arxiv: 2605.04771 · v1 · submitted 2026-05-06 · 💰 econ.EM

Recognition: unknown

It's complicated: A Non-parametric Test of Preference Stability between Singles and Couples

Stefan Hubner

Pith reviewed 2026-05-08 16:25 UTC · model grok-4.3

classification 💰 econ.EM

keywords preference stabilityrevealed preferencecollective householdnon-parametric testconsumption datahousehold formation

0 comments

The pith

A non-parametric test rejects the assumption that singles and couples have stable preferences using consumption data alone.

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

The paper develops a method to test whether preferences stay the same when people form couples, working only with observed consumption choices of singles and couples and without needing information on how resources are split inside households or how couples form. It shows that stable preferences would require the separate choice patterns of singles and couples to fit together as if drawn from a shared random-utility structure over all possible pairings. This requirement is violated in three large panel datasets, so the stability hypothesis is rejected. The finding matters because many models of taxes, welfare, and family decisions treat preferences as fixed across marital status.

Core claim

Under the preference-stability hypothesis, consumption choices from an endogenously matched population admit a conditional random-utility representation over counterfactual pairings of couples and singles. Preference stability is testable as a feasibility restriction on the observed marginal choice distributions, and this restriction is violated in consumption data from the Dutch LISS, Russian RLMS, and Spanish ECPF panels.

What carries the argument

The conditional random-utility representation over counterfactual pairings of couples and singles, which converts the stability hypothesis into a feasibility restriction on marginal choice distributions.

If this is right

Singles' data can be used for revealed-preference analysis of couples only when the feasibility restriction holds.
Rejection implies that collective-household models assuming fixed preferences across marital status are misspecified.
The test works without observing intra-household allocations or marriage transitions.
Policy predictions based on stable preferences between singles and couples require revision.

Where Pith is reading between the lines

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

Rejection may reflect bargaining or coordination effects that alter effective preferences after coupling.
The same feasibility approach could be applied to check stability around other transitions such as parenthood.
If instability is common, structural models should allow preferences to depend on household composition rather than treat them as primitive.

Load-bearing premise

Under stable preferences, consumption choices from an endogenously matched population of singles and couples can be represented as conditional random utilities over all possible counterfactual pairings.

What would settle it

A new panel dataset in which the observed marginal distributions of singles' and couples' choices satisfy the feasibility restriction derived from the conditional random-utility model would fail to reject stability.

Figures

Figures reproduced from arXiv: 2605.04771 by Stefan Hubner.

**Figure 1.** Figure 1: Block structure of N ˆ N extended assignment matrix (l.h.s.) with finite sample counterparts (r.h.s.). Factual couples Zeus & Hera Counterfactual couples (τ f) Zeus & Athena Counterfactual single men (τm) Zeus Counterfactual couples (τm) Apollo & Hera Counterfactual couples (τm, τ f ) Apollo & Athena Factual single men Apollo Counterfactual s. women (τ f) Hera Factual single women Athena dummy couples N2 N… view at source ↗

**Figure 2.** Figure 2: Illustration of individual choices under two different budgets. view at source ↗

read the original abstract

This paper develops a method to use singles' data in a non-parametric revealed preference setting of collective household choice. We use it to test the controversial assumption of preference stability between singles and couples, without data on intra-household allocation or marital transitions. We show that, under the preference-stability hypothesis, consumption choices from an endogenously matched population admit a conditional random-utility representation over counterfactual pairings of couples and singles. Preference stability is testable as a feasibility restriction on the observed marginal choice distributions. We reject the hypothesis using consumption data from the Dutch LISS, the Russian RLMS, and the Spanish ECPF panels.

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.

Desk Editor's Note private letter to a colleague

New non-parametric test rejects preference stability between singles and couples, but the derivation under endogenous matching needs scrutiny.

read the letter

The main takeaway is that the paper develops a non-parametric test for whether preferences remain stable when people move from being single to being in a couple, using only the marginal consumption distributions from singles and couples. It rejects the stability hypothesis in data from the Netherlands, Russia, and Spain. What the paper does well is to embed the collective household revealed preference conditions into a conditional random utility representation over all possible counterfactual pairings. This turns the stability assumption into a check on whether the observed marginals are feasible under some joint distribution, without needing data on who is paired with whom or how resources are allocated inside the household. That's a practical step forward for applied work that often lacks those details. The potential issue is whether this representation actually follows from preference stability once you allow for endogenous matching. The stress test raises a fair point: if the same unobserved heterogeneity that affects matching also influences the utilities, then the matching probabilities aren't independent of the shocks in the way the model seems to assume. The paper claims the conditional representation holds, but I'd want to see the explicit steps showing that the endogeneity doesn't break the feasibility restriction. The rejections are interesting, but without more on data cleaning, how the test statistic is computed, and checks for robustness to different specifications, it's hard to gauge how strong the evidence is. This kind of work is aimed at researchers in household economics who use revealed preference or collective models. Someone looking for new ways to test maintained assumptions with limited data would find it useful. It deserves serious refereeing because the method is original and the empirical results challenge a standard assumption, even if the derivation needs careful verification. Recommendation: Yes, send it out for review with attention to the matching endogeneity concern.

Referee Report

2 major / 2 minor

Summary. The paper develops a non-parametric revealed-preference test of preference stability between singles and couples in a collective-household model. It shows that, under the stability hypothesis, observed marginal consumption distributions of singles and couples must satisfy a feasibility restriction derived from embedding collective revealed-preference conditions inside a conditional random-utility representation over counterfactual pairings. The authors reject the stability hypothesis on consumption data from the Dutch LISS, Russian RLMS, and Spanish ECPF panels.

Significance. If the derivation is valid and the empirical implementation is robust, the paper supplies a practical, data-light way to test a core maintained assumption in household economics without requiring intra-household allocation data or observed marital transitions. A credible rejection across three distinct panels would imply that standard models assuming stable preferences across marital status may be misspecified, with consequences for equivalence-scale estimation, welfare analysis, and policy counterfactuals. The non-parametric character and multi-country application are clear strengths.

major comments (2)

[theoretical derivation of the feasibility restriction] The central theoretical claim (the section deriving the conditional random-utility representation and the associated feasibility restriction) does not supply an explicit argument that the representation survives endogenous matching. The construction treats matching probabilities as separable from the utility shocks that rationalize choices, yet the same stable preferences govern both; no proof is given that the observed marginals remain feasible when this separability fails. This is load-bearing for the validity of the test.
[empirical application and results] Empirical section reporting the rejections: the manuscript states that the hypothesis is rejected on the three panels but supplies no details on sample construction, consumption aggregation, or the precise numerical check of the feasibility restriction (e.g., how the joint support over counterfactual pairings is discretized or how measurement error is handled). Without these, it is impossible to verify that the reported rejections are not artifacts of data-processing choices.

minor comments (2)

[theoretical setup] Notation for the marginal distributions of singles versus couples could be made more explicit to avoid confusion when the feasibility condition is stated.
[abstract] The abstract would benefit from a brief statement of sample sizes or number of observations per panel.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed comments. We address each major comment below and describe the revisions we will implement.

read point-by-point responses

Referee: [theoretical derivation of the feasibility restriction] The central theoretical claim (the section deriving the conditional random-utility representation and the associated feasibility restriction) does not supply an explicit argument that the representation survives endogenous matching. The construction treats matching probabilities as separable from the utility shocks that rationalize choices, yet the same stable preferences govern both; no proof is given that the observed marginals remain feasible when this separability fails. This is load-bearing for the validity of the test.

Authors: We acknowledge that the current manuscript does not contain an explicit lemma establishing that the conditional random-utility representation and resulting feasibility restriction continue to hold when matching probabilities are permitted to depend on the same utility shocks that rationalize observed choices. The derivation in the paper integrates over the endogenous matching distribution induced by stable preferences, but the argument would be strengthened by a formal demonstration that the marginal feasibility condition is preserved without separability. We will add such a lemma to the theoretical section in the revision. revision: yes
Referee: [empirical application and results] Empirical section reporting the rejections: the manuscript states that the hypothesis is rejected on the three panels but supplies no details on sample construction, consumption aggregation, or the precise numerical check of the feasibility restriction (e.g., how the joint support over counterfactual pairings is discretized or how measurement error is handled). Without these, it is impossible to verify that the reported rejections are not artifacts of data-processing choices.

Authors: We agree that the empirical section lacks sufficient detail for full verification. The revised manuscript will expand this section with explicit descriptions of sample construction for each panel (Dutch LISS, Russian RLMS, Spanish ECPF), the aggregation of consumption items, the discretization procedure used for the joint support over counterfactual pairings, and the treatment of measurement error together with associated robustness checks. Supplementary materials containing the relevant code and summary statistics will also be provided. revision: yes

Circularity Check

0 steps flagged

No significant circularity; feasibility restriction derived from hypothesis and tested on independent data

full rationale

The paper states that under the preference-stability hypothesis, choices admit a conditional random-utility representation, making stability testable as a feasibility restriction on observed marginal distributions. This restriction is then applied to reject the hypothesis using consumption data from the Dutch LISS, Russian RLMS, and Spanish ECPF panels. No equation or step reduces a claimed prediction to a fitted input by construction, nor does any load-bearing claim rely on self-citation chains or imported uniqueness theorems. The derivation is self-contained against external benchmarks, with the empirical rejection using data independent of the theoretical construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the revealed-preference framework for collective households and the assumption that stable preferences imply a conditional random-utility representation; no free parameters or invented entities are described in the abstract.

axioms (1)

domain assumption Consumption choices from an endogenously matched population admit a conditional random-utility representation over counterfactual pairings under preference stability
Invoked to turn preference stability into a testable feasibility restriction on marginal choice distributions.

pith-pipeline@v0.9.0 · 5389 in / 1152 out tokens · 29783 ms · 2026-05-08T16:25:12.006007+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

75 extracted references · 1 canonical work pages

[1]

Becker , journal =

Gary S. Becker , journal =. A Theory of the Allocation of Time , volume =
[2]

1981 , doi =

Becker, Gary S , journal =. 1981 , doi =

1981
[3]

W. M. Gorman , booktitle =. Tricks with Utility Functions , year =
[4]

Taxation and the Household , volume =

Patricia Apps and Ray Rees , journal =. Taxation and the Household , volume =
[5]

1994 , doi =

Browning, Martin and Bourguignon, François and Chiappori, Pierre-André and Lechene, Valérie , journal =. 1994 , doi =

1994
[6]

Browning, Martin and Chiappori, Pierre-André , journal =
[7]

2006 , doi =

Chiappori, Pierre-André and Ekeland, Ivar , journal =. 2006 , doi =

2006
[8]

2009 , doi =

Chiappori, Pierre-André and Ekeland, Ivar , journal =. 2009 , doi =

2009
[9]

2012 , doi =

Chiappori, Pierre-André and Donni, Olivier and Komunjer, Ivana , journal =. 2012 , doi =

2012
[10]

2013 , url =

Browning, Martin and Chiappori, Pierre-André and Lewbel, Arthur , journal =. 2013 , url =

2013
[11]

2008 , doi =

Lewbel, Arthur and Pendakur, Krishna , journal =. 2008 , doi =

2008
[12]

Identification of Semiparametric Model Coefficients, with an Application to Collective Households , volume =

Arthur Lewbel and Xirong Lin , journal =. Identification of Semiparametric Model Coefficients, with an Application to Collective Households , volume =. 2022 , url =

2022
[13]

Inefficient Collective Households: Cooperation and Consumption , volume =

Arthur Lewbel and Krishna Pendakur , journal =. Inefficient Collective Households: Cooperation and Consumption , volume =. 2022 , doi =

2022
[14]

Is Shelter Really a Public Good Within Households? , year =

Lewbel, Arthur and Pendakur, Krishna , note =. Is Shelter Really a Public Good Within Households? , year =
[15]

2014 , doi =

Mazzocco, Maurizio and Ruiz, Claudia and Yamaguchi, Shintaro , journal =. 2014 , doi =

2014
[16]

2015 , doi =

Voena, Alessandra , journal =. 2015 , doi =

2015
[17]

Gayle, George-Levi and Shephard, Andrew , journal =
[18]

Commitment and the Dynamics of Household Labour Supply , volume =

Alexandros Theloudis and Jorge Velilla and Pierre-Andr. Commitment and the Dynamics of Household Labour Supply , volume =. The Economic Journal , month =. 2025 , doi =

2025
[19]

Marriage, Labor Supply and the Dynamics of the Social Safety Net , year =

Hamish Low and Costas Meghir and Luigi Pistaferri and Alessandra Voena , institution =. Marriage, Labor Supply and the Dynamics of the Social Safety Net , year =
[20]

2009 , url =

Chiappori, Pierre-André and Donni, Olivier , journal =. 2009 , url =

2009
[21]

2020 , doi =

Chiappori, Pierre-André , journal =. 2020 , doi =

2020
[22]

2018 , doi =

Kitamura, Yuichi and Stoye, Jörg , journal =. 2018 , doi =

2018
[23]

1967 , url =

Afriat, Sydney , journal =. 1967 , url =

1967
[24]

1982 , url =

Varian, Hal , journal =. 1982 , url =

1982
[25]

2007 , url =

Cherchye, Laurens and De Rock, Bram and Vermeulen, Frederic , journal =. 2007 , url =

2007
[26]

Cherchye, Laurens and De Rock, Bram and Vermeulen, Frederic , journal =
[27]

2011 , doi =

Cherchye, Laurens and De Rock, Bram and Vermeulen, Frederic , journal =. 2011 , doi =

2011
[28]

2012 , url =

Cherchye, Laurens and De Rock, Bram and Vermeulen, Frederic , journal =. 2012 , url =

2012
[29]

2014 , doi =

Adams, Abi and Cherchye, Laurens and De Rock, Bram and Verriest, Ewout , journal =. 2014 , doi =

2014
[30]

2016 , doi =

Brugler, James , journal =. 2016 , doi =

2016
[31]

1997 , doi =

Banks, James and Blundell, Richard and Lewbel, Arthur , journal =. 1997 , doi =

1997
[32]

2023 , doi =

Hubner, Stefan , journal =. 2023 , doi =

2023
[33]

Identification of time-varying transformation models with fixed effects, with an application to unobserved heterogeneity in resource shares , volume =

Botosaru, Irene and Muris, Chris and Pendakur, Krishna , journal =. Identification of time-varying transformation models with fixed effects, with an application to unobserved heterogeneity in resource shares , volume =
[34]

Economies of Scale to Consumption in Collective Households , year =

Hsin-Yuan Hsieh and Arthur Lewbel and Krishna Pendakur , journal =. Economies of Scale to Consumption in Collective Households , year =
[35]

Intra-Household Welfare Inequality and Household Public Goods , year =

Pierre-Andr. Intra-Household Welfare Inequality and Household Public Goods , year =. Cowles Foundation Discussion Paper , month =
[36]

2013 , url =

Dunbar, Geoffrey and Lewbel, Arthur and Pendakur, Krishna , journal =. 2013 , url =

2013
[37]

2021 , isbn =

Dunbar, Geoffrey and Lewbel, Arthur and Pendakur, Krishna , journal =. 2021 , isbn =

2021
[38]

2022 , url =

Lechene, Valérie and Pendakur, Krishna and Wolf, Alex , journal =. 2022 , url =

2022
[39]

Why Are Older Women Missing in India? The Age Profile of Bargaining Power and Poverty , volume =

Rossella Calvi , journal =. Why Are Older Women Missing in India? The Age Profile of Bargaining Power and Poverty , volume =. 2020 , doi =

2020
[40]

Individual Consumption in Collective Households: Identification Using Repeated Observations with an Application to PROGRESA , volume =

Senay Sokullu and Christine Valente , journal =. Individual Consumption in Collective Households: Identification Using Repeated Observations with an Application to PROGRESA , volume =. 2022 , doi =

2022
[41]

2014 , doi =

Hoderlein, Stefan and Stoye, Jörg , journal =. 2014 , doi =

2014
[42]

and Stoye, Jörg , journal =

Deb, Rahul and Kitamura, Yuichi and Quah, John K.H. and Stoye, Jörg , journal =
[43]

1991 , url =

McFadden, Daniel and Richter, Marcel , booktitle =. 1991 , url =

1991
[44]

2005 , url =

McFadden, Daniel , journal =. 2005 , url =

2005
[45]

Nonparametric Analysis of Random Utility Models: Computational Tools for Statistical Testing , volume =

Smeulders, Bart and Cherchye, Laurens and De Rock, Bram , journal =. Nonparametric Analysis of Random Utility Models: Computational Tools for Statistical Testing , volume =. 2021 , doi =

2021
[46]

and Kashaev, Nail , journal =

Aguiar, Victor H. and Kashaev, Nail , journal =. Stochastic Revealed Preferences with Measurement Error , volume =. 2021 , doi =

2021
[47]

A Dual Approach to Nonparametric Characterization for Random Utility Models , year =

Nobuo Koida and Koji Shirai , journal =. A Dual Approach to Nonparametric Characterization for Random Utility Models , year =
[48]

An Alternative Approach for Nonparametric Analysis of Random Utility Models , year =

Christopher Turansick , journal =. An Alternative Approach for Nonparametric Analysis of Random Utility Models , year =
[49]

1971 , url =

Hurwicz, Leonid and Uzawa, Hirofumi , booktitle =. 1971 , url =

1971
[50]

, edition =

Rotman, Joseph J. , edition =. An Introduction to the Theory of Groups , year =
[51]

2021 , url =

Galichon, Alfred , number =. 2021 , url =

2021
[52]

1973 , doi =

Becker, Gary S , journal =. 1973 , doi =

1973
[53]

Foundations of Modern Probability , year =

Kallenberg, Olav , publisher =. Foundations of Modern Probability , year =. doi:10.1007/978-1-4757-4015-8 , isbn =

work page doi:10.1007/978-1-4757-4015-8
[54]

Funzione caratteristica di un fenomeno aleatorio , volume =

de Finetti, Bruno , journal =. Funzione caratteristica di un fenomeno aleatorio , volume =
[55]

, journal =

Hewitt, Edwin and Savage, Leonard J. , journal =. Symmetric Measures on Cartesian Products , volume =. 1955 , doi =

1955
[56]

1988 , url =

Chiappori, Pierre-André , journal =. 1988 , url =

1988
[57]

1992 , url =

Chiappori, Pierre-André , journal =. 1992 , url =

1992
[58]

Non-parametric Analysis of Time-Inconsistent Preferences , volume =

Laura Blow and Martin Browning and Ian Crawford , journal =. Non-parametric Analysis of Time-Inconsistent Preferences , volume =. 2021 , doi =

2021
[59]

1971 , doi =

Shapley, Lloyd Stowell and Shubik, Martin , journal =. 1971 , doi =

1971
[60]

Duality and Modern Economics , year =

Cornes, Richard , publisher =. Duality and Modern Economics , year =
[61]

The Recoverability of Consumers' Preferences from Market Demand Behavior , volume =

Andreu Mas-Colell , journal =. The Recoverability of Consumers' Preferences from Market Demand Behavior , volume =. 1977 , url =

1977
[62]

On Revealed Preference Analysis , volume =

Andreu Mas-Colell , journal =. On Revealed Preference Analysis , volume =. 1978 , doi =

1978
[63]

, journal =

Andrews, Donald W.K. , journal =. 2000 , doi =

2000
[64]

2005 , url =

Franc, Vojtech and Hlavac, Vaclav and Navara, Mirko , journal =. 2005 , url =

2005
[65]

2006 , url =

Johansson, Bjoern and Elfving, Tommy and Kozlov, Vladimir and Censor, Yair and Forssen, Per-Erik and Granlund, Goesta , journal =. 2006 , url =

2006
[66]

2011 , doi =

Beatty, Timothy and Crawford, Ian , journal =. 2011 , doi =

2011
[67]

Barten, Anton Peter , edition =
[68]

2017 , doi =

Cherchye, Laurens and Demuynck, Thomas and De Rock, Bram and Vermeulen, Frederic , journal =. 2017 , doi =

2017
[69]

Production Heterogeneity in Collective Labor Supply Models with Children , year =

Gauthier, Charles , journal =. Production Heterogeneity in Collective Labor Supply Models with Children , year =
[70]

2009 , doi =

Imbens, Guido and Newey, Whitney , journal =. 2009 , doi =

2009
[71]

and Rees, Ray , journal =

Apps, Patricia F. and Rees, Ray , journal =. Collective Labor Supply and Household Production , volume =
[72]

, number =

Kechris, Alexander S. , number =. Classical Descriptive Set Theory , year =
[73]

Optimal Transport: Old and New , year =

Villani, C\'. Optimal Transport: Old and New , year =
[74]

1959 , isbn =

Block, Henry David and Marschak, Jacob , journal =. 1959 , isbn =

1959
[75]

2016 , url =

Dette, Holger and Hoderlein, Stefan and Neumeyer, Natalie , journal =. 2016 , url =

2016