📈 economics

Quality at Stale Prices

This paper demonstrates that price rigidity can sustain high-quality production by preserving a reputationally valuable demand environment where sales are difficult, showing that a strategic firm's quality incentives are maximized at "stale" prices rather than immediately after a price reset.

Original authors: Georgy Lukyanov, Konstantin Shamruk, Ekaterina Logina

Published 2026-08-14

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Georgy Lukyanov, Konstantin Shamruk, Ekaterina Logina

Original paper licensed under CC BY 4.0 (http://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

The Hidden Game Behind the Price Tag

Imagine you are walking down a street where every shop has a sign with a fixed price on it. You don't know if the shop owner is a "good" person who always makes perfect products, or a "strategic" person who sometimes cuts corners to save money. You can't see what they are making; you only see if you decide to buy something. This is the world of reputation: a game where people try to figure out who is trustworthy based on limited clues.

In economics, there's a classic idea that if a shop lowers its price, more people will buy, and the owner will be happy. But there's a twist: if everyone buys easily, a single sale doesn't tell you much about the owner's quality. It's like a teacher giving a test where everyone gets an A; the grade doesn't prove anyone studied hard. However, if the price is high and only the most confident customers buy, a sale becomes a huge signal that the product is great. This paper explores a strange corner of this game: what happens when a shop is stuck with an old, high price that no longer makes sense, but that very "stale" price accidentally forces the owner to work harder to keep their reputation?

The Story of the Stale Price

This paper, written by Georgy Lukyanov, Konstantin Shamruk, and Ekaterina Logina, asks a counter-intuitive question: Why would a company improve the quality of its product at a price it would never choose to set today?

Usually, we think of prices as just numbers that balance supply and demand. But the authors show that a price tag is also a filter. It decides who buys and how much information that purchase reveals to the world.

The Setup: The Invisible Clock and the Price Desk

Imagine a long-lived company with two possible "personalities":

The Commitment Type: A robot that always makes high-quality products, no matter what.
The Strategic Type: A human who can choose to make high-quality products (which costs money) or low-quality ones (which is free).

Customers arrive randomly. They get a private hint about the quality (like a gut feeling or a sample) and then decide whether to buy. Here is the catch: The company cannot see who didn't buy. If a customer walks away, the company only knows "someone didn't buy," but not who or why. The only public news is a sale.

Inside the company, there is a "Pricing Desk." This desk is like an autopilot that can only change the price tag at random moments (like a clock ticking). Between those ticks, the price is stuck. The desk sees all the sales and the history, but it doesn't know if the company is the "Commitment" robot or the "Strategic" human. It just wants to maximize the company's total value.

The Twist: The "Stale" Price

Here is where the magic happens. Suppose the company started with a high price when its reputation was great. Time passes. No one buys for a while. Because there are no sales, the public starts to doubt the company's quality. The reputation slips.

But the price is stale. It's stuck at that old, high number because the "Pricing Desk" hasn't had a chance to change it yet.

Now, the situation is weird:

The price is high, so few people want to buy.
The reputation is low, so even fewer people want to buy.
But, the few people who do buy are the ones who are absolutely sure the product is good.

Because the buyers are so selective, a single sale becomes a massive signal. It screams, "We must be great!" to the rest of the world. For the "Strategic" human running the show, this is a golden opportunity. If they spend a little extra money to make a high-quality product right now, they might get a sale. That sale will boost their reputation so much that it's worth the cost.

The paper proves that there is a specific "threshold" of cost. If the cost to make good stuff is just a tiny bit lower than the maximum possible gain from that reputation boost, the strategic human will actually choose to make high-quality products.

Why Flexibility Kills Effort

The most surprising part of the paper is what happens when the "Pricing Desk" finally wakes up and changes the price.

If the desk sees the low reputation, it will naturally cut the price to attract more customers.

Before the cut: The price was high and stale. Sales were rare but very informative. The human worked hard to get that rare, valuable sale.
After the cut: The price is lower. More people buy, even if they aren't sure about the quality. A sale now happens easily, so it doesn't prove anything. The "signal" is weak.

The paper shows that once the price is reset to this new, lower level, the incentive to work hard disappears. The human stops making high-quality products because a sale no longer repairs their reputation. The "stale" price was actually the only thing keeping them honest.

The Math of the "Small Branch"

The authors don't just guess this; they prove it with heavy math. They show that if the cost of quality is just slightly below the breaking point (let's call the difference $\delta$ ), the company will produce high quality, but only in a very specific, narrow situation:

The quality boost is small (proportional to $\delta$ ).
It only happens when the reputation is in a very specific, narrow range (proportional to $\sqrt{\delta}$ ).
The total amount of "good stuff" produced is tiny (proportional to $\delta^{3/2}$ ).

Think of it like a flickering light. It only turns on when the battery is at a very specific, critical level. If you change the price even a little bit, the light goes out.

What This Means for the Real World

This paper suggests that price rigidity (prices that don't change often) isn't always a bad thing. In a world where customers learn from each other, a "stale" price can act as a discipline device. It creates a high-stakes environment where a sale is hard to get but very valuable.

If companies could change prices instantly, they might lower them to boost sales, but in doing so, they would destroy the very incentive that made them want to be good in the first place. The "stale" price preserves a difficult environment where effort pays off.

The authors are careful to say this is a theoretical proof, not a simulation of a specific market. They show that under these specific rules, a "bifurcation" happens: a new path of behavior (working hard) branches off from the old path (slacking off) right at the edge of where it becomes profitable.

So, the next time you see a shop with a price tag that seems too high or too old, don't just think it's a mistake. It might be the only thing keeping the owner from cutting corners. The price isn't just a number; it's a mirror that reflects how hard the seller is willing to work.

Technical Summary: "Quality at Stale Prices"

Problem and Motivation
The paper investigates a fundamental question in reputation dynamics: Why would a strategic firm supply high quality at a price it would no longer choose if it could reset that price immediately? In standard reputation models, high prices often signal quality or sustain effort through high margins. However, this paper posits that a posted price serves a dual function: it allocates current demand and determines how private customer information becomes public.

The authors study a long-lived firm with a fixed type (either a "commitment" type that always supplies high quality or a "strategic" type that chooses quality) facing a stream of short-lived customers. Customers observe a private signal of quality and an idiosyncratic taste shock before deciding to buy. Crucially, individual non-purchases are unobserved; only purchases and price resets are public. The firm's price is set by a "pricing desk" that observes the public history but not the firm's hidden type or production actions. This desk can reset the price only at Poisson-distributed opportunities (Calvo friction).

The core mechanism explored is that when reputation falls due to a lack of sales, a "stale" (inherited) high price makes a subsequent purchase both difficult and highly informative. If a customer buys at a high price despite low reputation, it implies they received strong private evidence of quality. This makes the purchase a valuable reputational repair. Conversely, if the price were reset to a lower level immediately, demand would rise, but the purchase would become less diagnostic, destroying the incentive for the strategic type to supply quality.

Methodology
The authors formalize this as a continuous-time reputation game with the following components:

Public State: The posterior probability $\mu$ that the firm is the commitment type. The current price $p$ is part of the public state.
Dynamics: Reputation drifts downward during periods of silence (no sales) and jumps upward upon a sale, governed by a purchase-only Bayesian filter derived from point-process hazards.
Decision Makers:
- The Strategic Type chooses quality $x \in \{0, 1\}$ to maximize expected firm value, weighing the cost of quality against the immediate revenue and the reputational jump from a sale.
- The Pricing Desk chooses a price from a finite menu $\mathcal{P}$ to maximize the posterior expected firm value $U(\mu, p)$ . It is information-constrained, observing only public history, ensuring prices are not signals of the firm's hidden type.
Equilibrium Concept: A stationary assessment consisting of a quality policy, a desk reset selector, and value functions satisfying Bellman equations and obedience conditions.

The analysis proceeds in two main stages:

Zero-Quality Benchmark: The authors first solve for the exact threshold cost $c^*$ above which the strategic type supplies zero quality. They distinguish between the global maximum of the quality gain over all possible current prices ( $c^*$ ) and the maximum restricted to prices the desk would choose immediately after a reset ( $c^*_{reset}$ ).
Bifurcation Analysis: For costs slightly below the threshold ( $c = c^* - \delta$ ), the authors construct a full-state stationary equilibrium. They utilize a "discounted occupation" argument to show that the strategic quality response is localized in a narrow belief interval. This localization ensures that the feedback effects on values and the desk's pricing policy are of a lower order ( $O(\delta^{3/2})$ ) than the quality incentive itself ( $O(\delta)$ ), allowing for a rigorous fixed-point argument using the Dugundji extension theorem and Kakutani–Fan–Glicksberg theorem.

Key Contributions and Results

The Stale-Price Wedge: The paper derives an exact cost threshold for zero strategic quality. A key finding is the existence of a "stale-price wedge": there exists a range of costs where supplying quality is unprofitable immediately after a price reset but profitable at an inherited (stale) price. This occurs because the inherited price preserves the high sensitivity of sales to actual quality (the "diagnostic" power of a sale), whereas a reset price lowers this sensitivity even if it raises demand.
Exact Threshold and Global Comparison: Unlike local approximations, the authors provide a global comparison between the unrestricted quality gain threshold and the reset-restricted threshold. They identify primitive conditions (Proposition 6) under which the inherited price lies between the reset price and the price that maximizes the sensitivity of sales to quality, creating the incentive for quality at the stale price.
Existence of a Bifurcating Equilibrium: The main theorem (Theorem 2) proves that for costs just below the zero-quality threshold, a non-zero stationary equilibrium exists. In this equilibrium:
- The strategic type supplies positive quality only at the inherited price $p^*$ and within a specific belief interval around a critical reputation $\mu^*$ .
- The quality response has a height of $O(\delta)$ and a support width of $O(\sqrt{\delta})$ , resulting in a total mass of $O(\delta^{3/2})$ .
- The pricing desk continues to reset to a different price $p_r \neq p^*$ whenever the opportunity arises, effectively terminating the quality spell.
- The equilibrium is robust to the feedback loops between quality, reputation, and pricing because the feedback effects are of lower order than the primary incentive.
Role of Price Rigidity: The paper demonstrates that price rigidity (the inability to reset prices immediately) is not merely a friction but a mechanism that sustains quality. By preventing the desk from lowering the price when reputation falls, the stale price maintains an environment where a sale is difficult and therefore highly informative, preserving the reputational return to effort.

Significance and Claims
The paper claims to provide a rigorous analytic bifurcation theorem for reputation games with endogenous quality and sticky prices. Its significance lies in:

Decoupling Roles: It explicitly separates the roles of type reputation (public state), expected current quality (customer inference), and posted price (monitoring technology).
New Channel: It identifies a novel channel where price rigidity sustains quality not through static margins, but by preserving the informativeness of customer actions. A price change improves current matching but can weaken future discipline by making sales less diagnostic.
Technical Robustness: The authors claim to solve the technical complications of exact price maximization (non-uniform convergence of selectors) by allowing the desk to mix among exact maximizers and using weak-star topology arguments.
Modest Scope: The paper is modest about its claims, stating it does not prove uniqueness, does not claim the effect is large at the bifurcation point, and does not establish existence far below the threshold. It establishes the existence of a mechanism where a stale price can sustain quality precisely because an immediate adjustment would destroy the monitoring environment.

The work connects the literature on reputation (Kreps & Wilson, Milgrom & Roberts) with observational learning and price rigidity, suggesting that in markets where customers learn from one another's choices, the inherited price is an integral part of the monitoring technology.