Concave Rationalization with an Ideal Point: An Afriat Theorem and an Application to Survey Design

Avner Seror

arxiv: 2401.03876 · v2 · submitted 2024-01-08 · 💰 econ.TH

Concave Rationalization with an Ideal Point: An Afriat Theorem and an Application to Survey Design

Avner Seror This is my paper

Pith reviewed 2026-05-24 04:51 UTC · model grok-4.3

classification 💰 econ.TH

keywords Afriat theoremideal pointconcave utilityrationalizationsingle-peaked preferencesconsistency indexsurvey designrevealed preference

0 comments

The pith

A system of Afriat inequalities with the unknown peak as a virtual observation characterizes concave rationalization with an ideal point.

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

The paper shows that choices from linear budget sets anchored at different corners can be explained by a continuous concave utility with an unknown ideal point exactly when a modified system of Afriat inequalities holds, treating the peak as an extra observation that delivers the highest utility. A stronger version of the inequalities requires that supergradients at each choice point toward the peak, which is needed for single-peaked behavior. The resulting characterization yields a consistency index that finds the largest share of data jointly explainable by one ideal point and supplies the foundation for a priced survey method in which the same questions are asked under different budget constraints to recover ideal answers and weights.

Core claim

A system of Afriat inequalities - where the unknown peak enters as a virtual observation with the highest utility - is necessary and sufficient for the existence of a continuous concave utility with an ideal point that rationalizes choices from linear budget sets anchored at different corners of the choice space. A stronger characterization adds the requirement that supergradients at observed choices point coordinatewise toward the peak, a necessary condition for single-peaked rationalizability.

What carries the argument

The peak-oriented Afriat system, in which the ideal point is inserted as a virtual observation with maximal utility.

If this is right

The peak-oriented system yields a Houtman-Maks consistency index that measures the largest fraction of observations jointly rationalizable with a common ideal point.
The characterization supplies the theoretical foundation for the Priced Survey Methodology in which respondents face the same questions under different linear constraints.
A parametric single-peaked specification then produces estimates of ideal answers and importance weights.
The method is applied to political preferences in a sample of French respondents.

Where Pith is reading between the lines

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

The same inequality system could be used to test single-peaked rationalizability in experimental choice data outside surveys.
Varying budget constraints in other elicitation settings might improve recovery of ideal points when preferences are known to be concave.
The consistency index offers a way to compare the prevalence of ideal-point behavior across different populations or question formats.

Load-bearing premise

Choices are generated from linear budget sets anchored at different corners of the choice space.

What would settle it

A dataset of choices from such budget sets that satisfies the peak-oriented inequalities yet cannot be rationalized by any continuous concave utility with an ideal point would falsify sufficiency.

Figures

Figures reproduced from arXiv: 2401.03876 by Avner Seror.

**Figure 2.** Figure 2: (a) Round 1 (b) Round 2 (c) Round 3 “buying” goods with their budget when they move from the default [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗

**Figure 3.** Figure 3: Revealed Preferences in the PSM One key aspect of the PSM is that ≥ does not constitute an exogenous pre-order of the set of possible survey answers. Indeed, respondents have ideal points when answering surveys, so the axiomatization of choice used in the consumer choice environment cannot be applied. My working assumption is that since q 0 never belongs to the choice sets after round 0, in each coordinate… view at source ↗

**Figure 4.** Figure 4: Dataset satisfying GARP (left panel) Indifference curves for single-peaked utility [PITH_FULL_IMAGE:figures/full_fig_p013_4.png] view at source ↗

**Figure 5.** Figure 5: Four regions comprehensive if nonempty in the coordinate system with origin c. Indeed, u c reaches its minimum value for xc = 0, so if set r c is nonempty, it necessarily includes x ∈ X such that xc = 0. Remark 1 For any x ∈ X, there exists a region rc such that x ∈ rc. If x ∈ int(rc), then x /∈ rw for any w ∈ C(X) \ c. This remark means that X = S c∈C(X) rc and that the different regions do not overlap in… view at source ↗

**Figure 6.** Figure 6: Examples of budget sets 4One alternative consists in reducing the available budgets in round k when q 0 i,ok is too close from o k . This option is however not possible when q 0 i = c for some c ∈ C(X). 5For each pair of rounds starting from a given corner, there are two symmetric price vectors: p k = (1, 2), and p l = (2, 1) for k, l ∈ {1, . . . , 8}. 20 [PITH_FULL_IMAGE:figures/full_fig_p020_6.png] view at source ↗

**Figure 7.** Figure 7: Indifference curves for Respondent 65 (left panel) and Respondent 94 (right [PITH_FULL_IMAGE:figures/full_fig_p024_7.png] view at source ↗

**Figure 8.** Figure 8: Correlation between b1 and q 0 1 (upper panel), and b2 and q 0 2 (lower panel). 27 [PITH_FULL_IMAGE:figures/full_fig_p027_8.png] view at source ↗

read the original abstract

This paper develops an Afriat-type characterization of concave rationalization with an unknown ideal point. We show that a system of Afriat inequalities - where the unknown peak enters as a virtual observation with the highest utility - is necessary and sufficient for the existence of a continuous concave utility with an ideal point that rationalizes choices from linear budget sets anchored at different corners of the choice space. A stronger characterization adds the requirement that supergradients at observed choices point coordinatewise toward the peak, a necessary condition for single-peaked rationalizability. The resulting peak-oriented Afriat system provides the basis for a Houtman--Maks consistency index that measures the largest fraction of observations jointly rationalizable with a common ideal point. This characterization provides the theoretical foundation for the Priced Survey Methodology (PSM), in which respondents complete the same survey under different linear constraints. A parametric single-peaked specification then sharpens identification into estimates of ideal answers and importance weights. We apply the PSM to study political preferences in a sample of French respondents.

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

This paper adds an ideal point to Afriat via a virtual observation and uses it to justify a priced survey format, with the core characterization holding under the corner-anchored budget assumption.

read the letter

The main thing here is a characterization that treats the unknown peak as a virtual bundle with highest utility, then gives Afriat-style inequalities that are necessary and sufficient for continuous concave rationalization. That setup is new relative to standard Afriat and to most single-peaked work. The paper also shows how the same inequalities support a Houtman-Maks index and a parametric single-peaked model for the Priced Survey Methodology, which they illustrate with French political-preference data. Both the theorem and the survey application look usable if the data really come from linear budgets anchored at different corners of the choice space. The stronger supergradient condition is presented as optional, which keeps the basic result clean. The maintained assumption on how the data are generated is restrictive, but the paper is explicit about it and designs the survey around it, so the limitation is not hidden. No internal contradictions appear in the stated logic, and the application supplies a concrete check on whether the inequalities can be implemented. This is the kind of revealed-preference result that people working on preference elicitation or on single-peaked models would want to see. It is worth sending to referees because the theorem is stated clearly enough to check and the survey method is a direct, testable output.

Referee Report

2 major / 1 minor

Summary. The paper develops an Afriat-type characterization for choice data generated from linear budget sets anchored at different corners of the choice space. It shows that a modified system of Afriat inequalities, treating an unknown ideal point as a virtual observation with the highest utility level, is necessary and sufficient for rationalization by a continuous concave utility attaining its maximum at that point. A stronger version requires that supergradients at observed points point coordinatewise toward the peak. The characterization underpins a Houtman-Maks-style consistency index and the Priced Survey Methodology (PSM) for eliciting ideal points and weights; the paper applies PSM to French respondents' political preferences.

Significance. If the stated necessity and sufficiency hold, the result extends classical Afriat theory to concave utilities with bliss points under a specific budget-set geometry, supplying a directly testable inequality system and a practical consistency index. The PSM application converts the theorem into a survey design tool that can identify ideal answers and importance weights from repeated responses under varying linear constraints, which is potentially valuable for applied work on single-peaked preferences in political economy and consumer theory.

major comments (2)

[Abstract] Abstract and introduction: the necessity and sufficiency claim is stated without any proof sketch, error analysis, or verification steps. Because the central result is a mathematical characterization, the absence of even an outline of the argument in the provided text prevents assessment of whether the virtual-observation construction correctly captures both directions of the equivalence.
[Abstract (paragraph describing the rationalization setting)] The theorem is derived under the maintained data-generating process that all observations come from linear budget sets anchored at different corners. If this geometry does not hold, the necessity direction fails; the paper should state explicitly whether the result is intended only for this class of budgets or whether it extends more generally.

minor comments (1)

Notation for the virtual peak observation and the supergradient condition should be introduced with explicit equation numbers in the main text rather than only in the abstract.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and constructive comments. We address each major comment below.

read point-by-point responses

Referee: [Abstract] Abstract and introduction: the necessity and sufficiency claim is stated without any proof sketch, error analysis, or verification steps. Because the central result is a mathematical characterization, the absence of even an outline of the argument in the provided text prevents assessment of whether the virtual-observation construction correctly captures both directions of the equivalence.

Authors: The abstract is intentionally brief, as is conventional. The full necessity and sufficiency proof, including the virtual-observation construction and verification of both directions via the Afriat inequalities, appears in Section 3. We will add a concise proof outline to the introduction to aid assessment without altering the abstract. revision: partial
Referee: [Abstract (paragraph describing the rationalization setting)] The theorem is derived under the maintained data-generating process that all observations come from linear budget sets anchored at different corners. If this geometry does not hold, the necessity direction fails; the paper should state explicitly whether the result is intended only for this class of budgets or whether it extends more generally.

Authors: The result is derived specifically for linear budget sets anchored at different corners, as stated in the abstract and Section 2; necessity uses this geometry to ensure the virtual peak observation interacts correctly with the budgets. We will insert an explicit clarifying sentence in the introduction stating that the characterization applies to this budget class and does not claim generality beyond it. revision: yes

Circularity Check

0 steps flagged

No significant circularity; mathematical characterization is self-contained

full rationale

The paper states a necessity-and-sufficiency theorem extending Afriat's inequalities to concave utilities with an unknown ideal point, where the peak enters as a virtual observation. This is a direct mathematical equivalence between the modified inequality system and the existence of a continuous concave rationalizing utility under the maintained data-generating process of linear budget sets anchored at corners. No step reduces a prediction or result to a fitted parameter, self-definition, or load-bearing self-citation; the theorem is proved from first principles in revealed-preference theory without importing uniqueness results or ansatzes from the authors' prior work. The Houtman-Maks index and PSM application follow from the characterization rather than presupposing it.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Ledger constructed from abstract only; full paper may contain additional assumptions or parameters.

axioms (2)

domain assumption Utility function is continuous and concave
Invoked as the target class of functions in the characterization.
domain assumption Observed choices arise from linear budget sets anchored at different corners
Stated as the data-generating environment for which the inequalities are necessary and sufficient.

pith-pipeline@v0.9.0 · 5709 in / 1300 out tokens · 26828 ms · 2026-05-24T04:51:33.399261+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

3 extracted references · 3 canonical work pages

[1]

x ≽R y and yok ≥ zok ⇒ x ≽R z

work page
[2]

x ≽R y and yok > z ok ⇒ x ≻R z

work page
[3]

Assume that ( ≽R ∪ ≥ , >) is not acyclic and let Q =≽R ∪ ≥

x ≻R y and yok ≥ zok ⇒ x ≻R y. Assume that ( ≽R ∪ ≥ , >) is not acyclic and let Q =≽R ∪ ≥. There exists a sequence of observations in D such that x1Q . . . QxL and xL oL > x 1 o1. Without loss of generality, the cycle can be rewritten as x1 o1 ≥ x2 o1, x2 o2 ≥ x3 o2, x3 o3 ≥ x4 o3, . . . xp−1 op−1 ≥ xp op−1, xp ≽R · · · ≽R xL, and xL oL > x 1 oL. 33 If p ...

work page 2016

[1] [1]

x ≽R y and yok ≥ zok ⇒ x ≽R z

work page

[2] [2]

x ≽R y and yok > z ok ⇒ x ≻R z

work page

[3] [3]

Assume that ( ≽R ∪ ≥ , >) is not acyclic and let Q =≽R ∪ ≥

x ≻R y and yok ≥ zok ⇒ x ≻R y. Assume that ( ≽R ∪ ≥ , >) is not acyclic and let Q =≽R ∪ ≥. There exists a sequence of observations in D such that x1Q . . . QxL and xL oL > x 1 o1. Without loss of generality, the cycle can be rewritten as x1 o1 ≥ x2 o1, x2 o2 ≥ x3 o2, x3 o3 ≥ x4 o3, . . . xp−1 op−1 ≥ xp op−1, xp ≽R · · · ≽R xL, and xL oL > x 1 oL. 33 If p ...

work page 2016