pith. sign in

arxiv: 2511.00764 · v2 · submitted 2025-11-02 · 🧮 math.PR · q-fin.RM

Further Developments on Stochastic Dominance for Convex Combinations of Infinite-Mean Random Variables

Keyi Zeng , Zhenfeng Zou , Yuting Su , Taizhong Hu This is my paper

Pith reviewed 2026-05-18 02:01 UTC · model grok-4.3

classification 🧮 math.PR q-fin.RM
keywords stochastic dominanceinfinite meanheavy-tailed distributionscompound binomialconvex combinationsfirst-order stochastic ordernonnegative random variables
0
0 comments X

The pith

For specific classes of infinite-mean distributions, convex combinations of iid nonnegative random variables are stochastically larger than any single one.

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

This paper studies stochastic dominance for convex combinations of iid nonnegative random variables with infinite means. It examines the properties and inclusion relations among previously identified distribution classes that include many heavy-tailed laws. The analysis extends earlier results on when weighted sums dominate individual variables in first-order stochastic dominance. A central part derives necessary and sufficient conditions for the dominance to be preserved when each variable follows a compound binomial distribution. The work matters for settings like risk analysis where infinite means make moment-based comparisons impossible and orderings must rely on tail behavior instead.

Core claim

The paper establishes that for random variables belonging to certain classes of nonnegative distributions with infinite means, any convex combination is larger than a single variable in the first-order stochastic dominance order. It systematically investigates the properties and inclusion relationships among these classes and extends prior results to more practical scenarios. For the case in which each random variable follows a compound binomial distribution, necessary and sufficient conditions are given under which the stochastic dominance relation continues to hold.

What carries the argument

The classes of nonnegative infinite-mean distributions for which first-order stochastic dominance is preserved under convex combinations of iid copies.

If this is right

  • The dominance relation extends to compound binomial distributions under explicit parameter conditions.
  • Inclusion relations among the distribution classes permit identification of additional members that satisfy the property.
  • The results apply to more practical modeling scenarios involving compound distributions.
  • Stochastic comparisons become possible without requiring finite moments.

Where Pith is reading between the lines

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

  • In portfolio or insurance contexts the ordering implies that spreading exposure across such variables does not produce a smaller outcome in stochastic size.
  • The classes could serve as a basis for comparing extreme-value models where conventional expectations diverge.
  • Similar dominance checks might be explored for other dependence structures or for higher-order stochastic orders.

Load-bearing premise

The random variables are independent and identically distributed nonnegative members of the identified infinite-mean distribution classes.

What would settle it

A concrete counterexample in which a convex combination of iid copies from one of the classes fails to be stochastically larger than a single copy, or a compound binomial instance that violates the derived necessary and sufficient conditions.

Figures

Figures reproduced from arXiv: 2511.00764 by Keyi Zeng, Taizhong Hu, Yuting Su, Zhenfeng Zou.

Figure 1
Figure 1. Figure 1: The function η1(x) Example 3.3. (H ⊊ V). Consider a distribution function F such that F(x) = 0 for x < 1, and F(x) = 1 − 3(1/x − 1)2 + 1 x , x ≥ 1. Denote g(x) = xF(x). Then g(x) = x for x ∈ [0, 1], and g(x) = 3(1/x − 1)2 + 1 for x > 1. It is easy to see that g ′ (x) = 6 x 2  1 − 1 x  ≥ 0, x ≥ 1, implying g(x) is increasing in x ∈ (1, ∞). Thus, F ∈ V. Denote η(x) = F(1/x). Then η(x) = 3x 3 −6x 2 + 4x for… view at source ↗
Figure 2
Figure 2. Figure 2: The function η2(x) Example 3.5 (G ̸⊂ H). Let F be a Log-Cauchy distribution, that is, F(x) = arctan(log x) π + 1 2 , x ∈ R++. Then the density function of F is f(x) = 1 πx[1 + (log x) 2] , x ∈ R++. According to [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Venn diagram illustrating the relationships among the classes [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Venn diagram illustrating the relationships among four classes [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗
read the original abstract

In recent years, stochastic dominance for independent and identically distributed (iid) infinite-mean random variables has received considerable attention. The literature has identified several classes of distributions of nonnegative random variables that encompass many common heavy-tailed distributions. A key result demonstrates that the weighted sum of iid random variables from these classes is stochastically larger than any individual random variable in the sense of the first-order stochastic dominance. This paper systematically investigates the properties and inclusion relationships among these distribution classes, and extends some existing results to more practical scenarios. Furthermore, we analyze the case where each random variable follows a compound binomial distribution, establishing necessary and sufficient conditions for the preservation of the aforementioned stochastic dominance relation.

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

1 major / 2 minor

Summary. The paper extends prior results on first-order stochastic dominance for weighted sums of iid nonnegative infinite-mean random variables. It first catalogs inclusion relations among the relevant distribution classes and then derives necessary and sufficient conditions under which the dominance relation is preserved when each summand follows a compound binomial distribution, relying on the standard tail-integral characterization of stochastic dominance together with the infinite-mean property.

Significance. If the central claims hold, the work provides useful extensions to the theory of stochastic orders for heavy-tailed infinite-mean distributions, including explicit conditions for the compound binomial case that may aid applications in risk management and insurance. The systematic cataloging of distribution classes and use of standard characterizations strengthen the foundation for further research in this area.

major comments (1)
  1. §4, Theorem 4.1: the necessity part of the condition for preservation of FSD under compound binomial summands is stated in terms of the tail behavior parameter; however, the proof sketch appears to assume a specific form of the compounding distribution that is not explicitly verified to be without loss of generality for the infinite-mean regime.
minor comments (2)
  1. The notation for the distribution classes (e.g., the symbols used for the regularly varying and subexponential classes) is introduced without a consolidated table; adding one would improve readability when comparing inclusion relations.
  2. Several references to prior work on infinite-mean stochastic dominance are cited but the precise statements of the baseline theorems being extended could be restated more explicitly in the introduction for self-contained reading.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the careful reading of our manuscript and the constructive comment. We address the major comment below.

read point-by-point responses

  1. Referee: §4, Theorem 4.1: the necessity part of the condition for preservation of FSD under compound binomial summands is stated in terms of the tail behavior parameter; however, the proof sketch appears to assume a specific form of the compounding distribution that is not explicitly verified to be without loss of generality for the infinite-mean regime.

    Authors: We thank the referee for this observation. The necessity condition in Theorem 4.1 is formulated in terms of the tail behavior parameter, which is the key quantity governing the infinite-mean regime via the tail-integral characterization of first-order stochastic dominance. The proof considers a representative compounding distribution for the compound binomial case. While this form is chosen for notational simplicity, the argument extends to general compounding distributions because the infinite-mean property ensures that the relevant tail integrals are determined solely by the tail behavior parameter, independent of additional details of the compounding law (provided the compounding distribution has positive mass on the positive integers). To make this generality explicit, we will add a short clarifying remark in the revised version of §4. This constitutes a partial revision. revision: partial

Circularity Check

0 steps flagged

No significant circularity identified

full rationale

The paper catalogs inclusion relations among distribution classes for nonnegative infinite-mean RVs and derives necessary and sufficient conditions for preservation of first-order stochastic dominance under compound binomial sums. These steps rest on the standard tail-integral characterization of stochastic dominance together with the infinite-mean property, both external to the present work. References to prior results on the same topic exist but are not load-bearing; the central claims retain independent mathematical content and do not reduce by construction to fitted parameters, self-definitions, or self-citation chains. The derivation is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on standard mathematical properties of first-order stochastic dominance and the definition of the distribution classes for infinite-mean nonnegative random variables.

axioms (1)
  • standard math Standard properties of first-order stochastic dominance for nonnegative random variables.
    The paper relies on the definition and properties of stochastic orders which are standard in probability theory.

pith-pipeline@v0.9.0 · 5645 in / 1078 out tokens · 41723 ms · 2026-05-18T02:01:41.858847+00:00 · methodology

discussion (0)

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

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

29 extracted references · 29 canonical work pages

  1. [1]

    and Samaniego, F.J

    Amiri, L., Khaledi, B.E. and Samaniego, F.J. (2011). On skewness and dispersion among convolutions of independent gamma random variables.Probability in the Engineering and Informational Sciences,25, 55-69. 29

    work page 2011

  2. [2]

    Arab, I., Lando, T., and Oliveira, P. E. (2025). Convex combinations of random variables stochastically dom- inate the parent for a large class of heavy-tailed distributions.Electronic Communications in Probability, 30, articl no. 65, 1-11

    work page 2025

  3. [3]

    and Proschan, F

    Barlow, R.E. and Proschan, F. (1981).Statistical Theory of Reliability and Life Testing. To Begin With, Silver Spring, MD

    work page 1981

  4. [4]

    and Perlman, M.D

    Bock, M.E., Diaconis, P., Huffer, F.W. and Perlman, M.D. (1987). Inequalities for linear combinations of Gamma random variables.The Canadian Journal of Statistics,15, 387-395

    work page 1987

  5. [5]

    (1997).Actuarial Mathematics, 2nd ed

    Bowers, N.L., Gerber, H.U., Hickman, J.C., Jones, D.A., Nesbitt, C.J. (1997).Actuarial Mathematics, 2nd ed. The Society of Actuaries, Illinois

    work page 1997

  6. [6]

    and Zou, Z

    Chen, Y., Hu, T., Shneer, S. and Zou, Z. (2025b). Stochastic dominance for linear combinations of infinite- mean risks.arXiv: 2505.01739

    work page arXiv

  7. [7]

    and Shneer, S

    Chen, Y. and Shneer, S. (2025). Risk aggregation and stochastic dominance for a class of heavy-tailed distributions.Astin Bulletin, doi:10.1017/asb.2025.10053

    work page doi:10.1017/asb.2025.10053 2025

  8. [8]

    and Wang, R

    Chen, Y. and Wang, R. (2025). Infinite-mean models in risk management: Discussions and recent advances. Risk Sciences,1, 100003

    work page 2025

  9. [9]

    Dhaene, J., Denuit, M., Goovaerts, M.J., Kaas, R., Vyncke, D. (2002). The concept of comonotonicity in actuarial science and finance: theory.Insurance: Mathematics and Economics,31, 3-33

    work page 2002

  10. [10]

    and Walkup, D.W

    Esary, J.D., Proschan, F. and Walkup, D.W. (1967). Association of random variables with applications. Annals of Mathematical Statistics,44, 1466-1474

    work page 1967

  11. [11]

    and Straumann, D

    Embrechts, P., McNeil, A. and Straumann, D. (2002) Correlation and dependence in risk management: Properties and pitfalls. InRisk Management: Value at Risk and Beyond(eds. Dempster), pp. 176-223. Cambridge University Press

    work page 2002

  12. [12]

    (2005).New majorization theory in economics and martingale convergence results in econo- metrics

    Ibragimov, R. (2005).New majorization theory in economics and martingale convergence results in econo- metrics. Ph.D. dissertation, Yale University, New Haven, CT

    work page 2005

  13. [13]

    Ibragimov, R. (2009). Portfolio diversification and value at risk under thick-tailedness.Quantitative Finance, 9(5), 565-580. 30

    work page 2009

  14. [14]

    and Kato, T

    Imamura, Y. and Kato, T. (2025). A note on subadditivity of value at risks (VaRs): a new connection to comonotonicity.Journal od Applied Probability, doi: 10.1017/jpr.2025.31

    work page doi:10.1017/jpr.2025.31 2025

  15. [15]

    (2009).An introduction to the theory of functional equations and inequalities: Cauchy’s equation and Jensen’s inequality

    Kuczma, M. (2009).An introduction to the theory of functional equations and inequalities: Cauchy’s equation and Jensen’s inequality. Katowice: Silesian University Press. Springer

    work page 2009

  16. [16]

    Ma, C. (1998). On peakedness of distributions of convex combinations.Journal of Statistical Planning and Inference,70, 51-56

    work page 1998

  17. [17]

    Ma, C. (2000). Convex orders for linear combinations of random variables.Journal of Statistical Planning and Inference,84(1-2), 11-25

    work page 2000

  18. [18]

    and Hu, T

    Mao, T., Pan, X. and Hu, T. (2013). On orderings between weighted sums of random variables.Probability in the Engineering and Informational Sciences,27(1):85-97

    work page 2013

  19. [19]

    and Arnold, B

    Marshall, A.W., Olkin, I. and Arnold, B. (2011).Inequalities: Theory of Majorization and Its Applications. Second edition. Springer, New York

    work page 2011

  20. [20]

    Matkowski, J. (1993). Subadditive functions and a relaxation of the homogeneity condition of seminorms. Proceedings of the American Mathematical Society,117(4), 991-1001

    work page 1993

  21. [21]

    and ´Swi¸ atkowski, T

    Matkowski, J. and ´Swi¸ atkowski, T. (1993). On subadditive functions.Proceedings of the American Mathe- matical Society,119(1), 187-197. M¨ uller, A. (2025). Some remarks on the effect of risk sharing and divesification for infinite mean risk.Astin Bulletin, doi:10.1017/asb.2025.10054

    work page doi:10.1017/asb.2025.10054 1993

  22. [22]

    and Hu, T

    Pan, X., Xu, M. and Hu, T. (2013). Some inequalities of linear combinations of independent random variables: II.Bernoulli,19(5A):1776-1789

    work page 2013

  23. [23]

    Proschan, F. (1965). Peakedness of distributions of convex combinations.The Annals of Mathematical Statis- tics,36(6):1703-1706

    work page 1965

  24. [24]

    Rosenbaum, R. A. (1950). Sub-additive functions.Duke Mathematical Journal,17, 227-247

    work page 1950

  25. [25]

    and Shanthikumar, J

    Shaked, M. and Shanthikumar, J. G. (2007).Stochastic Orders. Springer, New York

    work page 2007

  26. [26]

    Vincent, L. (2025). Diversification and stochastic dominance: When all eggs are better put in one basket. arXiv: 2507.16265v2

    work page arXiv 2025

  27. [27]

    and Hu, T

    Xu, M. and Hu, T. (2011). Some inequalities of linear combinations of independent random variables: I. Journal of Applied Probability,48(4):1179-1188

    work page 2011

  28. [28]

    Yu, Y. (2011). Some stochastic inequalities for weighted sums.Bernoulli,17(3):1044-1053

    work page 2011

  29. [29]

    (1964).Convex Transformations of Random Variables

    Zwet, W.R. (1964).Convex Transformations of Random Variables. Mathematisch Centrum Amsterdam. 31

    work page 1964