.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for the supportive report, accurate summary of the contribution, and recommendation of minor revision. No specific major comments were raised in the report.

Circularity Check

0 steps flagged

No circularity: standard axiomatic characterization

full rationale

The paper states a characterization result in which a consistency condition called separability, together with monotonicity and continuity, is shown to be necessary and sufficient for random coefficients logit. This is a direct mathematical derivation from explicitly stated axioms to the target functional form; the abstract and title present no fitted parameters renamed as predictions, no self-citation load-bearing steps, and no self-definitional reductions. The derivation chain is therefore self-contained against external benchmarks and receives the default non-circularity finding.

Axiom & Free-Parameter Ledger

0 free parameters · 3 axioms · 0 invented entities

Based solely on the abstract, the central claim rests on three conditions: the newly proposed separability consistency, plus monotonicity and continuity. No free parameters or invented entities are mentioned.

axioms (3)
  • domain assumption Separability consistency: choice probabilities in a composite menu equal those from isolated components
    The key consistency condition proposed in the abstract
  • standard math Monotonicity
    Standard background assumption in stochastic choice theory
  • standard math Continuity
    Standard background assumption in stochastic choice theory

pith-pipeline@v0.9.0 · 5582 in / 1210 out tokens · 26551 ms · 2026-05-24T05:29:27.784914+00:00 · methodology

discussion (0)

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

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Learning to Coordinate over Networks with Bounded Rationality

    eess.SY 2026-04 unverdicted novelty 7.0

    Bounded-rational agents on networks reach higher stationary coordination probability when connectivity is uniform, with the probability increasing in rationality parameter beta and the partition function approximable ...

Reference graph

Works this paper leans on

49 extracted references · 49 canonical work pages · cited by 1 Pith paper

  1. [1]

    Agranov and P

    M. Agranov and P. Ortoleva. Stochastic choice and preferences for randomization. Journal of Political Economy, 125 0 (1): 0 40--68, 2017

    work page 2017

  2. [2]

    Agranov and P

    M. Agranov and P. Ortoleva. Revealed preferences for randomization: An overview. In AEA Papers and Proceedings, volume 112, pages 426--430. American Economic Association 2014 Broadway, Suite 305, Nashville, TN 37203, 2022

    work page 2014

  3. [3]

    S. P. Anderson, A. De Palma, and J.-F. Thisse. Discrete choice theory of product differentiation. MIT press, 1992

    work page 1992

  4. [4]

    Azrieli, C

    Y. Azrieli, C. P. Chambers, and P. J. Healy. Incentives in experiments: A theoretical analysis. Journal of Political Economy, 126 0 (4): 0 1472--1503, 2018

    work page 2018

  5. [5]

    Barberis, M

    N. Barberis, M. Huang, and R. H. Thaler. Individual preferences, monetary gambles, and stock market participation: A case for narrow framing. American economic review, 96 0 (4): 0 1069--1090, 2006

    work page 2006

  6. [6]

    G. M. Becker, M. H. Degroot, and J. Marschak. Probabilities of choices among very similar objects: An experiment to decide between two models. Behavioral science, 8 0 (4): 0 306--311, 1963

    work page 1963

  7. [7]

    J. Besag. Spatial interaction and the statistical analysis of lattice systems. Journal of the Royal Statistical Society: Series B (Methodological), 36 0 (2): 0 192--225, 1974

    work page 1974

  8. [8]

    H. D. Block and J. Marschak. Random orderings and stochastic theories of response. 1959

    work page 1959

  9. [9]

    M. K. Camara. Computationally tractable choice. In Proceedings of the 23rd ACM Conference on Economics and Computation, pages 28--28, 2022

    work page 2022

  10. [10]

    Cerreia-Vioglio, D

    S. Cerreia-Vioglio, D. Dillenberger, P. Ortoleva, and G. Riella. Deliberately stochastic. American Economic Review, 109 0 (7): 0 2425--2445, 2019

    work page 2019

  11. [11]

    Cerreia-Vioglio, P

    S. Cerreia-Vioglio, P. O. Lindberg, F. Maccheroni, M. Marinacci, and A. Rustichini. A canon of probabilistic rationality. Journal of Economic Theory, 196: 0 105289, 2021 a

    work page 2021

  12. [12]

    Cerreia-Vioglio, F

    S. Cerreia-Vioglio, F. Maccheroni, M. Marinacci, and A. Rustichini. Axiomatic tests for the boltzmann distribution. Journal of Statistical Mechanics: Theory and Experiment, 2021 0 (1): 0 013406, 2021 b

    work page 2021

  13. [13]

    Cerreia-Vioglio, F

    S. Cerreia-Vioglio, F. Maccheroni, M. Marinacci, and A. Rustichini. Multinomial Logit Processes and Preference Discovery: Inside and Outside the Black Box . The Review of Economic Studies, 08 2022. ISSN 0034-6527. doi:10.1093/restud/rdac046. URL https://doi.org/10.1093/restud/rdac046. rdac046

    work page doi:10.1093/restud/rdac046 2022

  14. [14]

    Chamberlin and J

    E. Chamberlin and J. Wolfe. Multiplicative homomorphisms of matrices. Proceedings of the American Mathematical Society, 4 0 (1): 0 37--42, 1953

    work page 1953

  15. [15]

    C. P. Chambers, Y. Masatlioglu, and C. Turansick. Correlated choice. arXiv preprint arXiv:2103.05084, 2021

    work page arXiv 2021

  16. [16]

    G. Debreu. Review of individual choice behavior by R D L uce. American Economic Review, 50 0 (1): 0 186--188, 1960

    work page 1960

  17. [17]

    Echenique, K

    F. Echenique, K. Saito, and G. Tserenjigmid. The perception-adjusted luce model. Mathematical Social Sciences, 93: 0 67--76, 2018

    work page 2018

  18. [18]

    Falmagne

    J.-C. Falmagne. A representation theorem for finite random scale systems. Journal of Mathematical Psychology, 18 0 (1): 0 52--72, 1978

    work page 1978

  19. [19]

    R. P. Feynman, R. B. Leighton, and M. Sands. The Feynman lectures on physics, Vol. I: The new millennium edition: mainly mechanics, radiation, and heat, volume 1. Basic books, 2011

    work page 2011

  20. [20]

    G. B. Folland. Real analysis: modern techniques and their applications, volume 40. John Wiley & Sons, 1999

    work page 1999

  21. [21]

    Frick, R

    M. Frick, R. Iijima, and T. Strzalecki. Dynamic random utility. Econometrica, 87 0 (6): 0 1941--2002, 2019

    work page 1941

  22. [22]

    Fudenberg and T

    D. Fudenberg and T. Strzalecki. Dynamic logit with choice aversion. Econometrica, 83 0 (2): 0 651--691, 2015

    work page 2015

  23. [23]

    F. Gul, P. Natenzon, and W. Pesendorfer. Random choice as behavioral optimization. Econometrica, 82 0 (5): 0 1873--1912, 2014

    work page 1912

  24. [24]

    D. H. Hyers. On the stability of the linear functional equation. Proceedings of the National Academy of Sciences, 27 0 (4): 0 222--224, 1941

    work page 1941

  25. [25]

    D. H. Hyers and T. M. Rassias. Approximate homomorphisms. Aequationes mathematicae, 44 0 (2-3): 0 125--153, 1992

    work page 1992

  26. [26]

    S.-M. Jung. Hyers-Ulam-Rassias stability of functional equations in nonlinear analysis, volume 48. Springer Science & Business Media, 2011

    work page 2011

  27. [27]

    Kashaev, C

    N. Kashaev, C. Gauthier, and V. H. Aguiar. Dynamic and stochastic rational behavior. arXiv preprint arXiv:2302.04417, 2023

    work page arXiv 2023

  28. [28]

    M. Kuczma. An introduction to the theory of functional equations and inequalities. (No Title), 2009

    work page 2009

  29. [29]

    L. D. Landau and E. M. Lifshitz. Statistical Physics: Volume 5, volume 5. Elsevier, 1951

    work page 1951

  30. [30]

    R. Li. Dynamic random choice. arXiv preprint arXiv:2102.00143, 2021

    work page arXiv 2021

  31. [31]

    R. D. Luce. Individual Choice Behavior: A theoretical analysis. John Wiley and Sons, Inc, 1959

    work page 1959

  32. [32]

    R. D. Luce and P. Suppes. Preferences, utility and subjective probability. Handbook of mathematical psychology, pages 249--410, 1965

    work page 1965

  33. [33]

    M. J. Machina. Stochastic choice functions generated from deterministic preferences over lotteries. The economic journal, 95 0 (379): 0 575--594, 1985

    work page 1985

  34. [34]

    Mat e jka and A

    F. Mat e jka and A. McKay. Rational inattention to discrete choices: A new foundation for the multinomial logit model. American Economic Review, 105 0 (1): 0 272--298, 2015

    work page 2015

  35. [35]

    L. Mattner. Cumulants are universal homomorphisms into hausdorff groups. Probability theory and related fields, 130: 0 151--166, 2004

    work page 2004

  36. [36]

    Mattsson and J

    L.-G. Mattsson and J. W. Weibull. Probabilistic choice and procedurally bounded rationality. Games and Economic Behavior, 41 0 (1): 0 61--78, 2002

    work page 2002

  37. [37]

    McFadden

    D. McFadden. Conditional logit analysis of qualitative choice behavior. Frontiers in Econometrics, pages 105--142, 1974

    work page 1974

  38. [38]

    McFadden

    D. McFadden. Economic choices. American economic review, 91 0 (3): 0 351--378, 2001

    work page 2001

  39. [39]

    McFadden and M

    D. McFadden and M. K. Richter. Stochastic rationality and revealed stochastic preference. Preferences, Uncertainty, and Optimality, Essays in Honor of Leo Hurwicz, Westview Press: Boulder, CO, pages 161--186, 1990

    work page 1990

  40. [40]

    McFadden, W

    D. McFadden, W. B. Tye, and K. Train. An application of diagnostic tests for the independence from irrelevant alternatives property of the multinomial logit model. Institute of Transportation Studies, University of California Berkeley, 1977

    work page 1977

  41. [41]

    R. D. McKelvey and T. R. Palfrey. Quantal response equilibria for normal form games. Games and economic behavior, 10 0 (1): 0 6--38, 1995

    work page 1995

  42. [42]

    K. Saito. Axiomatizations of the mixed logit model. 2018

    work page 2018

  43. [43]

    C. E. Shannon. A mathematical theory of communication. The Bell system technical journal, 27 0 (3): 0 379--423, 1948

    work page 1948

  44. [44]

    C. E. Shannon. Coding theorems for a discrete source with a fidelity criterion. IRE Nat. Conv. Rec, 4 0 (142-163): 0 1, 1959

    work page 1959

  45. [45]

    C. A. Sims. Implications of rational inattention. Journal of monetary Economics, 50 0 (3): 0 665--690, 2003

    work page 2003

  46. [46]

    Steiner, C

    J. Steiner, C. Stewart, and F. Mat e jka. Rational inattention dynamics: Inertia and delay in decision-making. Econometrica, 85 0 (2): 0 521--553, 2017

    work page 2017

  47. [47]

    K. E. Train. Discrete choice methods with simulation. Cambridge university press, 2009

    work page 2009

  48. [48]

    R. Webb. The (neural) dynamics of stochastic choice. Management Science, 65 0 (1): 0 230--255, 2019

    work page 2019

  49. [49]

    Woodford

    M. Woodford. Stochastic choice: An optimizing neuroeconomic model. American Economic Review, 104 0 (5): 0 495--500, 2014

    work page 2014