Information Intermediaries in Monopolistic Screening

arXiv:2604.09343 · econ.TH · Submitted 2026-04-10 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Economics. The day in economics: theory, mechanisms and the evidence beneath them.

Ruth: We're looking at the paper: "Information Intermediaries in Monopolistic Screening".

Warren: Here's the gist of the paper. We investigate the relationship between product offerings, information dissemination, and consumer decision-making in a monopolistic screening environment in which consumers lack information about their valuation of quality-differentiated products <ref:2604.09343#pg2>.

Ruth: Our results show that as intermediaries place greater emphasis on consumer surplus over product quality, sellers are prompted to strategically expand their product range <ref:2604.09343#pg8>.

Warren: The role of information intermediaries proves pivotal in shaping consumer welfare, market profitability, and overarching economic efficiency <ref:2604.09343#pg2>.

Ruth: Model Setup The model involves a monopolist selling goods of varying quality to a potential buyer whose value is given by a random variable θ ∈

0, 1: distributed according to the cumulative distribution function (CDF) F0 <ref:2604.09343#pg3>.

Paper summary: Warren: So wrapping up this paper "Information Intermediaries in Monopolistic Screening", we see two main outcomes depending on how much the intermediary cares about consumer surplus versus quality.

Iris: The key finding is that as the intermediary prioritizes consumer surplus over product quality, sellers are incentivized to increase the number of products they offer.

Ruth: But there’s a catch: this expanded product variety actually reduces overall economic efficiency when compared to scenarios where sellers provide information directly to consumers.

Warren: It’s like if you have too many options because the person helping you is focused on making sure you're happy, it ends up making the whole system less efficient overall.

Iris: The paper suggests that even though intermediaries can encourage sellers to broaden their offerings, there is a reduction in economic efficiency under their presence compared to when sellers are directly providing information.

Ruth: And this means that the average quality level achieved isn't as high as it could be under a standard screening model, but it's also lower than if consumers just relied on direct seller-to-consumer information.

Warren: It sounds like the presence of these intermediaries creates a trade-off: you get enhanced buyer surplus, but you sacrifice some average product quality.

Iris: The paper also points out that the consumer’s rents and the total surplus aren't moving in a straight line with how much emphasis is placed on the intermediary's bias term.

Ruth: That non-monotonic relationship between consumer surplus and that bias term suggests that even a slight change in what information is available to buyers can actually reduce their payoffs.

Warren: And the authors point out that the monopolist has a way to mitigate this loss by being able to increase the bias, which helps offset some of the efficiency hits from expanding product offerings.

Iris: Ultimately, the paper says that when these intermediary's preferences line up with consumer and seller preferences, like in the Mussa-Rosen menu case, things align nicely.

Ruth: But if the bias is strong—if it's very large—the optimal menu structure ends up looking almost exactly like what a seller would offer if she were providing direct information to the buyer herself.

Warren: So, while intermediaries are important for market dynamics, their influence on efficiency depends entirely on how they balance their goals and how much variety they introduce into the market.

Conclusion: Ruth: So, to wrap up this session on "Information Intermediaries in Monopolistic Screening," we're looking at what these authors actually found about how product variety changes when there's an outside information source involved.

Warren: They basically showed that if the intermediary cares more about making the consumer happy than they do about the quality of the stuff, then sellers will start offering way more products.

Iris: That expansion in product offerings, though, comes with a cost to overall economic efficiency compared to when sellers are just giving direct information to buyers.

Ruth: It means that while we get more variety from these intermediaries, the average quality level ends up lower than it would be under a standard setup or if consumers just went straight to the seller.

Warren: I see that trade-off playing out—more buyer happiness versus less efficient market outcomes. It makes you wonder how much of that lost efficiency is worth gaining extra buyer surplus.

Iris: The paper also points out that this relationship between consumer happiness and the intermediary's bias isn't a straight line; it goes up and down depending on the numbers involved.

Ruth: That means even small changes in what information is shared can actually hurt consumer payoffs sometimes, which is something we need to watch carefully when thinking about how these systems work.

Warren: And the authors suggest that the monopolist has some leverage here, because if she can increase that bias term, it helps mitigate some of those efficiency losses from offering a wider range.

Iris: It boils down to whether this intermediary is helping or hurting the market structure in the long run, and we need to see how those incentives play out across different scenarios.

Panagiotis Kyriazis, Edmund Y. Lou

European University Institute · University of Auckland

econ.TH

Submitted: 2026-04-10

Updated: 2026-11-02

License: http://creativecommons.org/licenses/by-nc-nd/4.0/

Importance score: 76/100

The gist: We investigate the relationship between product offerings, information dissemination, and consumer decision-making in a monopolistic screening environment in which consumers lack information about…

Key concepts

Monopolistic Screening Environment
A market structure where a single seller offers products of varying qualities to consumers who cannot directly assess the true quality or value of those products. The goal is for the seller to use product variety to signal quality.
Information Intermediary Bias (b)
The intermediary provides information about the consumer's valuation ($ heta$) but has a built-in bias ($b$). This parameter represents the intermediary's preference, influencing how much weight they give to consumer surplus versus the actual product quality when advising buyers.
Optimal Menu Expansion
The seller's optimal strategy is to choose the number of products in their menu. The model shows that as the intermediary's bias decreases, the optimal number of products increases, meaning sellers offer a wider variety when intermediaries are less biased toward consumer surplus.

Terminology

Summary

We investigate the relationship between product offerings, information dissemination, and consumer decision-making in a monopolistic screening environment in which consumers lack information about their valuation of quality-differentiated products <ref:2604. p. 0

Where the paper says it

Information Intermediaries in Monopolistic Screening ∗ Panagiotis Kyriazis† Edmund Y. Lou‡ March 2026 Abstract We investigate the relationship between product offerings, information dissemination, and consumer decision-making in a monopolistic screening environment in which consumers lack information about their valuation of quality-differentiated products.

09343#pg2>. Our results show that as intermediaries place greater emphasis on consumer surplus over product quality, sellers are prompted to strategically expand their product range <ref:2604. p. 0
Where the paper says it

We characterize the monopolist’s profit-maximizing finite-item menu. Our results show that as intermediaries place greater emphasis on consumer surplus over product quality, sellers are prompted to strategically expand their product range.

09343#pg8>. The role of information intermediaries proves pivotal in shaping consumer welfare, market profitability, and overarching economic efficiency <ref:2604. p. 0
Where the paper says it

Intriguingly, this augmented product variety decreases economic efficiency compared to scenarios where direct seller-to-consumer information provision is the norm. The role of information intermediaries proves pivotal in shaping consumer welfare, market profitability, and overarching economic efficiency.

09343#pg2>.

Model Setup

The model involves a monopolist selling goods of varying quality to a potential buyer whose value is given by a random variable θ ∈ [0, 1] distributed according to the cumulative distribution function (CDF) F0 <ref:2604.09343#pg3>. The game starts with the monopolist offering the buyer a contract M consisting of pairs of menu items (q, t) ∈ [0, q¯] × R+ <ref:2604. p. 9

Where the paper says it

3 Throughout the paper, we assume that F0 has an increasing hazard rate and, thus, is Myerson regular. The game starts with the monopolist offering the buyer a contract M consisting of pairs of menu items (q, t) ∈ [0, q¯] × R+.

09343#pg6>. Each menu item corresponds to a transfer t to be paid to the monopolist by the buyer and the quality q of the product the buyer gets in exchange <ref:2604. p. 9
Where the paper says it

The game starts with the monopolist offering the buyer a contract M consisting of pairs of menu items (q, t) ∈ [0, q¯] × R+. Each menu item corresponds to a transfer t to be paid to the monopolist by the buyer and the quality q of the product the buyer gets in exchange.

09343#pg6>. The utility of the buyer is given by U B(θ, q, t) = θq − t <ref:2604. p. 9
Where the paper says it

We assume the utility of the buyer is quasi-linear. That is, given the buyer’s type θ, his utility from the menu item (q, t) is given by U B (θ, q, t) = θq − t (1) The monopolist’s cost of providing quality q is given by c(q) where the function c : R+ → R+ is assumed to be strictly increasing, continuously differentiable, and strictly convex.

09343#pg6>. The monopolist’s cost of providing quality q is given by c(q) where the function c: R+ → R+ is assumed to be strictly increasing, continuously differentiable, and strictly convex <ref:2604. p. 9
Where the paper says it

We assume the utility of the buyer is quasi-linear. That is, given the buyer’s type θ, his utility from the menu item (q, t) is given by U B (θ, q, t) = θq − t (1) The monopolist’s cost of providing quality q is given by c(q) where the function c : R+ → R+ is assumed to be strictly increasing, continuously differentiable, and strictly convex.

09343#pg6>.

Intermediary Influence

The buyer has access to an informational intermediary (he) who can provide information about θ after observing the posted menu <ref:2604.09343#pg6>. The intermediary’s payoff if the buyer chooses the item (q, t) is given by U I(θ, q, t) = (θ + b)q − t where the parameter b captures the intermediary’s bias toward higher-quality products <ref:2604. p. 9

Where the paper says it

Specifically, after observing the posted menu, the intermediary can pick and commit to an information structure s : [0, 1] → ∆([0, 1]) whose realization is observed by the buyer before his choice of a menu item. We assume that the intermediary’s payoff if the buyer chooses the item (q, t) is given by U I (θ, q, t) = (θ + b)q − t (2) where the parameter b captures the intermediary’s bias toward higher-quality products and is commonly known to all players.

09343#pg7>. Given a realization s, the buyer’s expected value is denoted by w:= E(θs) <ref:2604. p. 9
Where the paper says it

Specifically, after observing the posted menu, the intermediary can pick and commit to an information structure s : [0, 1] → ∆([0, 1]) whose realization is observed by the buyer before his choice of a menu item. We assume that the intermediary’s payoff if the buyer chooses the item (q, t) is given by U I (θ, q, t) = (θ + b)q − t (2) where the parameter b captures the intermediary’s bias toward higher-quality products and is commonly known to all players.

09343#pg8>. The seller's problem is to maximize (q(w),t(w)),G∈MPC(F) ∫ 1 0 [(t − c(q))dG(w)] s.t. wq(w) − t(w) ≥ wq'(w′) − t'(w′) for all w, w′ ∈ supp G (B-IC), and q(wq)(w) − t(wq)(w) ≥ 0 for all w ∈ supp G (B-IR), where G is the distribution over posterior means <ref:2604.09343#pg9>.

Bias Effects on Menu Structure

The main result is that as the intermediary assigns greater weight to consumer surplus relative to product quality, the optimal menu expands to include a growing number of products <ref:2604.09343#pg2>. In terms of market outcomes of interest, while the presence of intermediaries can incentivize suppliers to broaden their product offerings, there is a reduction in overall economic efficiency under the intermediary’s presence compared to scenarios where sellers directly provide information to consumers <ref:2604. p. 5

Where the paper says it

Thus, it is as if the consumer has private information about their value, as in the classic monopolistic screening model of Mussa and Rosen (1978). In terms of market outcomes of interest, while the presence of intermediaries can incentivize suppliers to broaden their product offerings, thereby augmenting consumer benefits, there is a reduction in overall economic efficiency under the intermediary’s presence compared to scenarios where sellers directly provide information to consumers.

09343#pg8>. The average quality level is higher than what it would be under standard screening models yet lower than if consumers relied solely on direct seller-to-consumer information <ref:2604. p. 7
Where the paper says it

Our results highlight the importance of recognizing the information intermediaries’ role when assessing market inefficiencies. In our model, the average quality level is higher than what it would be under standard screening models yet lower than if consumers relied solely on direct seller-to-consumer information.

09343#pg2>.

Optimal Menu Characterization

The optimal number of items in the menu, N∗ b, is the highest possible number such that the lowest quality offered, obtained as the solution to (OPT-Q1), is positive <ref:2604.09343#pg8>. The optimal qualities are given by c'(q∗ i) = w∗ i − ∑ N-1 j=i (F0(w∗ j+1 − b) − F0(w∗ j − b)) (w∗ i+1 − w∗ i) for i = 1,..., N - 1 <ref:2604. p. 0

Where the paper says it

lou@auckland. 1 arXiv:2604.

09343#pg9>. The optimal number of items increases as the intermediary’s bias b decreases <ref:2604. p. 5
Where the paper says it

Moreover, by the same logic, the seller must include a successively higher number of options in the menu as the bias keeps decreasing. In the limiting case where the intermediary’s bias approaches zero so that the intermediary’s and the consumer’s preferences coincide, it is optimal for the seller to offer a continuum of items.

09343#pg8>.

Comparative Statics and Policy Implications

The consumer’s rents, the total surplus, and the average distortions in the market are non-monotone in the bias term <ref:2604.09343#pg2>. The presence of intermediaries leads to a “trade-off” of average product quality for enhanced buyer surplus <ref:2604. p. 4

Where the paper says it

In other words, we investigate how it evolves as the intermediary’s emphasis shifts between product quality and consumer surplus. We leverage our results to analyze the implications of the change in supply’s response to market outcomes of interest, specifically average quality, profits, consumer surplus, distortions, and overall economic efficiency.

09343#pg8>. This suggests that while some degree of information is undoubtedly better than none, even a slight alteration in the information made available to buyers might reduce consumer payoffs from the interaction <ref:2604. p. 6
Where the paper says it

In such cases, a seller’s ability to tailor products and prices to different consumer types introduces a non-monotonic change in consumer welfare, as we discussed. Consequently, while some degree of information is undoubtedly better than none, even a slight alteration in the information made available to buyers might reduce consumer payoffs from the interaction.

09343#pg2>. The capability of a monopolist to augment the bias diminishes her profit loss, and this expansion of product offerings will mitigate the profit loss due to the decrease in b <ref:2604. p. 6
Where the paper says it

happens because a decrease in the intermediary’s bias reduces the seller’s profits. While introducing a wider range of products mitigates the profit loss, it does not fully reverse it.

09343#pg8>.

Conclusion

The optimal menu coincides with the Mussa-Rosen menu if the intermediary’s and the consumer’s preferences coincide <ref:2604.09343#pg2>. Conversely, when the bias is large, it coincides with the one the seller would offer if she were providing information directly to the buyer <ref:2604. p. 6

Where the paper says it

Moreover, while the consumer’s payoff in the presence of the intermediary is higher relative to the case where the seller provides information to the consumer directly, it exhibits a non-monotonic relationship with the bias term. In essence, consumers favor an intermediary with a very low bias over one with a very high bias.

09343#pg2>. As the intermediary’s bias diminishes, consumer surplus exhibits a non-monotonic relationship with the intermediary’s bias <ref:2604. p. 6
Where the paper says it

Thus, profits decrease as the intermediary’s bias does, and this reduction dominates the gains in consumer payoffs. Moreover, while the consumer’s payoff in the presence of the intermediary is higher relative to the case where the seller provides information to the consumer directly, it exhibits a non-monotonic relationship with the bias term.

09343#pg8>. The seller’s ability to respond to more accurate buyer’s learning by offering a wider array of products proves beneficial for both the seller and the consumer and, therefore, increases the overall efficiency in the economy <ref:2604.09343#pg2>.

References

Arieli, Itai, Yakov Babichenko, Rann Smorodinsky, and Takuro Yamashita (2023), “Optimal persuasion via bi-pooling” <ref:2604.09343#pg9>. Bergemann, Dirk, Tibor Heumann, and Stephen Morris (2022), “Screening with persuasion” <ref:2604. p. 7

Where the paper says it

Our work closely aligns with Bergemann et al. (2022), who explore optimal pricing through both mechanism and information design.

09343#pg9>. Brynjolfsson, Erik, Yu (Jeffrey) Hu, and Michael D. p. 0
Where the paper says it

Our insights underscore the complexities ∗ We owe special thanks to Alessandro Pavan and Piotr Dworczak for their support and guidance throughout this project. We also wish to thank Zeinab Aboutalebi, Ian Ball, Sandeep Baliga, Özlem Bedre-Defolie, Georgy Egorov, Daniel Garcia, Maarten Janssen, Peter Klibanoff, Laurent Mathevet, Niko Matouschek, Jeffrey Mensch, Filippo Mezzanotti, Wojciech Olszewski, Harry Pei, Alvaro Sandroni, Karl Schlag, Marciano Siniscalchi, as well as various seminar audiences at Northwestern University, University of Vienna and European University Institute for valuable comments and suggestions.

Smith (2003), “Consumer surplus in the digital economy: Estimating the value of increased product variety at online booksellers” <ref:2604. p. 5
Where the paper says it

2 This 2For consumers, the increased product variety is beneficial. Especially in the contemporary digital economy, the benefits of expanded product variety in electronic markets to consumer welfare are widely acknowledged.

09343#pg8>. Deniz, Dizdar and Eugen Kovac (2020), “A simple proof of strong duality in the linear persuasion problem” <ref:2604. p. 8
Where the paper says it

From a methodological standpoint, our approach leans heavily on the Bayesian Persuasion literature. We leverage findings related to dual price functions and bi-pooling distributions derived by Dworczak and Martini (2019), Deniz and Kovac (2020), Kleiner et al.

09343#pg9>. Dixit, Avinash K. and Joseph E. Stiglitz (1977), “Monopolistic competition and optimum product diversity” <ref:2604. p. 0
Where the paper says it

Information Intermediaries in Monopolistic Screening ∗ Panagiotis Kyriazis† Edmund Y. Lou‡ March 2026 Abstract We investigate the relationship between product offerings, information dissemination, and consumer decision-making in a monopolistic screening environment in which consumers lack information about their valuation of quality-differentiated products.

09343#pg9>. Dworczak, Piotr and Giorgio Martini (2019), “The simple economics of optimal persuasion” <ref:2604. p. 8
Where the paper says it

From a methodological standpoint, our approach leans heavily on the Bayesian Persuasion literature. We leverage findings related to dual price functions and bi-pooling distributions derived by Dworczak and Martini (2019), Deniz and Kovac (2020), Kleiner et al.

09343#pg9>. Kleiner, Andreas, Benny Moldovanu, and Philipp Strack (2021), “Extreme points and majorization: Economic applications” <ref:2604. p. 8
Where the paper says it

We leverage findings related to dual price functions and bi-pooling distributions derived by Dworczak and Martini (2019), Deniz and Kovac (2020), Kleiner et al. (2021) and Arieli et al.

09343#pg9>. Kolotilin, Anton, Timofiy Mylovanov, and Andriy Zapechelnyuk (2022), “Censorship as optimal persuasion” <ref:2604. p. 4
Where the paper says it

This choice must be optimal for the intermediary in the sense that the intermediary must have no incentive to deviate and provide information in a different way than the one suggested by the monopolist. From a methodological perspective, we model the information provision stage as a Bayesian persuasion problem.

09343#pg9>. Malenko, Andrey and Anton Tsoy (2019), “Selling to advised buyers” <ref:2604. p. 8
Where the paper says it

They show that dynamic mechanisms dominate static ones and derive the optimal selling mechanism in many scenarios. Our paper differs in that the “expert” has commitment power relative to Malenko and Tsoy (2019) and, moreover, we study the problem of selling multiple quality-differentiated products to a single consumer instead of the allocation of a single object to many buyers.

09343#pg9>. Maskin, Eric and John Riley (1984), “Monopoly with incomplete information” <ref:2604. p. 3
Where the paper says it

In our model, a seller (she) offers goods of heterogeneous quality to screen different types of a consumer (he). In a standard screening framework, as in Mussa and Rosen (1978) or Maskin and Riley (1984), the buyer’s taste for quality is his private information.

09343#pg9>. Mensch, Jeffrey (2022), “Screening inattentive buyers” <ref:2604. p. 8
Where the paper says it

complement each other in Section 6. Mensch (2022) adopts a similar rational inattention approach focusing on the sale of a fixed-quality product to single or multiple buyers.

09343#pg9>. Mensch, Jeffrey and Doron Ravid (2022), “Monopoly, product quality, and flexible learning” <ref:2604. p. 8
Where the paper says it

More importantly, while Roesler and Szentes (2017) address optimal buyer learning–a situation we mimic when the intermediary has no bias– Ravid et al. (2022) presume consumers pay for information, akin to Thereze (2022) and Mensch and Ravid (2022), while we assume that the intermediary provides information to the consumer for free.

09343#pg9>. Mussa, Michael and Sherwin Rosen (1978), “Monopoly and product quality” <ref:2604. p. 7
Where the paper says it

Our screening setting is in the spirit of Mussa and Rosen (1978) and Maskin and Riley (1984). The key distinction is that information about the consumer’s preferences for quality or product characteristics that affect the match with the good is provided by an intermediary rather than the consumer having private information.

09343#pg9>. Perrigne, Isabelle and Quang Vuong (2019), “Econometrics of auctions and nonlinear pricing” <ref:2604.09343#pg9>. Ravid, Doron, Anne-Katrin Roesler, and Balázs Szentes (2022), “Learning before trading: On the inefficiency of ignoring free information” <ref:2604. p. 8
Where the paper says it

More importantly, while Roesler and Szentes (2017) address optimal buyer learning–a situation we mimic when the intermediary has no bias– Ravid et al. (2022) presume consumers pay for information, akin to Thereze (2022) and Mensch and Ravid (2022), while we assume that the intermediary provides information to the consumer for free.

09343#pg9>. Roesler, Anne-Katrin and Balázs Szentes (2017), “Buyer-optimal learning and monopoly pricing” <ref:2604. p. 8
Where the paper says it

This sequence is pivotal. More importantly, while Roesler and Szentes (2017) address optimal buyer learning–a situation we mimic when the intermediary has no bias– Ravid et al.

09343#pg9>. Spence, Michael A. (1980), “Multi-product quantity-dependent prices and profitability constraints” <ref:2604. p. 6
Where the paper says it

Such constraints introduce additional inefficiencies. The seller, in this context, cannot counteraverage selling prices, thereby boosting consumer surplus, Brynjolfsson et al.

09343#pg9>. Thereze, João (2022), “Screening costly information” <ref:2604. p. 7
Where the paper says it

We further characterize the exact number of items that the optimal menu must feature and how the intermediary’s presence affects market outcomes. A different approach to capturing information frictions in a monopolistic screening environment is pursued in Thereze (2022) and Mensch and Ravid (2022).

09343#pg9>.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Model dynamic product variety adjustment for intermediaries by simulating how sellers expand their product range in response to a decreasing intermediary bias (b). This allows AI systems to predict market behavior under conditions where information intermediaries shift from prioritizing quality to prioritizing consumer surplus, leading to strategic menu expansion.

  2. Develop an AI agent capable of optimizing the optimal finite-item menu based on the current level of intermediary bias (b) and external constraints on product variety. This system can determine the precise number of items (Nb) required for a seller to maximize profit while respecting constraints like a maximum allowable assortment size, as shown in Proposition 6.

  3. Create an AI evaluator that assesses market efficiency and quality distortions by comparing outcomes under different information provision scenarios—specifically, comparing an intermediary-driven screening environment against a scenario where the seller provides direct information to the consumer. This enables researchers to accurately estimate whether ignoring the intermediary leads to overestimating or underestimating quality degradation (Figure 7).

  4. Implement a mechanism for policy design centered on consumer learning that anticipates how intermediaries will disseminate information based on their objective function (consumer surplus versus product quality). This system can suggest optimal information transparency levels required by policymakers to enhance consumer welfare while managing the resulting impact on seller profitability and economic efficiency.

Abstract

We investigate the relationship between product offerings, information dissemination, and consumer decision-making in a monopolistic screening environment in which consumers lack information about their valuation of quality-differentiated products. An intermediary, who is driven by the objective of maximizing consumer surplus but is also biased towards high-quality products, provides recommendations after the monopolist announces the menu of product choices. We characterize the monopolist's profit-maximizing finite-item menu. Our results show that as intermediaries place greater emphasis on consumer surplus over product quality, sellers are prompted to strategically expand their product range. Intriguingly, this augmented product variety decreases economic efficiency compared to scenarios where direct seller-to-consumer information provision is the norm. The role of information intermediaries proves pivotal in shaping consumer welfare, market profitability, and overarching economic efficiency. Our insights underscore the complexities introduced by these intermediaries that policymakers and market designers must consider when designing policies centered on consumer learning and market information transparency.

Sources

Related papers