RESEARCH DOSSIER
Social Choice and Collective Preference Aggregation
Abstract. Collective preference aggregation transforms individual evaluations into collective comparisons and choices under specified institutional rules. This monograph investigates majority decision, cycles, Arrow's impossibility theorem, scoring and elimination, restricted domains, strategic incentives, individual rights and welfare information. Standard results are reconstructed under explicit assumptions, constructed examples and exhaustive finite enumeration demonstrate differences among procedures, and official electoral descriptions illustrate contrasting institutional outputs. The synthesis distinguishes social orderings, choice functions, seat allocation and welfare assessment, and explains the identification limits of ordinal information. Mathematical possibilities, observed frequencies and causal effects require different forms of evidence. No new theorem or empirical population estimate is asserted.
Keywords: Social choice; collective choice; preference aggregation; Arrow's theorem; majority rule; Condorcet cycle; voting systems; strategy-proofness; single-peaked preferences; welfare economics; individual rights; institutional design
Collective choice as an economic object
A collective economic choice determines a state of affairs that affects several persons. It may allocate a public budget, establish a cooperative's investment program, define a rule for shared infrastructure, or choose the members of a body authorized to decide on subsequent policies. The scientific object is the transformation of individual evaluations and institutionally admitted information into a collective output. Scarcity determines the feasible set; preferences express individual evaluations of its elements; the aggregation procedure establishes how those evaluations count.
An individual's coherent ranking of feasible bundles does not establish a coherent ranking for a society. In a collective decision, there are several orderings and a rule governing their interpretation. A social choice analyst must distinguish the feasible alternatives, decision units, admissible information, aggregation rule, type of output and economic consequences. The central thesis of this dossier is that a collective judgment inherits the informational and normative conditions of the procedure from which it is derived. Its welfare significance requires a separate assessment of the consequences and their distribution.
The genealogy of social choice connects Condorcet's preliminary discourse (1784/1785) to modern axiomatic analysis. The digitized discourse is catalogued as 1784, whereas the treatise as a whole is conventionally dated 1785. Arrow's 1950 article and Social Choice and Individual Values (1951; third edition 2012) provide the foundational framework for comparing logically defined requirements on social ordering. Sen's work expands the informational basis of normative analysis beyond rankings alone.
This research uses conceptual definition, mathematical reconstruction, counterexamples, finite enumeration, and a limited international institutional comparison. The examples are explicitly synthetic unless labeled as administrative evidence. The source corpus is targeted rather than exhaustive; the evidence record distinguishes primary text, abstract and bibliographic access. Neither theoretical universality nor causal effects are inferred from a case study. Economic Theory is the writing owner of ET-F02, with Mathematical Economics as a shared scientific membership. The object is ordinal aggregation and collective choice; broader research on distribution, public finance, political incentives and market design has its own scope.
Preference profiles and collective outputs
Let N be a finite set of n persons, and X a finite set of feasible alternatives. A weak preference ordering R_i on X is complete and transitive. Its strict part P_i ranks x above y when x R_i y holds and y R_i x does not. Its symmetric part represents indifference. A profile is the list of individual orderings; strict linear orders form a domain without indifference between distinct alternatives.
A social welfare function in the Arrovian sense maps admissible profiles to complete, transitive social preference relations. A social choice correspondence associates a profile with a nonempty subset of X. A resolute choice function selects a single member. A representative-allocation rule may instead return a vector of seats; a randomized rule can return a probability distribution over alternatives. These are distinct mathematical outputs. A unique winner does not entail a social ranking of all defeated alternatives, and an allocation of seats is not reducible to a winner rule over individual candidates.
List's Social Choice Theory (2022) situates these different aggregation objects within a common field. The choice of output matters when characterizing a rule. Independence of irrelevant alternatives for social comparisons is an across-profile requirement; a consistency condition on selections from different feasible menus is a separate notion. Incentive compatibility must similarly be defined with reference to the actual output and available reports.
For ordinal utility, any strictly increasing transformation of a numerical representation retains its ranking. Values 1, 2 and 3 therefore represent the same order as 1, 10 and 100. The numeric distances have no invariant significance under arbitrary increasing transformations. Independently rescaling cardinal utilities across persons can also change the comparison of sums. Interpersonal comparability depends on justified restrictions on such transformations, rather than the mere availability of numeric scores.
Write ABC for A preferred to B preferred to C. For strict ballots, the pairwise margin M(x,y) is the number ranking x above y minus the number ranking y above x. Thus M(x,y)=-M(y,x). If the number of strict voters is odd, each comparison has a strict majority; otherwise ties require explicit treatment. A majority relation ranks x above y if M(x,y)>0. It remains to determine whether this relation is transitive.
Eligibility and weights are additional inputs. Anonymity is relative to the defined units: equal weight per person, per member, per jurisdiction and per share describe different populations. A result about equal voter influence applies to the units over which anonymity is defined, not automatically to every affected person.
Binary choice and the majority principle
In a binary decision, code support for x as +1, support for y as -1 and indifference as zero. Simple majority returns the sign of the sum. With an odd number of strict voters, exactly one alternative wins. With an equal balance of opposing inputs under the trivalent convention, the collective result is a tie.
May's majority characterization (1952) identifies four requirements in a binary setting: a group decision defined over admissible inputs, anonymity, neutrality and positive responsiveness. Anonymity means that exchanging voters leaves the outcome unchanged. Neutrality means exchanging the alternatives reverses the decision. Positive responsiveness means a favorable movement in one person's comparison converts an otherwise favorable or tied decision into a favorable one.
The central symmetry is readily illustrated. For an even strict electorate, an exactly balanced profile must be a tie because swapping the alternatives is equivalent to a permutation of persons. A favorable flip from that profile must produce a strict victory; successive favorable flips preserve it. The reverse argument handles the losing side. In a trivalent domain, a balanced profile may contain indifferent persons; the same symmetry applies, and favorable changes establish the sign of the result. This reconstruction identifies the mechanism of the characterization without claiming to replace May's complete proof of independence of conditions.
Binary majority gives each formally equal voter one unit in the margin. It does not measure benefit or harm. A municipality choosing between two investment programs by majority establishes which program more voters prefer, but its net benefits and incidence require independent economic evidence.
A qualified majority changes the rule. If a proposal needs 60 percent support and receives 55 percent, the status quo survives. The rule privileges a designated default; neutrality between the proposal and the default is therefore sacrificed. This may express a legitimate institutional commitment to stability, and its desirability depends on the consequences of both adoption and nonadoption. A procedural property can be analyzed without assuming that the same property is ethically decisive in every setting.
Majority cycles and agenda dependence
Consider three voters with rankings ABC, BCA and CAB. A defeats B by 2 to 1; B defeats C by 2 to 1; C defeats A by 2 to 1. Every individual has a transitive ranking but the collective pairwise relation cycles. No alternative defeats every rival. The result is the Condorcet paradox.
In a sequential agenda using sincere pairwise comparisons, first comparing A with B gives A, which then loses to C. Beginning with B versus C yields B, which then loses to A. Beginning with C versus A yields C, which then loses to B. The same rankings therefore permit each final alternative under an appropriate sequence. This is a computation under sincere comparison, not a prediction about a strategic sequential game.
For a budget committee, the scheduling of amendments may determine which feasible compromise remains before the final vote. Equal weighting in each comparison leaves room for the agenda to affect the economic outcome. Agenda control is therefore part of the effective decision rule. A final vote does not identify what every alternative would have received under a different agenda.
An explicit synthetic probability model quantifies the cycle in a restricted setting. Three labeled voters independently select, with equal probability, one of the six strict rankings of three alternatives. This yields 216 equally likely ordered profiles. Exhaustive enumeration gives 12 cyclic profiles, or 1/18, approximately 5.56 percent. The 12 cases are the six permutations of each of the two cyclic orientations. The remaining 204 yield transitive strict majority tournaments.
The calculation is a probability under independent uniformly distributed rankings, not a rate in actual elections. Correlated preferences, aligned policy dimensions, strategic reporting and participation can alter the distribution of profiles. A single-peaked domain can exclude cycles entirely under the relevant assumptions. Three questions must remain distinct: whether a cycle is mathematically possible, its probability under an assumed or estimated distribution, and how a given institution resolves a realized cycle.
Arrow's impossibility theorem
In the finite Arrovian framework with at least two persons and three alternatives, five familiar requirements cannot all be satisfied: unrestricted domain of individual weak orderings, complete and transitive social ordering, weak Pareto, independence of irrelevant alternatives (IIA), and non-dictatorship. The formulation follows Morreau's statement of Arrow's theorem (2025), with the historical attribution to Arrow (1951/1963).
Weak Pareto means that strict unanimity on a pair implies the same strict social comparison. IIA means that if every individual retains the comparison of x with y between two profiles, the social comparison of that pair remains unchanged. Unrestricted domain allows every logically possible profile. Dictatorship means there is one fixed individual whose strict preferences over every pair always prevail socially. The theorem is about this uniform authority across all profiles, not an individual occasionally casting a pivotal vote.
An explanatory proof can be given with decisive coalitions. Say a coalition S is almost decisive for a pair x,y if society strictly prefers x to y when everyone in S strictly prefers x to y and everyone outside S strictly prefers y to x. By IIA, the social comparison then holds at every profile with that pairwise pattern. Pareto makes N almost decisive for each pair.
Field expansion uses a third distinct alternative z. Suppose S is almost decisive for x,y. Assign S the ranking x>y>z and outsiders y>z>x. Society ranks x above y by the assumption and y above z by unanimity; transitivity gives x above z. Hence S is almost decisive for x,z. Assign instead S z>x>y and outsiders y>z>x. Unanimity gives z above x, the assumption gives x above y, and transitivity yields z above y. Hence S is almost decisive for z,y. Repeated use of these two implications expands decisiveness to every ordered pair, including reversals. For instance, from x,y one gets x,z, then y,z, then y,x; the same expansions cover any additional alternatives.
Almost decisiveness then implies unconditional decisiveness. Let a,b be any pair for which S unanimously ranks a above b in a given profile, with arbitrary outside comparisons including ties. Construct another profile preserving all a-b comparisons: place a>c>b for members of S, and put c above both a and b for everyone outside S, retaining their original relative ranking of a and b. Almost decisiveness for a,c implies a above c socially, and strict unanimity gives c above b. Therefore a is socially above b in the constructed profile and, by IIA, the original one.
Select an almost-decisive coalition S of minimal size; it is nonempty by Pareto. Suppose S has two nonempty disjoint parts U and V, and T consists of outsiders. Give U x>y>z, V y>z>x, and T z>x>y. Society ranks y above z because S is decisive. If society strictly ranks x above z, U is almost decisive for x,z, as all others rank z above x. Otherwise completeness gives z at least as high as x. Transitivity with the strict y-over-z comparison yields y strictly above x. Only V ranks y above x, so V is almost decisive for y,x. Either case contradicts minimality. S is therefore a singleton, whose strict preferences prevail across profiles. Non-dictatorship fails.
The reasoning is a reconstruction of the standard decisive-coalition proof, not a new theorem. It relies on the admissibility of the constructed profiles and a complete transitive social relation. Its economic contribution is a precise identification of the trade-off: pairwise majority can violate collective ordering, scoring can violate IIA, domain restrictions exclude profiles, and richer informational frameworks alter the input structure. Nothing in the theorem makes a particular relaxation universally preferable. A proposal to use one rule must explain which requirements fit the intended decision.
Scoring elimination and pairwise methods
Plurality selects an alternative with maximal first-place count. Borda scoring for m alternatives assigns m-1 points to first place, m-2 to second, continuing to zero. A Condorcet winner defeats every other alternative in a strict pairwise majority contest. A Condorcet-consistent choice rule selects that winner whenever one exists. Instant-runoff voting eliminates an alternative with the lowest continuing first-preference count, transferring ballots to the next ranked continuing alternative according to the rules, until a winner emerges.
Constructed profile P consists of four ABC, three BCA and two CBA ballots. Plurality selects A with four first preferences. B defeats A by 5 to 4 and C by 7 to 2. Borda scores are A=8, B=12, C=7. Instant runoff eliminates C and both its ballots move to B, which wins 5 to 4. The same inputs therefore distinguish plurality from the other three rule outcomes.
Constructed profile Q consists of 40 ABC, 35 CBA and 25 BAC ballots. First preferences are A=40, C=35 and B=25. B defeats A by 60 to 40 and C by 65 to 35, but instant runoff eliminates B first, transfers its 25 ballots to A and selects A by 65 to 35. Borda scores are A=105, B=125 and C=70. This exhibits a failure of universal Condorcet consistency for instant runoff, even with complete strict ballots and no ties.
Borda's failure of IIA can be demonstrated by two profiles of five voters. In U, three voters submit ABC and two submit BCA. Scores A=6, B=7 and C=2 put B above A. In V, the same three voters submit ACB and the same two submit BAC. Now A=8, B=4 and C=3. Every A-versus-B comparison is preserved while their social score order reverses. Changing the position of C changes the score distance between A and B. The rule deliberately uses information disallowed by pairwise IIA.
A Kemeny-style ranking minimizes the number of pairwise disagreements between the collective ordering and all individual strict rankings. If a strict Condorcet winner is not first in a candidate ordering, moving it to first makes every affected pairwise term strictly better. Hence an optimal ordering starts with that winner. Cyclic profiles may yield several minimizers, in which case output multiplicity or a tie rule is required. For small alternative sets the objective can be evaluated by enumeration; large sets may require exact optimization or verified bounds.
Approval voting counts the number of voters approving each alternative. It permits several approvals. An ordinal ranking does not identify a voter's approval threshold, so two identical rankings may generate different approval ballots. Approval and ranking procedures consequently cannot be compared as if they always observe the same information.
An analyst should report exactly which information each rule uses, how it treats ties, what it returns and which properties it possesses over the stated domain. A pair of favorable examples is insufficient to establish universal superiority.
Restricted domains and the median
A common structural restriction places alternatives on a line. Each individual has a unique most preferred point, or peak; within either side of that peak, moving away strictly reduces preference. Cross-side comparisons may differ. The domain is single-peaked. Black's group decision-making study (1948) is a foundational reference for this approach.
With n=2k+1 strict voters, sort the peaks and let p be the median. For any feasible alternative x to its left, the k+1 or more voters with peaks at p or to its right rank p above x. For x to its right, the k+1 or more voters with peaks at p or to its left rank p above x. Thus p is the Condorcet winner. The derivation requires a common axis, unique peaks and strict monotonicity on each side, but does not require identical preferences across opposite sides.
For five preferred expenditure levels 10, 20, 30, 60 and 90, the median is 30. Every lower level loses to 30 among voters with peaks 30 or higher; every higher level loses among voters with peaks 30 or lower. The result identifies majority stability, not the maximum of aggregate net benefits.
Moulin's single-peaked voting analysis (1980) examines mechanisms based on reported peaks. An elementary strategy-proofness proof fixes the other 2k reports, sorted q_1 through q_2k for k at least one. The outcome of one's report t is t clamped to the interval [q_k,q_(k+1)]. If the actual peak is inside this interval, truthful reporting attains it. If it lies below, the left endpoint is best among attainable outcomes; if above, the right endpoint is best. Truthful reporting is consequently weakly dominant for this ordinary median rule. This establishes the specific result without claiming to reproduce Moulin's full characterization.
The common axis is an important economic assumption. A budget may vary by amount, composition, tax financing and geographic allocation. Agreement on spending levels may conceal disagreement about the associated distributional packages. Projecting multidimensional disagreement onto a single axis without empirical justification can manufacture an appearance of stability. The structure of the feasible alternatives and actual preferences must justify the restriction.
Strategic reporting and incentives
A rule computes an outcome from reported preferences. Individuals can have incentives to submit reports that differ from their true rankings. Strategy-proofness for a deterministic single-winner function requires truthful reporting to be weakly dominant against every profile of other reports, with outcomes evaluated by the individual's actual ranking.
Gibbard's manipulation theorem (1973) and Satterthwaite's strategy-proofness analysis (1975) establish that unrestricted strict preferences, resoluteness, onto range with at least three alternatives, and strategy-proofness imply dictatorship. The result is about a single choice and incentives; Arrow's theorem is about a social ordering and axioms. Their connections do not erase the distinction.
A four-alternative counterexample computed independently uses three voters with true rankings ABCD, ABCD and BCDA. Borda scoring 3,2,1,0 gives totals A=6, B=7, C=4, D=1; B is the unique winner. The first voter instead reports ACDB. The new totals are A=6, B=5, C=5, D=2. A is now the unique winner, which the first voter prefers to B. Neither profile involves a tie at the top. The example was selected by enumerating strict four-alternative profiles and unilateral changes.
The existence of manipulation is weaker than a claim that a given voter will manipulate. Actual behavior depends on beliefs about other reports, information, computational capacity, stakes, participation and preferences about honesty. Dominant-strategy truthfulness quantifies over every report profile, while Bayesian incentives refer to expectations under specified beliefs. Repeated interactions and coordinated coalition deviations introduce further conditions.
Randomization changes the output. Random dictatorship selects a voter according to fixed probabilities and implements that voter's top-ranked alternative. Truthful top reporting maximizes the outcome conditional on selection and has no effect otherwise, so it is individually strategy-proof under expected utility over outcomes. The realized result need not be the majority favorite. The example changes the deterministic-output premise rather than disproving the theorem.
A rule's mathematical performance on sincere rankings is therefore only one component of institutional assessment. If the reporting behavior changes when the mechanism changes, the counterfactual input profile also changes.
Welfare information and individual rights
The welfare significance of collective preference depends on admissible interpersonal information. In The Possibility of Social Choice (1999, pp. 356-357), Sen develops the relevance of interpersonal comparability and partial comparisons to constructive social welfare analysis. The informational assumptions of a welfare functional matter as much as its algebraic form.
Consider two feasible projects X and Y and three individuals. Assume, for illustration, comparable welfare changes from Y to X of +1,+1,-10. A majority prefers X, but the simple sum is -8. Retain the same preference directions and instead use +10,+10,-1. The sum is +19. Ordinal inputs are identical in the two scenarios. They identify the count of favorable comparisons but cannot identify the sign of a sum of comparable changes.
This does not establish that the utilitarian criterion should govern the choice. It establishes an identification limit conditional on that criterion. Common unit utility must be defended; otherwise the numeric sum can change with independently admissible rescalings. Distributional objectives may also depend on initial welfare levels, the incidence of gains and losses, or an explicit inequality criterion.
Partial information can nevertheless support robust comparisons. Suppose two comparable gains each lie between 3 and 5 and a loss lies between 1 and 4. The resulting sum is between 2 and 9, so its positive sign is identified under the interval assumptions. If the loss could reach 12 instead, the sum lies between -6 and 9; the sign is indeterminate. This simple construction identifies what additional evidence would resolve the assessment without demanding precise measurement of every quantity.
A Pareto improvement makes at least one individual strictly better off and leaves nobody worse off. Weak Pareto in Arrow's theorem refers specifically to strict unanimity. Pareto efficiency does not identify one uniquely fair allocation: several mutually conflicting allocations may be efficient.
Sen's Paretian liberal theorem (1970) concerns a distinct conflict between minimal personal decisiveness and a Pareto requirement. Let x mean only person 1 undertakes an activity, y only person 2, and z neither. Person 1 ranks z>x>y, and person 2 ranks x>y>z. If person 1 is decisive over x versus z and person 2 over y versus z, rights require z>x and y>z socially. Both people prefer x to y, so unanimity requires x>y. The result is a cycle x>y>z>x. This constructed instance exhibits the logical conflict without treating a welfare ranking as a substitute for rights.
An economic institution may implement personal rights as constraints on feasible alternatives, qualify a collective decision rule, or make explicit the domain on which decisions are shared. Each response alters a condition of the problem. Specification of authority cannot be inferred simply from the observed preferences of the majority.
Judgment aggregation and epistemic decisions
Preference aggregation concerns what persons rank more highly. Epistemic aggregation concerns judgments about a fact or uncertain state. The same group may perform both tasks: estimating economic consequences and selecting a policy based on their evaluation.
The elementary Condorcet jury result, reviewed by List (2022), assumes an odd number of independent judgments on a binary question with a common probability p of correctness greater than one-half. Under the model, the probability of correct majority judgment tends to one as the group grows. The argument is concentration of independent Bernoulli draws around a proportion strictly above one-half.
At p=0.6, three independent judgments give majority accuracy 3(0.6 squared)(0.4)+0.6 cubed=0.648. For five judgments the corresponding binomial sum is 0.68256. With perfectly correlated answers and marginal correctness 0.6, adding persons leaves majority accuracy at 0.6. The number of judgments does not measure the amount of independent information.
An economic forecasting panel may use overlapping data and similar models, or respond to common organizational incentives. Correlation can invalidate the independence assumption. A decision analysis should distinguish forecast quality from the normative evaluation of the consequences forecast.
Judgment aggregation also involves logical dependencies. Consider three individuals evaluating propositions P, Q, and their conjunction. Their coherent judgments are (true,true,true), (true,false,false), and (false,true,false). Separate majorities accept P and Q, but reject their conjunction. An institution must choose how to aggregate premises and conclusions, or select a consistent collective set. The problem illustrates that independent aggregation of logically interconnected questions can fail collective consistency. A coherent decision therefore may require an explicit specification of both the informational and logical structure of the task.
Institutional outputs in international comparison
Australia's House of Representatives count and New Zealand's mixed-member proportional system illuminate different institutional outputs. The purpose is comparative description, not estimation of effects on economic welfare or political behavior.
The Australian counting example reports 64,859 formal votes and an absolute-majority threshold of 32,430. The first-count tallies in the Australian Electoral Commission's table are Wilkie 13,788; Jackson 23,215; Barnes 856; Simpkins 14,688; and Couser 12,312. They sum exactly to the formal total. After transfers, the final counts are Wilkie 33,217 and Jackson 31,642, again summing to 64,859. The eventual winner begins behind the initial leader.
The winner's final share is 51.21 percent rounded to two decimals, and the vote margin is 1,575. The arithmetic is independently recomputed. The table illustrates preferential-count mechanics, but does not supply every bilateral preference comparison among the original candidates. It cannot identify the Condorcet winner by itself. The Commission also explains a distinction between the elected-candidate outcome and two-party-preferred statistics; those calculations need not name the same final pair.
The New Zealand Electoral Commission's explanation of MMP identifies two inputs: the electorate vote and party vote. The party vote largely determines a party's share of parliamentary seats, while the electorate vote concerns local representation. The official explanation specifies eligibility through at least five percent of the party vote or winning an electorate seat. This account reports the consulted administrative description, not exhaustive electoral law.
One procedure selects a particular officeholder within a district; the other combines local selection with the allocation of parliamentary representation by party. Evaluating the second requires a seat-allocation criterion, attention to thresholds and representation, and the relation between electorate seats and party entitlements. A single-winner axiomatic framework does not directly encompass all these objects.
Neither example controls for differences in electorate, institutions, economic circumstances, party systems or political history. Comparisons of economic outcomes would require additional data and a credible causal design. The narrower finding is that actual voting procedures operationalize different definitions of the collective output and different uses of the information available from voters.
Empirical identification and evaluation
A theorem identifies conditions under which formally defined properties can coexist. Empirical social choice investigates what those conditions imply over observed or plausible distributions of preferences. Regenwetter's empirical research agenda (2009) highlights the need to examine whether mathematical assumptions describe actual preference distributions and when alternative aggregation procedures agree.
Ballot format determines identification. First preferences identify plurality counts but generally not pairwise margins, Borda totals or counterfactual elimination orders. Truncated rankings reveal a subset of comparisons. Treating omitted alternatives as all equally last is an added convention, rather than a direct observation. Hypothetical surveys may observe a different behavioral object from high-stakes reported ballots.
For complete strict ballots, a reproducible study records ballot multiplicities, computes the antisymmetric pairwise margin matrix, detects a Condorcet winner or cycles, computes the specified scoring and elimination rules, and records which outputs differ. Tie conventions must be specified before comparison. An exact small-instance enumeration can validate more complex implementations.
For partial ballots, consider all completions consistent with observed comparisons. A property is identified if it holds across every admissible completion. One may instead use a statistical imputation model, but its point estimates depend on the assumptions governing unobserved comparisons. Sensitivity to alternative missingness conventions should be reported.
The denominator of any empirical cycle frequency matters: counting each election equally, weighting by turnout, or restricting to contests with specified candidate counts produces different estimands. A synthetic uniform-profile probability is a benchmark; an archive rate describes the covered cases. Neither establishes a universal probability of cycles in real democratic institutions.
Counterfactual analysis separates the mechanical effect of recounting fixed ballots from the behavioral effect of changing the institution. The latter can change candidate entry, turnout, campaigning and strategic reports. Applying a different formula to previously observed ballots identifies a conditional mechanical outcome, not necessarily the equilibrium outcome of an institutional reform. Estimation of a causal effect requires assumptions or research designs appropriate to that distinction.
The present study supplies exhaustive enumeration only for the small stated domain and one independently checked constructed manipulation example, along with administrative count arithmetic. It has not obtained microdata, estimated actual cycle prevalence or identified causal effects. A future study can apply these calculations to lawfully available anonymized profiles under a predeclared missing-data and sampling strategy.
Authorial synthesis and economic applications
The results can be integrated through a distinction between institutional construction, computation and economic assessment. Institutional construction identifies feasible alternatives, participation, authority and the admissible reports. Computation applies an aggregation rule to those inputs. Economic assessment considers implementation, costs, benefits, distribution and uncertainty. These ordinary analytical distinctions make visible the assumptions whose omission produces misleading interpretations; they do not constitute a newly named theory.
The original profile comparisons demonstrate that the same rankings can yield different winners through a change of rule, and that the same pairwise preferences can yield opposite Borda comparisons through a change in the placement of other options. The welfare illustration shows that identical majority support can correspond to opposite aggregate comparable welfare changes. These three differences concern distinct informational structures. A collective result must be interpreted with reference to its generating procedure.
For public-budget decisions, feasibility screening must determine which alternatives satisfy the resource constraint before the rule is applied. A median outcome can be stable on an expenditure axis, but the same expenditure may imply very different tax liabilities and service allocations. The economist should assess the consequences of the actual package and the validity of the common-axis approximation.
In a cooperative organization, the units of representation determine the meaning of procedural symmetry. Equal member votes, equal shares and weights tied to capital participation produce different influence structures. A separate economic analysis may consider exposure to losses, exit rights and the financing of common assets. Voting weights alone cannot establish a desirable allocation of risks or explain sustained participation.
For regulatory decisions, expert forecasts of policy effects can be combined with stakeholder judgments about trade-offs only after their meanings are specified. A score representing expected harm is epistemic evidence about an uncertain consequence; a distributional weight reflects a normative choice. Putting them into a composite without documenting normalization and scaling conceals that distinction. Robustness across defensible assumptions can identify whether a recommendation depends on a small or large change in valuations.
A social welfare assessment may use partial comparisons that support some rankings without fully resolving all possible comparisons. The set of well-supported welfare judgments is then different from a complete institutional rule that must produce a decision. A choice under uncertainty may be justified procedurally even when the welfare ordering remains incomplete. An explanation should state which of these claims it makes.
The economic significance of social choice lies in showing the conditions of legitimate inference. Binary majority expresses a symmetry between alternatives and voters; single-peaked restrictions support stable medians; scoring uses placement information; elimination uses ordered transfers; welfare measures need justified interpersonal information. Those mechanisms address distinguishable problems. Comparing them requires deciding which problem the institution is intended to solve.
The authorial contribution is the constructed profile P/Q/U/V comparison, independent finite enumeration, checked strategic and welfare counterexamples, reconstructed decisive-coalition and median arguments, official count arithmetic and integrated interpretation of identification limits. These are analytical and pedagogical contributions. The established theorems retain their historical attribution. No claim is made that the examples establish a novel theorem, a new social choice mechanism or an original general doctrine.
Scientific limitations and open questions
The source corpus combines accessible scholarly entries, selected original text extracts, publisher metadata, institutional abstracts and official administrative descriptions. Direct publisher full-text access was unavailable for several foundational studies; source-specific access limitations are recorded in the evidence ledger. The PDF page-image retrieval attempts for May and Sen failed, although text extracts were available. This research is not an exhaustive literature review and does not purport to inspect every original proof or edition.
The finite models assume fixed alternatives and persons, strict rankings in the computed examples, equal voter weights, and explicit tie assumptions. The decisive-coalition reconstruction handles weak individual orders by preserving the pairwise comparison in its extension step, but relies on full domain and a complete transitive social relation. The median result requires the common ordering of alternatives and strict single-peakedness. The cycle enumeration assumes a uniform distribution over independent rankings only in a three-person three-alternative setting.
Open empirical questions include how often realistic profiles admit a common single-peaked axis; whether different aggregation rules agree when ballots are incomplete; how much apparent reform effects derive from behavioral change rather than mechanical recounting; and which partial interpersonal comparisons can resolve specific policy judgments.
There is also a question of institutional and computational transparency. An exact method may require substantial computation as the alternative set grows. Researchers should report runtime, implementation conventions, and reproducibility of results. The choice of a tractable rule may be justified by the institution's constraints, but computational simplicity is not by itself a defensible theory of fairness.
These questions identify a continuing program in mathematical economics and economic theory. Its tasks are to clarify the inputs, prove conditional properties, measure relevant domains and evaluate the substantive economic consequences of actual decisions.
Conclusion
Collective preference aggregation transforms a plurality of individual orderings into decisions or social rankings by using explicitly defined inputs and procedures. Binary majority has a compelling symmetry characterization. Majority cycles arise when more alternatives are admitted. Arrow's theorem establishes an incompatibility among requirements for universal collective ordering, while the separate Gibbard-Satterthwaite theorem identifies limitations on deterministic strategy-proof choice. Single-peaked domains, modified outputs and richer welfare information establish possibilities by changing the assumptions under which the limitations arise.
The constructed examples isolate the differences between first-place counts, scoring, elimination, pairwise comparisons, agenda dependence and strategic behavior. Institutional examples establish the difference between selecting an officeholder and distributing representative seats. Welfare examples establish that the magnitude and distribution of consequences are not identified by ordinal support alone.
The central conclusion is that the validity of a collective economic judgment depends upon its specified domain, information, rule and output. A procedure can establish a legitimate choice within its authority while leaving the evaluation of welfare consequences open. The enduring contribution of social choice to economics is the rigorous articulation of those conditions and their implications for institutional decisions.
References
- Arrow, Kenneth J. (1951; 2012). Social Choice and Individual Values. Third edition, foreword by Eric S. Maskin. New Haven: Yale University Press. ISBN 9780300186987. Yale University Press
- Arrow, Kenneth J. (1950). A Difficulty in the Concept of Social Welfare. Journal of Political Economy, 58(4), 328-346. DOI: 10.1086/256963. Journal of Political Economy
- Morreau, Michael (2025). Arrow's Theorem. The Stanford Encyclopedia of Philosophy. Revised December 7, 2025. Stanford University. Stanford Encyclopedia of Philosophy
- List, Christian (2022). Social Choice Theory. The Stanford Encyclopedia of Philosophy. Revised October 14, 2022. Stanford University. Stanford Encyclopedia of Philosophy
- May, Kenneth O. (1952). A Set of Independent Necessary and Sufficient Conditions for Simple Majority Decision. Econometrica, 20(4), 680-684. DOI: 10.2307/1907651. University-hosted article
- Black, Duncan (1948). On the Rationale of Group Decision-making. Journal of Political Economy, 56(1), 23-34. DOI: 10.1086/256633. Journal of Political Economy
- Gibbard, Allan (1973). Manipulation of Voting Schemes: A General Result. Econometrica, 41(4), 587-601. DOI: 10.2307/1914083. Article transcription
- Satterthwaite, Mark Allen (1975). Strategy-proofness and Arrow's conditions: Existence and correspondence theorems for voting procedures and social welfare functions. Journal of Economic Theory, 10(2), 187-217. DOI: 10.1016/0022-0531(75)90050-2. Kellogg research record
- Moulin, Herve (1980). On strategy-proofness and single peakedness. Public Choice, 35(4), 437-455. RePEc article record
- Sen, Amartya (1999). The Possibility of Social Choice. American Economic Review, 89(3), 349-378. DOI: 10.1257/aer.89.3.349. University-hosted article
- Sen, Amartya (1970). The Impossibility of a Paretian Liberal. Journal of Political Economy, 78(1), 152-157. DOI: 10.1086/259614. Journal of Political Economy
- Regenwetter, Michel (2009). Perspectives on Preference Aggregation. Perspectives on Psychological Science, 4(4), 403-407. DOI: 10.1111/j.1745-6924.2009.01146.x. Journal article
- Australian Electoral Commission (n.d.). How the House of Representatives result is determined. Official counting explanation and worked example. Accessed November 17, 2026. Official counting explanation
- Electoral Commission of New Zealand (n.d.). What is MMP? Official institutional explanation. Accessed November 17, 2026. Official MMP explanation
- Condorcet, Jean-Antoine-Nicolas de Caritat, marquis de (1784/1785). Discours preliminaire, Essai sur l'application de l'analyse a la probabilite des decisions rendues a la pluralite des voix. Paris: Imprimerie royale. Electronic edition by Marc-Kevin Daoust, Les Classiques des sciences sociales, Universite du Quebec a Chicoutimi. Digital edition