📈 economics

Selective Visibility and Coordination Loss in Matching Markets:An Evolutionary Model of Aspirational Search

This paper employs an evolutionary game-theoretic model to demonstrate how truthful but selectively visible, aspirational outcomes in matching markets distort agents' perceptions of attainability, leading to excessive high-target search, reduced commitment, and a decline in durable match formation despite individual adaptiveness.

Original authors: yao zhang

Published 2026-09-03

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: yao zhang

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

In the digital age, we often assume that seeing more success stories is a good thing. If a platform shows us people finding love, landing dream jobs, or achieving great things, we naturally feel that these outcomes are within our reach. This intuition relies on two quiet assumptions: first, that the stories we see are true, and second, that they are a fair sample of reality. But what happens when the stories are true, yet the sample is skewed? This question sits at the intersection of how we judge our own worth and how markets function. For decades, psychologists have known that people evaluate themselves by comparing their lives to others, often looking upward to those who seem more successful. Economists have long studied matching markets, where two groups—like job seekers and employers, or people looking for partners—try to find each other. The puzzle this research addresses is how these two fields collide: what occurs when a digital platform truthfully highlights only the most exciting, self-relevant successes, and how that selective spotlight changes the behavior of everyone involved.

The study, conducted by an economist at Chongqing Technology and Business University, explores a scenario where a digital platform does not lie. It does not fabricate success stories. Instead, it simply chooses to show a specific type of success more often than it occurs in real life. Imagine a dating app that, instead of showing a random mix of profiles, disproportionately displays the few people who have found highly compatible, long-term partners quickly. Every single profile shown is real, and every success story is genuine. However, because the platform over-represents these exceptional cases, the average user begins to believe that finding such a partner is much easier and more common than it actually is. The researcher built a mathematical model to trace how this "truthful but unrepresentative" view trickles down from a simple change in what people see to a complex shift in how the entire market behaves.

The model reveals a chain reaction that starts with a change in perception. When users see a steady stream of exceptional, self-relevant successes, they unconsciously adjust their own sense of what is possible for them. They feel more capable and more likely to succeed than the statistics would suggest. This boost in confidence leads them to change their strategy: they stop looking for partners who are a realistic match and start aiming much higher, targeting people who are far out of their league. This is a rational move for an individual; if you think you have a better chance than you actually do, you should aim higher. But the problem arises when everyone does this at once.

As more people in the market shift their sights upward, the competition for those top-tier targets becomes fierce. The people who are actually being targeted—the high-value counterparts—notice this surge of interest. They realize that the pool of people pursuing them is now filled with individuals who are overestimating their own chances. Consequently, these high-value targets become less willing to commit. They feel that the attention they are receiving is not genuine or sustainable, so they pull back, lowering their level of engagement or commitment. The result is a breakdown in coordination. The first group is chasing a dream that is statistically unlikely, while the second group is retreating from the chase. The market becomes clogged with people searching for the wrong things, and the number of successful, lasting matches actually decreases.

The research confirms that this coordination loss happens even though no one is lying. The platform is simply doing what many algorithms do: showing content that is relevant and exciting to keep users engaged. The model shows that this individual adaptation—aiming higher because the view looks brighter—becomes a collective failure when the market cannot scale to meet those new, inflated expectations. The study uses a framework called evolutionary game theory, which tracks how strategies spread through a population over time, to prove that this outcome is not just a possibility but a stable state. Under general conditions, the more the platform highlights these exceptional cases, the more the market moves away from successful pairings.

Crucially, the paper also investigates a solution. It suggests that the problem is not the existence of successful stories, but the lack of context. The researchers propose a "representativeness restoration" intervention. This does not mean hiding the successful stories or censoring the truth. Instead, it means showing the successful stories alongside the full picture of reality. If a user sees a story about a perfect match, they should also see data showing how rare that outcome is, or how many people tried and failed before finding it. By restoring the statistical context, the platform can help users calibrate their expectations without removing the inspiration. The model demonstrates that this simple adjustment reverses the negative trend. When users see the full distribution of outcomes, they stop overestimating their chances, they stop aiming unrealistically high, and the market returns to a state where lasting matches can form.

The study goes a step further by asking why platforms might continue to show these skewed views in the first place. It models a scenario where the platform itself is an active participant, adjusting how much it highlights exceptional cases based on user activity. The findings suggest a dangerous feedback loop: when users aim higher, they browse more and interact more, which generates more traffic and content for the platform. This makes the selective highlighting strategy profitable in the short term. However, this short-term gain comes at the cost of long-term market health. The platform might be rewarded for the activity it generates, even as the actual success rate of its users declines. This creates a situation where the platform's incentives are misaligned with the users' ultimate goal of finding a lasting match.

The research does not claim that ambitious search is bad, nor does it argue that exceptional outcomes should be hidden. The core insight is that truthfulness is not enough. A sample can be entirely true and still be misleading if it is not representative. In rank-sensitive markets, where opportunities are scarce and people are uncertain about their own standing, the way information is presented matters as much as the information itself. The study concludes that for digital platforms to function well, they must do more than just verify facts; they must ensure that the stories they tell provide a calibrated view of reality. By helping users distinguish between what is possible and what is probable, platforms can prevent the collective drift toward unrealistic goals and preserve the conditions necessary for genuine connection. The work serves as a reminder that in a world of curated feeds, the most valuable service a platform can offer might be the honest context that allows us to see ourselves clearly.

Technical Summary: Selective Visibility and Coordination Loss in Matching Markets

Problem Statement
The paper addresses a paradox in rank-sensitive matching markets (e.g., online dating, labor platforms): market coordination can deteriorate even when the outcomes displayed to participants are entirely truthful. The core issue is not misinformation, but unrepresentative sampling. When platforms selectively over-sample genuine, high-salience outcomes (exceptional successes that are self-relevant to the user), participants may incorrectly infer that these outcomes are attainable for the general population. This leads to an "aspirational search" where agents revise their search targets upward. While individually adaptive, this collective shift can trigger a coordination failure: increased congestion in high-target search reduces the return on commitment for the counterpart population, leading to a collapse in durable match formation.

Methodology
The paper employs an evolutionary game theory framework involving adaptive agents who update strategies based on relative payoffs rather than solving complete dynamic programs. The model is structured in three progressive layers:

Microfoundations of Selective Exposure:
- The author distinguishes between a representative distribution of genuine outcomes ( $G_0$ ) and a visible distribution ( $G_z$ ) created by a platform's selection strategy.
- A selection kernel exponentially tilts $G_0$ toward cases with high self-relevance ( $s$ ) and positive outcome surprise ( $r$ ).
- Proposition 1 proves analytically that even with truthful data, this selective sampling ( $z > 0$ ) raises the population's average perceived attainable position ( $Q$ ) because the behavioral update function is increasing in the same self-relevance-surprise score used by the selection rule.
Two-Population Evolutionary Dynamics:
- The model embeds the belief channel into a two-population system: a focal population choosing between "higher-target search" ( $H$ ) and "calibrated search" ( $C$ ), and a counterpart population choosing between "long-term commitment" ( $I$ ) and "low commitment/exit" ( $E$ ).
- Payoff differences ( $F$ $F$ for search, $H$ $H$ for commitment) are defined by general monotonicity assumptions rather than specific functional forms initially:
  - $F_q > 0$ : Higher perceived position increases aspirational search.
  - $F_x < 0$ : Congestion reduces the return to high-target search.
  - $H_x < 0$ : Widespread aspirational search reduces the return to commitment.
- The system evolves via replicator equations. The paper establishes conditions for the existence, uniqueness, and global stability of an interior equilibrium.
Endogenous Visibility (Three-Population System):
- The share of selective presentation ( $z$ ) is endogenized as a third population evolving based on its own payoff difference ( $P$ ), which depends on traffic/content value and saturation effects.
- The stability of this three-population system is decomposed into an amplification loop (visibility $\to$ search $\to$ traffic) and a correction loop (saturation/commitment decline $\to$ reduced incentive for selective presentation).

Key Contributions

Formalizing "Truthful but Unrepresentative": The paper provides a probabilistic micro-foundation showing how truthful data can become unrepresentative through exponential tilting toward self-relevant surprises, explicitly separating truthfulness from representativeness.
General Cross-Population Interaction: Without assuming affine payoffs, the paper derives comparative statics showing that increased selective exposure ( $z$ ) raises higher-target search ( $x$ ) and lowers counterpart commitment ( $y$ ), thereby reducing durable match formation ( $M$ ).
Governance via Representativeness: The paper introduces a "representativeness-restoration" intervention ( $\tau$ ) that mixes the visible distribution back toward the representative baseline. It proves that this intervention improves coordination without suppressing successful outcomes, merely by correcting the statistical context.
Endogenous Feedback Stability: By modeling the platform's selection strategy as an adaptive variable, the paper identifies analytical conditions (via the Jacobian of the three-population system) where the feedback loop between visibility, search, and traffic can destabilize the market, distinguishing between amplification and correction mechanisms.

Key Results

Coordination Loss: Under general sign restrictions, greater selective exposure leads to a unique interior equilibrium where higher-target search increases and counterpart commitment decreases. Consequently, the durable-match formation index declines (Proposition 5).
Global Convergence: Under uniform inward boundary conditions (where strategies are not absorbed by the edges), the interior equilibrium is globally attracting (Proposition 4).
Restoration Effect: Increasing the correction intensity $\tau$ reverses the comparative statics: it lowers aspirational search, raises commitment, and improves durable matching (Proposition 6).
Affine Benchmarks: Using affine approximations, the paper derives closed-form thresholds for regime shifts (e.g., when commitment leaves the boundary $y=1$ ) and defines a "visibility-to-calibration ratio" ( $\Psi$ ) that measures the force of selective exposure against market congestion.
Stability Conditions: For the endogenous three-population system, local stability depends on a feedback boundary condition: $k(\nu\phi - \delta\omega) < \xi(\gamma\nu + \beta\delta)$ . If the amplification of visibility and traffic outweighs the damping from saturation and behavioral correction, the system may lose stability.

Significance and Claims
The paper claims to demonstrate that individually adaptive behavior can become collectively non-scalable in rank-sensitive markets when the information environment is selectively sampled.

Theoretical Scope: The mechanism is specific to environments with scarce upper-tail opportunities, uncertainty about relative position, and strategic reciprocity. It does not claim that ambitious search is inherently inefficient or that exceptional outcomes should be hidden.
Managerial Implication: The primary contribution is a design principle: platforms should evaluate information not just for truthfulness and relevance, but for representativeness and calibration. The paper argues that "restoring the denominator" (providing base-rate or percentile context) can improve market coordination without censoring success stories.
Limitations: The author explicitly notes that the model does not establish irreversible hysteresis (path dependence) without additional state variables like memory or sunk costs. Furthermore, the paper acknowledges that while empirical literature supports the individual components (subjective self-position, aspirational pursuit), the complete chain from selective sampling to bilateral strategy adaptation has not yet been empirically identified in a single study. The computational exercises are presented as illustrations of analytical regimes, not as empirical evidence of prevalence.

In conclusion, the paper posits that the divergence between "what is possible" (exceptional cases) and "what is probable" (representative cases) creates a coordination loss that can be mitigated by aligning the visible information environment with the underlying opportunity distribution.

Technical Summary: Selective Visibility and Coordination Loss in Matching Markets

More like this