Elena Istomina

Hi! I am a visiting research specialist at Princeton University and a visiting researcher at the Moroccan Center for Game Theory (at UM6P). I was previously a postdoctoral researcher at the University of Chicago (2024-2025), where I obtained my PhD in 2024.

My research interests are mostly in applied microeconomic theory.

Primary fields: Microeconomic Theory.

Secondary fields: Industrial Organization, Behavioral Economics.

Contact Information

Tel.: (312) 547-1340

E-mail: eistomina [AT] princeton [DOT] edu

Fisher Hall, Office B014
Princeton, NJ 08544

Research

Working Papers

Markdowns (JMP)

(Previously circulated as Sorting Stock Through Sales: Inventory Turnover and Outlets)

Abstract: I model markdown pricing as a tool for price discrimination by product quality when quality is only observable to consumers. A seller offers durable goods of uncertain quality. She introduces new inventory at high prices and gradually marks down unsold inventory. Consumers arrive sequentially, choose a price point at which to inspect a good, observe its quality, and decide whether to purchase. Their decisions endogenously sort products by quality across markdown levels, resulting in indirect price discrimination by quality. I characterize sorting equilibria: steady states in which consumers optimally choose prices for inspection and sustain the quality distribution through their purchases. Despite the richness of the equilibrium set, the main result shows that any equilibrium can be summarized by a single statistic. Consumer and seller payoffs depend only on the relative quality difference between the highest and lowest prices, while intermediate markdowns and the sorting path are payoff-irrelevant. I show that the seller faces a fundamental trade-off: she must sacrifice sales volume and efficiency to achieve quality-based price discrimination.

Costly Communication of Service Quality

Abstract: The paper extends the classical monopolistic screening model of Mussa & Rosen (1978), assuming the buyer cannot self-select in a menu. Instead, the seller price-discriminates by learning the buyer's preferences through communication. Communication is noisy and costly to the seller, and the buyer is strategic about his report. Under regularity restrictions on communication costs, I show the seller's problem of joint communication and menu design reduces to a rational inattention problem about the buyer's virtual type. The model delivers that costly communication may increase or decrease the efficiency of quality allocation, depending on the distribution of types and the convexity of production costs. Relative to the social planner facing the same communication costs, the seller chooses communication that is wastefully precise. The paper also establishes that adding ex-post constraints can be beneficial to the seller, as they can be a substitute for costly direct communication.

Closing Deals and Tipping Points

Abstract: This paper models an uninformed seller negotiating with informed buyers when completing the transaction requires time and effort. The buyer exerts costly effort to expedite the deal, while the seller learns about the buyer's enthusiasm and adjusts her pricing over time. Anticipating this, the buyer strategically speeds up or slows down the process. I show that the seller's beliefs about the buyer and the final price exhibit tipping points: the seller becomes more pessimistic over time as higher buyer types exert more effort, leading to abrupt shifts in beliefs and market activity. Under some conditions, the market comes to a freeze right before a burst of activity.

Inactive Research

Sequential Screening with Personal Selling

Abstract: I analyze a monopolistic screening model, where the buyer's type is initially unknown to both market sides. The seller engages in costless sequential communication with the buyer before presenting a final product offer. At each communication period, the seller selects a threshold and discloses to the buyer whether his type is above or below it. The optimal strategy for the seller is to gradually disclose information about the buyer's type, starting from the bottom. Compared to the standard monopolistic screening, this approach enables the seller to extract the entire surplus not only from the lowest served type but from a whole range of lower types. I also introduce an analog of a virtual type for a learning-buyer environment and examine the consequences of the buyer's limited knowledge for consumer welfare.

†This paper was anticipated by Heumann (2020) .