24th Annual International Industrial Organization Conference

April 10, 2026
Time	Location	Event
15:30 to 19:00	Pacific Foyer (3rd Floor)	Conference Registration
17:00 to 18:30		Rising Stars Sessions Sponsored by Alfred P. Sloan Foundation
18:30 to 20:30	Pacific Ballroom (3rd Floor)	Welcome Reception Sponsored by Analysis Group
April 11, 2026
Time	Location	Event
07:15 to 09:00	Pacific Ballroom (3rd Floor)	Breakfast
07:30 to 17:00	Pacific Foyer (3rd Floor)	Registration
08:00 to 10:00		Saturday Morning Sessions
10:00 to 10:15	Pacific Foyer (3rd Floor)	Coffee Break
10:15 to 12:15		Saturday Midday Sessions
12:20 to 14:00		Keynote and Awards Luncheon: Keynote Address: John Vickers, University of Oxford
14:15 to 16:15		Saturday Afternoon Sessions
16:15 to 17:00	Pacific Foyer (3rd Floor)	Afternoon Refreshment Break Sponsored by Bates White
17:00 to 18:30		Plenary Sessions
April 12, 2026
Time	Location	Event
07:15 to 09:00	Pacific Ballroom (3rd Floor)	Breakfast
07:30 to 11:00	Pacific Foyer (3rd Floor)	Registration
08:00 to 10:00		Sunday Morning Sessions
10:00 to 10:15	Pacific Foyer (3rd Floor)	Coffee Break
10:15 to 12:15		Sunday Midday Sessions

Summary of All Sessions
Click here for an index of all participants

#	Date/Time	Title/Location	Papers
1	April 10, 2026 17:00-18:30	Consumer Dynamics     Location: Salon A (3rd Floor)	3
2	April 10, 2026 17:00-18:30	Digital Platforms     Location: Salon B (3rd Floor)	3
3	April 10, 2026 17:00-18:30	Monopsony Power     Location: Salon C (3rd Floor)	3
4	April 10, 2026 17:00-18:30	Production Function Estimation and Applications     Location: Aegean (3rd Floor)	3
5	April 10, 2026 17:00-18:30	Real Estate Markets     Location: Thompson (4th Floor)	3
6	April 10, 2026 17:00-18:30	Antitrust and Competition     Location: Mediterranean (3rd Floor)	3
7	April 10, 2026 17:00-18:30	EVs and Green Energy     Location: Caspian (3rd Floor)	3
8	April 10, 2026 17:00-18:30	Platform Design and Information     Location: Brewster (4th Floor)	3
9	April 10, 2026 17:00-18:30	Regulating Externalities     Location: Georges (4th Floor)	3
10	April 10, 2026 17:00-18:30	Regulation and Quality     Location: Spectacle (4th Floor)	3
11	April 11, 2026 8:00-10:00	Advertising, Perceptions, and Market Competition     Location: Salon A (3rd Floor)	4
12	April 11, 2026 8:00-10:00	Health Care Markets: Bargaining, Integration, and Provider Performance (Sponsored by FTI Consulting)     Location: Atlantic 2 (3rd Floor)	4
13	April 11, 2026 8:00-10:00	Technology and Telecommunication     Location: Salon C (3rd Floor)	4
14	April 11, 2026 8:00-10:00	Advances in Demand Estimation     Location: Atlantic 1 (3rd Floor)	4
15	April 11, 2026 8:00-10:00	Regulation, Product Variety, and Quality     Location: Caspian (3rd Floor)	4
16	April 11, 2026 8:00-10:00	Experimentation, Learning, and Dynamic Investment     Location: Aegean (3rd Floor)	4
17	April 11, 2026 8:00-10:00	Digital Platforms and Regulation     Location: Georges (4th Floor)	3
18	April 11, 2026 8:00-10:00	Incentivizing Digital Content Creation     Location: Thompson (4th Floor)	3
19	April 11, 2026 8:00-10:00	Long-Term Care, Inspections, and Quality Measurement     Location: Salon B (3rd Floor)	4
20	April 11, 2026 8:00-10:00	Behavioral and Dynamic Bidding in Auctions     Location: Mediterranean (3rd Floor)	4
21	April 11, 2026 8:00-10:00	Merger Efficiencies (Sponsored by Charles River Associates)     Location: Atlantic 3 (3rd Floor)	4
22	April 11, 2026 8:00-10:00	Effects of Cartels     Location: Brewster (4th Floor)	4
23	April 11, 2026 10:15-12:15	Screening, Investment, and Organizational Incentives     Location: Thompson (4th Floor)	4
24	April 11, 2026 10:15-12:15	Contract Design and Enforcement in Vertical Relationships     Location: Mediterranean (3rd Floor)	3
25	April 11, 2026 10:15-12:15	Consumer Behavior and Information     Location: Georges (4th Floor)	4
26	April 11, 2026 10:15-12:15	Market structure and environmental policy     Location: Brewster (4th Floor)	4
27	April 11, 2026 10:15-12:15	Technology Adoption, Diffusion, and Industry Outcomes     Location: Caspian (3rd Floor)	4
28	April 11, 2026 10:15-12:15	Innovation and adoption     Location: Aegean (3rd Floor)	4
29	April 11, 2026 10:15-12:15	Insurance Design, Intermediation, and Selection     Location: Atlantic 3 (3rd Floor)	4
30	April 11, 2026 10:15-12:15	Market Power, Information, and Policy in Auctions     Location: Salon C (3rd Floor)	4
31	April 11, 2026 10:15-12:15	Financial Markets     Location: Salon B (3rd Floor)	4
32	April 11, 2026 10:15-12:15	Merger Evaluation (Sponsored by Cornerstone Research)     Location: Atlantic 2 (3rd Floor)	4
33	April 11, 2026 10:15-12:15	Cartel Stability     Location: Salon A (3rd Floor)	4
34	April 11, 2026 10:15-12:15	Invited Session: Industrial Organization and Education (Sponsored by Compass Lexecon)     Location: Atlantic 1 (3rd Floor)	4
35	April 11, 2026 14:15-16:15	Information Design and Disclosure     Location: Thompson (4th Floor)	4
36	April 11, 2026 14:15-16:15	Labor Market Power, Talent Allocation, and the Rise of Markups     Location: Salon B (3rd Floor)	3
37	April 11, 2026 14:15-16:15	Environment and Energy     Location: Brewster (4th Floor)	4
38	April 11, 2026 14:15-16:15	Empirical Approaches to Market Structure and Policy     Location: Salon C (3rd Floor)	4
39	April 11, 2026 14:15-16:15	Market Power and Intellectual Property     Location: Georges (4th Floor)	4
40	April 11, 2026 14:15-16:15	Platform design: Information     Location: Caspian (3rd Floor)	4
41	April 11, 2026 14:15-16:15	Dynamics and Market Design in Environmental IO     Location: Atlantic 2&3 (3rd Floor)	4
42	April 11, 2026 14:15-16:15	Hospital Performance, Pricing, and Provider Consolidation     Location: Aegean (3rd Floor)	4
43	April 11, 2026 14:15-16:15	Trade and IO (I)     Location: Spectacle (4th Floor)	4
44	April 11, 2026 14:15-16:15	Labor Antitrust     Location: Mediterranean (3rd Floor)	3
45	April 11, 2026 14:15-16:15	Merger Remedies (Sponsored by Analysis Group)     Location: Salon A (3rd Floor)	4
46	April 11, 2026 14:15-16:15	Invited Session: Algorithmic Pricing (Sponsored by Keystone)     Location: Atlantic 1 (3rd Floor)	4
47	April 11, 2026 17:00-18:30	Advances in Computational Methods in IO (Sponsored By Amazon)     Location: Atlantic 1 (3rd Floor)	0
48	April 11, 2026 17:00-18:30	Switching Costs in the Digital Economy (Sponsored by Econic Partners)     Location: Atlantic 2&3 (3rd Floor)	0
49	April 12, 2026 8:00-10:00	Personalization, Pricing, and Consumer Data     Location: Thompson (4th Floor)	4
50	April 12, 2026 8:00-10:00	Bargaining, Restraints, and Spillovers in Vertical Markets     Location: Salon B (3rd Floor)	4
51	April 12, 2026 8:00-10:00	Information and Regulation     Location: Salon C (3rd Floor)	4
52	April 12, 2026 8:00-10:00	Market Power: Measurement and Consequences     Location: Georges (4th Floor)	3
53	April 12, 2026 8:00-10:00	Production Function and Productivity Estimation     Location: Aegean (3rd Floor)	3
54	April 12, 2026 8:00-10:00	Platform design: Pricing     Location: Atlantic 2 (3rd Floor)	4
55	April 12, 2026 8:00-10:00	Environmental Policy Design and Market Outcomes     Location: Salon A (3rd Floor)	4
56	April 12, 2026 8:00-10:00	Pharmaceutical Innovation, Pricing, and Global Entry     Location: Atlantic 3 (3rd Floor)	4
57	April 12, 2026 8:00-10:00	Auctions: Entry, Scale, and Design     Location: Caspian (3rd Floor)	4
58	April 12, 2026 8:00-10:00	Digital Platforms, Market Dynamics, and User Behavior     Location: Mediterranean (3rd Floor)	4
59	April 12, 2026 8:00-10:00	Merger Thresholds     Location: Atlantic 1 (3rd Floor)	3
60	April 12, 2026 8:00-10:00	Trade and IO (II)     Location: Brewster (4th Floor)	4
61	April 12, 2026 10:15-12:15	Search, Platforms, and Consumer Learning     Location: Caspian (3rd Floor)	4
62	April 12, 2026 10:15-12:15	Production Networks and Supply Chain Dynamics     Location: Brewster (4th Floor)	4
63	April 12, 2026 10:15-12:15	Mergers, Market Structure, and Welfare     Location: Mediterranean (3rd Floor)	4
64	April 12, 2026 10:15-12:15	Pricing Strategies     Location: Salon A (3rd Floor)	3
65	April 12, 2026 10:15-12:15	Platforms and Innovation (Sponsored by Brattle)     Location: Atlantic 3 (3rd Floor)	4
66	April 12, 2026 10:15-12:15	Innovation, Competition, and Market Structure     Location: Salon B (3rd Floor)	4
67	April 12, 2026 10:15-12:15	Platform Design: Supply-Side Incentives     Location: Salon C (3rd Floor)	4
68	April 12, 2026 10:15-12:15	Technology, Spatial Access, and Public Program Design     Location: Georges (4th Floor)	4
69	April 12, 2026 10:15-12:15	Merger Retrospectives     Location: Aegean (3rd Floor)	4
70	April 12, 2026 10:15-12:15	Firm Relationships and Policy     Location: Thompson (4th Floor)	4
71	April 12, 2026 10:15-12:15	Panel Discussion: Doing IO at Antitrust Consulting Firms     Location: Atlantic 1 (3rd Floor)	0
72	April 12, 2026 10:15-12:15	The Impact of Generative AI on Content, Information, and Society     Location: Atlantic 2 (3rd Floor)	4

 

72 sessions, 257 papers, and 0 presentations with no associated papers

24th Annual International Industrial Organization Conference
Detailed List of Sessions

    

Session 1: Consumer Dynamics April 10, 2026 17:00 to 18:30 Location: Salon A (3rd Floor)
Session Chair: Tobias Salz, MIT
The Last Mile to First Treatment: Searching for Opioid Use Disorder Medication Abstract Despite growing policy support for buprenorphine, the leading medication for opioid use disorder, treatment initiation remains low. Comprehensive insurance data from Washington State show that fewer than half of first-time patients fill their prescriptions, well below rates for other chronic conditions. Initiation is hindered not only by the physical cost of visiting pharmacies (e.g., distance) but also by uncertainty about availability. I estimate their joint effect using a structural sequential search model in which patients search across pharmacies with unknown availability. Normalizing the default (most frequently visited) pharmacy’s search cost to zero, I identify search costs by comparing non-default choices (utility net of search costs) to default choices (utility only). I find that search costs account for 70% of observed treatment failures. Counterfactuals show that simple prescriber guidance on pharmacy availability could raise initiation share by 17%.
By Lance Gui; University of Arizona
Presented by: Lance Gui, University of Arizona
Discussant:   Brad Larsen, Washington University in St. Louis
State Dependence and Complementarities in Smartphone Choices Abstract The recent antitrust case between the Department of Justice (DOJ) and Apple centers on what drives consumers' persistent choices of smartphones within the same operating system. This paper empirically examines this state dependence in smartphone choices and its implications for consumer surplus, firm profits, and the DOJ's potential remedies. I disentangle multiple factors that can drive repeated smartphone choices but have not been considered jointly in the literature: (1) consumer heterogeneity in brand preferences, (2) one-time costs of switching operating systems, (3) complementary value from using a smartphone and smartwatch within the same ecosystem, and (4) installment contracts with mobile carriers. Using a novel consumer panel survey, I estimate a dynamic demand model where consumers make joint purchase decisions for smartphones and smartwatches. The estimation results indicate that switching costs are significant and similar in magnitude for both operating systems. The cost of switching from iOS to Android is estimated at approximately 92.9 USD. Counterfactual simulations suggest that while switching costs are a primary driver of state dependence for iOS users, persistent choices also arise from differences in product characteristics, pricing, and consumer heterogeneity. Reducing switching costs increases average consumer surplus, although price responses may change the direction of welfare effects.
By Seungwhan Chun; UNC Chapel Hill
Presented by: Seungwhan Chun, UNC Chapel Hill
Discussant:   Judith Chevalier, Yale University
Sale Pricing and Demand Dynamics Abstract I study imperfect price discrimination in markets where consumers can store goods for future consumption. I develop a dynamic model in which consumer decisions incorporate inventory costs. I show that when sellers know the distribution of inventory costs, they can discriminate between consumers with heterogeneous price sensitivities. Using store-level data, I estimate the model parameters and identify the share of less price-sensitive consumers who do not store versus more price-sensitive consumers who take advantage of lower sale price. I find that (i) static models understate own-price elasticities by up to 50 percent, (ii) storing consumers exhibit higher cross-price elasticities, highlighting the primacy of price over brand loyalty, (iii) sellers increase profits by up to 10 percent when combining volume discount pricing with inter-temporal pricing strategy, and (iv) while sale price primarily benefit storers, incorporating inventory costs substantially tempers these gains.
By Muhammad Shabanpour; Northeastern University
Presented by: Muhammad Shabanpour, Northeastern University
Discussant:   Fei Li, University of North Carolina
Session 2: Digital Platforms April 10, 2026 17:00 to 18:30 Location: Salon B (3rd Floor)
Session Chair: Ginger Jin, University of Maryland at College Park
Making Talk Cheap: Generative AI and Labor Market Signaling Abstract Large language models (LLMs) like ChatGPT have significantly lowered the cost of producing written content. This paper studies how LLMs, through lowering writing costs, disrupt markets that traditionally relied on writing as a costly signal of quality (e.g., job applications, college essays). Using data from Freelancer.com, a major digital labor platform, we explore the effects of LLMs’ disruption of labor market signaling on equilibrium market outcomes. We develop a novel LLM-based measure to quantify the extent to which an application is tailored to a given job posting. Taking the measure to the data, we find that employers had a high willingness to pay for workers with more customized applications in the period before LLMs were introduced, but not after. To isolate and quantify the effect of LLMs’ disruption of signaling on equilibrium outcomes, we develop and estimate a structural model of labor market signaling, in which workers invest costly effort to produce noisy signals that predict their ability in equilibrium. We use the estimated model to simulate a counterfactual equilibrium in which LLMs render written applications useless in signaling workers’ ability. Without costly signaling, employers are less able to identify high-ability workers, causing the market to become significantly less meritocratic: compared to the pre-LLM equilibrium, workers in the top quintile of the ability distribution are hired 19% less often, workers in the bottom quintile are hired 14% more often.
By Anais Galdin; Dartmouth College, Tuck Business Schools    Jesse Silbert; Princeton University
Presented by: Jesse Silbert, Princeton University
Discussant:   Leon Musolff, Wharton School of the University of Penn
The Transmission of Trade Shocks in Digital Markets Abstract This paper studies how trade policy shocks propagate through digital marketplaces. Using high-frequency, product-level data from a major e-commerce platform, we show that heightened trade uncertainty during the 2025 U.S. presidential transition reduced product turnover, while a sharp increase in U.S. tariffs on Chinese imports raised prices and significantly altered non-price dimensions of competition, including delivery performance, platform-assigned badges, sales outcomes, consumer ratings and reviews, and the composition of brands and manufacturers. We also provide evidence of incomplete tariff pass-through in digital markets. Estimating a structural demand model, we recover consumer utilities and document declines in consumer surplus in product categories with greater exposure to Chinese imports.
By Jorge Klinnert; University of Maryland
Presented by: Jorge Klinnert, University of Maryland
Discussant:   Yanyou Chen, University of Toronto
The Value of Intermediation in Bikeshare Abstract Transportation markets often face imbalances between demand and supply when agents do not internalize the system-wide effects of their choices. Intermediaries arise in such markets to ameliorate these imbalances. This paper studies the role of intermediation in New York City's Citi Bike system, where the operator rebalances bikes between locations to mediate demand between current and future consumers. I develop and estimate a spatial matching model using granular trip and bike availability data from 2018–2024. A unique aspect of this setting is that it allows me to consider the impact of network expansion, as the system expanded rapidly during this period. The model consists of consumer demand for traveling by bike, matching with bikes at the origin and docks at the destination, and a steady state equilibrium distribution of consumers and bikes in the system. This allows me to quantify how consumer travel patterns change as a result of changes to waiting times and travel times. I find that in 2022, removing rebalancing movements decreases the number of consumers using the system during rush hour by 19 percent.
By Anya Tarascina; University of Wisconsin-Madison
Presented by: Anya Tarascina, University of Wisconsin-Madison
Discussant:   Tianli Xia, University of Rochester
Session 3: Monopsony Power April 10, 2026 17:00 to 18:30 Location: Salon C (3rd Floor)
Session Chair: Robert Clark, University of Toronto
Do Middlemen Raise Drug Costs? Countervailing Market Power Meets Agency Frictions Abstract Many markets exhibit vertical layers with successive monopolists. One rationale for the existence of large middlemen is their ability to exercise countervailing market power to upstream monopolists. However, having market power may also exacerbate misaligned incentives between middlemen and consumers. In this work, I study the trade-off between countervailing market power and agency frictions in the US prescription drug market. Pharmacy Benefit Managers (PBMs) are large middlemen who negotiate with drugmakers on behalf of insurers for rebates, which may in turn inflate list prices – sticker prices set by drugmakers – and make drugs less affordable. At the same time, high rebates may benefit consumers through lower insurance costs. To assess the role of PBMs, I estimate a vertical model of drug and insurance demand, rebate negotiation and list price setting for oral anticoagulants in Medicare Part D. I find that PBMs reduce dispersion in rebates, which helps smaller insurers. Moreover, policy solutions that remove agency frictions while preserving – or even enhancing – countervailing market power may improve consumer welfare by up to 17% of annual premiums.
By Catherine Che; UC Berkeley
Presented by: Catherine Che, UC Berkeley
Discussant:   Ashley Swanson, University of Wisconsin-Madison
Why Do Contract Workers Earn Less? Evidence from India’s Auto Industry Abstract Contract workers constitute half of employment in India’s automotive industry but earn substantially less than permanent workers. Using data from the Annual Survey of Industries (2002-2019), I develop an estimator of labor supply and demand schedules to explain this wage premium. The model features nested CES production with distinct worker types, discrete choice supply functions with worker type-specific wage sensitivity, and differentiated market conduct—Nash-Bertrand competition for contract workers versus plant-level union bargaining for permanent workers. I find that the wage premium stems entirely from permanent workers’ higher productivity rather than differential monopsony power or unionization advantages.
By Davide Luparello; Penn State
Presented by: Davide Luparello, Penn State
Discussant:   Michael Rubens, UCLA
Milking Market Power: How Dairy Cooperatives Raise Rivals' Costs Abstract This paper studies how vertically integrated farmer-owned cooperatives leverage their control over the supply of raw milk to distort competition in downstream milk markets. Using a structural model of demand and supply, I examine whether cooperatives have both the incentive and ability to raise rivals’ costs. Results show cooperatives raised upstream markups on rival processors by 12–15% and were driven by the Herd Retirement Program, a supply contraction initiative. Counterfactual analysis implies that retail prices were 33% higher and processor markups were 14.8% lower than they would have been under Nash-Bertrand competition, and raw milk prices were 15.46% higher than the monopoly price. I relate these results to documentary evidence indicating that cooperatives raised rivals’ costs.
By Shivam Agrawal; The Ohio State University
Presented by: Shivam Agrawal, The Ohio State University
Discussant:   Sofia Villas-Boas, University of California, Berkeley
Session 4: Production Function Estimation and Applications April 10, 2026 17:00 to 18:30 Location: Aegean (3rd Floor)
Session Chair: Jordi Jaumandreu, Boston University
Firm in the clouds: Supplier market power and firm adoption in the cloud computing industry Abstract I study the effects of technology adoption on firm productivity in the European cloud computing services market, a setting with falling adoption costs, switching costs among providers, and bundling complementarities. I study how firms make technology bundling choices when faced with these adoption frictions, and analyze the effects of a proposed data egress fee ban in the EU’s Data Act of 2024. I find that adopting cloud computing increases firm productivity from 0.7% to 3.3%, with heterogeneous effects across sectors and a larger impact for firms that purchase services from multiple cloud providers. To estimate the effects of an egress fee ban on consumer welfare, I estimate a model of industry dynamics, in which firms produce output and make IT input bundle choices. I find substantial entry and switching costs. Simulating the effects of an egress fee ban, I find welfare increases of 7.2 million EUR per year. By comparison, a proposed mandate increasing software interoperability among cloud providers increases welfare by 612 million EUR per year.
By Nicholas Emery-Xu; University of California, Los Angeles
Presented by: Nicholas Emery-Xu, University of California, Los Angeles
Discussant:   Jihye Jeon, Boston University
Weak Identification Robust Methods for Production Function Estimation Abstract This paper revisits control-function estimation of production functions. We show that, in empirically relevant environments, the structural parameters can be weakly identified even when they are formally point identified. Casting control-function estimators in a GMM framework, we characterize the consequences of weak identification for estimation and inference, including nonstandard asymptotic behavior and unreliable Wald inference. We then develop and implement identification-robust inference procedures for control-function estimators and provide practical guidance for diagnosing weak identification and reporting inference that remains valid when identification is weak. Finally, we illustrate how weak identification propagates to economically relevant objects constructed from production-function estimates by studying firm-level markups under the production approach.
By Jorge de la Cal Medina; University of Amsterdam, Tinbergen Institute
Presented by: Jorge de la Cal Medina, University of Amsterdam, Tinbergen Institute
Discussant:   Haiqing Xu, University of Texas
Nonparametric Identification and Estimation of Production Functions Invariant to Productivity Dynamics Abstract Abstract We establish the nonparametric identification of both the production function and the distribution of unobserved productivity without imposing dynamic restrictions on the productivity process. Unlike standard proxy variable estimators that rely on a first-order Markov assumption, our strategy exploits the static covariance structure of three flexible inputs. By treating input demands as proxies subject to optimization errors, we map the production system to a measurement error model with auxiliary variables. This framework identifies the model primitives using only the joint distribution of inputs within a single period. We propose a Generalized Method of Moments estimator and verify its consistency through Monte Carlo simulations, particularly in non-stationary environments where conventional dynamic methods exhibit bias. An empirical application to Japanese manufacturing firms yields elasticity estimates that differ from standard benchmarks, providing evidence of time-varying production technologies and alternative implications for allocative efficiency.
By Rentaro Utamaru; The University of Tokyo
Presented by: Rentaro Utamaru, The University of Tokyo
Discussant:   Devesh Raval, Federal Trade Commission
Session 5: Real Estate Markets April 10, 2026 17:00 to 18:30 Location: Thompson (4th Floor)
Session Chair: Katja Seim, Yale
The Expansion and Dynamic Equilibrium Effects of Institutional Landlords Abstract This paper studies how dynamically formed cost efficiencies from scope and density drive institutional landlords’ expansion and, in turn, alter the distribution of welfare across heterogeneous households in single-family housing markets. Institutional landlords convert owner-occupied homes into large, spatially clustered rental portfolios. They constrain households' access to homeownership while expanding rental opportunities. This leads households to reoptimize between buying and renting, as buyers may face higher prices while renters may benefit from expanded choice sets. We build a dynamic equilibrium model of landlord investment with three key features: (i) oligopolistic landlords' investment determines the evolution of housing supply structure, (ii) portfolio size and density introduce endogenous variation in landlord costs, and (iii) households substitute within and across buying and renting in an integrated choice set. We estimate the model using firm-property-level data from 2013 to 2022 in the Atlanta metropolitan area. We find that institutional landlords' expansion achieved a 60.03% reduction in maintenance cost from economies of scope and density. Households' total welfare increased, with varying effects across renters and buyers. The majority of renters gained from expanded rental supply, while a small fraction of renters, together with most buyers, lost from diminished access to affordable homeownership. Our findings have significant policy implications for regulating institutional landlords’ expansion in the single-family home market.
By Zhichun Wang; Yale University    Daojing Zhai; Yale
Presented by: Zhichun Wang, Yale University
Discussant:   David Genesove, Hebrew University of Jerusalem
Targeting and Price Pass-Through in Housing Voucher Design Abstract Housing vouchers are a common policy for expanding access to homeownership, yet their effectiveness is called into question due to concerns over subsidies increasing housing prices and being poorly targeted. I study the equilibrium and distributional effects of homeownership vouchers through the case of Santiago's DS1 program, which subsidizes 7% of the city's transactions and is the largest such program in the OECD. I build an equilibrium model of a housing market with targeted and rationed homeownership vouchers. The model features endogenous voucher take-up and supply responses through existing unit sales and new construction. I estimate the model using novel data on voucher applications and usage, linked to the universe of real estate transactions and new development surveys. I evaluate the equilibrium impacts of the program relative to a scenario without the program, finding that it increases homeownership rates while raising prices. New development plays a significant role in dampening price inflation by increasing the supply of affordable units. Overall, I estimate that each dollar spent yields 61 cents in surplus. Although half of policy spending is transferred to beneficiaries, pecuniary externalities harm non-beneficiaries, reducing net consumer transfers to 25 cents per dollar spent. Counterfactual policies reveal a trade-off between targeting and price pass-through: policies that reduce price pass-through worsen targeting, as assistance goes to households more likely to become homeowners without the program.
By Fernando Ochoa; NYU
Presented by: Fernando Ochoa, NYU
Discussant:   Jean-Francois Houde, University of Wisconsin-Madison
Selling Fast or Selling Junk: Is iBuying Sustainable? Abstract This paper examines challenges with algorithmic intermediation on the real estate market and evaluates strategies to mitigate adverse selection when private information about product quality is intertwined with private information about preferences. I examine these issues in the context of iBuyers—firms that offer instant home purchases using big-data-driven pricing models—and analyze why they have struggled to achieve sustainable profitability. I develop a model in which home sellers choose between selling to an iBuyer and listing on the open market based on two dimensions of private information: unobserved house quality and the hassle costs of traditional selling. Sellers may select an iBuyer either to avoid the time and effort of listing or because the iBuyer’s offer exceeds their expected market price, with the latter case generating adverse selection against the iBuyer. Using detailed transaction and listing data, I estimate the joint distribution of these factors, identified from repeated sales and seller choice following iBuyer entry. Counterfactual analyses show that a revenue-sharing contract mitigates adverse selection by improving selection incentives, while incorporating a fine-tuned LLM-based text score derived from past unstructured listing data further reduces informational frictions by providing a signal of unobserved house quality. Together, these mechanisms enhance the viability of algorithmic intermediation in the housing market.
By So Hye Yoon; Princeton University
Presented by: So Hye Yoon, Princeton University
Discussant:   Zach Brown, University of Michigan
Session 6: Antitrust and Competition April 10, 2026 17:00 to 18:30 Location: Mediterranean (3rd Floor)
Session Chair: Robert Porter, Northwestern University
Vertical Integration and Regulated Profits in Pharmacy Benefits Markets Abstract This paper studies the effects of vertical integration between insurers, pharmacy benefit managers (PBMs), and pharmacies on drug prices and insurance premiums. I construct an empirical model of pharmacy pricing, insurer premium setting, and consumer demand for insurance plans and pharmacies in Medicare Part D, a government program that provides subsidized drug insurance to older adults in the United States. I estimate the model using prescription drug claims data, which I combine with novel information on insurer-PBM relationships and pharmacy ownership. In equilibrium, vertically integrated insurers reduce premiums and increase internal prices for prescription fills, shifting profits to their pharmacies. Two institutional features motivate this profit-shifting strategy: consumer cost-sharing, which allows firms to retain profits on integrated prescription fills; and regulatory caps on insurer profits, which incentivize firms to “tunnel” excess profits to pharmacies through higher drug prices. My estimated model predicts that the divestiture of vertically integrated pharmacies would reduce drug prices by 7.3% and increase annual consumer surplus for Medicare enrollees by 8.1%.
By Eric Yde; University of Virginia
Presented by: Eric Yde, University of Virginia
Discussant:   Keith Ericson, Boston University
Collective Decision-Making in Global Antitrust Enforcement: Evidence from Shipyard Mergers Abstract This paper examines how factors beyond market impact and strategic interaction among authorities shape antitrust decision-making in shipyard mergers. Modeling merger reviews as unanimous voting games, we rationalize two empirical patterns: authorities may become more lenient due to trade or political considerations, with many large ship-buying countries rubber-stamping the mergers. Structural estimation shows that removing non-market considerations or incentives to free-ride on other authorities leads to stricter enforcement. As the probability of obtaining unanimous approval falls by 0.64–1.04 percentage points, anti-competitive mergers become less likely to materialize, resulting in consumer welfare gains equivalent to 4.1–5.8% of annual global ship-sales revenue.
By Chuyue Tian; Stanford University, Graduate School of Business    Hsin-Tien Tiffany Tsai; National University of Singapore
Presented by: Chuyue Tian, Stanford University, Graduate School of Business
Discussant:   Myrto Kalouptsidi, Harvard University
Airline Industry Dynamics 1993–2022 Abstract I examine the U.S. airline industry from 1993 to 2022, a period marked by financial distress and turbulent dynamics in demand, costs, and industry structure. I estimate a structural model with semiparametric demand, Nash–Bertrand competition, and nonparametric marginal costs for each year to capture flexible relationships among product characteristics and their evolution over time. The results show that markups and producer surplus fell over the period, with most of the decline occurring in the late 1990s and early 2000s. Counterfactual decompositions indicate that increased consumer price sensitivity was the dominant driver of the declines, outweighing increased costs and other factors, while the effects of mergers and low-cost carrier expansion on markups largely offset each other.
By Sunmin Kim; The Ohio State University
Presented by: Sunmin Kim, The Ohio State University
Discussant:   Charles Murry, University of Michigan
Session 7: EVs and Green Energy April 10, 2026 17:00 to 18:30 Location: Caspian (3rd Floor)
Session Chair: Lawrence White, Stern School of Business, New York Unive
Investment and the Transfer of Power: Dynamic Effects of Transmission in Electricity Markets Abstract Renewable resources are essential for energy transition, but their unequal geographic distribution limits broader adoption. Long-distance transmission offers a potential solution by connecting renewables-rich areas to high-demand markets. I examine how expanded transmission capacity affects investment decisions by generators. I estimate a dynamic model of generator behavior consisting of a short-run optimal dispatch problem incorporating line losses and transmission constraints and a long-run dynamic game for capacity investment. Using an originally constructed dataset combining EIA, ISO, and proprietary data to estimate the model from 2018--2023, I find that upgrading all inter-zonal transmission to modern HVDC standards would reduce average wholesale prices by 6\%, generating approximately \$4 billion in annual welfare gains. Regional effects are heterogeneous: the Northeast experiences the largest price declines while the Midwest sees modest increases. In the long run, adding major transmission projects such as the Grain Belt Express increases solar and wind adoption in the Eastern Interconnection by over 10\% by 2050. I use these estimates to evaluate proposed policies including the Inflation Reduction Act, and the Bipartisan Infrastructure Law.
By Dana Golden; Stony Brook University
Presented by: Dana Golden, Stony Brook University
Discussant:   Jonathan Elliott, Johns Hopkins University
Estimating High-Dimensional Dynamic Games: Innovation Policy in the EV Industry Abstract Fine-grained economic study in dynamic settings often requires specifying many state variables, such as when multiple agents are involved or when a rich description is necessary. Yet this brings the curse of dimensionality that has long posed a major estimation challenge in the dynamic modeling literature. To address this issue, I develop a layered-local-iteration (LLI) method that exploits the layered structure of state spaces to perform within-layer parallel computation on GPUs, achieving speedups by orders of magnitude in solving the Bellman equation. By dramatically reducing computation time, LLI substantially expands the class of dynamic models that can be feasibly estimated in empirical work. I apply LLI to study the innovation dynamics of the Chinese electric-vehicle industry. To capture the competitive mechanisms, e.g., knowledge spillovers and infrastructure free-riding, among multiple heterogeneous firms, a high-dimensional state space is necessary. Counterfactual analysis highlights the importance of incorporating firm heterogeneity in innovation policy design: the welfare-maximizing subsidy scheme is a stage-specific mix of R&D subsidies and public-charging subsidies that steers firms toward investments aligned with their comparative advantages, amplifying knowledge spillovers and infrastructure compatibility.
By Zikun Liu; Yale University
Presented by: Zikun Liu, Yale University
Discussant:   Adam Dearing, Cornell University
Innovation Path Choices in China’s Electric Vehicle Battery Industry Abstract Green technologies that are equally clean can differ substantially in their cost structures and welfare implications, yet little is known about whether market forces direct innovation toward the socially optimal technological path when multiple green options coexist. This paper studies innovation path choices in China’s electric vehicle (EV) battery industry, asking whether market driven innovation aligns with the social planner’s preferred path and how industrial policies can correct potential distortions. The paper contributes by quantifying distortions in firms’ technology path choices between two emerging clean technologies and by evaluating how alternative policy instruments shape the direction of green innovation. I focus on two dominant battery technologies, Lithium-Iron-Phosphate (LFP) and Nickel-Cobalt-Manganese (NCM), which differ sharply in innovation difficulty and production costs. I develop and estimate a finite-horizon dynamic structural model in which battery suppliers choose how much to innovate along each technological path by improving energy-density frontiers. They then bargain over battery prices with downstream EV manufacturers, who compete in prices for consumers choosing among differentiated vehicles. The framework compares firms' profit-maximizing innovation decisions to those of a social planner, who maximizes total surplus, encompassing consumer surplus, producer surplus, and environmental benefits. The model is estimated using a comprehensive panel dataset covering China’s passenger vehicle market from 2017 to 2023, combining vehicle-level sales, detailed vehicle characteristics, supplier relationships, battery input prices, and policy variables. I estimate consumer demand using a nested logit model, recover battery marginal costs through a bilateral bargaining framework, and identify innovation cost parameters for leading battery suppliers. Counterfactual analyses reveal substantial distortions in innovation direction. Relative to the social optimum, the market undertakes only one-fourth as much innovation in LFP, a technology with high sunk innovation costs but low marginal production costs, and twice as much in NCM, which has the opposite cost structure. This divergence is driven primarily by vertical separation and downstream competition, not uninternalized spillovers or environmental benefits. Policy simulations indicate that R&D subsidies are more effective than performance-based consumer subsidies in aligning innovation with the socially preferred path, highlighting industrial policy’s role in shaping both the level and direction of green innovation.
By Qian Wang; UMD
Presented by: Qian Wang, UMD
Discussant:   Paul Scott, Massachusetts Institute of Technology
Session 8: Platform Design and Information April 10, 2026 17:00 to 18:30 Location: Brewster (4th Floor)
Session Chair: Jeanine Miklos-Thal, U Rochester
Platform Governance and Information Disclosure in Markets with Add-Ons Abstract This paper investigates how ad-valorem platform commissions and governance strategies determine market transparency in the presence of consumer myopia. We model sellers who strategically shroud add-ons to sell them off-platform, thereby avoiding platform fees. We analyze market outcomes under a laissez-faire policy and compare them against disclosure mandates. Our analysis reveals a new mechanism: ad-valorem commissions act as a competition multiplier. Sellers pass their commission-free add-on profits through to consumers via lower base good prices, effectively subsidizing the base good. Consequently, a market with hidden fees can yield higher consumer surplus than a transparent one. These findings highlight an unintended consequence of transparency regulation: mandating disclosure may harm consumers by eliminating these implicit subsidies to the base good price.
By Huixin She; Toulouse School of Economics
Presented by: Huixin She, Toulouse School of Economics
Discussant:   Nima Haghpanah, Yale
Targeted Advertising Platforms: Data Sharing and Customer Poaching Abstract E-commerce platforms are rolling out ambitious targeted advertising initiatives that rely on merchants sharing customer data with each other via the platform. Yet current platform designs fail to address participating merchants' concerns about customer poaching. This paper proposes a model of designing targeted advertising platforms that incentivize merchants to voluntarily share customer data despite poaching concerns. I characterize the optimal mechanism that maximizes a weighted sum of platform's revenues, customer engagement, and merchants' surplus. In sufficiently large platforms, the optimal (indirect) mechanism consists of three markets: a buying and a selling market in which the platform buys and sells data (ads) at a posted price, respectively, and an exchange market in which merchants exchange all their data in return for higher quality ads. The model is broad in scope with applications in other combinatorial exchange settings such as community time banks and greenhouse gas credit market.
By Klajdi Hoxha; Stanford GSB
Presented by: Klajdi Hoxha, Stanford GSB
Discussant:   Heski Bar-Isaac, University of Toronto
Optimizing Multi-Stage Personalization in the Customer Journey Abstract Firms increasingly leverage personalization to influence product discovery and engagement throughout the customer journey. However, implementing effective multi-stage personalization is challenging because each stage’s impact depends on customer intent and attention, and cross-stage spillovers may reduce overall effectiveness. This paper examines personalization at two key stages: an early, firm-driven recommendation stage and a later, customer-initiated query stage. Using data from a field experiment on a major e-commerce platform, I find that recommendation-stage personalization increases immediate revenue but reduces revenue at the query stage, resulting in no net gain. I then develop and deploy a personalized search-ranking algorithm in a subsequent field experiment. The results show that query-stage personalization increases total transactions without cannibalizing revenue from the recommendation stage. To identify the revenue-maximizing multi-stage personalization strategy across platform designs and customers, I build a structural search model. I exploit experimental variation and estimate the model using a neural network approach to address computational challenges. Counterfactual simulations reveal that personalizing only at the query stage results in higher total revenue than personalizing at both stages or only at the recommendation stage. Moreover, applying personalization only to customers with consistent preferences further improves revenue. These findings identify the conditions under which personalization is most effective and offer firms guidance on optimizing it across stages and customers.
By Yu Song; University of Michigan
Presented by: Yu Song, University of Michigan
Discussant:   Vivek Bhattacharya, Northwestern University
Session 9: Regulating Externalities April 10, 2026 17:00 to 18:30 Location: Georges (4th Floor)
Session Chair: Joel Waldfogel, University of Minnesota
Environmental Regulation with Irreversible Investments: Evidence from High Plains Aquifer Depletion Abstract Many of the world’s major aquifers are being rapidly depleted from agricultural irrigation, generating dynamic common-pool externalities by raising future extraction costs. Entry restrictions are commonly used to limit depletion because well drilling is easily monitored, but they are second-best compared to Pigouvian taxes that directly target the intensive margin of water use. When policies cannot be tailored to heterogeneous users, however, the relative effectiveness and political feasibility of entry fees and water-use taxes become theoretically ambiguous, depending crucially on the correlation between water users' productivity and externalities. To study this question, we develop a dynamic model of farmers’ joint well-drilling and water-use decisions, integrated with a physically realistic model of groundwater flows, and estimate it using field-level data on aquifer levels, water use, and crop production in the Kansas High Plains Aquifer from 1959 to 2022. We find that field-level productivity and water-use externalities are strongly positively correlated due to the spatial concentration of high-productivity fields, leading uniform taxes to outperform entry fees in terms of aggregate welfare. Nevertheless, entry fees are preferred by most users because the optimal uniform tax exceeds the marginal social cost of water use for all but the most productive fields. However, driven by irreversible well investments that lock in depletion from high-externality early entrants, the effectiveness and popularity of entry fees decline rapidly over time. These findings highlight how heterogeneity and irreversibility jointly shape the efficiency and political feasibility of environmental regulation.
By Aaron Berman; MIT    Nathaniel Hickok; MIT
Presented by: Nathaniel Hickok, MIT
Discussant:   Luming Chen, University of Michigan
Navigating the Commons Abstract The overuse of open-access resources is a classic example of externalities. Inefficiencies arise not only from resource use by existing participants but also from their investment in capacity and the entry of new firms. Standard models of externalities, however, typically abstract from firms’ entry, exit, and capital accumulation. This paper develops a model of strategic firm dynamics with production externalities, in which firms interact through stock depletion and congestion. I estimate the model using firm-level panel data from the American whaling industry (1804–1909), an unregulated global commons. With the estimated model, I quantify the shadow prices of externalities and propose a tractable framework for optimal policy design. The results show that per-unit Pigouvian taxes substantially improve welfare yet fall short of the first best: they correct stock externalities but leave congestion unpriced, leading to persistent overcapacity. Optimal regulation combines per-unit taxes with state-dependent lump-sum fees that vary with vessel capacity and productivity to internalize firms’ dynamic interactions. The welfare effects of these policies depend critically on technology, demand, and resource regeneration, underscoring the importance of adaptive policy design.
By Yangkeun Yun; UCLA
Presented by: Yangkeun Yun, UCLA
Discussant:   Mitsuru Igami, University of Toronto
Strategic Decision on Network Intensity: Evidence from U.S. Firm-to-Firm R&D Collaboration Network Abstract Firm-to-firm R&D collaborations have become increasingly common, and firms have been assigning more inventors to these collaborative projects. Despite this trend, existing research on collaboration subsidies has mainly focused on the existence of collaborations rather than their intensity. In this paper, I develop a model in which firms jointly choose their R&D collaborators, the intensity of each collaboration, and their R&D investment. Accounting for heterogeneity in intensity is central to the optimal subsidy design. Using U.S. patent and accounting data, I find that increasing collaboration intensity from below-median to above-median raises collaboration costs by 23% but increases benefits from technological complementarities by more than twelvefold. Counterfactual analysis shows that, under a constrained government budget of less than $26 billion, targeting high-intensity collaborations is more cost-effective than subsidizing all collaborations uniformly. Although targeting high-intensity collaborations increases network congestion and attracts less R&D-efficient firms, it accelerates technological complementarities.
By Sinjeong Kim; Pennsylvania State University
Presented by: Sinjeong Kim, Pennsylvania State University
Discussant:   Quan Le, Harvard
Session 10: Regulation and Quality April 10, 2026 17:00 to 18:30 Location: Spectacle (4th Floor)
Session Chair: Alan Sorensen, University of Wisconsin - Madison
The Welfare Effects of Accountable Care Organizations Abstract This paper studies the welfare effects of Accountable Care Organizations (ACOs), a policy whose goal is to reduce healthcare spending for Traditional Medicare patients. ACOs are groups of healthcare providers whose performance is evaluated by comparing a spending benchmark to the total outpatient and inpatient spending of assigned patients. Spending below the benchmark results in a financial reward given to the ACO, while spending above the benchmark results in a penalty. The welfare impact of this policy is theoretically ambiguous, as lower spending may reduce quality of care. To quantify these effects, I develop a structural model of supply and demand in the Medicare outpatient facility market where facilities compete in outpatient spending and quality. The model incorporates endogenous quality provision and spillovers onto Medicare Advantage patients, whose healthcare spending is not directly targeted by the program. I estimate the model using the universe of hospital-based claims from New York State. A counterfactual simulation that removes ACOs from the market shows that ACOs increase welfare by $1.09 billion. Spillovers onto Medicare Advantage patients account for 32% of the consumer welfare gains.
By James Gluzman; Stony Brook University
Presented by: James Gluzman, Stony Brook University
Discussant:   Paul Grieco, Penn State
Delayed Information and Behavioral Spillovers: Evidence from U.S. Drinking Water Violations Abstract Timely public disclosure is a central component of environmental regulation, yet little is known about how delays in mandated notifications affect protective behavior and its spillovers across communities. This paper studies how the timing of public notification influences household avoidance behavior following acute microbial drinking water violations in the United States. I combine administrative data on drinking water violations and enforcement actions with high-frequency retail scanner data on bottled water sales from 2006–2019. I show that avoidance behavior responds sharply to timely public notification but is substantially muted when notification is delayed or missing. Event-study evidence indicates that households do not respond to the onset of contamination itself, but rather to the arrival of salient public information. I further document spillover effects in neighboring communities that are not directly affected by violations. These spillovers arise primarily within shared media markets and dissipate with geographic and informational distance, highlighting the role of local media in transmitting risk information. Alternative channels such as social connectedness and socio-demographic similarity do not generate comparable spillovers in the absence of media exposure. Together, the results demonstrate that the effectiveness of right-to-know policies depends critically on implementation, particularly the timing of information release, and that information disclosure can induce behavioral responses beyond the targeted population.
By Raushan Baizakova; Indiana University
Presented by: Raushan Baizakova, Indiana University
Discussant:   Peter Newberry, University of Georgia
Improving Quality through Regulation or Reputation: Theory and Evidence from the Environmental Impact Assessment Industry Abstract Ensuring reliable provision of high-quality goods and services remains a central challenge in many markets. This paper studies how reputation interacts with entry regulation and shapes optimal regulatory policies using empirical, theoretical, and quantitative approaches. Empirically, I show that an entry deregulation reform in the Environmental Impact Assessment (EIA) industry leads to a persistent decline in service quality. I also document that negative reputation shocks sharply reduce firms' market share by half, indicating that both regulation and reputation are powerful devices. Theoretically, I develop a firm entry game in which reputation partially substitutes for regulatory screening, weakening the quality-improving effect of entry barriers. Structurally, I estimate a dynamic oligopoly model within a Moment-based Markov Equilibrium framework that incorporates endogenous reputation formation. Counterfactual simulations reveal that accounting for reputation substantially alters the optimal stringency and timing of entry regulations. These results underscore that effective regulation in emerging markets must be designed jointly with the market’s reputation forces.
By Yucheng Quan; The University of Hong Kong
Presented by: Yucheng Quan, University of Hong Kong
Discussant:   Mo Xiao, University of Arizona
Session 11: Advertising, Perceptions, and Market Competition April 11, 2026 8:00 to 10:00 Location: Salon A (3rd Floor)
Session Chair: Aljoscha Janssen, Singapore Management University
Advertising for Dropout Consumers Abstract We analyze advertising for consumers who are imperfectly informed about quality and prices and for whom participation in the market is costly. When seller costs and quality are persistent, a seller's price may induce consumers to abstain in the next period, causing the seller to lose sales and the ability to demonstrate that costs have dropped and quality increased. Consumers may then continue to remain pessimistic and abstain thereafter. In equilibrium this tendency of consumers to “drop out" leads to non-uniform advertising dynamics known as “pulse advertising".
By Arthur Fishman; Bar Ilan University
Presented by: Arthur Fishman, Bar Ilan University
Discussant:   Marcel Preuss, Cornell University
Evidence for the Illusion of Competition Abstract In most consumer markets, firms sell similar products under multiple brand names. Hence, there are typically more brands for sale than competing firms selling them. This holds in over 77% of 358 consumer markets tracked by MRI Simmons. In the median market, there are 27% fewer firms than major brands. We document, using a novel survey design, that consumers severely underestimate brand co-ownership and overestimate the number of competing firms across eight studied markets. We call this bias the illusion of competition. If the number of brands exceeds the number of firms by five, on average subjects overestimate the number of firms by four. Further, if two brands are co-owned, subjects realize this only 8.1% of the time. As a result, subjects’ responses imply substantially underestimated measures of market concentration. Using another novel experimental design, we show that correcting the illusion of competition causes subjects to believe markets are less competitive and to be more supportive of antitrust policies. Documenting the prevalence of the illusion of competition is important because our companion paper shows that the bias can lead to higher equilibrium prices (Grubb and Westphal, 2026).
By Michael Grubb; Boston College    Ryan Westphal; Brandeis University
Presented by: Ryan Westphal, Brandeis University
Discussant:   Quinn Maingi, USC Marshall School of Business
Uncertainty in Dynamic Pricing Environments Abstract We develop a model of consumer choice with dynamic pricing and valuation uncertainty, and allow for purchase delay. Simulating the model, we find that delay is less likely when the expected increase in price is relatively large, even with high uncertainty. Using novel data that include information on trip characteristics, passenger demographics, fares, and timing of purchases from the Survey of International Air Travelers, we document delay patterns and price heterogeneity in the airline industry. We then perform a series of empirical tests to study selection into an arrival date, the propensity to delay, and the consequences of delay. Although trip and passenger characteristics drive much of the arrival behavior, the evolution of prices plays a role in the choice to delay. Consistent with our model, consumers in flight segments with relatively steep price paths are less likely to delay. Finally, controlling for trip and passenger characteristics and the purchase period, passengers who delay pay slightly more on average, though this effect disappears if we consider longer delay periods. This may suggest that consumers are willing to delay and pay more in order to reduce their uncertainty, but that they are unwilling to delay long periods of time if prices continue to increase.
By Garrett Scott; University of Mississippi
Presented by: Garrett Scott, University of Mississippi
Discussant:   Johannes Kandelhardt,
Patriotic Spots, Polarized Markets: Bud Light’s Advertising Elasticity During a Nationwide Boycott Abstract This article examines how Bud Light’s 2023 boycott and subsequent change to a patriotic, untargeted TV advertising campaign reshaped advertising effectiveness across U.S. markets. Using a triple-difference border design comparing bordering counties split by Designated Market Areas before and after the boycott, I find that the elasticity of sales to TV advertising collapsed nationwide. Post-boycott, advertising became counterproductive in liberal counties—turning the elasticity negative—while stabilizing around zero in conservative areas. Event-study evidence confirms flat pre-trends and dates the structural break precisely to April 2023. Despite increased advertising spend, sales declined broadly, suggesting that uniform messaging in a polarized environment exacerbated losses. Counterfactual reallocations show that shifting up to half of post-boycott ads from liberal to conservative markets could have mitigated the shortfall by about 26%, while a temporary full pause of TV ads would have reduced losses by over 50%. As boycotts and reputational risks grow more frequent, the findings show the strategic importance of agile, data-driven, and politically informed media planning—sometimes the best response may be to do less.
By Aljoscha Janssen; Singapore Management University
Presented by: Aljoscha Janssen, Singapore Management University
Discussant:   Yunhao Huang, University of Southern California
Session 12: Health Care Markets: Bargaining, Integration, and Provider Performance (Sponsored by FTI Consulting) April 11, 2026 8:00 to 10:00 Location: Atlantic 2 (3rd Floor)
Session Chair: Yi Wang, Purdue University
Dynamic Bargaining between Hospitals and Insurers Abstract This paper quantifies the role of government-driven benchmark rate inflation in spending on behalf of privately insured patients. When contracts are formed simultaneously, anticipated price inflation has no effect in net present value terms. When contracts are multiyear and staggered and negotiators discount future payments, anticipated inflation is passed through to real spending due to the asymmetric discounting from the perspective of negotiation and the market. I leverage panel data on hospital–insurer contracts from West Virginia to show that contracts are multiyear and staggered, with even short-lived contracts remaining in place for three years or more. I estimate a structural model of bargaining with staggered contracts to characterize the degree to which negotiators discount future profits. I find that negotiators were substantially, but incompletely, forward-looking: I reject the null hypothesis of myopia and estimate an annual discounting rate of β = 0.899. I use the estimated dynamic model to quantify the forward-looking response to a proposed government-driven reform that would have increased private payment inflation between negotiations. The reform would not have any effect under a static model. I find that the reform would increase private spending after nine years by $4.98 billion, while a myopic model lacking forward-looking responses would overestimate the effect by $2.35 billion and miss short-term dynamics, including the possibility of payment decreases.
By Jacob Dorn; Cornell University
Presented by: Jacob Dorn, Cornell University
Discussant:   Giulia Sabbadini, Düsseldorf Institute for Competition Economics
Vertical Integration and Regulatory Compliance in U.S. Health Insurance Markets Abstract I analyze how vertical integration between insurers and providers affects the compliance with Medical Loss Ratio regulation and the consumer welfare in the U.S. health insurance market, using a rich dataset from 2018 to 2023. Reduced-form evidence suggests that vertical integration raises the medical loss ratios of their subsidiary insurers and enables previously non-compliant insurers to achieve regulatory compliance. Motivated by these reduced-form results, I propose and estimate a model of insurer–provider bargaining and find that removing vertical integration leads to lower negotiated prices and premiums, improving consumer welfare. In contrast, stricter regulations raise both prices and total spending. Therefore, vertical integration can weaken the effectiveness of Medical Loss Ratio regulation and reduce its intended consumer welfare.
By Rui Li; SUNY Albany
Presented by: Rui Li, SUNY Albany
Discussant:   Yonggeun Jung, University of Florida
Estimating Production Functions with Latent Team Structures: An Analysis of Nursing Homes Abstract I consider robust specification and estimation of production functions when the researcher observes a disaggregated vector of endogenous labor inputs. Drawing on personnel and organizational economics, I develop a latent model of matching teams of worker types with bundles of tasks under time constraints and costly team formation. I adapt its implications into a penalized and shape-constrained GMM estimator and establish its consistency. Applying it to the US nursing home industry, I estimate revenue generation and predict counterfactual labor demand and health outcomes under a proposed targeted minimum staffing mandate. I find that the policy improves care quality for long-stay patients but has mixed effects for short-stay patients: it narrows disparities and raises bottom-decile quality, but reduces mean and top-decile quality.
By Nihal Mehta
Presented by: Nihal Mehta,
Discussant:   Liang Zhao, University of Cambridge
Testing Vertical Relationships in the U.S. Infant Formula Market: Implications for Government Costs and Welfare Abstract The U.S. Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) provides free infant formula to low-income households, serving around 39% of U.S. infants. As WIC’s single most expensive benefit, infant formula accounts for around half of the total WIC food costs. To reduce government costs, WIC awards exclusive contracts to manufacturers offering the lowest net price (i.e., wholesale price minus rebate) to the government in each state via public auction, effectively creating monopolies in the WIC market. The broader implications of this policy hinge on vertical relationships between manufacturers and retailers, which remain poorly understood. I identify the vertical relationship as best characterized by two-part tariffs (TPT), where retailers decide retail prices and pay fixed fees to manufacturers and wholesale markups are zero. This finding challenges the common but untested resale price maintenance (RPM) assumption in the literature. Counterfactual simulations show that TPT is more efficient than RPM, yielding higher consumer and total surplus while reducing government costs through lower total markups at equivalent net prices. Specifically, under RPM, average retail prices would be 3.8% higher, consumer surplus would be 7.4% lower, and producer surplus would be 2.2% higher under the current design of WIC. These findings demonstrate that conduct assumptions critically shape welfare outcomes.
By Yi Wang; Purdue University    Juan Sesmero; Purdue University    Meilin Ma; Purdue University
Presented by: Yi Wang, Purdue University
Discussant:   Benjamin Rosa, University of Michigan
Session 13: Technology and Telecommunication April 11, 2026 8:00 to 10:00 Location: Salon C (3rd Floor)
Session Chair: Audrey Tiew, New York University
Regulatory Reform and Market Evolution in the Spectrum Commons Abstract How does a decline in regulatory burdens shape product introduction, firm participation, and competitive dynamics in technology markets? This study examines the Federal Communications Commission’s transformation of wireless device regulation between 1997 and 2001. Using administrative records covering nearly 200,000 equipment authorization applications between 1990 and 2015, the study documents market, firm, and product outcomes following a decline in regulatory costs. The study documents four broad patterns. First, regulatory processing times fell sharply, followed by a sustained expansion in product introductions and firm participation, especially in unlicensed spectrum applications. Second, this expansion was driven disproportionately by new entrants, who contributed extensively to product variety and experimentation. Third, despite heightened entry and competition, established firms retained persistent survival advantages. Fourth, the combination of widespread entry and incumbent persistence is difficult to reconcile with strong forms of regulatory capture. Taken together, the findings suggest that well-designed regulatory reforms can simultaneously enable entrant-led experimentation while preserving incumbent persistence. The findings inform contemporary debates on regulating emerging technologies, in which similar privatized certification models are being considered.
By Haiyang Zhang; Harvard Business School    Shane Greenstein; Harvard Business School    Roberto Fontana; University di Pavia    Do Yoon Kim; Boston College
Presented by: Haiyang Zhang, Harvard Business School
Discussant:   Anil Jain, Federal Reserve Board
Secrecy, Spectrum, and Certification: Evidence from Wireless Electronic Devices Abstract This paper examines the strategic decisions of firms to withhold information in innovation-intensive markets, highlighting that secrecy contains distinct temporal dimensions. The study illustrates the thesis using wireless products, the majority of which utilize unlicensed spectrum. Using all FCC equipment filings from 2001 to 2021, we analyze firms' choices between long-term confidentiality (LTC), which permanently conceals technical information such as block diagrams, and short-term confidentiality (STC), which temporarily delays the public release of marketing information, including user manuals. We argue that the former acts to preserve "trade secrets," the traditional role of IPR. In contrast, the latter acts as "commercialization insurance," a previously unrecognized role for secrecy to ensure against launch and coordination risks in the commercialization process. We examine evidence consistent with this view. We find that the determinants for LTC and STC differ sharply. LTC use reflects appropriability concerns and institutional complexities—technically complex products and competitive markets are 12 to 35% more likely to adopt it. Traditional forms of IPR, such as patenting, are a substitute for LTC, especially in the case of WiFi. By contrast, STC is driven by the commercialization timing of branded products by large firms. Large, domestic, patent-owning firms with regulatorily complex products are more likely to use STC. Technical aspects of the product do not predict STC usage. Usage is consistently higher among incumbents for STC, while it quickly converges for LTC, even after controlling for all other variables. An exposure-based event study further shows that firms with a high ex-ante propensity to use STC bring products to market faster after the policy becomes available, with no corresponding changes in product complexity or volume. Together, these results show that permanent and temporary confidentiality serve distinct strategic purposes.
By Ruby Zhang; Harvard University    Roberto Fontana; University di Pavia    Shane Greenstein; Harvard Business School    Do Yoon Kim; Boston College
Presented by: Ruby Zhang, Harvard University
Discussant:   Jonas Boeschemeier, Technical University of Munich
Broadband Deployment in Equilibrium: Subsidy Design for the Digital Divide Abstract Broadband access has become a near necessity, yet many U.S. households remain without high-speed access or significant choice among Internet service providers. This article examines the potential effects of policies that have been proposed to ameliorate the digital divide. Combining data from a survey of Seattle households’ broadband subscriptions and broadband deployment data from the FCC, I estimate a dynamic structural model of Seattle’s broadband market, which allows me to quantify the effect of competition on broadband availability and quality in equilibrium. I find that, of recent policies proposed to address the digital divide, a demand-side subsidy program increasing broadband affordability for low-income households is significantly more cost effective than a supply-side policy that subsidizes increased broadband deployment.
By Andrew Kearns; Federal Trade Commission
Presented by: Andrew Kearns, Federal Trade Commission
Discussant:   Hengyi Huang, Tilburg University
Industrial Policy, Location Choice, and Firm Performance in High-Tech Manufacturing Abstract National policies from the U.S. and China- driven by concerns over security and industrial self-reliance - now heavily influence investment incentives in high-tech manufacturing industries. This paper examines the effects of industrial policy on contract manufacturing capacity investment tradeoffs in the global semiconductor industry. We assemble a novel dataset that combines quarterly facility-level capacity investments with global contract manufacturing orders from 2004 to 2015. Using these data, we estimate a structural model of contracting between semiconductor manufacturers and their clients, recovering key cost parameters that capture the differences in cost advantages across firms, geographies, and relative buyer locations. We apply the model to a detailed case study of a major semiconductor manufacturer evaluating whether to locate a large fabrication facility in the U.S. or remain in its home region amid shifting industrial policies. In counterfactual simulations, we find that locating in the U.S. would require an additional investment of $1.2 billion compared to the home region, which could roughly be offset by lump-sum subsidies comparable (percentage-wise) to those provided under the CHIPS and Science Act. However, the profit reduction caused by cost disadvantages and reduced competitiveness at the U.S. location has a greater longer-run impact. Across various policy scenarios - including a no-policy baseline, U.S. export controls alone, Chinese tax subsidies alone, or a combination of both - locating in the U.S. consistently results in an additional $1.6 to $1.8 billion profit loss compared to staying in the home region. This loss is concentrated in later years as the facility's technology matures. Across all simulated scenarios, industrial policies reduce manufacturing profits by $3.1 to $10.6 billion relative to the no-policy baseline, but U.S. import tariffs substantially change the pattern of these losses: under our conservative tariff rate calibration, relocating to the U.S. increases the firm's profit by $5.9 billion relative to remaining in its home region, primarily by offsetting the cost disadvantages of a U.S. facility and enhancing the firm's competitiveness in capturing U.S. demand, particularly for mature technologies.
By Audrey Tiew; New York University    Ran Zhuo; University of Michigan, Ross School of B
Presented by: Audrey Tiew, New York University
Discussant:   Rachel Nam, USI Lugano/Swiss Finance Institute
Session 14: Advances in Demand Estimation April 11, 2026 8:00 to 10:00 Location: Atlantic 1 (3rd Floor)
Session Chair: Zach Brown, University of Michigan
Identifying Preference Heterogeneity in the BLP Model with Micro Data Abstract We non-parametrically assess the identifying content of micro data in the “BLP” random coefficients logit model of aggregate demand. First, we show that the Jacobian of inverse demand w.r.t. market share encodes all information about preference heterogeneity in this model. We then show that three types of micro data identify various aspects of the Jacobian. Demographic-driven variation in data on one purchase per individual identifies reduced-rank transformations of the Jacobian. Second-choice data identifies the Jacobian up to diversion ratios. Longitudinal data, with multiple purchase events by each individual, identifies the entire Jacobian. Thus, longitudinal data contains the most identifying content, while second-choice data's identifying content is particularly relevant for measuring product rivalry. We also show how micro data can identify hidden nesting structures. Finally, we discuss the added value of micro data that is compatible with market-level data.
By Adam Dearing; Cornell University
Presented by: Adam Dearing, Cornell University
Discussant:   Felipe Barbieri, Dartmouth Tuck
Sequential algorithm for structural estimations with equilibrium constraints Abstract This study investigates sequential algorithms exhibiting the Zero Jacobian Property (ZJP) for estimating structural models with equilibrium constraints. For the Maximum Likelihood Estimation (MLE) and the Generalized Method of Moments (GMM), the current study shows that they attains fast (near-quadratic) local convergence in large samples to the solution of the constrained optimization problem. If the initial consistent estimate of the parameters are available, the algorithms yield asymptotically efficient estimator even after one iteration. It then proposes a novel algorithm called Sequential Linearly Constrained (SLC) algorithm, which is applicable to a broader class of structural models than existing methods. SLC is attractive in that it can be implemented without explicitly computing the Jacobian of the equilibrium constraints and can be several times faster than the Nested Fixed Point (NFXP) approach. The current study demonstrates the good performance through two numerical experiments: a dynamic discrete game with time-varying unobserved heterogeneity and a dynamic demand model.
By Takeshi Fukasawa; Waseda University
Presented by: Takeshi Fukasawa, Waseda University
Discussant:   Tianyi Li, The Pennsylvania State University
Leaving the Nest: Simulated Cardell Errors for Flexible Dependence in Panel Discrete Choice Abstract We develop a panel discrete-choice demand framework that allows flexible correlation in unobserved utility across products and over time while preserving logit-form choice probabilities. A key property of our framework is a variance-components construction based on simulated draws from the Cardell distribution. Our approach exploits a convolution property, the sum of a Cardell shock and scaled type-I extreme-value term is itself distributed type-I extreme-value. Using this property, we obtain a logit kernel conditional on draws. The model nests standard logit-based specifications, including nested logit, and also accommodates overlapping nests and serial correlation. We propose a parsimonious AR--persistent specification and estimate persistence in long panels via a pairwise (composite) likelihood that avoids high-dimensional integration. Monte Carlo experiments show accurate recovery of price sensitivity and persistence and reliable elasticity estimates. We also show bias in welfare estimates when ignoring persistence. We illustrate the approach using linked household-panel and scanner data on over-the-counter medicines.
By Aaron Kaye; Boston University    Zach Brown; University of Michigan
Presented by: Zach Brown, University of Michigan
Discussant:   Anders Munk-Nielsen, University of Copenhagen
Salience and Buyer's Remorse: Optimal Nonlinear Pricing with Cognitively Constrained Consumers Abstract Nonlinear pricing theory predicts that firms can extract surplus by inducing heterogeneous consumers to self-sort across price contract offers that are ex-post optimal for them. We study subscription pricing when the frictionless sorting assumption fails. Using large-scale subscription experiments conducted by Lyft, we document systematic deviations from optimal self-selection: many high-demand consumers decline subscriptions that would have saved them money, while some subscribers fail to break even. We develop a structural model of intensive-margin demand in which consumers may exhibit salience failures, forecast errors about future demand, or impulsivity. We show that subscription uptake can be recast as onesided noncompliance in a binary-instrument framework, allowing us to leverage LATE methods to identify counterfactual outcome distributions and a novel “uptake function” linking baseline outcomes to compliance behavior. Combining experimental price variation with this identification strategy, we recover utility primitives, demand heterogeneity, and behavioral parameters. Salience failures and forecast errors play quantitatively important roles. Counterfactual analyses show that optimal subscription pricing generates substantial gains relative to linear pricing, but these gains are highly sensitive to consumer deviations from ex-post optimal choice. Implementing nonlinear pricing therefore requires not only optimal contract design for consumer screening, but also coordinated efforts to mitigate behavioral frictions.
By Gregory Sun; Washington University in St. Louis
Presented by: Gregory Sun, Washington University in St. Louis
Discussant:   Sebastien Cerles, Paris-Saclay University
Session 15: Regulation, Product Variety, and Quality April 11, 2026 8:00 to 10:00 Location: Caspian (3rd Floor)
Session Chair: Jonathan Scott, University of Texas, Dallas
Opening Hours and Consumer Behavior: Evidence from GPS Data and Deregulation Abstract In 2019, North Dakota repealed its Sunday closing law, which had required most non-grocery stores to close between midnight and noon. Using this policy change and consumer GPS data, we study the impact of opening hours on shopping behavior and welfare. We compare visits before and after the repeal in North Dakota and neighboring states using difference-in-differences and event-study designs. The repeal caused a large increase in Sunday morning visits, originating partly from intertemporal, store-type, and cross-border substitution. The closing law’s welfare loss is equivalent to increasing the travel distance to affected stores by about 1.4 miles per consumer.
By Javier Donna; University of Miami    Marit Hinnosaar; University of Nottingham    Toomas Hinnosaar; University of Nottingham    Andre Trindade; Nova SBE
Presented by: Andre Trindade, Nova SBE
Discussant:   Jonathan Scott, University of Texas, Dallas
The Price of Portability: How Interstate Licensing Transforms Mental Health Markets Abstract Interstate occupational licensure compacts are an increasingly popular way of addressing provider shortages in mental healthcare markets. I study the effects of licensure compacts on the size and composition of the psychologist workforce. I develop an empirical model of entry for psychologists differentiated by whether they provide in-person or virtual care. Results show that the Psychologist Interjurisdictional Compact increased overall supply by 2.9 psychologists per market but decreased in-person supply by 0.59 psychologists. Fifty-three percent of this decline is attributable to increased competitive pressure from out-of-state therapists. While geographic license portability increases supply, it reduces access to in-person care.
By Cici McNamara; Georgia Institute of Technology
Presented by: Cici McNamara, Georgia Institute of Technology
Discussant:   Conor Ryan, Penn State University
Vertical Integration and E-commerce Competition: Evidence from Amazon Marketplace Abstract Many e-commerce platforms are vertically integrated and compete directly with third-party (3P) sellers. This raises potential competition concerns, as platforms may leverage their market power in ways that harm 3P-sellers and consumers. To analyze the effects of vertical integration on competition and welfare, I build a structural model that captures both the pricing and the product offering decisions of the platform and 3P-sellers. Using data from Amazon, I estimate the model and examine the impact of a policy prohibiting Amazon from selling products on its own marketplace. The counterfactual analysis reveals that while the policy decreases consumer welfare by 18.6%, subsequent entry by other sellers recovers up to 4.4 percentage points of this loss.
By Luca Bennati; Central Bank of Mexico
Presented by: Luca Bennati, Central Bank of Mexico
Discussant:   Imke Reimers, Cornell University
The Long Run Costs of Short Run Mispricing: How Dynamic Generation Costs and Local Traffic Patterns Shape the Charging Network Abstract The spatial expansion of public electric vehicle (EV) charging infrastructure increasingly interacts with electricity markets in which marginal generation costs and environmental damages vary sharply over the day. This paper shows that time-of-use (TOU) pricing can shape long-run charging station investment by interacting with predictable, exogenous local traffic patterns. When charging demand concentrates during low marginal cost hours in a particular market, firms face higher margins and a stronger incentive to enter. We exploit quasi-experimental changes in TOU rate schedules across California utilities and their interaction with exogenous, intra-day foot traffic patterns across local markets to estimate these incentives in an entry setting. Incorporating hourly marginal damages from electricity generation allows us to characterize the socially efficient spatial distribution of public charging infrastructure and to assess how closely current TOU designs align private incentives with social marginal costs.
By Molin Qin; University of Wisconsin, Madison    Jonathan Scott; University of Texas, Dallas
Presented by: Jonathan Scott, University of Texas, Dallas
Discussant:   Anette Boom, Copenhagen Business School
Session 16: Experimentation, Learning, and Dynamic Investment April 11, 2026 8:00 to 10:00 Location: Aegean (3rd Floor)
Session Chair: Mohaddeseh Heydari Nejad, Indiana University
Over-Experimentation Under Delegation: Forward–Reverse Stopping Contracts Abstract We study a dynamic delegation problem in which a long-lived principal assigns experimentation tasks to short-lived agents, each active for only one period. The principal values experimentation more due to her longer horizon. We model this as a contextual bandit problem with type-dependent Poisson outcomes: one agent type has a known success probability, while the other is uncertain (either high or low) and is gradually learned. When the principal observes the agent’s type, the first-best policy features experimentation cycles, with phases of engagement, abandonment, and revival driven by belief dynamics and finite horizons. When the agent’s type is privately known, the optimal delegation rule takes the form of a forward–reverse stopping contract (FRSC) that uses the timing of the risky task as a screening device. Relative to the principal’s first best, the optimal FRSC generates both over- and under-experimentation, as the principal trades off current efficiency against future learning.
By Liang Zhong; The University of Hong Kong
Presented by: Liang Zhong, The University of Hong Kong
Discussant:   Tommaso Alba, KU Leuven
Job Security and Innovation Abstract We develop a theory in which a lower economic cost of unemployment increases workers’ willingness to join risky startups, thereby depressing negotiated wages relative to safer firms. These lower wages incentivize endogenous experimentation by young firms, activities that are risky but hold the potential for exceptionally high productivity, ultimately boosting aggregate productivity. Using Danish employer-employee matched data and exploiting geographical variation, we empirically test this mechanism and show that wages at experimenting startups are lower relative to non-experimenting firms in labor markets with higher job-finding rates—a pattern that holds both across firms and within firms that hire workers across multiple local labor markets.
By renato faccini; bank of england    Renato Faccini; Danmarks Nationalbank    Seho Kim; Danmarks Nationalbank    Javier Miranda; IWH/Friedrick-Schiller University
Presented by: Javier Miranda, IWH/Friedrick-Schiller University
Discussant:   Furkan Kilic, University of Chicago
Investment, Productivity, and Selection in the U.S. Shale Boom Abstract Why was the U.S. shale oil and gas revolution so revolutionary? As the U.S. Energy Information Administration quipped in 2024, ``the U.S. produces more crude oil than any country, ever''. The current, record, rate of production has been achieved though increases in the industry's oil and gas production per well, which have out-weighed a decrease in the drilling of new wells since the onset of the boom in 2010. This project asks how, why, and when the industry achieved these gains. Our primary focus, at least for now, is on decomposing the evolution of output per well into changes due to drilling site selection versus changes due to firms' adoption of improved technologies or fracking inputs. Site selection could be positive (better geologic locations are drilled first), negative (firms learn over time which locations are best), or some of both. We evaluate these possibilities by developing and estimating a joint model of oil production and drilling decisions. While the model is tailored to the shale oil and gas setting, its core ideas are applicable to other settings in which the productive outcome of an investment is a function of both the investment's location and the investing firm's skill in executing the project, conditional on location. The model uses local variation in land leasing difficulty as identifying variation that shifts the timing of firms' first well drilled in each location. And it accounts for firms' ability to learn from previously drilled wells' production realizations before deciding whether to drill additional wells in the same location. Using data from the Bakken Shale in North Dakota, we find evidence of positive selection early in the boom and negative selection later, but these effects are swamped by a large increase (~0.5 log points) in output per well that is driven by changes in firms' inputs and application of technology, conditional on location. Most of this increase occurred shortly after the sharp fall in oil prices and drilling activity in late 2014, consistent with ``slack time'' theories of innovation.
By Thomas Covert; AirBnb    Konan Hara; Michigan State University    Ryan Kellogg; University of Chicago    Richard Sweeney; Boston College
Presented by: Konan Hara, Michigan State University
Discussant:   Jordi Jaumandreu, Boston University
A Structural Model of Mentorship in Startup Accelerators: Matching, Learning, and Value Creation Abstract Breakthrough innovations are a key driver of long-run growth, yet early-stage innovation markets are characterized by high uncertainty and scarce resources, making it difficult to identify and support ventures with transformative potential. Intermediaries such as startup accelerators play a central role in identifying high-potential ventures and allocating scarce resources toward them. This paper studies the mechanisms through which mentorship creates value in the Creative Destruction Lab (CDL), a global mentorship-driven accelerator. I develop and estimate a dynamic structural model in which mentorship generates value through two channels: (i) a direct effect, in which mentors’ inputs improve startup quality, and (ii) a screening effect, in which mentors learn about latent venture quality and reallocate attention toward higher-potential ventures. The model allows for information spillovers across mentors, capturing how quality signals generated in mentorship interactions diffuse through the mentor network. The model enables counterfactual analysis that quantifies the value added by mentors when they actively shape startups’ strategic direction, relative to a regime in which mentors primarily support the execution of entrepreneurs’ original strategies. The estimates show that mentorship produces substantial gains through both channels. Implementing mentor-provided objectives raises startup quality by 67\%, while mentor learning improves the identification of high-quality ventures over time. Although individual mentors learn slowly about underlying quality, the results reveal large information spillovers across the mentor pool, implying that mentorship generates externalities within the accelerator. Counterfactual simulations show that the optimal mentorship design varies across sectors: active strategic guidance is most valuable in established sectors, whereas in emerging sectors, such as quantum, a more passive mentorship approach yields better outcomes as uncertainty resolves over time. These findings highlight the role of intermediaries in early-stage innovation markets and inform the design of mentorship-driven accelerator programs.
By Mohaddeseh Heydari Nejad; Indiana University
Presented by: Mohaddeseh Heydari Nejad, Indiana University
Discussant:   Sabien Dobbelaere, Vrije Universiteit Amsterdam
Session 17: Digital Platforms and Regulation April 11, 2026 8:00 to 10:00 Location: Georges (4th Floor)
Session Chair: Filomena Garcia, North Carolina State University
When does quality beat data? Gaining momentum in dynamic platform competition Abstract We consider a dynamic model with two horizontally differentiated platforms that serve heterogeneous users. In each period, platforms collect data from users that improves their services in future periods. The incumbent platform starts with a larger initial data base, while the entrant platform has a higher base quality. We find that when the value that users derive from data is low, the entrant gains momentum–gradually increasing its market share over time and eventually dominating the market. This occurs even if the entrant has an initial lower market share than the incumbent. When the value of data is high, the opposite occurs and the incumbent strengthens its position. We further find that greater product differentiation benefits the entrant, making it more likely for it to gain momentum and dominate the market. Finally, a policy that mandates reciprocal data sharing between platforms leads to long-run coexistence, with both platforms being active in the long run.
By Sarit Markovich; Northwestern University
Presented by: Sarit Markovich, Northwestern University
Discussant:   Jay Lee, University of Florida
A Spatial Model of Dynamic Pricing, Matching Frictions, and Taxi Search Abstract Dynamic pricing is widely used to manage matching frictions in platform markets. With spatial frictions, price variation shifts both passenger demand and driver locations, which in turn changes waiting times and matching efficiency. This paper develops and estimates a spatial equilibrium model of taxi driver location choice in the presence of dynamic pricing by Transportation Network Providers (TNPs). The model combines a spatial matching technology, a taxi demand system in which TNP prices shift demand toward taxis, and a dynamic location choice model for taxi drivers, estimated using Chicago trip data. Two counteracting forces arise from the interaction of pricing and spatial frictions. In sparse areas, despite higher TNP prices under surge, riders shift away from taxis because of longer taxi waiting times. In dense areas, higher TNP prices shift riders toward taxis, reducing search times. However, because TNPs often lower prices in dense areas, taxi demand can remain limited even though drivers gravitate toward them. Counterfactual simulations show that removing dynamic pricing reduces taxi demand by 51% and decreases the number of active drivers by 10–15% in central areas, while increasing taxi demand in peripheral regions. Dynamic pricing by TNPs draws taxi drivers toward dense regions, improving taxi market outcomes there through spatial reallocation.
By Jay Lee; University of Florida
Presented by: Jay Lee, University of Florida
Discussant:   Filomena Garcia, North Carolina State University
When Do Platform Mergers Benefit Users? The Role of Multi-Homing, Network Effects and Post-Merger Strategies Abstract This paper examines mergers in a setting where two differentiated platforms connect users and remain active after the merger. We highlight the critical interplay between network effects, multi-homing, and post-merger strategies. When users on both sides are single-homing, mergers reduce user surplus, and network effects do not alter this conclusion. By contrast, when one side is multi-homing, network effects become decisive: strong network effects can flip the welfare impact from negative to positive. We further examine post-merger strategies such as integration and bundling, showing that they can reduce—or even fully eliminate—the potential welfare gains from mergers.
By Filomena Garcia; North Carolina State University
Presented by: Filomena Garcia, North Carolina State University
Discussant:   Sarit Markovich, Northwestern University
Session 18: Incentivizing Digital Content Creation April 11, 2026 8:00 to 10:00 Location: Thompson (4th Floor)
Session Chair: Fei Li, University of North Carolina
Supporting Content Creators on Two-Sided Markets: Experimental Evidence from a Short-Form Video Platform Abstract A few top content creators capture most of the impressions on digital platforms, discouraging grassroots users from creating new content, and thereby threatening the platform ecology. This concentration presents a dilemma to the platform: whether a platform should capitalize on established content creators’ popular content in the short run or promote content creation from amateurs in the long run. To quantify this tradeoff, I study a two-sided experiment by a short-form video platform that exogenously increases the impressions of treated amateur-generated content to treated viewers. Although viewer usage time decreased in the short run, the program successfully fostered the production of higher-quality and more diverse content from amateur creators. Overall, incentivizing amateur content creation offered net benefits to platforms in three months.
By Tianli Xia; University of Rochester
Presented by: Tianli Xia, University of Rochester
Discussant:   Rubing Li, New York University
Algorithmic Attention and Content Creation on Social Media Platforms Abstract This paper develops a theory framework to examine how an ads-funded social media platform allocates attention through recommendation algorithms and how this in turn shapes content creation and consumption. Producers, on one side of the platform, are differentiated both in the topics they have talents and in their ability to produce high-quality content. Consumers, on the other side, are heterogeneous in topics they have interest and in appreciating the contents. The recommendation algorithm is designed to incentivize content production and ads consumption. The optimal algorithm filters out low-ability producers but guarantees a minimal readership for those who produce, while propagating viral content for high-ability ones. The algorithm deliberately assigns excessive consumer attention in some non-interested contents so as to leverage the two-sided network effect.
By Yi Chen; Cornell University    Fei Li; University of North Carolina    Marcel Preuss; SC Johnson Graduate School of Management
Presented by: Fei Li, University of North Carolina
Discussant:   Yang Yu, Massachusetts Institute of Technology
Designing Self-Sustaining Markets: An Application to Content Platforms Abstract When supply incentives are driven by demand, the natural force of the market may suppress the provision of products that are valuable to some consumers, leading to market failure. This paper explores a solution. Focusing on content platforms, we design a recommendation mechanism that maximizes total user utility by strategically allocating demand to sustain valuable content production. Theoretically, we prove that the optimal recommendation mechanism considers not only consumption utility, as in standard recommender systems, but also creator sensitivity, which captures how easily a creator can be incentivized by demand, and creator contribution, which reflects a creator’s importance to users overall. Computationally, we develop PixSet, a novel data transformation framework that converts variable-length sets (e.g., numbers of creators of various types) into fixed-size tensors (e.g., density distribution over pixels in a type space), which simplifies estimation as a computer vision problem with 99.9 gain in computational efficiency. The proposed recommendation mechanism is validated by both an observational-data analysis and a field experiment on Tencent WeChat, one of the world's largest content platforms. This mechanism has since been deployed as the default recommendation system at WeChat, serving billions of users and tens of millions of creators each day.
By Lei Huang; MIT    Juanjuan Zhang; MIT
Presented by: Lei Huang, MIT
Discussant:   Hassan Sayed, Center for Global Development
Session 19: Long-Term Care, Inspections, and Quality Measurement April 11, 2026 8:00 to 10:00 Location: Salon B (3rd Floor)
Session Chair: Ashley Swanson, University of Wisconsin-Madison
Agreeing to Disagree: Political Partisanship and Team Performance in Healthcare Abstract This paper studies how political partisanship affects team performance. We link doctors' political affiliations from voter registration records to their patients' inpatient claims. In the context of heart attack treatment, we find that patient 30-day mortality rate is 8% lower when the doctor performing the procedure ("proceduralist") and the doctor providing inpatient care ("physician") are from different political parties. We exploit within-proceduralist variation in team political composition for identification and find no evidence that patient-provider selection or alternative team characteristics confound our estimates. We further show that politically mixed teams achieve better performance while using fewer medical resources, indicating improved productivity. Finally, we find that politically mixed teams choose more appropriate treatments for patients, suggesting reduced confirmation bias as a possible mechanism. Our results shed light on the potential costs of increased political segregation in the workplace.
By Haizhen Lin; Indiana University    Yanhao Wang; University of Alberta    Jia Xiang; Kelley School of Business, Indiana Unive
Presented by: Yanhao Wang, University of Alberta
Discussant:   Tong Liu, Massachusetts Institute of Technology
Competition-for-the-field in medical device procurement Abstract We examine "competition-for-the-field" procurement mechanisms in the US medical device industry, specifically for Cardiac Rhythm Management devices. Motivated by evidence that hospitals frequently limit suppliers and award dominant volume shares (70-80%) in exchange for discounts, we model a two-tier auction framework. In this setting, manufacturers bid aggressive "Tier 1" prices to secure a committed dominant share, while simultaneously bidding "Tier 2" prices for the residual, physician-sensitive market segment. We define the supplier's objective function and simulate equilibrium bidding strategies. Preliminary results indicate that increasing the committed preferred supplier fraction intensifies competition, significantly reducing equilibrium prices and markups compared to a status quo of simultaneous differentiated Bertrand competition.
By Julie Mortimer; Washington University in St. Louis    Charles Murry; University of Michigan
Presented by: Charles Murry, University of Michigan
Discussant:   Gaurab Aryal, Boston University
Predictably Unpredictable Inspections Abstract Inspections are a common tool for acquiring information and incentivizing compliance. Although inspections are typically unannounced, their timing often follows a predictable schedule. We study how this predictability shapes firm effort and patient outcomes in U.S. nursing homes, leveraging detailed administrative data on staffing, care, and health outcomes. Nursing homes "slack" in the low-risk period following an inspection and ramp up effort as their next inspection approaches. Patient survival mirrors this pattern, suggesting that these fluctuations in effort have meaningful consequences for the quality of patient care. We embed these estimates in a dynamic model capturing how inspection regimes incentivize effort and provide information about quality. Our estimates indicate that moving to unpredictable inspections could induce as much additional effort as increasing the frequency of inspections by 12%, while only minimally reducing their informational value.
By Ashvin Gandhi; UCLA    Andrew Olenski; Lehigh University    Maggie Shi; University of Chicago
Presented by: Ashvin Gandhi, UCLA
Discussant:   Yiyi Zhou, Stony Brook University
The Measurement and Production of Quality in Long-Term Care Facilities in the US: 2012–2017 Abstract The quality of long-term care facilities is believed to be poorly measured and widely variable. In this paper we develop a new measure of quality of care that reflects how nursing homes impact the health trajectories of long-term care patients. We then develop and estimate a model in which nursing homes jointly produce quality and quantity, such that quality is driven by production factors, productivity, and a residual term that captures facility-level preferences over producing quality vs. quantity. Descriptively, we present new evidence on the extent of variation in quality, productivity, and preferences—some of which is systematically related to geography and ownership. We then consider a counterfactual reallocation of patients to alternative facilities. This exercise highlights that capacity constraints following from, e.g., certificate of need laws and construction moratoria, are a first-order concern for delivered quality of care. For example, facilities in the most capacity-constrained markets exhibit the weakest preferences for quality, while facilities in the least constrained markets exhibit the strongest reallocation of patients to high-quality facilities.
By Matthew Backus; Berkeley    Andrew Olenski; Lehigh University    Ashley Swanson; University of Wisconsin-Madison
Presented by: Ashley Swanson, University of Wisconsin-Madison
Discussant:   Jia Xiang, Indiana University
Session 20: Behavioral and Dynamic Bidding in Auctions April 11, 2026 8:00 to 10:00 Location: Mediterranean (3rd Floor)
Session Chair: Miguel Alcobendas, Yahoo Research
{Dynamic Bidding under Limited Commitment: Evidence from Judicial Sales Abstract Auctioneers often try to resell an object at a lower reserve price that failed in a previous auction. As a result, potential bidders may engage in dynamic bidding games; a bidder can either bid in the current period or wait for a future round with a reduced reserve price. We document such behaviors in a large dataset from judicial auctions. We develop and estimate an empirical model of dynamic bidding when the seller has limited commitment. Through the estimated model, we quantify the inefficiency of such auction design in actual judicial sales.
By Hui Liu; Tianjin University    Yao Luo; University of Toronto    Ruli Xiao; Indiana University
Presented by: Ruli Xiao, Indiana University
Discussant:   Yuki Ito, Indiana University
Auctions with Imperfectly Competitive Resale: Evidence from Central Bank Auctions Abstract As the central bank adjusts the supply of reserves, it unevenly affects bank profits and redistributes rents across institutions. When reserves are scarce, supply is fine-tuned to manage overnight interest rates through auctions, followed by bilateral trading in imperfectly competitive interbank markets. This environment allows banks to use private information to extract rents from the central bank and from counterparties in interbank trading, splitting gains from trade among competitors. To quantify these rents, we extend Kastl (2011) to study bidding in multi-unit auctions with post-auction bilateral trading, relaxing the independent private values assumption to allow for trading externalities. Applying the model to Canada, we show secondary markets were crucial for managing liquidity and document substantial heterogeneity in information rents. These results highlight efficiency and distributional costs of scarce reserves with implications for central bank balance sheet size.
By James Chapman; Bank of Canada    Eric Richert; University of Chicago    Yu Zhu; Renmin University of China
Presented by: Eric Richert, University of Chicago
Discussant:   Nathalie Gimenes, Pontifical Catholic University of Rio de Janeiro
Auditing Collusion in Public Procurement Quality: An Empirical Study Abstract By using a unique dataset of public procurement auctions and construction audit scores, we investigate detected tacit collusion between auditors (supervisors) and construction firms (agents), in which auditors and firms collude to cut corners in quality audits, resulting in relatively low-quality construction. In this collusion, the auditors (municipal officers) partially neglect to carry out quality checkups to save their time and effort. Knowing that is the case, construction firms make fewer efforts to improve the quality of the construction. Eventually whistle-blowing happens, and a subsequent scrutinization by the principal halts this tacit collusion, as well as revealing the poor quality of past procurement constructions. After the whistle-blowing, the firms improve the quality of procured projects by 7 to 10.2 percent. However, as an unintended consequence, such quality improvement comes with an increase in procurement auction prices. The municipal’s expenditure on construction increases by 1.4 percent after the end of the auditor-firm collusion, indicating that, when it comes to public construction procurements, the old adage ``you get what you pay for’’ can apply.
By Rieko Ishii; Shiga University    Hisayuki Yoshimoto; University of Glasgow
Presented by: Hisayuki Yoshimoto, University of Glasgow
Discussant:   Ji Hyun (Kiara) Kim, Washington University in St. Louis
Mass of Bids at Reserve Abstract We document partial pooling of bids at the reserve-price level in a large market for digital display advertising that employs a first-price auction mechanism to sell individual impressions. Such partial pooling is unexpected because standard theory predicts that equilibrium bidding strategies in first-price auctions should not involve a substantial mass of bids at the reserve. To explain the mechanism underlying the partial pooling, we analyze a unique feature of our data - a direct observation of the actual bidding strategy of one of the bidders. The bidder whose strategy we observe employs a heuristic bidding strategy that mechanically produces the mass of bids at reserve as follows: before bidding begins, the bidder pre-calculates a candidate bidding strategy based on only his valuation of the impression but without considering the reserve price. When an opportunity to buy an impression arises, the bidder censors the bid at exactly the reserve price whenever the valuation of the opportunity exceeds the reserve but the pre-calculated bidding falls below it. While we do not directly observe the bidding strategies of the remaining bidders in our data, we test for a data pattern consistent with the Heuristic model: when the reserve shifts downward for an exogenous reason, the tail of the bidding distribution above the higher of the two reserve levels should be the same. We find that a majority of the bidders do indeed exhibit this pattern. To understand the economic consequences of the Heuristic Bidding strategy, we use a counterfactual analysis to compare the publisher's revenue and the bidders' surplus under the heuristic bidding strategy and the more theoretically grounded Bayes-Nash equilibrium bidding strategy. We find that the counterfactual Bayes-Nash equilibrium model involves generally higher bids than the Heuristic model, increasing the average auction revenue by 9\% and decreasing bidders' surplus by 5\%- both economically meaningful amounts. Widespread use of simplified bidding heuristics in online advertising can thus lead to suboptimal outcomes for publishers.
By Miguel Alcobendas; Yahoo Research    Robert Zeithammer; UCLA Anderson School of Management
Presented by: Miguel Alcobendas, Yahoo Research
Discussant:   Shunto Kobayashi, Questrom School of Business, Boston University
Session 21: Merger Efficiencies (Sponsored by Charles River Associates) April 11, 2026 8:00 to 10:00 Location: Atlantic 3 (3rd Floor)
Session Chair: Chun-Yu Ho, University at Albany, SUNY
Merger efficiency and coordinated effects: nothing to sneeze at? Evidence from cough and cold medicines in the Philippines Abstract This paper examines the competitive effects of a merger in the cough and cold medicines segment between GlaxoSmithKline (GSK) and Pfizer. The merger was cleared by competition authorities based on expected cost efficiencies. Using product-level data from the Philippines, we estimate a structural demand model to assess both efficiency and pricing effects. We find evidence of merger- specific efficiency gains, but these effects are asymmetric: prices and marginal costs decline for Pfizer products, while GSK prices increase. Further, the price of a close competitor, Sanofi, also increased following the merger, likely due to increased coordination between the merging parties and the competitor.
By Farasat Bokhari; Loughborough University    Sean Ennis; University of East Anglia    Carlos Vega; Philippine Competition Commission    Weijie Yan; University of East Anglia
Presented by: Farasat Bokhari, Loughborough University
Discussant:   Chun-Yu Ho, University at Albany, SUNY
Digital (Killer?) Acquisitions Abstract This paper examines the effects of 1,200 acquisitions by major technology firms on innovation. Using detailed patent and workforce data linked to technology acquisitions and a suite of event-study and difference-in-differences designs, we document four main findings. First, although most acquired startups hold no patents, those with patents tend to operate in technology areas where the acquirer already has a presence and which subsequently experience further acquisition activity. Second, innovation typically rises before an acquisition but only persists afterward when further acquisitions occur, suggesting that acquisitions respond to innovation trends rather than initiating them. Third, at the patent level, acquired patents receive significantly more citations after the acquisition than comparable patents, and these effects are not driven solely by self-citations from the acquiring firm. These post-acquisition citation effects are smaller when more employees from the acquired firm are retained, consistent with innovation spillovers occurring through employee mobility. Fourth, we document significant work force attrition exceeding 50% on average at target firms three years post-acquisition. Our results suggest that acquisitions by digital incumbents often amplify, rather than suppress, the diffusion and visibility of acquired technologies.
By Florian Ederer; Boston University    Regina Seibel; University of Toronto    Timothy Simcoe; Boston University
Presented by: Regina Seibel, University of Toronto
Discussant:   Benjamin Leyden, Cornell University
Competition, Procurement and Learning-by-Doing in the Space Launch Industry Abstract We estimate a dynamic model of the U.S. space launch industry. The model allows past launches to improve rocket reliability and lower launch costs. It also allows the government to make forward-looking procurement choices. The estimated model implies that market outcomes can be inefficient, and that there can be large benefits to innovation and mergers that generate learning synergies. We use the model to analyze policy-relevant issues in the recent history of the industry: the 2006 United Launch Alliance ``merger-to-monopoly''; innovation in the form of SpaceX's Falcon 9 and ULA's recent introduction of Vulcan; the costs and benefits of forward-looking procurements; and, the trade-offs between the advantages of centralized control and possible inefficiencies.
By Ruibing Su; University of Maryland, College Park    Andrew Sweeting; University of Maryland    Chenyu Yang; University of Maryland, College Park
Presented by: Chenyu Yang, University of Maryland, College Park
Discussant:   Vivek Bhattacharya, Northwestern University
Mergers and Quality Provision in Healthcare: Evidence from Nursing Homes Abstract This paper tests whether mergers between nursing home chains and independent facilities affect quality of care using facility-level data from 1999-2019. Staggered difference-in-differences estimates suggest that acquired facilities experience a 8% reduction in health deficiency citations 3 years post-merger. This improvement is specific to mergers between chains and independent homes and persists for at least four years. Quality effects are driven by mergers involving smaller, higher-quality and non-private-equity-owned chains. A structural model suggests that the quality effect is generated by enhanced cost efficiency achieved by facilities serving larger numbers of residents after mergers. This finding is supported by the reduced inpatient cost and nurse hours with an adjusted mix of nurse inputs post-merger.
By Pinka Chatterji; The University at Albany    Chun-Yu Ho; University at Albany, SUNY    Wenqing Li; University at Albany, State University of New York
Presented by: Chun-Yu Ho, University at Albany, SUNY
Discussant:   Yangyi Deng, The Chinese University of Hong Kong
Session 22: Effects of Cartels April 11, 2026 8:00 to 10:00 Location: Brewster (4th Floor)
Session Chair: Mitsuru Igami, University of Toronto
Is Cartel Enforcement Effective? Evidence from the Stainless Steel Industry Abstract This paper examines whether cartel enforcement restores competition or if collusion persists after a cartel’s formal dissolution. We analyze the European stainless steel cartel, where collusion operated through adjustments to a formula-based pricing system. The formulaic nature of pricing in this industry makes it possible to clearly distinguish collusive and competitive conduct. Following the cartel’s detection, the European Commission imposed fines and directed firms to revise their formulas to eliminate the collusive changes. Using a novel, hand-collected dataset of prices during and after the cartel, we test whether collusion ended after the cartel’s detection, as reflected in whether firms revised their pricing formulas as directed by the Commission. Results suggest that producers largely continued pricing according to the collusive formula after detection and, within a few years, modified the formula to further increase prices. Moreover, the adjustments to the pricing formula implemented in Europe during and immediately after the cartel appear to have later served as focal points for tacit collusion in the United States, where no cartel existed.
By Daniel Garcia; University of Florida    Douglas Turner; University of Florida
Presented by: Douglas Turner, University of Florida
Discussant:   Daniel Chaves, University of Western Ontario
The Impact of Cartels on Productivity: A Concrete Example from Japan Abstract We study the impact of cartels on productivity using a novel plant-level dataset from the Japanese ready-mixed concrete industry, where cartels are legally permitted. After estimating plant-level productivity, we adopt a difference-in-differences design to show that cartel collapse increases plant-level and market-level productivity, while cartel formation has no effect. Furthermore, a triple-difference analysis reveals that productivity gains are more pronounced for initially less productive plants and those in high-density markets. These results, combined with decomposition analyses showing that market-level improvements are driven by within-plant changes rather than reallocation or exit, suggest that the treatment effect of competition drives productivity gains.
By Satoshi Imahie; Toulouse School of Economics    Masahiro Nishida; University of Wisconsin-Madison    Kazuma Takakura; University of Maryland    Yasutora Watanabe; University of Tokyo
Presented by: Masahiro Nishida, University of Wisconsin-Madison
Discussant:   Camilo Rubbini, Florida State University
Market-wide Collusion between Suppliers and Retailers Abstract We study “market collusion”, where retailers and suppliers maximize their joint profits in a repeated game. Market collusion is price-increasing when product substitutability is high and price-decreasing when product substitutability is low. It is always stable when product substitutability is intermediate. Price-decreasing market collusion is more stable than successive collusion (supplier collusion followed by retailer collusion), while price-increasing market collusion is less stable than successive collusion. Successive collusion is more socially harmful than market collusion and firms prefer market collusion to successive collusion. The results can support a rule of reason approach toward market collusion when product substitutability is low.
By David Gilo; Tel Aviv University    Yaron Yehezkel; Tel Aviv University
Presented by: Yaron Yehezkel, Tel Aviv University
Discussant:   Margaret Levenstein, University of Michigan
Information Sharing vs. Collusive Coordination: A Bayes Correlated Equilibrium Approach to Algorithmic Pricing in US Multifamily Rentals Abstract We investigate whether rent increases facilitated by RealPage’s algorithmic pricing software reflect benign information aggregation (“information effect”) or anticompetitive coordination (“collusion effect”). While recent antitrust actions equate data pooling with price-fixing, economic theory suggests that information sharing can alter prices in non-collusive Bayesian Nash equilibria. We bridge this gap using a structural model robust to the unobserved information structure. Adopting Bayes Correlated Equilibrium (BCE) and a “Mean-Preserving Contraction (MPC)” constraint, we identify the set of marginal costs rationalizable under any static information environment. Applied to US multifamily housing data with demand estimates from Calder-Wang and Kim (2024), our approach distinguishes between the economic “gray zone” of sharpened information and price levels that necessitate dynamic punishment strategies.
By Sophie Calder-Wang; University of Pennsylvania    Mitsuru Igami; University of Toronto    Gi Heung Kim; Boston College    Takuo Sugaya; Stanford University
Presented by: Mitsuru Igami, University of Toronto
Discussant:   Jean-Francois Houde, University of Wisconsin-Madison
Session 23: Screening, Investment, and Organizational Incentives April 11, 2026 10:15 to 12:15 Location: Thompson (4th Floor)
Session Chair: Jorge Lemus, University of Illinois Urbana-Champaign
Delegated Screening: Evidence from Government-Backed Loans Abstract Delegated screening can leverage the informational advantage of intermediaries subject to two principal-agent frictions: (i) under-provision of screening effort and (ii) selection distortion tilting toward intermediaries' private benefits. This paper examines the power of dynamic incentives in mitigating both frictions. In the context of a large government-backed credit program in Colombia, we find evidence that lenders conduct less screening and take on greater risk with government-backed loans than with their traditional products. However, this tendency is mitigated by a dynamic quota-and-fee mechanism employed by the program. Initial calibration suggests that, absent dynamic incentives, banks would exert even less screening effort, thereby issuing fewer but riskier government-backed loans.
By Felipe Brugues; Instituto Tecnologico Autonomo de Mexico    Rebecca De Simone; University of Michigan Ross School of Business    Thi Mai Anh Nguyen; New York University    Juan Velez-Velasquez; Banco de la República
Presented by: Thi Mai Anh Nguyen, New York University
Discussant:   Seth Smith, University of Georgia
Credit Without Proximity: Informational Frictions and Unequal Gains from Technology Abstract We study how the organization of screening activity---and its endogenous response to economic and technological forces---affects informational efficiency, credit allocation, and the distribution of borrower risk. Using US administrative data linking loan officers to mortgage applications and loan performance, we document that local officers achieve higher screening precision and faster processing; the informational benefits of proximity accrue disproportionately to borrowers with higher observable risk; and lenders allocate labor elastically with respect to local wages, resulting in systematic spatial misallocation of underwriting capacity relative to mortgage demand. We develop and estimate a structural model in which lenders set prices before observing borrower-specific signals, borrowers self-select on posted rates, and loan officers of different screening precision generate information that determines loan approvals. Lenders compete in mortgage pricing and in labor markets for local and remote officers, endogenously allocating information production across markets. We find substantial baseline credit rationing—up to 15 percent in high-risk segments—with local officers eliminating roughly half of it while also reducing excessively risky approvals. A technology shock that raises the processing productivity of remote officers induces lenders to substitute away from local screening, lowering informational efficiency, increasing excessively risky approvals and expected defaults, and tightening rationing for marginal borrowers despite only modest reductions in interest rates.
By Erica Xuewei Jiang; UCLA Anderson    Yeonjoon Lee; The Federal Reserve Bank of Richmond    Quinn Maingi; USC Marshall School of Business
Presented by: Quinn Maingi, USC Marshall School of Business
Discussant:   Ryan Westphal, Brandeis University
Allocating Positional Goods: A Mechanism Design Approach Abstract I study the optimal allocation of positional goods in the presence of externalities arising from consumers’ concerns about relative consumption. Applications include luxury goods, priority services, education, and organizational hierarchies. Using a mechanism-design approach, I characterize the set of feasible allocations through a majorization condition. The revenue-maximizing mechanism possibly excludes some buyers and fully separates participants under Myerson’s regularity condition. The seller can guarantee at least half the maximum revenue by offering a single good. Without exclusion, offering more levels of goods decreases (increases) consumer surplus under increasing (decreasing) failure rates. Higher participation raises consumer surplus under increasing failure rates.
By Peiran Xiao; University of Southern California
Presented by: Peiran Xiao, University of Southern California
Discussant:   Arthur Fishman, Bar Ilan University
Selling (Un)Finished Products Abstract We investigate an innovator's choice between in-house product development and external acquisition from a seller. In-house development requires the buyer to make a non-refundable investment that may not yield a viable product. Alternatively, the buyer can acquire the product from a seller who can offer a prototype or a finished product. The buyer can privately learns the value of a finished product, but is uncertain about this value at the prototype stage. The seller's tradeoff is one of information rents versus search deterrence. We characterize conditions under which the seller benefits from offering a finished product rather than a prototype.
By Jorge Lemus; University of Illinois Urbana-Champaign    Francisco Poggi; University of Mannheim
Presented by: Jorge Lemus, University of Illinois Urbana-Champaign
Discussant:   Qiang Fu, National University of Singapore
Session 24: Contract Design and Enforcement in Vertical Relationships April 11, 2026 10:15 to 12:15 Location: Mediterranean (3rd Floor)
Session Chair: Leila Safavi, Pomona College
Why Suppliers Might Prefer Linear and Secret Vertical Contracts? Abstract We consider suppliers who endogenously choose the type of vertical contract they use with retailers, with whom they have interlocking relationships, when secret linear contracts are one of their options. If contracts are constrained to be linear, suppliers non-cooperatively choose secret contracts instead of public contracts if supplier substitutability is high for any level of retailer substitutability. If supplier substitutability is low, suppliers choose secret contracts when retailer substitutability too is low. When suppliers can choose among four contract types: secret linear, public linear, secret two-part tariffs and public two-part tariffs, secret linear contracts emerge as the unique equilibrium if supplier and retailer substitutability are high while public two-part tariffs emerge as the unique equilibrium if these substitutability levels are low.
By David Gilo; Tel Aviv University    Yaron Yehezkel; Tel Aviv University
Presented by: David Gilo, Tel Aviv University
Discussant:   Claire Chambolle, INRAE & CREST
Hybrid Contracting in Repeated Interactions Abstract Many business relationships begin with informal interactions and later transition to formal contracts. Using a repeated-games model with a finite horizon, we show that this hybrid-contracting approach can both prolong cooperation (intensive margin) and enable it across a broader range of settings (extensive margin). We model the contract as a “smooth-landing contract” that limits actions only near the end of the relationship. We show that this flexible design supports early cooperation and outperforms rigid contracts. Our findings are robust to changes in contracting costs and timing, with optimal contract length balancing profitability and implementability.
By Bernhard Ganglmair; University of Mannheim & ZEW Mannheim    Julian Klix; University of Mannheim    Dongsoo Shin; Santa Clara University
Presented by: Dongsoo Shin, Santa Clara University
Discussant:   Yang Yang, School of Economics and management, Tsinghua University
Contracting Decisions under Joint Liability: Evidence from the Hazardous Waste Industry Abstract Do firms factor in expected liability costs into their contracting decisions? This paper studies how joint and several liability affects contracting patterns, market shares, and the allocation of risk across firms in a setting where firms face potential liability for their suppliers’ actions. Joint liability arises in many markets: franchisors may be held liable for franchisee’s labor violations, and distributors may be held liable for manufacturers’ product defects. Economic theory predicts that when liability is shared between contracting partners and suppliers are heterogeneous in risk, firms have incentives to contract with high-quality suppliers. However, there is little empirical evidence on whether firms respond to liability regimes through supplier choice. I study this question in the hazardous waste disposal industry, a $9.6 billion market with over nine million transactions between 2003 and 2017 with the potential for toxic spills. Waste generators contract with treatment, storage, and disposal firms and may be held jointly liable for health and environmental damages caused by their disposal suppliers under the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA, or “Superfund”). This setting provides an ideal laboratory for studying liability and partner choice: suppliers are observably heterogeneous in environmental compliance, transactions are frequent and well-documented, and liability exposure varies both across space and over time. I exploit a natural experiment created by the Supreme Court’s 2009 Burlington Northern & Santa Fe Railway Co. v. United States decision. Prior to the ruling, federal courts were split over the scope of “arranger liability” under CERCLA, generating geographic variation in waste generators’ expected liability exposure. Burlington Northern resolved this circuit split by sharply narrowing arranger liability in some circuits where courts had previously interpreted the statute broadly (treated circuits), while leaving doctrine unchanged in circuits that had already adopted narrow interpretations (control circuits). This ruling reduced expected joint liability for generators in treated circuits. The analysis uses comprehensive administrative data from the EPA’s RCRA Biennial Reports from 2003–2017. I define markets at the industry-state level, which captures the economically relevant choice set faced by waste generators, who typically select among a limited number of disposal options conditional on waste type, regulatory constraints, and geographic proximity. The outcome of interest is suppliers’ market shares: the share of hazardous waste contracts originating from a given industry-state market in which the firm operates. A central challenge is measuring suppliers’ environmental risk in a way that reflects buyers’ information at the time of contracting. I address this by constructing supplier-level risk measures using a Bayesian learning framework in which waste generators update beliefs about a supplier’s accident or noncompliance risk based on its publicly observed history of regulatory violations. Specifically, I merge contract data with data on enforcement actions and documented violations from EPA’s ECHO database. This approach yields a continuous measure of perceived environmental risk that captures reputational differences across suppliers and allows them to be classified “dirty”, “neutral” or “clean” based on tercile cutoffs of mean baseline risk. I estimate a triple-differences research design comparing changes in market shares of dirty and clean suppliers relative to neutral suppliers, before and after Burlington Northern, across circuits affected and unaffected by the ruling. This design isolates the effect of weakened joint liability from confounding trends in waste generation, disposal capacity, or regulatory enforcement. Event-study estimates show no differential pre-trends between treated and control markets and reveal effects that emerge four after the ruling, consistent with 18-24 month contract durations documented in government procurement data rather than immediate price responses. The results show that weakening joint liability leads to a substantial reallocation of contracts toward higher-risk suppliers. In the preferred specification, the relative market share of dirty suppliers increases by 28.7 percent in treated markets following Burlington Northern, with largest gains accrue to the dirtiest suppliers. These effects are economically meaningful and robust across alternative market definitions, risk classifications, and specifications that control for disposal capacity, enforcement intensity, and generator composition. These findings suggest that weakening joint liability increased environmental risk by reallocating waste toward higher-violation disposal facilities. This paper contributes to the economic literature on tort reform, which largely focused on firms’ own precautionary choices and examined how changes in liability regimes affect firm-level accidents, litigation, and environmental outcomes. I highlight a new and empirically important mechanism of adjustment: contracting decisions. By exposing firms to liability for their trading partners’ actions, joint liability reallocates demand to high-quality suppliers, potentially reducing environmental risk, even when firms’ technologies and precautionary decisions remain unchanged. This paper also contributes to the literature on reputation and market composition, which studies how quality signals affect the competitive environment under information asymmetry. I provide empirical evidence that firms respond to reputation incentives for quality in an environmental context, demonstrating that weakening liability reduces the returns to maintaining a “clean” reputation. Overall, the findings demonstrate that liability rules shape market outcomes not only firms’ incentives to take precaution, but also by altering choice of contracting partner and the allocation of demand across heterogeneous suppliers.
By Leila Safavi; Pomona College
Presented by: Leila Safavi, Pomona College
Discussant:   Jack Collison, University of Wisconsin Madison
Session 25: Consumer Behavior and Information April 11, 2026 10:15 to 12:15 Location: Georges (4th Floor)
Session Chair: Qi Pan, The Chinese University of Hong Kong (Shenzhen)
Value of "Availability" during Demand Shocks: Evidence from U.S. Retailers during the COVID-19 Pandemic Abstract This article studies the role of large retailers in ensuring product access during demand shocks. Focusing on U.S. consumers shopping for personal cleaning products during the COVID-19 pandemic, I quantify the consumer value of different store types at selected points in time. Because stockouts are not directly observed, I develop an algorithm to infer store-level stockouts from sales data. High-Availability Retailers (HARs)—such as discount stores and warehouse clubs—consistently maintained greater inventories and more varieties throughout the shock, with inventory depth constituting the primary margin of advantage. The uneven distribution of HAR stores across markets contributed to inequality in product access, with shoppers in "availability deserts" (markets without HAR stores) generally experiencing lower welfare.
By Tianyi Li; The Pennsylvania State University
Presented by: Tianyi Li, The Pennsylvania State University
Discussant:   James Dana, Northeastern University
Quality Disclosure, Product-Harm Crises, and Word of Mouth: The Impact of Multiple Sources of Information on Automobile Purchase Abstract When considering the purchase of high-value durable goods such as automobiles, consumers often invest significant time and effort in gathering information from a variety of sources. Despite the prevalence of multi-source information environments, the influence of these diverse information streams on consumer purchasing behavior remains insufficiently understood, largely due to the complexities of information integration. This research addresses this gap by leveraging an individual-level car purchase dataset in conjunction with three primary information channels: online reviews, product recalls, and results from destructive testing. Employing a state space approach within a consumer choice model framework, our findings reveal that the influence of any single information source may be either amplified or diminished by the presence of others in shaping purchase decisions. Furthermore, counterfactual simulation analyses suggest that while online reviews exert a more immediate, short-term effect on consumer choices, information from destructive testing demonstrates a more enduring, long-term impact on sales outcomes.
By Qi Pan; The Chinese University of Hong Kong (Shenzhen)    Zhehao Hu; University of North Carolina at Chapel Hill
Presented by: Zhehao Hu, University of North Carolina at Chapel Hill
Discussant:   WALEED HASSAN, KU Leuven
Flexible Estimation of Random Coefficient Logit Models of Differentiated Products Demand Abstract Demand estimation is a crucial step in structural modeling. In order to catch fundamentals when simulating counterfactuals and provide welfare-enhancing policy advice, the present paper combines the Berry, Levinsohn, and Pakes (1995) model and the logit mixed logit model of Train (2016) to estimate random taste heterogeneity using highly fexible parametric distributions. The model operates solely on aggregate data but oers the option of incorporating micro data, closely aligning higher moments of the mixture distribution with the data. The common assumption of normally distributed preferences is shown to be a source of bias, resulting in non-trivially biased estimates of elasticities, market power, and welfare changes resulting from mergers.
By Johannes Kandelhardt
Presented by: Johannes Kandelhardt,
Discussant:   Yibo Fang, Fudan University
Reference Effect in The Dynamic Promotion Strategy: The Case of Ride-Hailing Service Abstract This study incorporates reference-dependent preferences into the analysis of customer decisions in ride-hailing services. Using individual-level data, we model and identify the reference effect as the average historical benefit a customer has received. This reference point subsequently influences both the customer's initial consideration of the service and their final purchase decision. Our empirical findings reveal two distinct behavioral mechanisms: a higher historical benefit increases the probability of considering the service, which we term the reference as memory RaM effect. Conversely, during the final decision stage, customers compare the current offered discount against this historical reference point, a mechanism we label reference as a comparison (RaC). These two effects, RaM and RaC, can either offset or reinforce one another, a dynamic that provides a strategic lever for optimizing promotional campaigns. In a subsequent counterfactual analysis, we leverage the estimated elasticities of both price and the reference effect to design a novel, reference-effect-based promotion allocation strategy.
By Qi Pan; The Chinese University of Hong Kong (Shenzhen)    Xianjie Zeng; The Chinese University of Hong Kong (Shenzhen)
Presented by: Qi Pan, The Chinese University of Hong Kong (Shenzhen)
Discussant:   Daniela Schmitt,
Session 26: Market structure and environmental policy April 11, 2026 10:15 to 12:15 Location: Brewster (4th Floor)
Session Chair: Conor Ryan, Penn State University
R&D Among Small Firms: Evidence From Genetically Engineered Crop Diversity after Regulatory Changes Abstract Small firms engage in not only different levels of research and development than big firms but different kinds. Even when these firms are less efficient and there are increasing returns to R&D, small firms may be more likely to create products for new markets. In this paper we study R&D in genetically engineered seeds. For over thirty years, most GE seeds in the United States were corn, cotton, or soybean, three of the largest crops grown. Most of the products were put out by four or five multinational agrichemical companies. In 2021, the United States Department of Agriculture streamlined its approval for GE crops. We describe how the set of deregulated articles changed under the new rule. We show three novel facts, most notably the growth in product diversity and the diversity of the researching institutions. We connect these facts to a model of research and development with diminishing returns. The model suggests that the regulatory change may have led to an increase in product variety at the cost of a crowd-out effect for more efficient producers.
By Samuel Bailey; Independent    Conor Ryan; Penn State University
Presented by: Conor Ryan, Penn State University
Discussant:   Marcos Barrozo, DePaul University
Anticipated Entry, Labor Market Competition, and Toxic Releases Abstract Do firms strategically invest in environmental quality to compete for workers? Using a quasi-experimental design that compares counties winning new manufacturing facilities with narrowly unsuccessful counterparts, we find that incumbent plants in winning counties reduce on-site toxic releases by approximately 9.3% shortly after the announcement of large plant openings, with no impact on off-site toxic releases. This effect is stronger for plants that share a similar labor pool with the new plants, in markets with tight labor competition, in industries with a higher proportion of skilled workers, in locations with an environmentally conscious population, and when the new plant is expected to be greener than the incumbent plant. Incumbent plants that reduce pollution experience smaller wage increases for skilled workers without a loss of workers after new plant openings. These results cannot be explained by alternative mechanisms such as regulatory pressure, knowledge spillovers, supply-chain relationships, investor pressure, and product-market competition. Our results are consistent with a simple game-theoretic model of preemptive environmental investment under expected heightened labor-market competition, in which firms face a strategic trade-off between emission reductions and wage increases. Our findings demonstrate that labor market competition creates market-based incentives for environmental improvement.
By Sangeun Ha; Copenhagen Business School    Manpreet Singh; Georgia Institute of Technology
Presented by: Sangeun Ha, Copenhagen Business School
Discussant:   Cici McNamara, Georgia Institute of Technology
Smog and Safety Checks: A Case for Monopoly Delegation Abstract Vehicle inspections—commonly known as smog and safety checks—are often delegated to private agents, much like other quality certification markets. When these agents compete, they face incentives to misreport quality. Theory and evidence from Chile’s concentrated vehicle-inspection markets show that these incentives respond to a coordination problem so severe that it can only be resolved by delegating each market to a single agent. Doing so compromises neither service quality nor the ex-ante competition for the market, while resulting in a permanent reduction in vehicle emissions of 35% relative to a scenario without inspections. An alternative to monopoly delegation is to allow multiple agents in the market while restricting consumer choice—assigning consumers to specific stations but permitting them to switch either by trading “location allowances” or by paying a “switching tax.”
By Nano Barahona; University of California, Berkeley    Juan-Pablo Montero; Pontificia Universidad Catolica de Chile    Juan-Pablo Montero; Pontificia Universidad Catolica de Chile    Pedro Skorin; MIT
Presented by: Juan-Pablo Montero, Pontificia Universidad Catolica de Chile
Discussant:   Andre Trindade, Nova SBE
Where is the Beef? Supply Chains and Carbon Emissions in the Amazon Abstract This paper studies environmental policy in settings where small and informal firms are prevalent, showing how policy shapes entry decisions and the resulting productivity distribution. I focus on the Brazilian Amazon’s cattle industry, the world’s leading driver of deforestation. Using rich data, I show that a large number of processing firms operate informally or only sell domestically, with notable spatial clustering by firm type. Using a quantitative model, I show that policies targeting only prominent (formal and exporting) firms can have adverse effects, reallocating production toward less efficient firms and high-emitting upstream suppliers. Ignoring this margin of firm heterogeneity overestimates emissions reductions and underestimates output losses more than two-fold. Alternative policies that better account for small and informal firms greatly reduce abatement costs.
By Marcos Barrozo; DePaul University
Presented by: Marcos Barrozo, DePaul University
Discussant:   Arifah Hasanbasri, University of Pittsburgh
Session 27: Technology Adoption, Diffusion, and Industry Outcomes April 11, 2026 10:15 to 12:15 Location: Caspian (3rd Floor)
Session Chair: John Turner, University of Georgia
AI Adoption, Market Outcomes, and Coordination Risks Abstract In this paper, we examine the impacts of artificial intelligence (AI) adoption in the U.S. airline industry, focusing on pricing, service quality, environmental externalities, and consumer welfare. We find that enterprise AI adoption improves on-time performance and local air quality while raising average fares and compressing price dispersion. For generative AI-enhanced pricing tools, the price effects of adoption are larger and span the fare distribution, but their magnitude and direction depend on how the technology is adopted. In particular, when rivals rely on the same algorithm vendor, price increases are muted or even reversed, whereas broader diffusion of AI pricing tools is associated with higher fares and reduced dispersion. Our findings point to a key trade-off in the adoption of AI technologies: while AI can enhance operational efficiency, widespread adoption in concentrated markets may raise coordination risks.
By Qi Ge; Vassar College    Minhae Kim; Oklahoma State University    Myongjin Kim; U Oklahoma
Presented by: Minhae Kim, Oklahoma State University
Discussant:   Dmitri Kirpichev, OECD
Do Patents Predict Innovation? Abstract There is a large academic literature that addresses whether patents are good measures of the innovations they cover. A related question that the literature has not addressed is whether existing patents are reliable predictors of future innovation in technological areas covered by the patents. The answer to this question is relevant to studies that might attempt to predict firm performance or the consequences of mergers or other conduct on innovation. We report data on patents that cover synthetic pesticides and compare the patent data to new pesticides that have been developed after the patents were filed. These patents are highly imperfect predictors of subsequent innovations. We explain the reasons for this result and why they apply to other industries to make patents unreliable predictors of innovation.
By Richard Gilbert; University of California-Berkeley    Sofia Villas-Boas; University of California, Berkeley
Presented by: Sofia Villas-Boas, University of California, Berkeley
Discussant:   Konan Hara, Michigan State University
Incomplete Patent Pools Abstract Technology standards often involve thousands of complementary standard-essential patents. Patent pools coordinate licensing yet they are often incomplete: patent owners may split across multiple pools or not join any pool. We build a model of technology adoption and licensing with endogenous pool formation. We show that incompleteness can be profit-maximizing when implementers wield buyer power. Two mechanisms drive this: free riding (licensing late, post-adoption) and fragmentation (creation of multiple pools). Both raise aggregate royalties by offsetting implementers' ability to bargain prices down. Network effects weaken but do not eliminate these effects. Courts' reliance on comparable licenses can be abused through hard-to-detect side transfers but, used carefully, can encourage larger pools.
By Erik Hovenkamp; Cornell University    Jorge Lemus; University of Illinois Urbana-Champaign    John Turner; University of Georgia
Presented by: John Turner, University of Georgia
Discussant:   Jay Pil Choi, Michigan State University
Stifling or Scaling? Acquisitions and their effect on start-up innovation Abstract This paper highlights a critical issue in the innovation ecosystem – whether acquisitions of start-ups by incumbent firms foster or stifle innovation. Using a unique firm-level dataset covering 60 countries for the period 2001–2021, we find that acquisition targets are highly innovative before the deal, and acquirors are generally innovation-oriented firms. Acquired start-ups are also technologically close to their acquirors and often share the same industries and countries, suggesting potential market synergies. However, post-acquisition, start-up patenting declines without a corresponding increase in the acquiror's innovation activity, raising concerns that such deals may serve anti-competitive purposes rather than enhancing innovation.
By Marius Berger; OECD    Sara Calligaris; OECD    Andrea Greppi; OECD    Dmitri Kirpichev; OECD
Presented by: Dmitri Kirpichev, OECD
Discussant:   Sofia Villas-Boas, University of California, Berkeley
Session 28: Innovation and adoption April 11, 2026 10:15 to 12:15 Location: Aegean (3rd Floor)
Session Chair: Chung-Ying Lee, National Taiwan University
Early adoption of generative AI: Users, uses, and behavioral change Abstract The launch of ChatGPT in November 2022 marked the beginning of widespread consumer adoption of generative artificial intelligence (GenAI). Using high-frequency app and browser usage data from a US consumer panel, we provide comprehensive evidence on early adoption of GenAI services: the extent and intensity of adoption, the characteristics of adopters, and the effects on broader digital behavior. We document that ChatGPT dominated the early market and that mobile app releases substantially increased usage intensity. Adopters are disproportionately male, younger, and oriented toward productivity and information-seeking in their pre-adoption digital behavior. Combining transition analysis, which captures task-level complementarity by measuring which services users access in sequence with GenAI, with an event-study design, we identify three distinct interaction patterns. First, productivity services exhibit high task-level complementarity and increased post-adoption usage. Second, leisure services show low task-level complementarity and declining usage. Third, education presents a hybrid pattern: high task-level complementarity but declining overall usage, a pattern that admits multiple interpretations including efficiency gains, substitution, and selection effects. Together, these findings reveal how GenAI is being embedded into consumers' digital usage flows during the first period of its widespread availability.
By Chiara Farronato; Harvard University    Dominik Rehse    Sebastian Valet; ZEW Mannheim
Presented by: Sebastian Valet, ZEW Mannheim
Discussant:   Emil Palikot, Northeastern University
Push and Pull Funding for Social Innovation Abstract If a mechanism designer would like to encourage firms to develop a socially beneficial innovation, is it cheaper to reward them ex-ante with push funding — such as research grants paid unconditionally — or ex-post with pull funding — such as prizes only paid upon successful innovation? We study this question in a multifirm setting where firms' probabilities of developing an innovation and costs of research and development are private information. We show that the most cost-efficient contract always involves a positive amount of pull funding, and derive a simple distributional condition under which a designer should also utilize a positive amount of push funding. While push funding screens firms based on their costs of research and development, because pull funding is paid out only upon the successful development of an innovation, it screens firms on both their costs and likelihoods of success. On the one hand, this selects in firms who are more likely to successfully develop the innovation. On the other, it allows firms to collect rents on two dimensions instead of one, generating a tradeoff. Additionally, we provide a simple condition characterizing when a pure pull contract is cheaper than a pure push contract, show that the form of pull funding that entails the lowest fundraising cost for the designer is a shared pot of money, and discuss how our results can be microfounded through a social planner's problem.
By Hassan Sayed; Center for Global Development    Christopher Snyder; Dartmouth College
Presented by: Hassan Sayed, Center for Global Development
Discussant:   Fei Li, University of North Carolina
Battery Swapping Networks and Electric Vehicle Adoption: Evidence from Taiwan’s E-Scooter Market Abstract This paper empirically analyzes the bidirectional feedback loop between electric two-wheeler (E-scooter) adoption and battery swapping infrastructure using granular panel data from Taiwan spanning 2019 to 2023. To overcome simultaneity bias, we employ a dual instrumental variable approach. We find that a 10\% increase in battery station availability causally drives a 4% increase in E-scooter sales, while a 10% expansion in the E-scooter installed base induces a 3% increase in station deployment. Demand responsiveness is strongest among prime-age commuters between the ages of 25 and 44 purchasing commute-oriented models. Crucially, infrastructure expansion actively displaces heavily polluting gas-powered scooters, particularly in higher-income and more educated districts. These findings demonstrate that rapid battery swapping technology effectively overcomes urban mobility bottlenecks, highlighting the necessity of targeted infrastructure investments alongside vehicle subsidies for decarbonizing high-density transit.
By Chung-Ying Lee; National Taiwan University    Sih-Yu Wei; University of Chicago
Presented by: Chung-Ying Lee, National Taiwan University
Discussant:   Leshui He, Bates College
On Benchmark Hacking in ML Contests: Modeling, Insights and Design Abstract Benchmark hacking refers to tuning a machine learning model to score highly on certain evaluation criteria without improving true generalization or faithfully solving the intended problem. We study this phenomenon in a generic machine learning contest, where each contestant chooses two types of effort: exploratory effort that improves model capability as desired by the contest host, and exploitative effort that only improves the model's fitness to the particular task in contest without contributing to true generalization. We establish the existence of a symmetric monotone pure strategy equilibrium in this competition game. It also provides a natural definition of benchmark hacking in this strategic context by comparing a player's equilibrium effort allocation to that of a single-agent baseline scenario. Under our definition, contestants with types below certain threshold (low types) always engage in benchmark hacking, whereas those above the threshold do not. Furthermore, we show that more skewed reward structures (favoring top-ranked contestants) can elicit more desirable contest outcomes. We also provide empirical evidence to support our theoretical predictions.
By Xiaoyun Qiu; Northwestern University    Haifeng Xu; University of Chicago    Yang Yu; Massachusetts Institute of Technology
Presented by: Yang Yu, Massachusetts Institute of Technology
Discussant:   Lei Huang, MIT
Session 29: Insurance Design, Intermediation, and Selection April 11, 2026 10:15 to 12:15 Location: Atlantic 3 (3rd Floor)
Session Chair: Yiyi Zhou, Stony Brook University
Breaking Bad Opioid Sorting Abstract In many markets, consumers rely on experts for access to goods and services. Consumer and expert preferences over outcomes are linked through consumers' choice of experts, or sorting. We investigate how consumer sorting contributes to observed differences in outcomes and its implications for expert-targeted policies. Using rich employer-sponsored health insurance claims data on opioid prescriptions for chronic pain, we first show that prescription intensity is highly dispersed across physicians. We decompose the variance of physicians' prescription decisions and find that patient sorting is more than three times as important as physicians' inherent prescription propensity for prescription intensity dispersion. Most of this sorting cannot be justified on medical grounds. We develop and estimate an equilibrium model of patient choice of physicians and physicians' prescription decisions. Patients optimally choose their physicians based on both their opioid preferences and their expectations of physicians' prescription decisions. Our counterfactual analysis shows that expert-targeted policies to curb non-medically grounded opioid prescribing will be severely attenuated by patient resorting. We propose an alternative policy that eliminates sorting based on non-medically grounded preferences, while preserving appropriate care for patients with medical needs for prescription opioids.
By Han Ng; Academia Sinica    Stephan Sagl; Indiana University    Jia Xiang; Kelley School of Business, Indiana Unive
Presented by: Stephan Sagl, Indiana University
Discussant:   Anran Li, University of Minnesota, Twin Cities
Going for Broker? Intermediation in Health Insurance Markets Abstract This paper studies how insurance brokers affect product choices, premiums, and welfare in the employer-sponsored insurance market. We compile a novel database of contracting relationships among employers, brokers, and insurers in New York State. Exploiting variations in commission schedules, we document two market distortions: First, brokers exhibit traditional agency frictions, steering employers towards more financially lucrative products. Second, commission levels affect ex-ante insurer-broker networks and, in turn, insurers' competitive pressure, leading to anti-competitive distortions. We develop and estimate a structural model of employer insurance demand, insurer pricing, and formation of broker-insurer contracting networks. We use the model to quantify the steering-competition tradeoff: higher commissions exacerbate steering but may also broaden broker networks, increase insurer competition, and lower premiums. We also explore optimal commission and network regulations for insurance brokers.
By Anran Li; University of Minnesota, Twin Cities    Tong Liu; Massachusetts Institute of Technology    Nicholas Tilipman; John Hopkins
Presented by: Anran Li, University of Minnesota, Twin Cities
Discussant:   Charles Murry, University of Michigan
The Impact of Limited Consideration on Market Outcomes in the Hospital Industry Abstract This paper introduces a random consideration sets model of hospital demand to analyze market interactions while relaxing the assumption that patients necessarily consider all options available from within their contracted networks of providers. A key contribution of the model is its ability to quantify the impact of bounded rationality on market outcomes when firms compete for consideration via advertising. The demand model identifies and estimates the probability that each patient considers and chooses each in-network hospital within a geographic radius of their home. On the supply side, hospitals compete in a two-stage game: (i) with each other for patients’ consideration through advertising expenditure, and (ii) with insurers through Nash-in-Nash bargaining over prices. Assuming full consideration in hospital demand biases distance elasticities downward in magnitude, as it fails to account for attention decreasing with distance. Counterfactual simulations show that limited consideration reduces healthcare access and artificially differentiates products, leading to higher average prices and greater price dispersion across hospitals. A retrospective merger is used as a natural experiment, providing evidence that accounting for bounded rationality in models of imperfect competition can improve the accuracy of counterfactual predictions.
By Jack Berger; University of Toronto, Rotman School of Management
Presented by: Jack Berger, University of Toronto, Rotman School of Management
Discussant:   Ashley Swanson, University of Wisconsin-Madison
Regulation of Contract Generosity in Selection Markets: The Case of Medicare Advantage Abstract In selection markets where serving different consumers incurs different costs, private firms may engage in cream-skimming by designing their contracts to attract profitable consumers and avoid unprofitable consumers, potentially harming consumer welfare. Policymakers have often tackled this issue by imposing direct restrictions on the level of generosity firms can offer. However, the impact of such regulations on welfare remains unclear, contingent on the market’s level of selection and other policies in place. Our study of the Medicare Advantage market reveals that consumers exhibit diverse healthcare needs and considerable moral hazard. The recent regulation on maximum out-of-pocket expenses, mandated by the Affordable Care Act, improves consumer welfare by 19% on average, especially benefiting those with high healthcare needs, while compressing private insurers’ markups. Government spending on the Medicare program decreases as the reduction in spending on Traditional Medicare outweighs the increased subsidies to private insurers.
By James Gluzman; Stony Brook University    Yiyi Zhou; Stony Brook University
Presented by: Yiyi Zhou, Stony Brook University
Discussant:   Shizhe Yu, University of Wisconsin-Madison
Session 30: Market Power, Information, and Policy in Auctions April 11, 2026 10:15 to 12:15 Location: Salon C (3rd Floor)
Session Chair: Brad Larsen, Washington University in St. Louis
Quantifying Bargaining Power Under Incomplete Information: A Supply-Side Analysis of the Used-Car Industry Abstract This study quantifies bargaining power in supply-side negotiations with incomplete information, where car dealers negotiate inventory prices with large wholesalers after an auction. We measure an agent's bargaining power by where the agent's expected surplus lies relative to a benchmark mechanism favoring the agent and one favoring the opponent. We consider second-best benchmarks, which account for information constraints, and first-best benchmarks, which do not, as well as benchmarks that account for the effect of competition on bargaining power. We propose a direct-mechanism method for estimating a seller's private value as the gradient of a menu from which she chooses a secret reserve price. Bargaining power weights offer insights about inefficiency, as bargaining is not a zero-sum game when agents have incomplete information. On average, dealers (buyers) have less bargaining power than sellers relative to a benchmark where dealers face no competition. Accounting for the direct effect of competition, dealers have more bargaining power than sellers, achieving close to the highest possible surplus given competition. This holds true when the seller is a manufacturer, a finding that is consistent with manufacturers' recent movement toward direct-to-consumer used-car sales.
By Brad Larsen; Washington University in St. Louis
Presented by: Brad Larsen, Washington University in St. Louis
Discussant:   Christopher Snyder, Dartmouth College
When help hurts: unintended consequences of set-aside auctions for small firms in Brazil. Abstract This paper studies the efficiency implications of Brazil’s set-aside procurement policy for divisible goods, under which purchases are split between an exclusive auction for small firms and a non-exclusive auction open to all bidders. Using data from Compras.gov.br, we exploit a distinctive institutional feature - random auction ending times - to identify cost synergies that arise when firms participate in paired auction sessions for the same good. We first present reduced-form evidence, based on a regression discontinuity design, showing that firms that win one auction become significantly more aggressive in the concurrent session, consistent with positive cost complementarities. We then develop a structural auction model that incorporates random ending times as bidding frictions, allowing us to recover firms’ cost distributions and estimate average synergy. Our estimates indicate that winning both auctions reduces production costs by approximately 49.6%. These findings suggest that the current set-aside framework may unintentionally hinder efficiency by preventing firms - particularly small ones - from exploiting economies of scale. The results highlight the potential efficiency gains from alternative procurement designs, such as bundling, that better account for production synergies.
By Nathalie Gimenes; Pontifical Catholic University of Rio de Janeiro    Lucas Lima; Pontifical Catholic University of Rio de janeiro    Marcelo Sant'Anna; FGV
Presented by: Nathalie Gimenes, Pontifical Catholic University of Rio de Janeiro
Discussant:   Eric Richert, University of Chicago
The Impact of Privacy Protection on Online Advertising Markets Abstract Online privacy protection has gained momentum in recent years and spurred both government regulations and private-sector initiatives. A centerpiece of this movement is the removal of third-party cookies, which are widely employed to track online user behavior and implement targeted ads, from web browsers. Using banner ad auction data from Yahoo, we study the effect of a third-party cookie ban on the online advertising market. We first document stylized facts about the value of third-party cookies to advertisers. Adopting a structural approach to recover advertisers' valuations from their bids in these auctions, we simulate a few counterfactual scenarios to quantify the impact of Google's plan to phase out third-party cookies from Chrome, its market-leading browser. Our counterfactual analysis suggests that an outright ban would reduce publisher revenue by 54% and advertiser surplus by 40%. The introduction of alternative tracking technologies under Google's Privacy Sandbox initiative would recoup part of the loss. In either case, we find that big tech firms can leverage their informational advantage over their competitors and gain a larger surplus from the ban.
By Miguel Alcobendas; Yahoo Research    Shunto Kobayashi; Questrom School of Business, Boston University    Ke Shi; Shanghai Advanced Institute of Finance (SAIF) at SJTU    Matthew Shum; California Institute of Technology
Presented by: Shunto Kobayashi, Questrom School of Business, Boston University
Discussant:   Hisayuki Yoshimoto, University of Glasgow
Round Bidding in Auctions Abstract Our analysis of novel data from hundreds of thousands of online auctions on a large platform operating in the Netherlands uncovers a tendency to bid round numbers. Round winning bids are higher than average for a given item and are eschewed by bidders as they gain experience. These findings lead us to hypothesize that, rather than delivering a strategic benefit (say, adding salience to a jump bid), round bidding is a symptom of a behavioral bias. We construct a structural model of behavioral bidding in auctions, estimated for each of a subsample of the most frequently auctioned items. Our median estimate is that 21% of bidders are prone to round-number bias, reducing their expected surplus by nearly 10%.
By Mark van Oldeniel; University of Groningen    Christopher Snyder; Dartmouth College    Adriaan Soetevent; University of Groningen
Presented by: Christopher Snyder, Dartmouth College
Discussant:   Brad Larsen, Washington University in St. Louis
Session 31: Financial Markets April 11, 2026 10:15 to 12:15 Location: Salon B (3rd Floor)
Session Chair: Carlos Canon, Bank of England
The Welfare Effect of Pricing Climate Risk in the US Mortgage Market Abstract I document that Government-Sponsored Entities (GSEs), as the primary public securitizers in the U.S. mortgage market, do not price sea-level rise (SLR) risk. Uniform GSE pricing influences lenders’ loan approval, interest rate setting, and securitization decisions across regions; more importantly, it generates a cross-subsidy from inland to coastal areas. Using a structural model of mortgage lending, I quantify this cross-subsidy and show that a small group of high-risk coastal borrowers receives sizable benefits. The cross-subsidy is regressive across the income distribution and becomes larger in more competitive markets. In counterfactuals where GSEs adopt risk-differentiated SLR pricing, nationwide welfare remains roughly unchanged, but the distribution shifts: inland regions experience modest gains in welfare and mortgage supply, low-risk coastal areas see modest declines, and high-risk coastal regions experience large reductions in welfare and mortgage supply along with substantially higher balance-sheet retention rates. These effects can be several times larger under higher levels of SLR risk. The results provide evidence on a core policy trade-off between distributional fairness across regions and incomes and the GSEs’ mandate to ensure broad access to affordable housing.
By Hengyi Huang; Tilburg University
Presented by: Hengyi Huang, Tilburg University
Discussant:   Audrey Tiew, New York University
Regulatory Hurdles and Costly Delay in Housing Development Abstract This paper quantifies the supply effects of discretionary permit review, a common but understudied regulatory friction that delays and adds uncertainty to housing development. Using the universe of permit applications in Seattle, Washington, during a period when a unit count threshold for undergoing discretionary permit review was in place, I show that undergoing discretionary review increased permit review time by 4 to 5 months, and that developers reduced unit counts and increased average unit size in order to avoid discretionary review. I then propose a novel extension of prevailing housing production models to multifamily developments that accounts for both the intensive and extensive margins of development. Using data that is freely and publicly available in most jurisdictions, I estimate the model and find that removing discretionary permit review would have increased the number of new units constructed by 5.5% while reducing the average size of new developments by nearly 2%. The estimated change comes largely from the intensive margin, as, in expectation, only two additional parcels would be redeveloped absent discretionary review.
By Daniel Gold; University of Wisconsin-Madison
Presented by: Daniel Gold, University of Wisconsin-Madison
Discussant:   Emma Du Puy, Columbia University
When Prices Begin to Reflect Risk: Evidence From A Reform in Hazard Insurance for U.S. Mortgages Abstract Changes in the price of risk can influence credit and investment even when the underlying risk is unchanged. We study this mechanism using a nationwide shift to more granular flood-insurance pricing that raised premiums sharply in some locations and lowered them elsewhere. Linking detailed insurance records to loan-level mortgage data, we compare census tracts experiencing larger premium increases with nearby tracts that did not. Higher-premium areas saw declines in new mortgages, smaller loan sizes, and shifts toward borrowers with weaker credit, alongside reductions in new construction. The results show how re-pricing long-mispriced risks affects household budgets and lender behavior, with broader implications for credit supply and local investment.
By Anil Jain; Federal Reserve Board
Presented by: Anil Jain, Federal Reserve Board
Discussant:   Bernhard Ganglmair, University of Mannheim & ZEW Mannheim
Repo dealer-driven bond mispricing Abstract This paper uses proprietary data sets from the UK bond and repo markets to analyse the effect of funding market frictions on bond prices and market-wide liquidity. Starting with the structure of the repo market, we demonstrate how individual dealer market power and dealer linkages generate frictions. Specifically, we demonstrate that frictions related to market power account for between 0.5 and 1.3 percentage points of bond price deviations, whereas the transmission of heterogeneously persistent shocks between dealers accounts for between 2 and 4 percentage points of price deviations
By Carlos Canon; Bank of England    Jozef Barunik; Institute of Economic Studies, Charles U    Eddie Gerba; Bank of England
Presented by: Carlos Canon, Bank of England
Discussant:   Anastasios Dosis, ESSEC Business School
Session 32: Merger Evaluation (Sponsored by Cornerstone Research) April 11, 2026 10:15 to 12:15 Location: Atlantic 2 (3rd Floor)
Session Chair: Joe Mazur, Purdue University
Mobile Game App Concentration and Cross-Border Merger Oversight Abstract Consumers often face numerous choices in differentiated product markets for digital goods. We investigate price effects of a merger in such a market using a spatial competition model and identify conditions under which a merger increases product prices. We then estimate the relationship between market concentration and price in the mobile gaming app market across 20 European countries, using variations in concentration following a lawsuit brought by Epic Games against Apple in 2020. We also demonstrate hypothetical merger effects across countries, based on post-merger concentration ratios and predicted price changes.
By Jin-Hyuk Kim; University of Colorado at Boulder    Liad Wagman; Rensselaer Polytechnic Institute
Presented by: Jin-Hyuk Kim, University of Colorado at Boulder
Discussant:   Regina Seibel, University of Toronto
Bridging Quasi-Experimental and Structural Approaches for Robust Evaluation of US Airline Mergers Abstract We bridge quasi-experimental and structural approaches for robust merger evaluation. First, we show that the difference-in-differences (DiD) equation is the “reduced form” of a structural model, where demand and cost parameters identify price effects of mergers even when the DiD approach faces identification challenges. Second, we propose a synthetic GMM approach by applying synthetic DiD weights to structural moment conditions to improve estimates when only a few treated markets are available. Applying this methodology to three airline mergers, we find modest efficiency gains entirely offset by increased coordination. The synthetic GMM refinement sharpens findings, uncovering anti-competitive effects standard approaches miss.
By Gaurab Aryal; Boston University    Anirban Chattopadhyaya; University of Virginia    Federico Ciliberto; U Virginia
Presented by: Federico Ciliberto, U Virginia
Discussant:   Devesh Raval, Federal Trade Commission
Concentration Indices, Welfare Distortions, and Misallocation in Oligopoly Abstract We study welfare distortions in a multiproduct-firm pricing game with constant elasticity of substitution (CES) or multinomial logit (MNL) demand. Using approximations both around small market shares and around monopolistic competition conduct, we identify sufficient statistics to gauge the extent of inefficiencies caused by oligopolistic market power. We find that, at a low order, the oligopoly distortions are proportional to the Herfindahl index of industry concentration. At a higher order, distortions also depend on the cubic Hannah-Kay concentration index. Additionally, we show that the welfare loss from resource misallocation is approximately proportional to the difference between the cubic Hannah-Kay index and the square of the Herfindahl index.
By Volker Nocke; University of Mannheim    Nicolas Schutz; University of Mannheim
Presented by: Nicolas Schutz, University of Mannheim
Discussant:   Michele Bisceglia, Yale University
Merger Effects on New Markets: Evidence from U.S. Airlines Abstract Can mergers affect markets where neither merging firm currently operates? Using data from four major U.S. airline mergers, we construct a continuous measure of merger-induced entry threat. Utilizing this measure, we design a two-stage difference-in-differences (DID) model and develop a doubly debiased machine learning (DML) estimator for the average partial effect (APE), where the treatment variable related to entry threat is flexibly estimated in the first stage. We present empirical evidence that mergers significantly affect peripheral markets through two opposing channels: by strengthening the merged firm's entry threat and by eliminating a potential entrant. The estimation reveals that entry threat dynamics generate heterogeneous effects across markets, resulting in an average pro-competitive price reduction of 2-3% in our setting. These findings offer new insights for antitrust policy and the empirical design of merger remedies.
By Myongjin Kim; U Oklahoma    Joe Mazur; Purdue University    Yongjoon Park; University of Massachusetts, Amherst    Jangsu Yoon; University of Kentucky
Presented by: Joe Mazur, Purdue University
Discussant:   John Kwoka, Northeastern University
Session 33: Cartel Stability April 11, 2026 10:15 to 12:15 Location: Salon A (3rd Floor)
Session Chair: Daniel Chaves, University of Western Ontario
Third-party information platforms and coordination: Evidence from the collapse of a cartel Abstract Information-sharing agreements among competitors present a challenge for antitrust analysis, as they can improve efficiency and market trans- parency but may also promote coordination among firms. This paper studies the role of information exchanges in facilitate collusive outcomes in the con- text of alleged cartels in Quebec’s retail gasoline industry. Between 2002 and 2006, cartels reportedly operated in four cities—Magog, Sherbrooke, Thet- ford Mines, and Victoriaville before collapsing following investigation by the Competition Bureau of Canada in May 2006. During this period, a mar- ket research firm operated an information-sharing platform through which gasoline stations could exchange sales figures. To test whether information sharing was an integral component of the collusive arrangements, we inves- tigate whether participation declined following the collapse of the cartels. Using station-level data from Kent Marketing and an event study design, we compare changes in participation in the affected markets to Montreal, where no collusion allegations existed. Our results suggest that participation on the platform fell sharply following the collapse, providing evidence that an important role of information-sharing platforms is to facilitate collusion.
By Jean-Francois Houde; University of Wisconsin-Madison
Presented by: Jean-Francois Houde, University of Wisconsin-Madison
Discussant:   Mitsuru Igami, University of Toronto
They’re Cheating on You: An Experimental Examination of Third-Party Information, Trust, and Collusive Stability Abstract Collusive arrangements can be broken up by numerous factors, including entry, cheating, exogenous shocks, and antitrust intervention. In all cases, trust among participants is important to cartel stability. Competition can, in principle, be promoted by discouraging trust among competing firms. In the policy arena, for example, amnesty or leniency policies are designed to create distrust among colluding firms. This paper explores another possibility – that third-party information can undermine trust among colluding parties, thus promoting competition. Three experiments test hypotheses regarding the impact of third-party interventions on trust. In these experiments, the “third-party intervention” is a report that a seller has offered a lower price than agreed. As this information appears to come from a buyer who has an obvious interest in obtaining a low price, it lacks credibility. Nevertheless, it may seed distrust. Within these experiments, we explore the influence of new information on trust and how disruption of trust affects collusion. Third-party communication is shown to affect both trust and collusive behavior; despite potential credibility concerns, reports of cheating lead to more competitive behavior. Many players, however, returned to collusion (offering a higher price) in the final round of the game.
By Margaret Levenstein; University of Michigan    Valerie Suslow; Johns Hopkins University    Brian Gunia; Johns Hopkins University
Presented by: Margaret Levenstein, University of Michigan
Discussant:   Douglas Turner, University of Florida
On Collusion and the Number of Firms Abstract Collusion persists in many markets, fueled by supra-competitive profits. A large body of theory predicts that cartels are less likely as the number of firms rises. We show, by contrast, that under broad conditions, incentives to collude can increase with firm count. The mechanism is simple: competitive profits fall faster than collusive profits as the number of firms grows, raising the reward from coordination. This challenges a core premise of U.S. and European antitrust policy and helps reconcile theory with empirical evidence of cartels with many participants. The result holds for price and quantity competition, whether or not self-enforcement binds.
By JOAQUIN COLEFF; UNIVERSIDAD NACIONAL DE LA PLATA    Camilo Rubbini; Florida State University    Bautista Vidal; Universidad Nacional de La Plata
Presented by: Camilo Rubbini, Florida State University
Discussant:   Yaron Yehezkel, Tel Aviv University
Hub-and-Spoke vs. Horizontal Collusion: Evidence from an automotive fuel cartel Abstract A hub-and-spoke cartel, where firms (spokes) limit competition with the help of an upstream supplier (hub), is a type of collusive arrangement observed across a variety of industries. In most cases, spokes reciprocate the hub's help by excluding its rivals. We study a hub-and-spoke cartel with an exclusion condition between gas stations and distributors in Brazil's Federal District gasoline market. Using a structural model of supply and demand, we estimate the gas stations' incentive to collude for different actions taken by the hub and quantify the pricing effect of the hub-and-spoke arrangement relative to a horizontal cartel in which the upstream hub is excluded.
By Daniel Chaves; University of Western Ontario    Marco Duarte; University of North Carolina at Chapel H
Presented by: Daniel Chaves, University of Western Ontario
Discussant:   Masahiro Nishida, University of Wisconsin-Madison
Session 34: Invited Session: Industrial Organization and Education (Sponsored by Compass Lexecon) April 11, 2026 10:15 to 12:15 Location: Atlantic 1 (3rd Floor)
Session Chairs:
Adam Kapor, Princeton University
Christopher Neilson, Yale University
General Equilibrium Effects of Quality Regulation in Higher Education Abstract General Equilibrium Effects of Quality Regulation in Higher Education
By Luis Armona; Harvard University    Felipe Brugues; Universitat Pompeu Fabra    Rebecca De Simone; University of Michigan Ross School of Business    Sebastián Otero; Columbia University
Presented by: Luis Armona, Harvard University
School Voucher Design and Strategic Pricing: Evidence from India Abstract This paper uses detailed government records from India to study the world’s largest voucher system for primary education, in which voucher levels are linked to private school tuition. Lottery results show that voucher recipients benefit from improved achievement measures and lower tuition payments. However, private schools respond to the tuition-linked voucher design by raising tuition fees up to 15%, with only modest improvements in quality. These strategic responses affect the 95% of children who do not receive vouchers. Using an empirical model of the education market, welfare estimates based on revealed preference suggest the program’s benefits exceed its costs. Failing to account for school responses, however, would have over- stated the benefit-cost ratio by a factor of two. An alternative voucher design – which pays a fixed amount regardless of tuition charged – would eliminate the distortionary price incentives and increase the benefit-cost ratio by 40%.
By Harshil Sahai; University of Chicago
Presented by: Harshil Sahai, University of Chicago
The Effects of Widespread Online Education on Market Structure and Enrollment Abstract We study the rapid expansion of Brazil’s private online higher-education sec- tor and its effects on market structure and college enrollment. Exploiting regional and field-specific variation in online education penetration, we find that online programs expand access for older students but divert younger students from higher-quality in-person programs. Greater competition lowers tuition prices but also reduces the supply of in-person degrees. Using an equilibrium model of college education, we show that in the absence of online pro- grams, total enrollment would be 14 percent lower, while in-person enrollment would rise by 33 percent. On net, aggregate labor-market value added declines by 1.4 percent. Online education raises value added for older students, who benefit from increased access, but lowers it for younger students, who shift toward lower-return online options. Counterfactual poli- cies that restrict online enrollment to older cohorts could increase value added for younger students without reducing gains for older cohorts.
By Nano Barahona; University of California, Berkeley    Caue de Castro Dobbin; Georgetown University    Sebastian Otero; Columbia University
Presented by: Caue de Castro Dobbin, Georgetown University
Recommendation Design in School Choice Platforms Abstract Recommendation Design in School Choice Platforms
By Claudia Allende; Stanford University    Adam Kapor; Princeton University    Christopher Neilson; Yale University    Fernando Ochoa; NYU
Presented by: Adam Kapor, Princeton University
Session 35: Information Design and Disclosure April 11, 2026 14:15 to 16:15 Location: Thompson (4th Floor)
Session Chair: Ralph Boleslavsky, Indiana University
Dynamic Disclosure with(out) Timestamps Abstract We study the role of timestamps in a dynamic disclosure game with an evolving state. At a random date, an agent obtains one piece of hard evidence of a hidden, binary state that evolves via Markov switching. When evidence carries a timestamp, the agent discloses good evidence immediately and bad evidence after a deterministic, timestamp-dependent delay. Without timestamps, equilibrium features a stockpiling phase in which the agent delays disclosure, followed by a purging phase with stochastic purging of evidence and then a phase of immediate disclosure of any good evidence; under some conditions, bad evidence is never disclosed. Timestamps prevent the agent from pretending good evidence is fresh and allow the agent to prove that bad evidence is old, thereby accelerating good-evidence disclosure and facilitating bad-evidence disclosure.
By Aaron Kolb; Indiana University Kelley School of Business    Beixi Zhou; University of Pittsburgh
Presented by: Beixi Zhou, University of Pittsburgh
Discussant:   Peiran Xiao, University of Southern California
Information Design with Seller Manipulation Abstract We study information design in a differentiated duopoly when sellers can manipulate consumer signals. Manipulation incentives decompose into two components: a demand-stealing effect that is invariant to information design, and a strategic price effect that increases with the competitiveness of the design. This asymmetry implies that competitive information structures---those maximizing marginal consumers---are most vulnerable to manipulation. Manipulation also generates implicit competitive softening: sellers collectively benefit from noise that raises prices, even absent coordination. Consequently, a consumer-oriented designer optimally retreats from competitive structures, sacrificing price discipline to limit manipulation harm. We characterize optimal design and examine asymmetric sellers and sequential timing.
By Byung-Cheol Kim; University of Alabama
Presented by: Byung-Cheol Kim, University of Alabama
Discussant:   Heiko Karle, Frankfurt School of Finance & Management
Informationally Robust Market Outcomes Abstract This paper examines the stability of consumer information structures in markets when sellers can disclose additional information to consumers. We introduce the solution concept of robust outcomes, defined by (i) stability, ensuring that sellers cannot profitably deviate by changing price or information provision, and (ii) when multiple information structures support the same stable price-surplus outcome, we select the one that maximally deters deviations. We establish that robust outcomes must feature interim censorship in consumer information: extreme values are fully revealed, while intermediate values remain hidden. We fully characterize the set of robust outcomes and derive their Pareto frontier in monopoly markets. We also extend the characterization results to oligopoly and vertically related markets.
By Wenji Xu; City University of Hong Kong
Presented by: Wenji Xu, City University of Hong Kong
Discussant:   Hongrui Zeng, Shanghai University of Finance and Economics
Limits of Disclosure in Search Markets Abstract This paper examines competitive information disclosure in search markets with a mix of savvy consumers, who search costlessly, and inexperienced consumers, who face positive search costs. Savvy consumers incentivize truthful disclosure; inexperienced consumers, concealment. With both types, the equilibrium features partial disclosure, which persists despite intense competition: in large markets, firms always conceal low valuations. Inexperienced consumers may search actively, but only in small markets. While savvy consumers benefit from increased competition, inexperienced consumers may be harmed. Changes in search costs have non-monotone effects: when costs are low, sufficient reductions increase informativeness and welfare; when costs are high, the opposite.
By Raphael Boleslavsky; Indiana University    Silvana Krasteva; Texas A&M University
Presented by: Ralph Boleslavsky, Indiana University
Discussant:   Jorge Lemus, University of Illinois Urbana-Champaign
Session 36: Labor Market Power, Talent Allocation, and the Rise of Markups April 11, 2026 14:15 to 16:15 Location: Salon B (3rd Floor)
Session Chair: Benjamin Rosa, University of Michigan
Monopsony and Backloaded Compensation: Theory and Evidence from Public Accountants Abstract In monopsony models, wage markdowns induce deadweight loss and are, therefore, inefficient. However, wage markdowns also arise in ’tournament’ models with back- loaded compensation, in which they are designed to induce effort and are, thus, ef- ficient. These two classes of models are not nested as they impose, respectively, ex- ogenous effort and perfect labor market competition. We build a model that combines monopsony power and backloaded efficiency pay in which wage markdown varia- tion can both induce misallocation and increase effort. We bring this model to the data by estimating a team production model for the public accounting industry, using novel and uniquely-rich data on 569 U.S. firms. Our estimates reveal substantial wage markdowns for junior acccountants, which turn into wage markups after 10+ years of experience, in line with tournament models of backloaded pay. However, we also find that long-run residual labor supply is not perfectly elastic, in line with monopsony models. Nevertheless, we find that most wage markdown variation is due to efficiency pay, rather than to variation in long-run labor supply elasticities.
By Michael Rubens; UCLA
Presented by: Michael Rubens, UCLA
Discussant:   Jacob Dorn, Cornell University
A Two-Sided Matching Model of the Market for Managerial Talent with an Application to the EPL Abstract Matching managerial talent to organizations is central to brand performance, yet studying the market for managerial talent is challenging because mutual selection between managers and organizations in equilibrium generates matches that are endogenous and inter-dependent. We develop a two-sided matching model to examine how managerial characteristics and their alignment with organizational attributes shape brand performance. Methodologically, we extend the two-sided matching maximum score estimator framework of Fox (2018) by explicitly incorporating match dissolution decisions (managerial dismissals) and realized performance outcomes. This extension enables estimation of manager-specific attributes and yields a more complete measure of match value. We assess the relative performance of the proposed estimator through Monte Carlo simulations. We apply the framework to manager–club matching in the English Premier League, a setting characterized by elite managerial talent and globally valuable sporting brands. Our results reveal diminishing marginal returns to interpersonal skills as squad size increases, increasing returns to tactical expertise with higher player quality, weaker effects of talent development in older squads, and negative performance effects of managerial misconduct that are partially mitigated in clubs with stronger historical performance. Managerial poaching across competing clubs has long been criticized, yet its performance consequences remain unclear. Using our structural framework, we evaluate two counterfactual policies, a poaching ban and a managerial transfer window, and show asymmetric effects across clubs. The proposed framework is broadly applicable to other two-sided contexts in marketing and management, including influencer–brand partnerships, intermediary selection, and executive recruitment.
By Ahmed Khwaja; University of Cambridge    Liang Zhao; University of Cambridge
Presented by: Liang Zhao, University of Cambridge
Discussant:   Phuong Ho, Massachusetts Institute of Technology
What Happens to Contractors After States Ban Affirmative Action? Abstract Using restricted Census business records, I explore how banning affirmative action in state contracting affects minority- and women-owned business enterprises (MWBEs). I find that ending affirmative action led MWBE contractors to gradually downsize, with the most pronounced reductions in force experienced by Black-owned businesses and larger MWBEs. Despite these labor force changes, existing MWBEs were no more likely to shut down than other businesses. New MWBEs were relatively less common after a state's ban, highlighting how bans can deter MWBE entrepreneurship.
By Benjamin Rosa; University of Michigan
Presented by: Benjamin Rosa, University of Michigan
Discussant:   Yi Wang, Purdue University
Session 37: Environment and Energy April 11, 2026 14:15 to 16:15 Location: Brewster (4th Floor)
Session Chair: Ignacia Mercadal, University of Florida
Unintended Environmental Policies: The Impact of European Agricultural Subsidies on Pollution Abstract I study the environmental impact of subsidies in polluting markets when producers differ in their propensity to pollute. I develop a dynamic model of heterogeneous producers where the elasticity of substitution between inputs determines the correlation between firm efficiency and pollution intensity. I use it to study EU agricultural subsidies, relying on the end of price support as a natural experiment. Estimation reveals that highly productive farms pollute more. Consequently, subsidy designs that alter market selection can reduce aggregate pollution. Counterfactuals show that while taxes maximize welfare, subsidies can achieve environmental gains without the large economic surplus losses of taxation.
By Emma Du Puy; Columbia University
Presented by: Emma Du Puy, Columbia University
Discussant:   Daniel Gold, University of Wisconsin-Madison
The Efficacy of Voluntary Overcompliance for Decarbonization: Evidence from California Abstract Agents subject to regulation occasionally perform above and beyond the minimum imposed standard; this voluntary overcompliance can be especially important in the context of climate change mitigation, where each agent’s additional contributions benefit society by reducing cumulative damages from emissions. I study the case of California, which has both aggressive decarbonization regulation and high participation in voluntary green power providers called Community Choice Aggregators (CCAs). I find that CCAs procure more green power for their customers than is required, that higher income and pro-environmental political attitudes are strong predictors of selection into CCAs, and that measures of higher willingness-to-pay for decarbonized power among communities correlate with higher voluntary green procurement. However, in assessing CCAs’ impacts on statewide decarbonization progress, I find that CCAs amount to a reshuffling of voluntary greenness rather than statewide additionality. This is because California’s incumbent utilities were already engaged in voluntary green procurement prior to large-scale CCA entry. After CCAs began serving a larger proportion of statewide load, stagnation or backsliding in other parts of the sector occurred, such that the state performed 5% above the RPS in 2017 but 0% above it in 2022, while CCA load share grew from 2% to 21% of statewide sales in the same timeframe. I also provide evidence that CCAs’ elevated levels of renewable energy are mostly attributable to resources originally procured on behalf of other incumbents, such that CCAs fare no better than other types of retailers at adding new renewable generators to the system on a per-kWh basis. These findings suggest that the primary effect of voluntary green power is to affect the distribution rather than the overall magnitude of decarbonization.
By Alison Ong; Harvard University
Presented by: Alison Ong, Harvard University
Discussant:   Haiyang Zhang, Harvard Business School
Designing second-best price zones in electricity markets Abstract In electricity markets, marginal costs vary substantially across space and time, implying welfare losses under spatially or temporally uniform pricing. Nevertheless, prices are typically aggregated into large zones with spatially uniform prices. This paper develops an empirical framework to quantify the welfare loss of zonal pricing and to design welfare-improving price zones. We propose the spatial R-squared as a measure of spatial market efficiency and show that economically motivated clustering methods recover efficiency-maximizing zones. Applying the framework to three U.S. wholesale electricity markets, we demonstrate its computational tractability and find that existing zones are misaligned with current network conditions.
By Jonas Boeschemeier; Technical University of Munich    Sebastian Schwenen; Technical University of Munich
Presented by: Jonas Boeschemeier, Technical University of Munich
Discussant:   Ting Liu, Stony Brook University
The Role of Strategic Behavior During Extreme Weather Events: The Case of Winter Storm Uri Abstract We study the behavior of individual electricity generators to identify the key factors affecting power system reliability during extreme weather events, focusing on the extreme cold weather event associated with Winter Storm Uri in February 2021.. Using data from the MISO South electricity market, we assess the relative contributions of generator strategic behavior, transmission interconnectivity, input cost shocks, and other factors to price spikes and system stress during periods of extreme weather. Our results indicate that strategic generator behavior played a limited role in the observed price spikes, compared with fundamentals such as input costs and scarcity conditions. At the same time, the findings suggest that the MISO South market is relatively competitive, and thus might not be representative of all deregulated markets.
By John Birge; University of Chicago    Ali Hortacsu; University of Chicago    Ignacia Mercadal; University of Florida    Michael Pavlin; Wilfrid Laurier University
Presented by: Ignacia Mercadal, University of Florida
Discussant:   Ruby Zhang, Harvard University
Session 38: Empirical Approaches to Market Structure and Policy April 11, 2026 14:15 to 16:15 Location: Salon C (3rd Floor)
Session Chair: Emily Cook, Texas A&M University
Employment relationships, wage setting, and labor market power Abstract We ask to what extent the quantification of labor market power depends on the modeling of the long-term worker-firm employment relationship. We develop an oligopsony model with dynamic wage contracts. Workers decide whether and where to work, choosing among firms providing different amenities and solving a dynamic discrete choice labor supply problem with firm-specific human capital. As a result, firms optimally choose wage-tenure contracts to attract and retain workers. We find that such contracts mitigate firms' incentives to impose large instantaneous wage markdowns—compared to standard static wage-setting models—thereby reducing the share of socially inefficient worker-firm separations. As a consequence, we show that the empirical approaches based on "sufficient statistics" tend to overestimate the extent of labor market power: low levels of firm-specific labor supply elasticities do not necessarily indicate rent extraction, but instead reflect firms' ability to retain workers by offering long-term value through human capital accumulation.
By Francesco Agostinelli; University of Pennsylvania    Domenico Ferraro; Arizona State University    Giuseppe Sorrenti; University of Lausanne    Leonard Treuren; KU Leuven
Presented by: Leonard Treuren, KU Leuven
Discussant:   Yushuo Pan, NY Fashion Innovation Center
Market Power and the Welfare Effects of Institutional Landlords Abstract In the last decade, large financial institutions in the United States have purchased hundreds of thousands of homes and converted them to rentals. This paper studies the welfare consequences of institutional ownership of single-family housing. We build an equilibrium model of the housing market with two sectors: rental and homeownership. The model captures two key forces from institutional purchases of homes: changes in rental concentration and reallocation of housing stock across sectors. To estimate the model, we construct a novel dataset of individual homes in metropolitan Atlanta, identifying institutional owners of each house and collecting house-level daily prices, rents, vacancies, web page views, and customer contacts from Zillow. Overall, we find that institutional acquisitions decrease rents and increase rental transactions,leading to large welfare gains for renters. This net benefit reflects two opposing forces: while higher concentration raises rents, higher rental supply lowers rents enough to more than offset the effect of concentration, pushing rents down overall. These renter gains come at the expense of homebuyers, whose welfare falls. On the supply side, institutional acquisitions benefit house sellers but harm the average landlord.
By Felipe Barbieri; Dartmouth Tuck    Gregory Dobbels; U.S. Department of Justice
Presented by: Felipe Barbieri, Dartmouth Tuck
Discussant:   Takeshi Fukasawa, Waseda University
Estimating the efficiency of the production line Abstract Understanding the efficiency of the production line is very important for both manufacturers and competition authorities. However, assessing the variations in firms’ productive efficiency based on publicly available data has been challenging for economists. This research paper addresses this issue by providing a clear definition of the efficiency of the production line and introducing a novel approach to estimate the efficiency. The proposed method combines discrete choice demand estimation with a new cost function, utilizing firms’ product prices and characteristics. This approach goes beyond traditional static demand estimation techniques by allowing for dynamic variations in demand coefficients over time. The study reveals that the price parameter of the demand function, α, also serves as an indicator of firms’ productive efficiency. To validate the effectiveness of the proposed method, the study applies it to real-world data from the U.S. automobile industry, resulting in firm-specific efficiency estimates and conducting an analysis of counterfactual scenarios.
By Yushuo Pan; NY Fashion Innovation Center
Presented by: Yushuo Pan, NY Fashion Innovation Center
Discussant:   Yingjun Su, Binghamton University
Optimal Aid Policies: Beyond College Access Abstract We study how college financial aid policies affect students across the income distribution in terms of their college enrollment, in-college decisions including study effort, in-college work, and borrowing, and ultimately graduation and labor market outcomes. We consider the optimal design of higher education policies taking into account that aid may relax students’ financial constraints, incentivize study effort, and result in policy responses from colleges in terms of admissions, quality investment, and institutional aid. We give special attention to the optimal design of government aid, including loan policies.
By Emily Cook; Texas A&M University    Chao Fu; University of Wisconsin - Madison    John Stromme; univ of wisconsin
Presented by: Emily Cook, Texas A&M University
Discussant:   Michael Sullivan, University of British Columbia
Session 39: Market Power and Intellectual Property April 11, 2026 14:15 to 16:15 Location: Georges (4th Floor)
Session Chair: Jordi Jaumandreu, Boston University
Measuring Market Power from Patents Abstract The patent system grants successful inventors temporary monopoly power, yet the extent to which patents translate into market power at the firm level remains quantitatively unclear. We quantify this relationship by developing and estimating a dynamic structural model linking firms’ patent renewal decisions to their production choices. Using data on publicly traded U.S. manufacturing firms that account for roughly 70% of sectoral revenue from 2000 to 2018, we find that expanding a firm’s patent portfolio by adding patents whose lifetime citations sum to 1,000 increases the firm’s markup by 0.0015 on average. Using the model, we show that patents account for 18.57% of revenue-weighted firm markups, with substantially larger effects in high technology sectors than in traditional industries. Moreover, the revenue-weighted contribution of patents to markups increased from 14% to 21% over the sample period.
By Xinghua Long; Shanghai University of Finance and Economics    Dongni Zhu; Shanghai University of Finance and Economics
Presented by: Dongni Zhu, Shanghai University of Finance and Economics
Discussant:   Mohaddeseh Heydari Nejad, Indiana University
Intangible Assets and Imperfections in Product and Labor Markets Abstract This paper develops a micro-founded framework linking price–cost and wage markups to intangible assets. Firms enter the intangible sector as innovators, hiring from unemployment to create match-specific intangibles, or as poachers, recruiting from innovators. A safe sector provides the outside option. Appropriability of innovation returns depends on worker leakability and firm retainability, generating hold-up and motivating rent sharing. Search frictions affect mobility. Under non-compete agreements, poached workers face start delays that weaken outside options. Using microdata from the Netherlands, we document higher price-cost and wage markups in more intangible-intensive firms and lower wages for workers under non-compete agreements in these firms, consistent with the model.
By Eric Bartelsman; Vrije Universiteit Amsterdam and Tinberg    Sabien Dobbelaere; Vrije Universiteit Amsterdam    Alessandro Zona Mattioli; Vrije Universiteit Amsterdam
Presented by: Sabien Dobbelaere, Vrije Universiteit Amsterdam
Discussant:   Giulio Gottardo, University of Oxford
Labor market power and innovation Abstract This paper examines how labor market power shapes how firms innovate and grow. We develop an endogenous growth model where firms optimize R&D spending to increase their future productivity while facing an upward-sloping labor supply curve, generating monopsony power. This creates two opposing distortions: (1) monopsonistic firms have stronger incentives to innovate and grow as they enjoy larger profits, but (2) firm growth increases (infra-)marginal labor costs by pushing firms up the labor supply curve, which reduces the returns to productivity-enhancing innovation. Theoretically, the first effect dominates for small firms, while the second is stronger for large firms. We test these predictions using rich firm-level data from the German manufacturing sector (1995–2018) to estimate firms' productivity and labor market power. Empirically, we find that, conditional on size, labor market power negatively correlates with R&D investment. Furthermore, small (large) firms in high-monopsony-power regions exhibit relatively high (low) R\&D spending, compared to competitive labor markets, which aligns with our model's predictions. When applying our model to the data, we find that East Germany's higher labor market power can explain 24.7% of the persistent productivity gap between East and West Germany and depresses overall GDP growth by 0.2% p.a.
By Richard Bräuer; Institute for Economic Research Halle    Jonathan Deist; IWH    Matthias Mertens; Massachusetts Institute of Technology
Presented by: Matthias Mertens, Massachusetts Institute of Technology
Discussant:   Hugo Molina, University of Paris-Saclay
Market Power and Technology in US Manufacturing Abstract We measure market power in product and labor markets for firms in US manufacturing from 1958 to 2018 using NBER-CES and Compustat data. Measurement is robust to any form of competition and accounts for Hicksian and labor-augmenting productivity to avoid biases in estimation. We estimate the long and short-run elasticities of scale and the wage markdown. These estimates allow us to infer the price-cost markup and evaluate contributions of product market power, labor market power, and technology to short-run profitability. Preliminary results show 36\% profitability, with product market power contributing 18 percentage points, technology 12, and labor market power 6.
By Sabien Dobbelaere; Vrije Universiteit Amsterdam    Jordi Jaumandreu; Boston University    Jacques Mairesse; ENSAE and Maastricht University
Presented by: Jordi Jaumandreu, Boston University
Discussant:   Stefan Weiland, Maastricht University
Session 40: Platform design: Information April 11, 2026 14:15 to 16:15 Location: Caspian (3rd Floor)
Session Chair: Joosung Lee, Sungkyunkwan University
Exposure Design for Two-Sided Platforms Abstract Online platforms choose exposure rules—who is shown to whom and how often—to speed up matches and raise flow surplus. Yet aggressively matching today’s best pairs can cannibalize future opportunities by thinning the effective option set for those who remain. I develop a two-sided sequential-search model with platform-controlled meeting propensities and define \emph{user value} as the aggregate continuation value of search on the platform, a natural objective for platforms that seek to grow and retain users. I show that maximizing flow match surplus generally does not maximize user value, and I propose a tractable algorithm to compute user-value–optimal exposure via entropic regularization, annealing, and Bregman–Dykstra projections. Applying the framework to a doctor–spot-job platform, I estimate preferences under two exposure regimes and quantify the gains from redesigning exposure.
By Kei Ikegami; University of Tokyo
Presented by: Kei Ikegami, University of Tokyo
Discussant:   Joosung Lee, Sungkyunkwan University
From Curation to Creation: How AI Summaries Shape User Reviews Abstract The increasing deployment of Generative AI tools to summarize User-Generated Content (UGC) raises critical questions regarding the influence on contribution incentives and the informational integrity of platforms. This research causally isolates the impact of AI-generated review summaries on subsequent user review behavior—specifically changes in quantity, valence, and content effort—by employing a Difference-in-Differences (DiD) strategy that exploits varied deployment of this feature across home improvement retailers using review data. We find that the AI summary feature acts as a positive information filter and influence, increasing the number of new reviews posted by approximately 28.5%, contrary to predictions that summarization suppresses UGC volume. Furthermore, the AI feature drives a significant upward shift in valence, resulting in a 3.3% increase in the average rating, primarily concentrated in 5-star reviews. Content analysis confirms that the summary successfully frames the affective discussion, leading to a 12.5% increase in Sentiment Agreement with highlighted aspects, while the thematic focus of reviews remains stable.
By Leshui He; Bates College    Imke Reimers; Cornell University    Benjamin Shiller; Brandeis University
Presented by: Leshui He, Bates College
Discussant:   Leon Musolff, Wharton School of the University of Penn
The Limits of Algorithmic Recommendation Abstract We study a bilateral posted-price market in which a platform cannot make transfers but can send private, obedient recommendations. We take Chatterjee and Samuelson (1983)'s double auction as a decentralized benchmark and ask how far informational algorithms can move outcomes. Our first result is a replication theorem: for any class-A sealed-offer equilibrium, a simple buyer-threshold recommendation exactly reproduces its trade set and interim payoffs type-by-type, so the double auction locus lies inside the recommendation-attainable set. Second, we characterize the upper Pareto frontier via a family of alpha-optimal threshold rules, spanning the seller-optimal efficient corner to the buyer-optimal corner. This frontier yields distribution-robust envelope theorems: at every seller-profit level achieved by sealed offers, some frontier recommendation weakly raises buyer surplus and maximizes welfare gains subject to the posted-price, no-transfer constraints. Third, we show that this boundary is robust to richer algorithms. Under a mild product structure on signals, two-sided information design has the same upper frontier as buyer-threshold recommendations, because seller payoffs depend only on the induced price -- demand curve and thresholdization is buyer-optimal at fixed demand. Budget-balanced ex-ante fees leave behavior unchanged and simply redistribute along the efficiency line. In this sense we identify the limits of algorithmic recommendation: every sealed-offer outcome can be replicated and is enveloped by threshold recommendations at the same seller-profit level, but no purely informational algorithm can cross the full-information welfare bound.
By Jaemin Son; Sungkyunkwan University    Joosung Lee; Sungkyunkwan University
Presented by: Joosung Lee, Sungkyunkwan University
Discussant:   Nicolas Bozzo Galleguillos, Keystone
The Economics of Algorithmic Personalization: Evidence from an Educational Technology Platform\ Abstract Can personalized recommendations improve engagement in educational technology? We design, test, and scale a collaborative filtering system for Freadom, an English-learning app for Indian children. A randomized controlled trial (RCT) shows that personalization, deployed in a single content section, increases engagement by 60% in that section and by 14% app-wide. We then exploit an eligibility threshold in a regression discontinuity design (RDD) to track effects over five months of deployment. For user cohorts receiving personalization during deployment, RDD estimates exceed RCT benchmark by a factor of 2.5, opposite of the ``voltage drop" typically observed in policy scale-ups. This provides evidence that, for algorithmic interventions, RCT estimates may be lower bounds on scaled impact rather than upper bounds. However, personalization benefits are front-loaded. Gains concentrate in users' first weeks, with diminishing returns thereafter. This pattern, combined with the sharp decline in predicted match quality as users exhaust their best content matches, suggests that content availability rather than algorithmic sophistication becomes the binding constraint.
By Emil Palikot; Northeastern University
Presented by: Emil Palikot, Northeastern University
Discussant:   Myongjin Kim, U Oklahoma
Session 41: Dynamics and Market Design in Environmental IO April 11, 2026 14:15 to 16:15 Location: Atlantic 2&3 (3rd Floor)
Session Chair: Kenneth Gillingham, Yale University
Battery Bidding and Market Efficiency in Energy Transition Abstract Utility-scale batteries play an increasingly active role in electricity markets by arbitraging energy across time, yet their efficiency depends critically on market design. This paper studies a prevalent feature of real-time electricity auctions: bid lead time, which requires batteries to commit bids in advance of dispatch. Using comprehensive 15-minute bidding and settlement data for all utility-scale batteries in ERCOT from 2018–2025, we document active dynamic bidding, rapid responses to price shocks, and the importance of recent price information for forecasting. We develop an equilibrium model of electricity markets with dynamic battery bidding and show that longer bid lead times substantially reduce operational efficiency and profits, and dampen complementarity between batteries and renewable energy. Allowing bids to depend on state of charge significantly mitigates these losses. Our results highlight bid lead time as a key determinant of battery performance and market outcomes.
By Hunt Allcott; Stanford University    Luming Chen; University of Michigan    Julia Park; Stanford University
Presented by: Luming Chen, University of Michigan
Discussant:   R. Andrew Butters, Indiana University
Cows and Trees Abstract The Brazilian Amazon plays a crucial role in regulating global climate and preserving biodiversity, yet it faces mounting pressures from deforestation, driven primarily by cattle ranching. The expansion of pastureland is shaped by cattle’s dual role as both output and capital stock, leading to nontrivial dynamic patterns. We develop a structural empirical model of ranchers’ cattle management and land use decisions that accounts for deforestation costs, herd dynamics, and price expectations. The model estimates reveal that deforestation is inelastic to temporary shocks to beef prices but highly elastic to persistent price changes, rationalizing existing estimates in the literature. Finally, we simulate various policies and discuss the implications of highly price-elastic deforestation.
By Paul Scott; Massachusetts Institute of Technology
Presented by: Paul Scott, Massachusetts Institute of Technology
Discussant:   Luming Chen, University of Michigan
Resale Markets for Differentiated Durable Goods: A Model of the Fashion Industry Abstract Abstract Despite the large environmental footprint of the rapidly expanding fashion market, few economic models address the structure of fashion and apparel markets. This paper examines the role of secondary markets by developing a simple dynamic model of a differentiated durable-goods market with infinitely lived consumers, extending the CES monopolistic competition framework of Dixit and Stiglitz (1977). Firms introduce new varieties in each period, and under free entry, the number of varieties is determined endogenously. With representative consumers, the option to resell acts as a demand subsidy: it lowers the prices of new goods, increases total durable-goods production, reduces product variety, and decreases consumer welfare, even though no trade actually occurs in the secondary market. When consumers have heterogeneous preferences over new and used goods, total durable-goods production still increases, driven by a shift in the population distribution toward a stronger preference for used goods. In contrast, when there are two types of products – high-quality durable goods and low-quality perishable goods – the introduction of secondary markets increases demand for secondhand high-quality goods and reduces total production by substituting low-quality perishable goods for high-quality durable goods. A shift in the population distribution toward stronger preferences for used goods has the same effect. These results suggest that policies encouraging the use of secondhand goods may be effective in reducing environmental harm when products differ in durability.
By Hideo Konishi; Boston College    Ying Wang; HSE University
Presented by: Hideo Konishi, Boston College
Discussant:   Paul Scott, Massachusetts Institute of Technology
Valuing Solar Subsidies Abstract Individuals trade present for future consumption across a range of economic behaviors, and this tradeoff may differ across socioeconomic groups. To assess these tradeoffs, we estimate a dynamic model of residential solar adoption and system sizing using household-level data that offer plausibly exogenous variation in the future benefits from adopting relative to upfront costs. We find implicit discount rates of 17.2%, 15.6%, and 10.9% for low-, medium-, and high-wealth households. This heterogeneity remarkably persists for those with high credit scores. Counterfactual simulations demonstrate opportunities to reduce the regressivity of solar adoption, increase policy cost-effectiveness, and improve welfare for low-wealth households.
By Bryan Bollinger; New York University    Kenneth Gillingham; Yale University    Justin Kirkpatrick; Michigan State University
Presented by: Kenneth Gillingham, Yale University
Discussant:   Takeaki Sunada, University of Rochester
Session 42: Hospital Performance, Pricing, and Provider Consolidation April 11, 2026 14:15 to 16:15 Location: Aegean (3rd Floor)
Session Chair: Jia Xiang, Indiana University
Targeting Efficiencies? Reducing Costs Through Hospital Consolidation Abstract Policymakers are concerned about the effects of consolidation in the hospital industry, much of it driven by the expansion of multi-unit hospital "systems." Prior work has shown that system acquisitions lead to net price increases at the target, but the mechanisms are not well understood. Price changes can occur through three channels: an increase in market power, the acquirer having higher pricing power conditional on market power, often interpreted as bargaining "ability," and the pass-through of cost synergies. We quantify the relative importance of these channels using novel data on transaction prices for hospital outpatient care and detailed administrative data on hospital costs from California. We first show that system ownership generates significant net price increases and cost savings for independent hospitals but not for targets that were already organized as systems. We then develop and structurally estimate a model of patient demand and hospital-insurer negotiations to quantify the different mechanisms. We find that systems target for acquisition independent hospitals with weak bargaining ability, suggesting a strategy of pricing arbitrage, and this explains most of the price increase observed at the target. We also detect significant aggregate cost savings from system ownership, of which approximately 50% is passed on to insurers in the form of price reductions. However, since the independent hospitals acquired are small relative to the acquirers, synergies do not meaningfully affect the system's negotiated prices.
By Jonathan Arnold; Cornerstone Research    Atul Gupta; University of Pennsylvania    Tong Liu; Massachusetts Institute of Technology    Alexander Olssen; The Wharton School, University of Pennsy
Presented by: Tong Liu, Massachusetts Institute of Technology
Discussant:   Yanhao Wang, University of Alberta
The Effects of Dialysis Firm Consolidation on Patient Health Abstract Like much of healthcare, the U.S. dialysis industry has become increasingly consolidated. Yet, evidence on the mechanisms and magnitude of consolidation’s impact on patient health remains limited. Using 30 years of administrative data covering more than 800 mergers and a stacked event-study design, we show that mergers trigger facility closures and short-term treatment disruptions that harm patients. Patients receive fewer dialysis sessions, and mortality rises by roughly 700 deaths per 100,000 patients in the merger year. Over time, patient health improves, with fewer hospitalizations, ICU days, and blood transfusion events, even as effects on laboratory biomarkers remain mixed. These merger effects are not driven by changes in market concentration, and hold across Medicare payment regimes with differing reimbursement structures and quality incentives. Mergers’ health costs, through short-run mortality increases, outweigh their benefits, through long-run reductions in hospitalization. Yet, targeted antitrust safeguards against post-merger treatment disruptions have the potential to curtail harms without sacrificing gains.
By Francisco Garrido; ITAM    Anwita Mahajan; University of California, San Diego    Adrian Rubli; ITAM
Presented by: Anwita Mahajan, University of California, San Diego
Discussant:   Jack Berger, University of Toronto, Rotman School of Management
Surprising Facts about the No Surprise Act Abstract The No Surprises Act (NSA), enacted in 2022, protects patients from unexpected out-of-network bills through an independent dispute resolution (IDR) process. Yet arbitration volumes have far exceeded expectations, with outcomes heavily favoring providers, raising concerns about higher premiums and narrower networks. This project investigates what drives excessive arbitration in emergency physician services and how outcomes shape provider-insurer negotiations, network participation, and patient costs. We construct a novel dataset that links federal IDR records to ACA Marketplace networks. Preliminary analyses show that NSA improves network participation, especially in states without any state laws that regulate. However, several procedures in certain states show a decline in network participation, implying strategic manipulations by institutional providers.
By Panle Jia Barwick; UW Madison and NBER    Anran Li; University of Minnesota, Twin Cities    Tianli Xia; University of Rochester    Shizhe Yu; University of Wisconsin-Madison
Presented by: Shizhe Yu, University of Wisconsin-Madison
Discussant:   Florian Ederer, Boston University
Race-Specific Provider Performance: Evidence from Black Lives Matter Protests Abstract This paper studies whether hospital performance for Black patients is malleable and how it shapes racial disparities in healthcare. Using local Black Lives Matter protests as an exogenous shock to providers’ racial awareness and social pressure, we find that exposed hospitals experienced significant declines in 30-day mortality among Black heart attack patients, with no detectable change for non-Black patients. Evidence from emergency department admissions and non-exposed patients supports a supply-side response. We observe no changes in resource utilization but substantial productivity gains, partly driven by an increased likelihood of Black patients being treated by high productivity physicians.
By Haizhen Lin; Indiana University    Yanhao Wang; University of Alberta    Jia Xiang; Indiana University
Presented by: Jia Xiang, Indiana University
Discussant:   Ashvin Gandhi, UCLA
Session 43: Trade and IO (I) April 11, 2026 14:15 to 16:15 Location: Spectacle (4th Floor)
Session Chair: Felix Montag, NYU Stern School of Business
Gravity beyond CES Abstract We derive a linear structural gravity equation that allows for rich substitution patterns based on observable characteristics. To achieve this, we take advantage of recent econometric work to linearize an import demand system with mixed CES preferences. Compared to traditional gravity models, the resulting equation features additional regressors that capture heterogeneity in the patterns of substitution across exporters. Importantly, this equation can be easily estimated through two-stage least squares (2SLS) and without additional data requirements relative to traditional gravity. We implement this method using bilateral trade data and find that the data strongly rejects the Independence of Irrelevant Alternative (IIA) assumption implied by standard trade models: we find an important role for vertical and geographical differentiation so that exporters with similar prices, or originating from similar regions, are closer substitutes. We show that this pattern has important implications in the context of the recent US-China trade war, in which our model can correctly predict which countries benefit the most from the reallocation of trade flows due to US tariffs on Chinese imports.
By gabriel smagghue; banque de france
Presented by: gabriel smagghue, banque de france
Discussant:   Luca Lorenzini, UCLA Anderson
Mergers and the Demand for Protectionism Abstract When import competition is strong, domestic mergers can strengthen the incentives to seek trade protection. However, merger control treats rivals’ production location as irrelevant. While intense foreign competition may support merger clearance under current practice, existing enforcement does not consider how mergers alter the merging parties’ incentives to petition for tariffs. I develop a model to quantify how mergers affect the merging firms’ demand for tariffs. I show that mergers between domestic producers increase their incentive to petition for tariffs and can generate merger-induced consumer harm through the trade-policy channel, whereas cross-border mergers are unlikely to have this effect. I apply the framework to the Whirlpool–Maytag merger in the U.S. washer market and show that the merger substantially amplified the profitability of tariffs for Whirlpool, resulting in consumer harm via the trade-policy channel that is comparable in magnitude to the direct harm from increased market power.
By Felix Montag; NYU Stern School of Business
Presented by: Felix Montag, NYU Stern School of Business
Discussant:   Jonathan Elliott, Johns Hopkins University
Agglomeration, Productivity, and Exports: A Unified Framework for Firm Dynamics Abstract This paper develops a unified empirical framework to study the interdependent effects of agglomeration on firm productivity and export behavior. While previous research has separately established that spatial proximity can enhance firm productivity through knowledge spillovers, and that exporting decisions are influenced by local networks, existing models typically treat productivity and exporting as independent outcomes. In contrast, this study recognizes the bidirectional relationship between productivity and exports—where more productive firms are more likely to export, and exporting can in turn enhance firm productivity. By jointly modeling these outcomes within the context of regional agglomeration economies, we aim to correct potential biases from partial frameworks that do not fully explore the nexus between firm pro- ductivity and export. The proposed approach allows for a richer understanding of how firms evolve in clusters, leveraging local knowledge flows, infrastructure, and collective market experiences. This contributes to the literature by offering a more comprehensive perspective on how spatial economic forces shape firm dynamics and international engagement.
By Yulu Wang; Binghamton University    Shunan Zhao; Oakland University    Man Jin; Oakland University    Mingyang Li; University of Connecticut
Presented by: Yulu Wang, Binghamton University
Discussant:   Yannick Bormans, KU Leuven
Foreign Import Competition and Process Innovation Abstract I examine how Chinese import penetration changes the type of innovation U.S. firms pursue. Using an instrumental variable strategy, I find that rising imports from China lead firms to shift their innovative efforts toward process innovation, which focuses on cost reduction rather than the creation of innovative products. This shift is stronger when product differentiation is difficult, when firms operate in ex-ante riskier environments, and for more labor-intensive firms. In contrast, the effect is weaker for firms that already have alternative ways to reduce costs. I also show that process innovation buffers firms against the adverse effects of import competition. Firms with higher process innovation experience smaller declines in profitability and in the growth of sales, assets, employment, and capital when import penetration rises. When Chinese import competition is high, market participants reward firms with more process-oriented patents, especially in industries in which product differentiation is limited. Together, these findings show that product market structure and firm characteristics play a central role in shaping how firms modify their investments in innovation in response to foreign competitive pressure.
By Nima Nematian; University of Arizona
Presented by: Nima Nematian, University of Arizona
Discussant:   gabriel smagghue, banque de france
Session 44: Labor Antitrust April 11, 2026 14:15 to 16:15 Location: Mediterranean (3rd Floor)
Session Chair: David Benson, Federal Reserve Board
Labor Market Power and Collusive Behavior Abstract This paper develops a theory of collusion in the presence of labor market power. In an oligopoly-oligopsony setting, a firm needs to increase its wage offers to recruit more workers and expand production, which dampens incentives to deviate from a collusive outcome. As a result, labor market power increases firms' ability to collude, and collusion harms consumers and workers, underlining the need for antitrust authorities to monitor collusive behavior also in labor markets. However, if only wage collusion is monitored, or is prevented by enforcing a minimum wage, firms fiercely collude on prices, leaving consumers worse off than under unconstrained collusion. By creating externalities across independent product markets, labor market power also engenders cross-market collusion and implies that conglomerate mergers produce anticompetitive multimarket-contact effects. No-poaching and non-compete agreements, preventing a firm from hiring its rivals' workers, act as facilitating practices; pay-equity regulations similarly discourage deviations and facilitate collusion.
By Michele Bisceglia; Yale University
Presented by: Michele Bisceglia, Yale University
Discussant:   Zhonglin Li, National University of Singapore
Common Ownership, Firm Hiring and the Allocation of Human Capital Abstract We study how common ownership shapes the allocation of human capital across firms. We construct a novel dataset combining LinkedIn-based employment histories with data on venture capital (VC) investments in U.S. startups. Exploiting the staggered adoption of liability waivers that allow investors to hold stakes in competing business opportunities as a plausibly exogenous shock to common ownership, we show that ventures with shared investors employ more workers, pay higher wages, and experience greater turnover than independent ones. At the same time, common ownership increases job-to-job mobility within investor portfolios, suggesting that shared investors facilitate the reallocation of workers across portfolio firms. Aggregating to the local market level, we find that common ownership significantly increases worker mobility without depressing overall employment or wages, though these positive effects weaken in more concentrated labor markets. Taken together, the results are consistent with coordination synergies that relax matching frictions outweighing potential monopsony effects.
By Mario Leccese; Boston University
Presented by: Mario Leccese, Boston University
Discussant:   Jaehak Lee, University at Albany, SUNY
Quantifying the Distortions of Labor Market Power: U.S. Coal Mines 2001-2019 Abstract I quantify how labor market power distorts the broader production process, combining evidence from merger event studies and a structural oligopsony model in which wages, employment, capital, and output are jointly determined in equilibrium. Using administrative data from the U.S. coal industry, I show that rent extraction from the labor market not only lowers wages and employment, but also creates a scale effect that reduces the firm’s demand for capital, suppresses output, and diminishes aggregate productivity. These “knock-on” distortions of labor market power are four times larger, by value, than the direct welfare loss to workers. My results suggest that labor market outcomes can significantly underestimate the welfare costs of oligopsony, especially in capital-intensive industries like mining.
By David Benson; Federal Reserve Board
Presented by: David Benson, Federal Reserve Board
Discussant:   Leonard Le Roux, International Finance Corporation - World Bank Group
Session 45: Merger Remedies (Sponsored by Analysis Group) April 11, 2026 14:15 to 16:15 Location: Salon A (3rd Floor)
Session Chair: Vivek Bhattacharya, Northwestern University
Store Exit After Structural Merger Remedies in the U.S. Grocery Industry Abstract Asset divestitures are commonly used as structural remedies to mitigate the anticompetitive effects of mergers. However, their design often involves policy debates, and their effectiveness is understudied. This paper asks whether stores divested as structural remedies are more likely to exit the market. Using panel data on grocery stores in the United States, I examine exit behavior of divested stores relative to comparable non-divested stores. Conditional on controls, divested stores exhibit higher exit rates: one-year exit is 3.1–5.5 percentage points higher (about 37–58\% relative to baseline exit rates of 8–10\%), and three-year exit is 4.6–6.9 percentage points higher (about 58–69\%). Observed survival durations are about 2.6–3.5 years shorter (about 53\% relative to a baseline of about 6.5 years). These patterns are consistent with concerns about buyer viability in remedy design.
By Kosuke Shimamoto; Duke University
Presented by: Kosuke Shimamoto, Duke University
Discussant:   Flavia Roldan, Universidad ORT Uruguay
Antitrust on Aisle Five: Lessons from Decades of Supermarket Divestitures Abstract Antitrust authorities often require asset divestitures to address competitive concerns raised by mergers. This paper leverages detailed business data to evaluate whether such divestitures successfully preserved competition in the U.S. grocery sector. We find that many divested stores struggled to remain profitable. Divested stores exhibit poor sales performance and have high exit rates relative to a control group of stores. Our analysis explores the factors behind these outcomes that underscore the importance of assessing whether divestiture buyers are capable of competing as effectively as the original merging firms. We find evidence that pre-acquisition banner and firm strength are important factors in predicting the success of a divestiture.
By Xiao Dong; FTC    Paul Koh; Yonsei University    Devesh Raval; Federal Trade Commission    Dominic Smith; Bureau of Labor Statistics    Brett Wendling; Federal Trade Commission
Presented by: Devesh Raval, Federal Trade Commission
Discussant:   Marc Remer, Swarthmore College
The Welfare Effects of Structural Remedies Abstract We study the welfare effects of structural remedies in horizontal mergers when both divested products and acquirer are chosen by the merging parties subject to a regulatory screen. We model remedies as reallocation of products across firms that move a pair of indices—summarising unilateral pricing incentives of the merged firm and the buyer. In this setting, we characterise the welfare frontier and the planner’s minimally distortive, welfare-restoring remedy and the one proposed by the firms. We then show how competition authorities’ screening rules, which typically proxy welfare with concentration or other stylised thresholds, can be misaligned with this frontier and therefore accept miss-allocating remedies that either under- or over-correct the merger. Then, using a structural model, we find that the proposed divestitures in a merger between two of the main suppliers of spirits in Sweden, decreased the distortions absent remedies but was not enough to re-establish pre-merger welfare.
By Oscar Jara; Norwegian School of Economics    Pedro Rojas; Toulouse School of Economics
Presented by: Oscar Jara, Norwegian School of Economics
Discussant:   Margaret Loudermilk, Charles River Associates
A Large-Scale Evaluation of Merger Simulations Abstract Prospective merger simulations are a commonly used tool in antitrust, but evidence about their accuracy is limited. We study 92 consummated mergers in consumer packaged goods and compare the realizations of price changes with predictions from merger simulations. Predicted price changes are correlated with, but typically larger than, realized price changes. Merger simulations have limited predictive power but are still typically more predictive than relying solely on structural presumptions. We use the departure between realization and predictions to estimate supply-side changes after the merger and find a role for both synergies and a departure from Bertrand-Nash conduct post-merger.
By Vivek Bhattacharya; Northwestern University    Gaston Illanes; Northwestern University    Avner Kreps; Northwestern University    Jose Salas; Northwestern University    David Stillerman; American University, Kogod School of Bus
Presented by: Vivek Bhattacharya, Northwestern University
Discussant:   Gloria Sheu, Federal Reserve Board of Governors
Session 46: Invited Session: Algorithmic Pricing (Sponsored by Keystone) April 11, 2026 14:15 to 16:15 Location: Atlantic 1 (3rd Floor)
Session Chairs:
Matthijs Wildenbeest, University of Arizona
Jeanine Miklos-Thal, U Rochester
Algorithmic Pricing and Consumer Sensitivity to Price Variability Abstract Algorithmic Pricing and Consumer Sensitivity to Price Variability
By Madhav Kumar; Harvard Business School    Diego Aparicio; IESE Business School    Dean Eckles; MIT    Catherine Tucker; MIT Sloan
Presented by: Madhav Kumar, Harvard Business School
Third-Party Pricing Algorithms and Information Sharing Abstract Third-Party Pricing Algorithms and Information Sharing
By Jeanine Miklos-Thal; U Rochester    Stepan Aleksenko; Southern Methodist University
Presented by: Jeanine Miklos-Thal, U Rochester
Collusion with Optimal Information Disclosure Abstract Motivated by recent concerns surrounding the use of third-party pricing algorithms by competing firms, we study repeated Bertrand competition where market demand or the cost of serving the market is observed by an intermediary (or ìalgorithmî) that optimally discloses demand or cost information to maximize Örmsícollusive proÖt.
By Takuo Sugaya; Stanford University    Alexander Wolitzky; MIT
Presented by: Alexander Wolitzky, MIT
Algorithmic Collusion of Pricing and Advertising on E-commerce Platforms Abstract When online sellers use AI learning algorithms to automatically compete on e-commerce platforms, there is concern that they will learn to coordinate on higher than competitive prices. However, this concern was primarily raised in single-dimension price competition. We investigate whether this prediction holds when sellers make pricing and advertising decisions together, i.e., two-dimensional decisions. We analyze competition in multi-agent reinforcement learning, and use a large-scale dataset from Amazon.com to provide empirical evidence. We show that when consumers have high search costs, learning algorithms can coordinate on prices lower than competitive prices, facilitating a win-winwin for consumers, sellers, and platforms. This occurs because algorithms learn to coordinate on lower advertising bids, which lower advertising costs, leading to lower prices and enlarging demand on the platform. We also show that our results generalize to any learning algorithm that uses exploration of price and advertising bids. Consistent with our predictions, an empirical analysis shows that price levels exhibit a negative interaction between estimated consumer search costs and algorithm usage index. We analyze the platform's strategic response and find that reserve price adjustments will not increase platform profits, but commission adjustments will, while maintaining the beneficial outcomes for both sellers and consumers.
By Hangcheng Zhao; Rutgers University    Ron Berman; Wharton School
Presented by: Hangcheng Zhao, Rutgers University
Session 47: Advances in Computational Methods in IO (Sponsored By Amazon) April 11, 2026 17:00 to 18:30 Location: Atlantic 1 (3rd Floor)
Session Chairs:
John Rust, Georgetown University
Ulrich Doraszelski, University of Pennsylvania
Session 48: Switching Costs in the Digital Economy (Sponsored by Econic Partners) April 11, 2026 17:00 to 18:30 Location: Atlantic 2&3 (3rd Floor)
Session Chairs:
Gary Biglaiser, UNC-Chapel HIll
Luis Cabral, NYU Stern
Session 49: Personalization, Pricing, and Consumer Data April 12, 2026 8:00 to 10:00 Location: Thompson (4th Floor)
Session Chair: Heiko Karle, Frankfurt School of Finance & Management
Competitive Personalized Pricing with Multidimensional Characteristics Abstract Consumers differ in both their brand-dependent preferences (loyalty) and the intensity of their brand-independent preferences (choosiness). Firms produce horizontally differentiated products and, depending on data availability or competition policy, tailor their prices based on what they learn about consumer characteristics. Information along different dimensions of consumer characteristics yields contrasting implications for consumer welfare and industry profit. Either loyalty-based pricing---i.e., price discrimination based only on loyalty---or fully personalized pricing---i.e., price discrimination based on both loyalty and choosiness---maximizes consumer welfare, while partially personalized pricing based on choosiness always maximizes industry profit.
By Qiang Fu; National University of Singapore    Zenan Wu; Peking University    Yuxuan Zhu; Peking University
Presented by: Qiang Fu, National University of Singapore
Discussant:   Beixi Zhou, University of Pittsburgh
Non-Linear Pricing under Quantity Restrictions and Taxes : Market for Fountain Soda Abstract In the sugar-sweetened beverage market, firms use nonlinear menu pricing to sell different drink sizes, distorting quantities for low-valuation consumers. These distortions change how taxation affects supply, making a Pigovian tax calibrated to the externality from excess sugar consumption welfare-suboptimal. A size restriction, similar to New York City’s 2012 proposed but unimplemented maximum drink size, offers an alternative approach by directly constraining portion sizes. I compare welfare under these two approaches, accounting for firms’ optimal menu responses. I identify two testable conditions under which size restrictions yield higher welfare than taxes: (1) the externality is small enough that a planner would not further reduce the smaller drink size, or (2) the monopolist’s distortion for low-type consumers is so large that the planner prefers a pooling allocation. Using convenience store data, I find that 76 to 90 percent of stores would be better served by a size restriction, driven by the fact that the modal store sells four larges for every small.
By Seth Smith; University of Georgia
Presented by: Seth Smith, University of Georgia
Discussant:   Thi Mai Anh Nguyen, New York University
Personalizing Two Interdependent Treatments: Structural Estimation and Evidence from Social E-Commerce Abstract In contemporary marketing, personalization has become central to firm strategy. While existing literature has predominantly examined the personalization of individual marketing variables in isolation, this research demonstrates the limitations of such approaches. We focus on the personalization decision of two interdependent marketing variables: the level of a promotional price discount, and whether to require a consumer to broadcast the promotion on social media in order to receive the discount. We develop a two-stage structural model, which captures two sequential customer decisions: sharing and purchasing. Using large-scale field experiment data from a leading Chinese grocery e-tailer, we estimate the model primitives (willingness to pay, sharing costs, referral generation capacity, and implied referral value) and conduct counterfactual profit comparisons of different personalization schemes. The counterfactual analysis demonstrates that price personalization alone increases profits by only 2.4%, whereas personalizing the broadcasting requirement increases profits by 27%. In contrast, an integrated approach that personalizes both price and broadcasting requirements increases profit by 40.3%, exceeding the sum of the profit increases from personalizing these requirements in isolation. These results highlight the value of coordinating across interdependent marketing variables. The paper also provides a framework for future work on multi-dimensional personalization.
By Xin Chen; Singapore Management University    Yunhao Huang; University of Southern California    Matthew Osborne; University of Toronto
Presented by: Yunhao Huang, University of Southern California
Discussant:   Max Schnidman, Microsoft
Context Effects and Price Dispersion in Search Markets Abstract We analyze search market models in which consumers exhibit context effects, i.e., they become less sensitive to price variations of fixed size when the price level or range inncreases. Context effects impose an upper bound on the value of search that is independent of the product value level. The price distribution implied by a search market model further tightens this bound. Small search costs therefore can generate "Varian-type" equilibria in which the option to search no longer disciplines prices and small reductions in search costs have no effect on the equilibrium price distribution. We use Varian-type equilibria to rationalize the observed price dispersion in several examples from the empirical literature with small search costs and to provide a framework that unifies different types of competition models with informational frictions.
By Heiko Karle; Frankfurt School of Finance & Management    Marcel Preuss; SC Johnson Graduate School of Management    Heiner Schumacher; University of Innsbruck
Presented by: Heiko Karle, Frankfurt School of Finance & Management
Discussant:   Byung-Cheol Kim, University of Alabama
Session 50: Bargaining, Restraints, and Spillovers in Vertical Markets April 12, 2026 8:00 to 10:00 Location: Salon B (3rd Floor)
Session Chair: Giulia Sabbadini, Düsseldorf Institute for Competition Economics
Cross-category Portfolio and Bargaining Power: Evidence from the US Men's Razor Market Abstract This paper studies how a manufacturer’s dominance in one product category affects its bargaining outcomes in another. Using data on the US shaving market, we develop a Nash bargaining framework and estimate a two-stage empirical model in which demand estimation informs the bargaining stage. We find that greater cross-category dominance, defined as the share of unit sales in non-focal categories, systematically increases negotiated shelf share, while larger retailers offset this effect. A counterfactual merger simulation between multi-category manufacturers shows that portfolio expansion reduces consumer welfare, with part of the loss stemming from reduced product variety. The results suggest that cross-category portfolio power can influence vertical negotiations beyond traditional price channels.
By Yonggeun Jung; University of Florida
Presented by: Yonggeun Jung, University of Florida
Discussant:   Rui Li, SUNY Albany
Resale Price Maintenance and Consumer Search Abstract We provide a novel pro-competitive rationale for resale price maintenance (RPM). We consider a model where some consumers are fully informed about downstream prices while other consumers are not. When an upstream manufacturer imposes a floor on downstream prices, this qualitatively changes downstream competition-influencing not just the level, but also the dispersion, of prices. The manufacturer optimally imposes a price floor which just eliminates all downstream price dispersion, and this leads to both higher (aggregate) consumer surplus and higher total welfare as compared to the case without RPM.
By Andrew Rhodes; Toulouse School of Economics    Yang Yang; School of Economics and management, Tsinghua University
Presented by: Yang Yang, School of Economics and management, Tsinghua University
Discussant:   David Gilo, Tel Aviv University
Timing Assumptions in Vertical Bargaining Abstract We develop an empirical framework to analyze vertical relationships with manufacturer–retailer bargaining. Our key innovation is the introduction of a novel Nash-in-Nash bargaining model that incorporates uncertainty in retailers’ pricing responses to wholesale prices. This model extends existing Nash-in-Nash frameworks by relaxing assumptions about the timing of wholesale and retail price setting. We show that our model can be micro- founded by a two-stage noncooperative game with delegated negotiations. We propose a two-step strategy that separably identifies bargaining and responsiveness parameters and implies a Generalized Method of Moments estimation procedure. A preliminary application to Dominick’s beer market indicates that the retailer is likely to be responsive in retail pricing, with estimates of the responsiveness parameter close to one.
By Hugo Molina; University of Paris-Saclay    Ao Wang; Warwick University
Presented by: Hugo Molina, University of Paris-Saclay
Discussant:   Unni Pillai, University at Albany, SUNY
Vertical Mergers, Customer Access, and Spillovers in Production Networks Abstract This paper analyses how vertical mergers affect non-integrating firms in the production networks of merging entities. Using novel data across a wide range of industries and countries, we document that, following a vertical merger, non-integrating suppliers experience a drop in revenues and in the number of customers, as well as in the probability of selling to the vertically integrating firms. Importantly, these effects are observed not only among rival suppliers -- consistent with classic theories of customer foreclosure-- but also among other non-integrating upstream firms. This suggests that vertical mergers can trigger substantial restructuring within firm-to-firm networks. We provide several pieces of evidence to explain the mechanisms behind these spillover effects. Furthermore, we show that the probability of acquiring new customers and the overall profitability of non-integrating firms decrease post-merger, pointing to stronger competitive pressure. These results have significant implications for competition authorities, highlighting the importance of considering the broader structure of firm-to-firm networks when developing theories of harm in the context of vertical mergers.
By Katharina Erhardt; University of Duesseldorf    Giulia Sabbadini; Düsseldorf Institute for Competition Economics    Joel Stiebale; DICE, Heinrich Heine University Düsseldorf
Presented by: Giulia Sabbadini, Düsseldorf Institute for Competition Economics
Discussant:   Michael Rubens, UCLA
Session 51: Information and Regulation April 12, 2026 8:00 to 10:00 Location: Salon C (3rd Floor)
Session Chair: Ting Liu, Stony Brook University
Collateral, Information and Welfare: Implications for Open Banking Abstract Data sharing (such as open banking initiatives) and improvements in data analytics enhance the capabilities of financial technology (fintech) companies and have the potential to reduce information asymmetries in credit markets. This is generally believed to alleviate adverse selection and improve welfare. Traditional lenders (such as banks), however, respond to increasing competition by offering more attractive products that involve costly collateral. We uncover a novel trade-off between improved information and destructive competition due to increased collateralization. For instance, we show that open banking or advances in data analytics may harm not only social welfare but also fintechs themselves. We also examine alternative institutional arrangements—such as the allocation of property rights and the creation of data markets—that can outperform open banking. Our results contribute to the ongoing policy debate on the welfare implications of open banking and data-sharing initiatives.
By Anastasios Dosis; ESSEC Business School
Presented by: Anastasios Dosis, ESSEC Business School
Discussant:   Andrew Kearns, Federal Trade Commission
Technology, Online Banks, and Credit Market Segmentation Abstract How does online bank development (digital-only depository institutions that originate loans without human intermediation) affect consumer credit market structure? Using loan-level data from Germany, we show that online banks cherry-pick low-risk borrowers, generating adverse selection for traditional banks. We develop and test a model in which online banks have lower costs but weaker screening because they rely solely on hard information. Online banks offer substantially lower rates to low-risk borrowers, but this advantage declines with credit risk, creating a crossover point beyond which traditional banks become more competitive. Using historical branch density as an instrument, we show that supply-driven screening differences contribute to this pattern. Extending the framework to fintech lenders reveals market segmentation: online banks serve the lowest-risk borrowers, traditional banks the medium-risk segment, and fintechs the highest-risk segment. Over time, the traditional online rate gap widens, consistent with deteriorating borrower pools at traditional banks. Our findings highlight that technological development in credit markets can generate important distributional consequences.
By Chiara Farronato; Harvard University    Rachel Nam; USI Lugano/Swiss Finance Institute    Loriana Pelizzon; Leibniz Institute for Financial Research
Presented by: Rachel Nam, USI Lugano/Swiss Finance Institute
Discussant:   Carlos Canon, Bank of England
Regulatory Capacity and Hard-to-Enforce Requirements Abstract In an asymmetric-regulation model as a simultaneous-move game, a firm can comply with two regulatory targets, and a regulator can audit the firm to verify compliance. Inspection by the regulator is imperfect, and it assesses the firm's compliance with the targets with different success probabilities. In equilibrium, the firm fully complies only if compliance costs are low. Otherwise, the firm always prioritizes the requirement that is easier to enforce with a higher success probability. Expanding regulatory capacity positively affects compliance with the easy-to-enforce target; however, a higher capacity for the regulator, allowing it to enforce both targets, can harm compliance with the hard-to-enforce target. The analysis highlights that higher regulatory capacity can backfire if policymakers care about hard-to-enforce requirements but fail to align the regulator’s objectives with these priorities.
By Jacopo Gambato; Universität Wien, ZEW Mannheim    Bernhard Ganglmair; University of Mannheim & ZEW Mannheim    Julia Krämer; Erasmus University Rotterdam
Presented by: Bernhard Ganglmair, University of Mannheim & ZEW Mannheim
Discussant:   Alison Ong, Harvard University
Optimal Emissions Permits in Oligopolistic Markets Abstract This paper studies cap-and-trade regulation in oligopolistic markets where firms compete in quantities and emissions are tied to output. We develop a Cournot model with heterogeneous marginal costs, a competitive permit market, and a regulator who chooses the total supply of tradable permits. The Coase Theorem holds in the sense that aggregate welfare depends only on the total number of permits, not on their allocation. Allowing for a general inverse demand function, we show that the optimal permit supply depends critically on demand curvature. With linear demand, the regulator’s policy coincides with the benchmark condition equating price to average social cost. With nonlinear demand, welfare effects arise through both total output and the reallocation of production across firms, leading convex demand to justify a looser cap and concave demand a tighter one. Moreover, the optimal permit supply increases (decreases) with industry size and cost dispersion when demand is convex (concave).
By Ting Liu; Stony Brook University
Presented by: Ting Liu, Stony Brook University
Discussant:   Ignacia Mercadal, University of Florida
Session 52: Market Power: Measurement and Consequences April 12, 2026 8:00 to 10:00 Location: Georges (4th Floor)
Session Chair: Yingjun Su, Binghamton University
Market power and Profitability in the EU over 2007-2022 Abstract This paper examines the evolution of market power in European firms between 2007 and 2022 using firm-level data from the Orbis database. We apply two complementary estimation frameworks: the De Loecker et al. (2020) method (DLEU), which infers markups under the assumption of fully variable intermediate inputs, and the Abraham et al. (2024) approach (ABKR), which jointly estimates price–cost margins, fixed costs, and excess profits. Both methods indicate that average markups in the EU have been broadly stable or only moderately rising over the period, contrasting with the pronounced upward trends documented for the United States. We find that within-firm dynamics, rather than reallocation or entry–exit, drive the modest aggregate increases. Firm-level heterogeneity is substantial: large, listed, and high-tech manufacturing firms exhibit higher DLEU markups but lower ABKR price–cost margins due to lower fixed cost ratios, resulting in higher net profitability. Ownership plays a smaller role in market power differences than firm size.
By WALEED HASSAN; KU Leuven    Yannick Bormans; KU Leuven    Elizaveta Archanskaia; EU Comission Brussels    Anna THUM-THYSEN; EU Comission Brussels    Maria Garrone; EU Comission Brussels
Presented by: WALEED HASSAN, KU Leuven
Discussant:   Zhehao Hu, University of North Carolina at Chapel Hill
Risk and Return in Asset Demand Systems Abstract We develop a characteristic-based asset demand model in which cross-asset risk-return trade-offs vary with asset characteristics. The model relaxes the uniform substitution structure of the multinomial logit (MNL), accommodates large price elasticities, and enables recovery of investor-specific primitives, including alphas and factor loadings, from structural demand estimates. Applied to U.S. institutional equity holdings from 2000 to 2022, the model reveals meaningful deviations from MNL substitution patterns, particularly along the market equity dimension. The estimated average own-price elasticity is 77 percent higher than under the MNL, driven largely by investors whose portfolios imply cross-asset complementarity. Nonetheless, both elasticity estimates are substantially lower than those implied by CAPM calibrations. The model also uncovers heterogeneity in investor alphas: hedge funds earn near-zero alphas, while brokers earn up to five basis points annually.
By Ozan Akbas; University of Warwick    Ao Wang; University of Warwick
Presented by: Ozan Akbas, University of Warwick
Discussant:   Johannes Kandelhardt,
The Geography of Market Power: Evidence from the Chinese Steel Industry Abstract This paper examines how the geographic distribution of supply and demand shapes market power in the Chinese steel industry. Drawing on novel data, we develop and estimate an equilibrium model that accommodates spatial demand variations and rich firm heterogeneity—encompassing differences in location, product quality, production coefficients, and cost efficiencies. Using this framework, we simulate the impact of shifts in downstream demand and evaluate the welfare implications of mergers and acquisitions under various market frictions—an issue central to China’s industrial policy.
By Loren Brandt; University of Toronto    Feitao Jiang; Chinese Academy of Social Sciences    Yao Luo; University of Toronto    Yingjun Su; Binghamton University
Presented by: Yingjun Su, Binghamton University
Discussant:   Emily Cook, Texas A&M University
Session 53: Production Function and Productivity Estimation April 12, 2026 8:00 to 10:00 Location: Aegean (3rd Floor)
Session Chair: Giulio Gottardo, University of Oxford
Robust Production Function Estimation when there is Market Power Abstract The production function is an engineering relationship, but recent estimators use firm’s optimal choices that depend on market power. Researchers often become puzzled: the estimator dynamic panel (DP), which is robust to market power because it does not use any FOC, often produces unsatisfactory outcomes; the estimators known as OP/LP, which are deemed inconsistent in the presence of market power, typically improve. We prove that the coincidence of DP and OP/LP, except by sampling error, is a necessary condition for consistency, and show how the improvements relate to the production function specification. We derive a novel estimator, robust to arbitrary forms of market power, based on a version of OP/LP that proxies for MC. Using this estimator, we propose a test for market power and a test for the specification, the latter based on the smaller set of assumptions used by DP.
By Jordi Jaumandreu; Boston University
Presented by: Jordi Jaumandreu, Boston University
Discussant:   Dongni Zhu, Shanghai University of Finance and Economics
A Theory of Inefficiency Distributions Abstract We develop a simple theory of the distribution of technical inefficiency. In our model, the joint forces of innovation, imitation and the refined use of existing technologies determine the shape and spread of inefficiency in the cross section. The theory yields clear patterns. When innovation arrives discontinuously, the inefficiency distribution transitions through a zero-inflated truncated normal shape and ultimately collapses to a point mass. Imitation and convergence mainly change the overall dispersion but rarely the qualitative shape class. However, when innovation is continuous, then the inefficiency distribution converges to a (left-truncated) normal form. These results matter for empirical work: in stochastic frontier analysis the assumed inefficiency distribution is not an innocuous convenience but an identifying choice that influences estimated technology, returns to scale, and technical change. We provide diagnostics and a theory-based menu of distributions that help researchers avoid misspecification.
By Jaap Bos; Maastricht University School of Business and Economics    Stefan Weiland; Maastricht University
Presented by: Stefan Weiland, Maastricht University
Discussant:   Javier Miranda, IWH/Friedrick-Schiller University
Production Functions with Noisy Data: A Flexible Cost Share Approach Abstract I introduce a new method to estimate heterogeneous output elasticities, markups, and revenue productivity using standard firm-level data. The approach avoids common assumptions on demand or productivity dynamics required by proxy methods. It exploits firms' optimizing behavior, and the fact that output elasticities are proportional to cost shares adjusted for frictions. I develop a two-stage semi-parametric procedure to address the key challenge: measurement error in unobserved capital costs (e.g., from adjustment frictions). The first stage non-parametrically purges this error from noisy proxies; the second uses these purged costs to estimate firm-specific elasticities, markups, and productivity. Monte Carlo simulations confirm the estimator's accuracy. Applying the method to Compustat, I find substantive differences with the past literature on the distribution of output elasticities, markups, and revenue productivity, with the treatment of Selling, General, and Administrative (SGA) expenses playing a critical role. When SGA is treated as a sunk cost, markup dispersion drives revenue productivity dispersion, and smaller firms exhibit higher markups and productivity. These results are reversed if SGA is treated as a production input.
By Giulio Gottardo; University of Oxford
Presented by: Giulio Gottardo, University of Oxford
Discussant:   Jordi Jaumandreu, Boston University
Session 54: Platform design: Pricing April 12, 2026 8:00 to 10:00 Location: Atlantic 2 (3rd Floor)
Session Chair: Hanna Halaburda, New York University
Reliability and Pricing in Cloud Computing Abstract To match volatile demand with fixed capacity, cloud computing platforms employ tiered reliability—offering discounted spot compute services from which users can be “evicted” (i.e., interrupted) with little warning when capacity tightens. We study this market design using proprietary data from a major cloud platform, exploiting a price experiment and the quasi-random nature of evictions to estimate a structural model. The price elasticity of demand is -0.5, and evictions persistently reduce usage by 40%, indicating a strong revealed preference for reliability. More usage increases eviction rates, consistent with congestion. We interpret these facts through a model where heterogeneous users choose the compute reliability for each workload, while learning about eviction risk through experience. On the supply side, evictions arise endogenously given fixed capacity. Preliminary counterfactual results suggest that tiered reliability provides Pareto gains relative to simply allowing the market to clear through congestion.
By James Brand; Microsoft    Juan Camilo Castillo; University of Pennsylvania    Chinmay Lohani; University of Pennsylvania    Leon Musolff; Wharton School of the University of Penn
Presented by: Leon Musolff, Wharton School of the University of Penn
Discussant:   Wenxuan Xu, National University of Singapore Business School
``Walled Gardens'': Exclusive Complementarity and Consumer Lock-in in Multiproduct Oligopoly Markets Abstract This paper studies how exclusive complementarity affects price competition among multiproduct firms operating in two sequentially available markets: a mature product market and an emerging product market. We show that exclusive complementarity reduces competition in the emerging market under log-concave consumer preferences, regardless of the number of firms. In the mature market, however, competition is unambiguously intensified only in duopoly; when the number of firms exceeds two, competition in the mature market can be relaxed. When exclusive complementarity generates true social benefits, we establish conditions under which it may harm consumer welfare. By characterizing these competitive effects, this paper provides qualitative guidelines for when ``walled gardens'' should be subject to antitrust scrutiny.
By Shiyun Xia; Tianjin University    Bei Yang
Presented by: Shiyun Xia, Tianjin University
Discussant:   Kei Ikegami, University of Tokyo
Decentralization and the Law of the Jungle: An Empirical Investigation of Ethereum’s Market Mechanism Abstract Blockchain technology aims to disintermediate traditional platforms by replacing centralized governance with decentralized market mechanisms. However, this shift introduces a tradeoff: while decentralization removes the platform’s ability to capture value, it also eliminates platform-wide mechanisms—such as subsidies, curation, or pricing strategies—that can support long-term platform performance. Using Ethereum as a case study, this paper examines how its market-based transaction validation system affects the allocation of transaction capacity and shapes platform dynamics. Specifically, we estimate demand elasticities across more than 1,500 decentralized applications (dApps) and evaluate the effects of transaction fees on various application categories. To address endogeneity concerns, we leverage Ethereum’s “difficulty bomb,” a protocol feature that periodically reduces transaction throughput, as an instrumental variable. This method provides exogenous variation in transaction costs, allowing us to identify differences in demand elasticities across dApps. Our analysis shows that during periods of high fees, low-elasticity transactions—such as those associated with exploitative miner extractable value (MEV) activities—tend to dominate the network. This crowding out effect disproportionately impacts fee-sensitive categories of applications—such as gaming, social, and utility dApps—reducing their viability and shifting transaction capacity toward applications that prioritize short-term profitability over long-term platform performance. This dynamic helps explain why, despite repeated promises to disintermediate platforms like Uber, decentralized platforms may struggle to support fee-sensitive but socially valuable use cases.
By Hanna Halaburda; New York University    Daniel Obermeier; NOVA School of Business and Economics
Presented by: Hanna Halaburda, New York University
Discussant:   Byoungmin Yu, University of Florida
Health Insurance Contracts and Providers Networks Abstract This paper examines the design of optimal health insurance menus amid varying provider quality for various diseases, focusing on monopolistic and competing insurers. I theoretically show that a monopolist offers broader (more efficient) networks at higher premiums, while competing insurers prefer narrower networks to limit competition. Using a rich dataset on Chile’s private health insurance system, I develop a version of the theory model suitable for estimation. Focusing on the allocation of consumers to providers, I found that consumers’ choices closely resemble the most efficient allocation due to regulations mandating insurers to cover at least what the public insurer does. In light of the theoretical results, this resemblance would not persist in an unregulated environment, underscoring the crucial role of regulation in these markets.
By Nicolas Bozzo Galleguillos; Keystone
Presented by: Nicolas Bozzo Galleguillos, Keystone
Discussant:   Federico Navarra, Charles River Associates
Session 55: Environmental Policy Design and Market Outcomes April 12, 2026 8:00 to 10:00 Location: Salon A (3rd Floor)
Session Chair: R. Andrew Butters, Indiana University
BEV vs. HEV purchase subsidies: Impacts on greenhouse-gas emissions in South Korea Abstract We evaluate and compare greenhouse gas (GHG) emissions under the current battery electric vehicle (BEV)-focused purchase subsidy policy and an alternative hybrid electric vehicle (HEV)-focused policy in South Korea’s new passenger vehicle market. Our demand estimation reveals that while higher-mileage consumers prefer fuel-efficient vehicles, HEVs exhibit greater substitutability with traditional internal combustion engine (ICE) vehicles than BEVs do. Counterfactual analysis shows, first, that HEV subsidies expand the new vehicle market to a greater extent than BEV subsidies. Second, a budget-neutral reallocation of the BEV subsidy to HEVs would reduce total cumulative GHG emissions by more than 600 thousand tonnes of CO2e during the 2020–2023 period. Third, ceteris paribus, a decrease of 40% or more in the country’s current GHG emission factor (453 g CO2e/kWh) through a cleaner electricity generation mix is required for the BEV subsidy policy to match the environmental efficiency of the HEV policy. Our findings suggest that an HEV-focused subsidy may not only facilitate a smoother transition to electrification for countries where legacy automakers play a vital role in the economy but also serve as a valuable tool to attain meaningful emission reductions in the interim.
By Youngjin Hong; University of Michigan-Ann Arbor    In Kyung Kim; Sogang University    Frank Verboven; KU Leuven
Presented by: Youngjin Hong, University of Michigan-Ann Arbor
Discussant:   Hideo Konishi, Boston College
Binding Emission Standards in the European Car Market Abstract The European Union (EU) has significantly tightened fleet emissions standards for new passenger vehicles. Consequently, now most firms are directly affected by the regulation. We adapt the standard differentiated product model to account for the emission standard, which introduces a kink in the profit function. Our method smoothes the kink by introducing uncertainty, allowing us to estimate the marginal cost of production. We demonstrate that tighter standards led to substantial price changes, varying across manufacturers, with many vehicles being sold at a price below their marginal cost. Counterfactuals show that the policy reduced overall emissions at the expense of overall consumer welfare, with substantial heterogeneity across EU countries.
By Thomas Wiedenhofer; University of Mannheim
Presented by: Thomas Wiedenhofer, University of Mannheim
Discussant:   Youngjin Hong, University of Michigan-Ann Arbor
From Pumps to Plugs: Value of Time and Charging Policies Abstract Designing an electric vehicle (EV) charging infrastructure requires understanding how drivers value charging speed, price, and convenience. We provide the first field-based estimates of drivers' value of time (VOT) in a fast-charging environment, using rich, high-frequency vehicle usage data from Shanghai in 2024. We estimate a random-coefficient discrete-choice model of drivers' charging decisions. The estimated mean value of time is 7.2 Yuan per hour (29% of the average wage), implying lifetime charging time costs of about 1.8% of the vehicle price. Counterfactual analyses reveal that the city's charging infrastructure reduces the lifetime charge-time cost by 4% of vehicle price and accounts for 10.7% of new adoptions. We find positive 10-year ROIs for station speed investments, and the importance of coordinating investment on both station density and speed to achieve an efficient charging infrastructure.
By Takeaki Sunada; University of Rochester    Tianli Xia; University of Rochester
Presented by: Takeaki Sunada, University of Rochester
Discussant:   Thomas Wiedenhofer, University of Mannheim
Pollution Pricing in Equilibrium: Production, Reallocation, and Aggregate Impacts Abstract We develop a framework for quantifying the equilibrium effects of pollution pricing using commonly available data on firm financials and emissions. We use the framework to estimate marginal abatement cost curves at the firm, industry, and economy level, which provide a clear decomposition of aggregate emissions reductions into within-firm abatement, reallocation across firms, and technological change. Applying the framework to the European Union Emissions Trading System (2005--2021), we find that firm-level abatement costs are steep in the short run: holding technology fixed, firms cannot easily reduce emissions without cutting output. Yet industry-level and economy-wide abatement cost curves are substantially flatter because production is reallocated toward cleaner firms. We document a 65 percent improvement in clean technology over our sample period, shifting abatement cost curves downward. Together, reallocation and technological change reconcile steep firm-level abatement costs, large aggregate emissions reductions, and small effects on output and employment.
By R. Andrew Butters; Indiana University    Jackson Dorsey; University of Texas at Austin    Ivan Rudik; Cornell University
Presented by: R. Andrew Butters, Indiana University
Discussant:   Kenneth Gillingham, Yale University
Session 56: Pharmaceutical Innovation, Pricing, and Global Entry April 12, 2026 8:00 to 10:00 Location: Atlantic 3 (3rd Floor)
Session Chair: Gaurab Aryal, Boston University
Antibiotic Resistance, Drug Prices, and Entry Abstract Antibiotics lose their power to kill microbes through excessive use, commonly known as antibiotic resistance. We analyze how market structure affects antibiotic resistance and consumer welfare. Antibiotic resistance is modeled by future cost increase from current use. The first best considers cost externality due to current consumption. Low prices under competition lead to high consumption and resistance, because consumers ignore externality. Competitive managed care plans use rationing contracts, which mitigate resistance by restricting use. Such contracts must offer benefits to compensate consumers for accepting rationing, so some market failure remains. A drug monopolist does fully internalize the resistance damage. Even though the externality is internalized, the monopolist's price remains high. We derive necessary and sufficient conditions for consumer surplus to be higher under monopoly than competition. A monopoly drug plan eliminates resistance inefficiency altogether, but fully extracts consumer surplus. If there is a potential entrant, an incumbent may consider entry deterrence and accommodation strategies. For entry deterrence, the incumbent may reduce current sales. If the reduction is so much that its future cost is sufficiently low, an entrant cannot recover fixed cost post entry. Entry deterrence reduces antibiotic resistance. For entry accommodation, two considerations are important. First, the incumbent will share the market with the entrant. The incumbent has reduced incentives to internalize the resistance effect, so raises production and exacerbates resistance. Second, the incumbent makes more profit when its future cost is low, so reduces production and mitigates resistance. An entry accommodation strategy's effect on antibiotic resistance is ambiguous.
By Miaoqing Jia; Weill Cornell Medical College    Ching-to Albert Ma; Boston University
Presented by: Ching-to Albert Ma, Boston University
Discussant:   Hongyuan Xia, Cornell University
Disclosure and the Pace of Drug Development Abstract Policies that mandate disclosure of innovative project outcomes aim to increase innovation by limiting wasteful duplicative innovation. Yet, such policies change not only the ex-post information environment but also firms' ex-ante innovation incentives. Firms may slow down their own innovation efforts in anticipation of increased disclosure by others. We examine the innovation-related impacts of the 2017 FDA Final Rule amendment, which mandates disclosure of clinical trial results for pharmaceutical firms. We show that the policy hastened and increased disclosure of results for clinical trials post-completion, but also increased the time to completion of clinical trials, the time between early phases of clinical trials, and delays in development-related investments. We provide evidence consistent with mandated disclosure leading firms to wait to learn from their competitors. Our results suggest that mandating disclosure may slow innovation when there is value to waiting.
By Colleen Cunningham; Eccles School of Business, University of    Florian Ederer; Boston University    Charles Hodgson; Yale University    Zhichun Wang; Yale University
Presented by: Florian Ederer, Boston University
Discussant:   Federico Ciliberto, U Virginia
Expected Insurance Coverage and Pharmaceutical Innovation: Evidence from China's National Drug Price Negotiation Policy Abstract How can developing countries foster pharmaceutical innovation when drug price negotiations risk eroding firms’ incentives? China’s National Health Insurance Drug Price Negotiation addresses this challenge by pairing substantial price cuts with expanded insurance coverage for innovative drugs, enlarging the effective market size. We examine its impact on innovation using a difference-in-differences design that compares clinical trials of new drugs (treated) with those of vaccines (unaffected). The policy increases the number of trials by 0.56 per disease per year. The effects are especially pronounced for more novel innovations and similar for domestic and foreign firms. We also find that the policy induces more R&D pharmaceutical firms, especially those of small size, to enter the market. Finally, we document increased collaboration and outsourcing across firms.
By Chenyuan Liu; Tsinghua University    Yi Lu; Tsinghua University    Ronghe Sun; Tsinghua University    Wanyu Yang; Dongbei University of Finance and Economics
Presented by: Chenyuan Liu, Tsinghua University
Discussant:   Lucy Xiaolu Wang, University of Massachusetts Amherst; Max Planck Institute for Innovation and Competition
Valuing Pharmaceutical Drug Innovations Abstract We propose a methodology to estimate the market value of pharmaceutical drugs. Our approach combines an event study with a model of discounted cash flows and uses stock market responses to drug development announcements to infer the values. We estimate that, on average, a successful drug is valued at $1.62 billion, and its value at the discovery stage is $64.3 million, with substantial heterogeneity across major diseases. Leveraging these estimates, we also determine the average drug development costs at various stages. Furthermore, we explore applying our estimates to design policies that support drug development through drug buyouts and cost-sharing agreements.
By Gaurab Aryal; Boston University    Federico Ciliberto; U Virginia    Leland Farmer; University of Virginia    Ekaterina Khmelnitskaya; University of British Columbia
Presented by: Gaurab Aryal, Boston University
Discussant:   Rebekah Dix, Yale
Session 57: Auctions: Entry, Scale, and Design April 12, 2026 8:00 to 10:00 Location: Caspian (3rd Floor)
Session Chair: David Genesove, Hebrew University of Jerusalem
Entry Deterrence in Procurement Auctions Abstract If a firm can make a credible announcement of its intent to enter a market, it may be able to deter rival firms from entering. We study procurement auctions conducted by Montana Department of Transportation, where a designated online Q&A forum serves as an entry disclosure device. We specify and estimate a model of procurement auctions with costly entry, in which firms have the option to disclose entry during a period before the bids are submitted. We find that disclosure deters entry from other firms, but leads remaining entrants to bid more aggressively. Overall, disclosure is beneficial for a firm if they can disclose at an early period. The availability of disclosure device decreases the auctioneer’s payment by 6.3%, while increasing the winner’s construction costs by 4.5% and decreasing the total entry costs by 11.1%.
By Yuki Ito; Indiana University
Presented by: Yuki Ito, Indiana University
Discussant:   Ruli Xiao, Indiana University
The Effects of Pre-Auction Nominations: Evidence from Oil & Gas Lease Sales Abstract Auctions receiving zero bids are not uncommon when sellers face uncertain demand over large inventory. While standard mechanisms like reserve prices or entry fees help extract revenue from interested bidders, they fail to solve the extensive margin problem of which items to offer, leading to costly failed auctions. Pre-auction nomination, where potential buyers pay a small fee to propose what should be auctioned, is a popular solution in the mineral rights market, but is not well studied. This paper shows that nomination serves dual purposes when private value of bidders are positively correlated and both the seller and the bidders do not know their private value until costly exploration: demand screening for sellers and quality screening for buyers, while creating incentives for bidders to free-ride on tracts others have identified as valuable instead of exploring and nominating themselves. Using a two-stage game with Poisson arrivals, I characterize equilibrium exploration and nomination decisions and structurally estimate exploration costs using data from North Dakota oil and gas auctions (2017--2018). Counterfactual analysis reveals that raising the nomination threshold beyond the current single-nomination level produces coordination failure: no firms nominate and no auctions are triggered. Varying the nomination fee reveals higher fees increase conditional bid quality through selection effect while leaving auction trigger rates nearly unchanged. Comparing nomination fee with reserve price, both screening bidder quality but at different stages, reveals taxing later during allocation stage generates higher auction trigger rate but lower expected profit per triggered auction.
By Ji Hyun (Kiara) Kim; Washington University in St. Louis
Presented by: Ji Hyun (Kiara) Kim, Washington University in St. Louis
Discussant:   Miguel Alcobendas, Yahoo Research
Allocating Resources under Strategic Misrepresentation Abstract We study how to allocate resources to participants who can strategically misrepresent their deservingness at a cost. A principal assigns item(s) (or money) among multiple agents on the basis of their costly signals. Each agent's signal reflects his private type, absence of misrepresentation, but can be inflated above his true type at a cost. The principal is a social planner that aims to maximize the weighted average of matching efficiency and a utilitarian objective. Strategic misrepresentation introduces novel incentive-compatibility constraints, under which we characterize the optimal mechanism. We apply our characterization to two kinds of markets, distinguished by resource scarcity, and show that the principal strictly benefits from randomizing the allocations based on costly signals when the population of participants is large enough. Interestingly, in large markets with scarce resources, the format of the optimal mechanism converges to a winner-takes-all contest; however, there is a non-diminishing value of randomizing allocations to middle types as the population of participants grows.
By Yingkai Li; National University of Singapore    Xiaoyun Qiu; Dartmouth College
Presented by: Xiaoyun Qiu, Dartmouth College
Discussant:   David Genesove, Hebrew University of Jerusalem
Auction Identification with Unobserved Rejected Offers Abstract We study identification of buyer and seller primitives in second-price auctions when sellers may reject the highest bid but only successful transactions are observed. In our framework a sale occurs only if the top bidder value exceeds a seller-chosen threshold that is unobserved by the econometrician. We first show that point identification using variation in the number of bidders is often achievable but may fail: in the two-point case the model has set-identified data regions in which price distributions can be rationalised by two pairs of value and threshold distributions. For general supports we reduce identification to a fixed point in the ratio of sale probabilities across bidder counts; uniqueness yields point identification. We then show that upper-tail support separation between the buyer value and threshold distributions is a sufficient condition for point identification. A conduct assumption maps thresholds into seller values; absent conduct we derive bounds based on assumptions similar to \cite{haile_tamer_2003}. We quantify biases from ignoring rejected offers, at the population level, for evaluation of counterfactual mergers and entry/exit episodes, and in Monte Carlo experiments, for the distributions themselves. The Monte Carlo experiments validate our approach.
By David Genesove; Hebrew University of Jerusalem
Presented by: David Genesove, Hebrew University of Jerusalem
Discussant:   Xiaoyun Qiu, Dartmouth College
Session 58: Digital Platforms, Market Dynamics, and User Behavior April 12, 2026 8:00 to 10:00 Location: Mediterranean (3rd Floor)
Session Chair: Andrea Mantovani, TSE
Online Travel Agencies and beyond: The role of sales channels for hotels and consumers Abstract This paper examines the impact of online travel agencies on hotel pricing strategies, consumer behavior, and market dynamics within the hospitality sector. Using channel-level proprietary data from major hotel chains across eight European countries, we adopt a structural approach to estimate demand and supply, and simulate policy counterfactuals. Our findings reveal that online travel agencies expand demand without exerting significant competitive pressure on market prices, due to limited substitutability between sales channels. We assess potential regulatory interventions. A fee cap would benefit hotels in the sample and consumers, while hurting outside competitors. Provisions that facilitate direct communication between hotels and customers, in the spirit of the disintermediation allowed by the DMA, would be successful in shifting some sales from the platform to the hotel website while reducing margins overall.
By Andrea Mantovani; Toulouse Business School    Laura Lasio; European Commission    Jack (Peiyao) Ma; University of Oxford    Carlo Reggiani; University of Manchester    Nestor Duch-Brown; Joint Research Centre (JRC), Seville
Presented by: Andrea Mantovani, TSE
Discussant:   Yanyou Chen, University of Toronto
How Do Past Privacy Choices Shape the Future? Abstract Consumers face frequent and consecutive digital privacy choices, but each choice is not necessarily independent. This paper demonstrates that past privacy choices affect consumers’ current privacy choices. Such state-dependent choices suggest that privacy choices can have externalities within a platform in which one app's data requests can affect the ability of other apps to collect data. Specifically, I use an individual-level consumer panel to investigate data consent decisions by consumers on Alipay, a major digital platform that connects users and third-party apps. Leveraging a natural experiment that encourages users to accept data requests, I find that the probability of rejecting the next request declines 15%. This effect decays over time; it is larger when preferences for the app at that moment are relatively weak, categorized by large language models (LLMs); and for users whose privacy preferences are relatively weak. The effect does not differ by whether the specific data requested in consecutive data consent decisions is the same. Overall, these results suggest that the externalities arising from state-dependent data consent choices are temporary. Nevertheless, this temporary effect incentivizes platforms to encourage apps to provide consumer-friendly data request designs.
By Verina Que; Nanyang Technological University
Presented by: Verina Que, Nanyang Technological University
Discussant:   Mark Jamison, University of Florida
Empowering Inclusive Work Abstract Can AI improve workplace outcomes for workers with disabilities? We examine the relative performance of hearing-impaired workers on one of China's largest food-delivery platforms. Pre-AI, disabled workers are slightly slower than non-disabled workers and have worse customer ratings, although they supply more hours to the platform and are much less likely to quit. Midway through our data, the platform suddenly introduces an AI-based intelligent outbound call system designed to improve communication for hearing-impaired workers. AI significantly increases the productivity of disabled workers, especially on tasks involving interaction with customers; reduces late deliveries and negative customer reviews; and increases labor supply and retention. AI eliminates over half of the disability hourly pay gap and substantially increases the profitability of disabled workers for the platform.
By Yanyou Chen; University of Toronto    Mitchell Hoffman; University of California, Santa Barbara    Huilan Xu; Zhejiang University    Zhe Yuan; Zhejiang University
Presented by: Yanyou Chen, University of Toronto
Discussant:   Verina Que, Nanyang Technological University
To Bid or Not to Bid? Entry, Bid Shading, and Bias in Sponsored Search Auctions Abstract We examine bidder participation in sponsored search auctions when platform rankings include both organic and sponsored positions. We develop a model in which bidders decide whether and how aggressively to bid for sponsored positions. The platform's ex post revenue maximizing ranking induces bid shading. The platform can increase its revenue by biasing rankings toward sponsored positions, even at the cost of lower-quality search results. We test the model using data from Amazon’s search results and find evidence consistent with the model's predictions. Our findings highlight a fundamental tension in platform design between revenue maximization, search results' quality, and bidder participation.
By Eeva Mauring; University of Bergen    Anastasiia Parakhoniak; Durham University    Cole Williams; University of Nebraska-Lincoln
Presented by: Cole Williams, University of Nebraska-Lincoln
Discussant:   Andrea Mantovani, TSE
Session 59: Merger Thresholds April 12, 2026 8:00 to 10:00 Location: Atlantic 1 (3rd Floor)
Session Chair: Margaret Loudermilk, Charles River Associates
Merger Effects and Efficiencies Under Uncertainty: Evidence from Supreme Court Decisions on Banking Abstract We study four seminal United States Supreme Court decisions from the 1960s on bank mergers. We interpret these decisions in a modern economic framework and find that merger simulations support enjoining these mergers, but only under certain assumptions about efficiencies and the distribution of consumer welfare. We augment the framework to incorporate uncertainty in the realization of efficiencies and to allow for differential consumer effects. We derive the critical probability of cognizability needed for efficiencies to eliminate expected harm to (1) aggregate consumer welfare, (2) total welfare, (3) subsets of consumers by income, and (4) all consumer income groups.
By David Benson; Federal Reserve Board    Aaron Garner; University of Pennsylvania    Gloria Sheu; Federal Reserve Board of Governors    Charles Taragin; Federal Reserve Board of Governors
Presented by: Gloria Sheu, Federal Reserve Board of Governors
Discussant:   Nicolas Schutz, University of Mannheim
Merger Screening Thresholds under Model Uncertainty Abstract Merger enforcement often relies on concentration screens to identify potentially anticompetitive deals. We show that screening thresholds are not conduct-neutral: alternative models of competition will translate identical structural changes into significantly different welfare effects. Estimating Cournot, Bertrand, and auction models for 929 U.S. bank mergers, we find that post-merger price effects under Cournot conduct that are typically only half as large as effects estimated with Bertrand or auctions. This model uncertainty rationalizes divergent thresholds for the same legal standard: permissive screens implicitly assume Cournot, while stricter screens guard against localized auction-style competition. We show that a tiered screening framework that combines safe-harbor and stop-gap thresholds with secondary concentration indicia improves policy robustness to uncertain conduct.
By David Benson; Federal Reserve Board    Margaret Loudermilk; Charles River Associates    Charles Taragin; Federal Reserve Board of Governors
Presented by: Margaret Loudermilk, Charles River Associates
Discussant:   Oscar Jara, Norwegian School of Economics
Interaction-Adjusted Concentration: A Quadratic Generalization of the HHI Abstract Conventional concentration measures like the Herfindahl-Hirschman Index (HHI) assume product homogeneity, failing to distinguish markets with varying degrees of substitutability and complementarity. We axiomatically develop a quadratic concentration index that explicitly accounts for pairwise product interactions and generalizes the HHI. A key insight from applying the index to differentiated-product markets is its implication for surplus distribution: for substitutes, higher concentration correlates with a higher producers’ share—mirroring the classic HHI result in homogeneous-goods settings—but the relationship reverses for complements. This suggests a stricter scrutiny for substitute-heavy markets and greater leniency for complementary ones. We further develop a parametric family of quadratic welfare indices that coincides with the weighted total surplus, allowing tailored welfare standards. Applying to merger review, these indices suggest a new screening rule: output-increasing cases warrant minimal review; when output may fall, screening tightens beyond current HHI-based guidelines.
By Yangyi Deng; The Chinese University of Hong Kong    Chiu Yu Ko; The Chinese University of Hong Kong
Presented by: Yangyi Deng, The Chinese University of Hong Kong
Discussant:   Kensuke Kubo, Keio University
Session 60: Trade and IO (II) April 12, 2026 8:00 to 10:00 Location: Brewster (4th Floor)
Session Chair: Jonathan Elliott, Johns Hopkins University
Aggregate Outcomes of Nonlinear Prices in Supply Chains Abstract We study the welfare implications of nonlinear pricing in supply chains. Using population-level firm-to-firm transactions from Chile, we find indicative evidence that sellers engage in quantity-dependent and buyer-specific pricing strategies. We develop a general equilibrium model where firms pay and charge nonlinear prices. Under standard assumptions, we show that the optimal pricing scheme takes the form of a two-part tariff—comprising a flat fee and a marginal price—consistent with the price schedules observed in the data. Nonlinear pricing increases output per firm but distorts firm entry because flat fees redistribute profits unevenly across firms. Quantifying the model, we find that welfare under nonlinear prices reaches about 75\% of the efficient benchmark. In a counterfactual policy that bans all price discrimination—constraining firms to uniform pricing, a single, quantity-invariant price for all buyers—welfare falls to about 49\% of the efficient benchmark. Firms constrained to uniform pricing raise marginal prices that compound along supply chains, amplifying deadweight losses through markup accumulation. When interpreting the same data as uniform pricing, rather than nonlinear pricing, measured welfare is about 57\% of the efficient benchmark. These results indicate that prohibiting price discrimination can be welfare-reducing and that the measured aggregate welfare impact of market power in supply chains depends meaningfully on the extent to which firms use nonlinear pricing.
By Luca Lorenzini; UCLA Anderson    Antonio Martner; UCLA
Presented by: Luca Lorenzini, UCLA Anderson
Discussant:   Yulu Wang, Binghamton University
Input and output market power in the presence of fixed costs Abstract This paper introduces a unified framework to jointly estimate market power in input and output markets in the presence of fixed costs and a scale parameter. By combining publicly available data on (1) the primal and dual revenue-based and cost-based Solow residuals and (2) firms’ capital components, we are able to simultaneously estimate price-cost margins, the share of fixity for each input (i.e. capital, labor and intermediate inputs) and mark-downs in the labor and intermediate input market. Using a panel of Belgian firms (2000-2022), we show that output market profitability (3.7% of operating revenue) adds to labor market profitability (0.2%) and intermediate input market profitability (0.2%), however, this hides substantially sectoral heterogeneity in magnitude as well as in sign. For example, labor market profitability ranges from monopsony power in Construction (+3.0%) to rent sharing in Media & IT (-1.0%), and respectively reinforces and offsets output market profitability. Not accounting for input market power would lead to a bias in the estimation and interpretation of output market power.
By Yannick Bormans; KU Leuven
Presented by: Yannick Bormans, KU Leuven
Discussant:   Nima Nematian, University of Arizona
Critical Minerals, Geopolitics, and the Green Transition Abstract The green energy transition will be fueled by the mining and processing of lithium, nickel, and cobalt, which are critical for the production of advanced batteries. These minerals are concentrated geographically but traded globally. We study the geopolitical implications of industrial policy in key mining countries, and we discuss the consequences for green technology adoption worldwide. We highlight how upstream resource concentration and downstream technology choices give rise to supply chain vulnerability, policy spillovers across mineral markets, and the potential for mineral cartels.
By Tomas Dominguez-Iino; University of Chicago    Jonathan Elliott; Johns Hopkins University    Allan Hsiao; Stanford University
Presented by: Jonathan Elliott, Johns Hopkins University
Discussant:   Jevgenijs Steinbuks, The World Bank
Estimating Willingness-to-Pay for Emerging Technologies: A Study of Hydrogen Demand in the Ammonia Industry Abstract This study introduces a novel method to quantitatively assess the willingness to pay (WTP) for emerging technologies, such as hydrogen, as substitutes for fossil fuels in industrial production. A three-step framework is developed to derive the WTP function based on industrial competition and market entry theory, relying exclusively on pre-entry market information. First, a system of equations is specified linking domestic consumption, production, and prices to fossil input prices, which proxy marginal production costs. Second, market equilibrium parameters required for numerical WTP estimation are empirically estimated using industry-level data. Third, an industrial competition model incorporating entry by producers adopting new technology is constructed, allowing WTP to be expressed as a function of conventional input costs, operational efficiency, and demand conditions. The framework is applied to hydrogen use in ammonia production, using consumption and trade data from 2000–2024 for sixteen major fertilizer-producing countries across four regions. Results highlight substantial cross-country heterogeneity, a binding hydrogen price threshold for large-scale adoption, and the limited effectiveness of carbon policies in accelerating hydrogen uptake.
By Svetlana Ikonnikova; Technical University of Munich    Jevgenijs Steinbuks; The World Bank
Presented by: Jevgenijs Steinbuks, The World Bank
Discussant:   Felix Montag, NYU Stern School of Business
Session 61: Search, Platforms, and Consumer Learning April 12, 2026 10:15 to 12:15 Location: Caspian (3rd Floor)
Session Chair: Marcel Preuss, Cornell University
The Informational Content of Consumer Choice in Differentiated Product Markets Abstract We study the impact of consumer inattention on market outcomes for the US ready-to-eat cereal market by estimating a discrete-type mixed logit model with heterogeneous consideration sets within and between consumer types. The full information benchmark model is statistically rejected against all limited consumer attention specifications. Under the full information assumption own-price elasticities are inflated and cross-price elasticities are an order of magnitude smaller than those of our most preferred limited consumer attention specification. Product-level markups are higher under limited attention and are estimated by all models to increase over the period from 2006 to 2020. The consideration proxy that best fits the observable data has on average six products, while there are on average 153 products in the market. While consumer behavior is best explained by limited attention, our model selection tests indicate that firms on average expect consumers to be fully informed when setting prices.
By Johannes Kandelhardt    Andre Romahn; University of Hamburg    Christine Zulehner; University of Vienna
Presented by: Johannes Kandelhardt,
Discussant:   Aljoscha Janssen, Singapore Management University
Sequential Search and Exposure Dependence Design Abstract This paper studies an information design problem in a sequential consumer search environment. By choosing the dependence across a limited sequence of exposures, an intermediary shapes how information is revealed, trading off a confirmatory design that reinforces early signals against an exploratory design that broadens consumer appeal. We show that consumers' search behavior features multiple endogenous cutoffs, rather than a single reservation rule, determining whether to continue within an intermediary, switch away, or convert. Equilibrium exposure dependence is jointly shaped by the structure of search frictions and the informativeness of early exposures about subsequent ones.
By Zhuozheng Li; Shanghai University of Finance & Economics    Hongrui Zeng; Shanghai University of Finance and Economics
Presented by: Hongrui Zeng, Shanghai University of Finance and Economics
Discussant:   Ralph Boleslavsky, Indiana University
Playing to the Algorithm: How Spotify’s Recommendations Shape Music Production Abstract I examine how recommender systems have influenced the music industry and shaped music production. Using a structural model of the recorded music industry, I analyze consumer behavior, platform recommendations, and rightsholder release decisions. I estimate a fixed cost of $170,000 for songs that enter Spotify’s Top 200. Counterfactual analysis shows that with randomized recommendations, fewer songs would enter the market, reducing consumer welfare by 4%. The songs that do enter would be 33 seconds longer on average and more heterogeneously long. Popularity-based recommendations that do not account for individual taste would generate a superstar effect—increasing gross profit margins for songs that enter the market to 40%—but reducing consumer welfare by 13%. Although recommender systems have reduced overall variety in music, they have also enabled additional entry and increased consumer welfare.
By Max Schnidman; Microsoft
Presented by: Max Schnidman, Microsoft
Discussant:   Garrett Scott, University of Mississippi
Optimally Informative Rankings and Consumer Search Abstract This paper investigates the optimal information policy of an online platform (or multi-product firm) when ranking products in response to a consumer search query. The informativeness of rankings ranges from full information to full obfuscation, and consumers learn their match values with the products by engaging in costly sequential search. Invoking continuous match value distributions allows us to establish a novel result about consumer search. While consumers buy products with high match values and continue searching when they encounter low match values, they abort search without buying a product for intermediate ones. For a large class of distributions, the optimal strategy of a platform maximizing the probability of the consumer buying a product is to provide either full or no information at all. As a result, platform and consumer welfare are either fully aligned or at odds with each other.
By Maarten Janssen; U Vienna    Thomas Jungbauer; Cornell University    Marcel Preuss; Cornell University    Cole Williams; Durham University
Presented by: Marcel Preuss, Cornell University
Discussant:   Wenji Xu, City University of Hong Kong
Session 62: Production Networks and Supply Chain Dynamics April 12, 2026 10:15 to 12:15 Location: Brewster (4th Floor)
Session Chair: Jack Collison, University of Wisconsin Madison
The Rise and Fall of Input Suppliers Abstract This paper presents a model of vertical market structure that integrates multiple determinants previously studied in isolation. It shows how the nature of market growth, whether through existing or new applications of an input, and the type of fixed costs, whether exogenous or endogenous, affect the emergence of external suppliers for the input. The role of such suppliers depends on the input’s technological generalizability, its potential for improvement, and the breadth and differentiation among downstream applications. Together, these factors shape whether a vertical market structure persists, and can potentially drive multiple transitions between vertical integration and outsourcing.
By Unni Pillai; University at Albany, SUNY
Presented by: Unni Pillai, University at Albany, SUNY
Discussant:   Dongsoo Shin, Santa Clara University
Markups, Markdowns and Bargaining in a Vertical Supply Chain Abstract This article bridges monopoly, monopsony, and countervailing power theories to analyze their welfare implications in a vertical supply chain. We develop a bilateral monopoly model with bargaining that accommodates upstream monopsony and downstream monopoly power. In equilibrium, the “short-side rule” applies: the quantity exchanged is determined by the firm willing to trade less. Welfare is maximized when each firm’s bargaining power exactly countervails the other’s market power. Otherwise, double marginalization arises in the form of double markdownization under excessive downstream bargaining power, or double markupization under excessive upstream bargaining power. We offer novel insights for price regulation (e.g., price floors) and competition policy.
By Claire Chambolle; INRAE & CREST
Presented by: Claire Chambolle, INRAE & CREST
Discussant:   Matthias Mertens, Massachusetts Institute of Technology
Copper Supply Chain Abstract Copper is a critical input for renewable energy technologies and essential infrastructure. Yet, the global supply chain is increasingly exposed to geopolitical risk due to the geographic concentration of mining, and especially refining, production and rising trade tensions. This paper develops a structural model of the global copper market that integrates dynamic extraction decisions by heterogeneous mines with bilateral Nash-in-Nash bargaining between mines and smelters over treatment and refining charges. Mines extract copper concentrate subject to capacity constraints and ore-grade depletion dynamics. Smelters refine concentrate and compete in Cournot, subject to capacity, selling cathode on behalf of mines into a global market. Using detailed mine- and smelter-level data covering about 70\% of world production from 2015–2023, I estimate production cost and benefit functions, the gains from trade for each bilateral relationship, and the pair-specific bargaining weights. Initial results reveal substantial heterogeneity across country pairs, with pronounced leverage for China on average. Future work will use the model to evaluate critical policy scenarios, including trade measures and disruptions at key production sites.
By Phuong Ho; Massachusetts Institute of Technology    Christopher Knittel; MIT    Nicholas Vreugdenhil; ASU
Presented by: Phuong Ho, Massachusetts Institute of Technology
Discussant:   Leila Safavi, Pomona College
Inventories and Merger (In)efficiencies In the Petroleum Industry Abstract Despite the prevalence and importance of inventory in many markets, it is largely ignored in competition policy. Recent vertical consolidation of oil producers and refineries is often motivated by hedging crude oil price volatility, which in turn alters incentives to hold inventories. In this paper, I develop a dynamic model in which refineries optimally choose inventories of inputs and outputs, production, and quantities in the presence of oil price and demand uncertainty. The model highlights how vertical integration reshapes intertemporal tradeoffs by changing both exposure to cost uncertainty and strategic inventory incentives. I compile a novel database of refinery-level inventories, production, and sales in the petroleum industry. Using these data, I provide granular evidence on how refineries use inventories of both inputs and outputs to smooth supply and demand shocks. Following vertical mergers, I find that refineries decrease crude oil inventories by nearly 10 percent, consistent with a reduced need for inventory-based hedging. The insights of this paper offer guidance to policymakers on a range of topics, including merger guidelines and inventory mandates.
By Jack Collison; University of Wisconsin Madison
Presented by: Jack Collison, University of Wisconsin Madison
Discussant:   Nihal Mehta,
Session 63: Mergers, Market Structure, and Welfare April 12, 2026 10:15 to 12:15 Location: Mediterranean (3rd Floor)
Session Chair: Michael Sullivan, University of British Columbia
Differentiated product demand estimation with secondary markets Abstract How do primary and secondary markets interact? We propose a structural model of car choice by forward-looking consumers that features endogenous price equilibrium for both new and used cars. We estimate the model using Danish register data and conduct two counterfactual exercises to study how the primary and secondary markets interact: First, we simulate the effects of rising marginal costs and second, a hypothetical merger between the two largest car manufacturers. We derive two main findings: First, we show that an active secondary market in equilibrium leads to more market power since firms internalize that secondary market prices respond to prices in the primary market. Therefore, by not explicitly modeling the secondary market in durable goods industries, researchers would be led to underestimate market power and for example the competitive harm from mergers. Second, we characterize how a dynamic model allows for substantially more flexible substitution patterns than a static model.
By Maximilian Blesch; Humboldt Universität Berlin    Kenneth Gillingham; Yale University    jonas hansen; University of Copenhagen    Fedor Iskhakov; Australian National University    Nikolaj Moll; University of Copenhagen    Anders Munk-Nielsen; University of Copenhagen    John Rust; Georgetown University    Bertel Schjerning; University of Copenhagen
Presented by: Anders Munk-Nielsen, University of Copenhagen
Discussant:   Leonard Treuren, KU Leuven
Retail Mergers and Uniform Pricing: Evidence from the U.S. Yogurt Industry Abstract We investigate the impact of uniform pricing and retail mergers in differentiated product markets. While existing research primarily examines uniform and local pricing strategies independently, our work develops a new empirical framework that allows both strategies to coexist. Using the Nielsen-Kilts dataset, we develop a novel methodology to classify retailers according to their pricing strategy. We find that uniform pricing leads to higher average prices than local pricing, though the magnitude of this effect varies significantly across markets. Merger simulations further indicate that uniform pricing mitigates price increases in markets where merging firms have initially high market shares, but exacerbates price increases elsewhere. These findings underscore the critical need for competition authorities to incorporate heterogeneous pricing strategies into merger evaluations.
By Sebastien Cerles; Paris-Saclay University    Hugo Molina; French National Institute for Agriculture, Food and the Environment (INRAE)
Presented by: Sebastien Cerles, Paris-Saclay University
Discussant:   Sunmin Kim, The Ohio State University
Multiproduct Multimarket Price Competition under Capacity Constraints Abstract Firms often face hard capacity constraints determined by their machinery and workforce. This paper establishes the existence and uniqueness of the Bertrand–Nash equilibrium in a multimarket, multiproduct oligopoly with capacity constraints. We generalize the aggregate game framework—under which the payoff of each firm depends on rivals’ actions only through a low-dimensional aggregator—to this setting with linear or convex cost functions. Relative to previous work featuring a single-layer vector fixed point, our equilibrium characterization yields a nested fixed point structure with an across-market layer induced by capacity constraints and a within-market layer governing pricing. We conclude by discussing the implications of our results for merger analysis.
By Bingyao Liu; University of Toronto    Yao Luo; University of Toronto
Presented by: Bingyao Liu, University of Toronto
Discussant:   Ozan Akbas, University of Warwick
Demand with Network Externalities: Identification and an Application to the Dating Websites Industry Abstract This paper characterizes the identifiability of demand models with network externalities. Such models are generally not identifiable with market-level data, although microdata linking consumers’ decisions and characteristics permit identification under plausible conditions. Identification relies on instrumental variables reflecting across-market variation in the distribution of consumer characteristics or in product characteristics. Guided by the identification analysis, I empirically evaluate how network externalities shape the effects of consolidation in the US dating websites industry. The results suggest that the aggregate welfare loss from monopolization, which reflects pricing effects, is attenuated by welfare gains owing to network externalities.
By Michael Sullivan; University of British Columbia
Presented by: Michael Sullivan, University of British Columbia
Discussant:   Adam Dearing, Cornell University
Session 64: Pricing Strategies April 12, 2026 10:15 to 12:15 Location: Salon A (3rd Floor)
Session Chair: James Dana, Northeastern University
Pricing and Consumption in Subscription Settings Abstract This paper investigates how subscription pricing affects consumption intensity, a key performance driver for firms operating under subscription-based business models. We analyze data from an online news publisher, a setting in which promotional pricing is commonly employed to attract new subscribers, though its broader effects remain ambiguous. We document that promotional subscribers, on average, consume substantially more than those paying regular price, even after accounting for differences in churn behavior. This pattern points to the importance of taking unobserved heterogeneity into account. We develop and estimate an empirical model of subscription and consumption behavior, showing that, because subscription costs are sunk at the time of consumption, it is possible to recover the correlation between consumption levels and consumers’ unobserved willingness to pay. We use the model to recover the underlying consumer parameters and to evaluate the impact of alternative pricing policies on both subscription revenues (via customer acquisition) and advertising revenues (via subsequent consumption). Our findings highlight the economic value of understanding how price shapes not only who subscribes, but also how much they engage with the product.
By Pedro Gardete; Nova School of Business and Economics    Daniela Schmitt    Florian Stahl; University of Mannheim
Presented by: Daniela Schmitt,
Discussant:   Qi Pan, The Chinese University of Hong Kong (Shenzhen)
Personalized vs. Uniform Algorithm Design: The Unintended Consequence of Restrictions in Data Access Abstract This paper examines the unintended consequences of limiting third-party algorithm providers' access to firms' private cost data. We develop a model in which a monopoly provider supplies pricing algorithms to competing firms facing heterogeneous marginal costs and demand uncertainty. Our analysis compares two regimes: "Personalized Design", where the provider uses private cost data to customize algorithms for each firm, and "Uniform Design" where data access restrictions lead the provider to offer a single standardized algorithm to all firms. We identify a key trade-off for the algorithm provider: compared to personalized design, uniform design generates a competition-softening effect that tends to increase the provider's profit, but introduces a cost-misalignment effect for heterogeneous firms, reducing their willingness to pay. The provider's design choice depends on the balance of these two effects. However, if data access regulations force a shift from personalized to uniform design, adopting firms benefit at the direct expense of consumer and social welfare. These results provide important policy implications.
By Yibo Fang; Fudan University    Dingwei Gu; School of Management, Fudan University
Presented by: Yibo Fang, Fudan University
Discussant:   Gregory Sun, Washington University in St. Louis
Competition, Price Discrimination, and Experience Goods: A Theory of Frequent Buyer Discounts Abstract We show that customer loyalty programs, and more specifically, discounts for frequent buyers, are an outcome of price discrimination by firms selling experience goods in imperfectly competitive markets. We analyze oligopoly models with search and learning frictions. Ex ante, consumers differ only in their purchase frequency. Consumers’ valuations are drawn from the same distribution and have identical learning costs. However, more frequent buyers have greater incentives to try new products; thus, in equilibrium, they have better outside options, try new products more often, have higher ex post expected valuations, and pay lower prices. Even when purchase frequency is private information, firms can offer more frequent buyers lower prices using loyalty programs, which screen consumers based on their purchase frequency.
By James Dana; Northeastern University    Johannes Horner; Toulouse    Anna Sanktjohanser; Toulouse School of Economics
Presented by: James Dana, Northeastern University
Discussant:   Zach Brown, University of Michigan
Session 65: Platforms and Innovation (Sponsored by Brattle) April 12, 2026 10:15 to 12:15 Location: Atlantic 3 (3rd Floor)
Session Chair: Imke Reimers, Cornell University
From Complaint to Action: Technology-Enabled Quality Improvement from Consumer Reviews Abstract This paper studies how the adoption of an automated review monitoring system (ARMS) enables firms to translate consumer reviews into operational quality improvements. ARMS provides automated reminders for negative reviews and facilitates back-end work-ticket management, reducing the costs of identifying, prioritizing, and addressing operational problems revealed by customer feedback. Using restaurant-level data on ARMS adoption and consumer reviews, we find that adoption leads to higher average star ratings, a lower share of negative reviews, and more positive sentiment in review text. We find no evidence that these improvements are driven by strategic manipulation of reviews. Instead, the effects are greater in operational dimensions that had been flagged by consumers prior to adoption, consistent with ARMS improving firms’ responsiveness to informative feedback. We further show that ARMS adoption crowds out front-end managerial responses to reviews, highlighting a substitution between operational remediation and public reply. The effects are strongest when staff exhibit a more reflective approach to handling negative reviews in the back end. Finally, ARMS adoption is associated with increased consumer engagement, as reflected in longer and more detailed reviews. Together, these findings show that online review systems, when complemented by operational technologies, can generate meaningful quality improvement rather than merely reputation management.
By Guangyu Cao; Peking University    Shenghao He; Peking University    Ginger Jin; University of Maryland at College Park
Presented by: Ginger Jin, University of Maryland at College Park
Discussant:   Luca Bennati, Central Bank of Mexico
Signaling the Value of an Innovation by (not) Litigating Patent Infringers Abstract We examine whether litigating or not litigating other firms who enter a market and infringe an existing patent for a previous innovation can signal the value of a patent. Without entry deterrence patent holder litigate, if the expected costs exceed the expected benefits. However, by avoiding litigation costs and relinquishing even larger litigation benefits, the patent holder can potentially signal a low value of a patent and deter further entry into its market by not litigating an infringing firm. We will derive the general conditions for this to happen and present some more specific examples.
By Anette Boom; Copenhagen Business School    Marek Giebel; Copenhagen Business School
Presented by: Anette Boom, Copenhagen Business School
Discussant:   Juan-Pablo Montero, Pontificia Universidad Catolica de Chile
AI and the quantity and quality of creative products: the case of books Abstract The arrival of LLMs has revolutionized the creation of books. The number of new books appearing for sale at Amazon each month doubled between 2023 and late 2025, driven almost entirely by new author entry. Using the number of ratings books have received as a measure of usage, or "quality," the post-2022 AI-heavy vintages have lower average quality; but the top 1,000 books per category have higher quality. The effect is larger in categories with faster growth in new titles. Despite delivering potentially low average quality, the AI influx is also raising the number of high-quality books and therefore the aggregate value of books coming to market. On balance, the advent of LLMs does not displace incumbent author activity; instead, it enhances their output.
By Imke Reimers; Cornell University    Joel Waldfogel; University of Minnesota
Presented by: Imke Reimers, Cornell University
Discussant:   Ginger Jin, University of Maryland at College Park
How Do Suppliers Choose in a Platform Market: A Case Study of the Game Industry Abstract This paper studies how exclusive contracts shape developers' platform choice decisions in the sixth-generation video game console market. We estimate a supply-side structural model in which developers jointly choose which platform bundle to release on and when to launch, recovering development costs, cross-platform synergies, and exclusivity incentives from observed revenues without estimating a demand system. Allowing developers to endogenously choose launch timing substantially reduces estimated exclusivity benefits---particularly for the dominant incumbent---indicating that static models conflate exclusivity compensation with the option value of waiting. Beyond the standard foreclosure motive, we find evidence of a risk mechanism: small developers require greater compensation to commit exclusively to a new entrant platform, consistent with platforms needing to compensate suppliers for bearing uncertainty over future consumer adoption.
By Arifah Hasanbasri; University of Pittsburgh    Timothy Derdenger; Carnegie Mellon University
Presented by: Arifah Hasanbasri, University of Pittsburgh
Discussant:   Sangeun Ha, Copenhagen Business School
Session 66: Innovation, Competition, and Market Structure April 12, 2026 10:15 to 12:15 Location: Salon B (3rd Floor)
Session Chair: Sarit Weisburd, The Hebrew University of Jerusalem
A Tale of Trolls and Fees: The Role of Fee-Shifting in Patent Litigation Abstract Patent Assertion Entities (PAEs) are often viewed as taxing innovative activity; we show how fee shifting in patent litigation can effectively deter their more frivolous assertions, increasing exposed firms' R&D, innovation output, and startup success. Our identification relies on the U.S. Supreme Court's 2014 ruling in Octane Fitness v. ICON Health & Fitness, which broadened courts' discretion to award attorneys fees against exceptionally weak infringement claims. Using a quasi-regression discontinuity design comparing cases pending at Octane's release with those filed and decided before or after, we find that the decision increased defendants' willingness to contest weak claims and prompted plaintiffs, especially PAEs, to file stronger ones. Pending cases saw more fee awards and lower settlement as well as plaintiff success rates; subsequent PAE cases involved stronger patents and higher success rates. A difference-in-differences analysis further shows that Octane boosted innovation: public firms that were particularly exposed to PAE assertions prior to Octane increased R&D and patenting in response, and private startup firms performed better in venture capital markets.
By Tommaso Alba; KU Leuven
Presented by: Tommaso Alba, KU Leuven
Discussant:   John Turner, University of Georgia
AI and Competition Abstract This paper investigates how competition affects the deployment timing of general-purpose AI technologies in the presence of safety risks. We begin by analyzing the monopoly case as a benchmark to delineate the effects of competition on deployment timing. To capture the homogenization of AI systems, we assume a positive correlation between the risk profiles of the competing technologies developed across firms. Our analysis reveals that competition can lead to two contrasting distortions, depending on the nature of competition, risk correlation, and the informativeness of beta testing: a race to the bottom or insufficient entry. The race to the bottom occurs when the first-mover advantage induces firms to deploy prematurely, even though it would be socially optimal to delay entry until the beta testing results are available. We find that as AI systems become more homogenized, the race to the bottom is more likely. Conversely, the incentive to free-ride on a rival’s experimentation may result in too little entry. We further discuss potential regulatory policies aimed at mitigating these distortions and improving the timing of AI technology deployment.
By Jay Pil Choi; Michigan State University    Doh-Shin Jeon; Toulouse School of Economics    Domenico Menicucci; Universita' degli Studi di Firenze
Presented by: Jay Pil Choi, Michigan State University
Discussant:   Liang Zhong, The University of Hong Kong
Technology Gaps, Competition, and Regional Convergence Abstract Even 30 years after the reunification, regions in East Germany (the former GDR) live in considerably different economic conditions, with the average GDP per capita still about 20 percent below the average level in the West German regions. In this paper, we explore the factors that impeded faster convergence despite massive support to the East with a particular focus on technological differences and firm behavior. In the immediate aftermath of the reunification, production in the former GDR exhibited a rapid catch-up with the West with a pick-up in labor productivity. But the convergence tapered off quickly thereafter, with a stark difference between East and West German firms’ product qualities persisting ever since. We build a quantitative model of innovation, competition, and regional integration that can mimic these dynamics and provides a suitable setting to evaluate alternative policies that could have altered these dynamics. We show that large initial technological differences depressed Eastern firms’ incentives to compete and invest in technology improvements, perpetuating initial gaps. Delaying reunification, that is, opening up to competition from the West, would not help Eastern firms build up capacity. Sustained support for R&D in the East from the West could have helped shrink persistent gaps in product quality and income, although more effective alternatives appear to be subsidies to Western firms via either R&D support, with knowledge spillovers lifting also Eastern technology, or direct income support to facilitate technology transfer to the East via licensing.
By Ufuk Akcigit; University of Chicago    Sina Ates; Federal Reserve Board    Furkan Kilic; University of Chicago    Matthias Mertens; Massachusetts Institute of Technology    Steffen Mueller; Halle Institute for Economic Research
Presented by: Furkan Kilic, University of Chicago
Discussant:   Sarit Weisburd, The Hebrew University of Jerusalem
Multinational Firms and Entrepreneurship Abstract Does globalization hinder entrepreneurship? This paper examines how the entry of foreign multinational corporations (MNCs) affects the formation and performance of new firms in the technology sector. Using comprehensive administrative data from Israel covering all firms and workers in the tech industry between 2005 and 2018, we find that MNC entry significantly reduces the likelihood that local employees found new startups. At the same time, wages are 15 percent higher at MNCs, and entry raises wages at nearby local firms by 7 percent, suggesting that higher opportunity costs suppress entrepreneurial activity. We also find suggestive evidence that startups founded after MNC entry have higher survival rates, consistent with stronger selection or positive spillovers. Taken together, our results highlight how MNC expansion dampens business dynamism while potentially shifting entrepreneurship toward higher-quality ventures.
By Itai Ater; Tel Aviv University    Noam Gruber; IMF    Sarit Weisburd; The Hebrew University of Jerusalem
Presented by: Sarit Weisburd, The Hebrew University of Jerusalem
Discussant:   Minhae Kim, Oklahoma State University
Session 67: Platform Design: Supply-Side Incentives April 12, 2026 10:15 to 12:15 Location: Salon C (3rd Floor)
Session Chair: Byoungmin Yu, University of Florida
Platform Design, Earnings Transparency and Minimum Wage Policies: Evidence from A Natural Experiment on Lyft Abstract We study the effects of a significant design and policy change at a major ridesharing platform that altered both provider earnings and platform transparency, examining how it altered outcomes for drivers, riders and the parent platform, and providing managerial insights about how to balance competing stakeholder interests while avoiding unintended consequences. In February 2024, Lyft introduced a policy guaranteeing drivers a minimum fraction of rider payments while also increasing per-ride earnings transparency. The rollout of this policy was staggered, first introduced in “major markets” that were more urban, providing a natural experiment to assess how platform transparency and earnings guarantees affect ridesharing availability, driver engagement and rider satisfaction. Using trip-level data from over 47 million rides in a major urban market and its neighboring suburban markets across six months, we applied dynamic staggered difference-in-differences models along with a geographic border strategy to measure the causal effects of these platform design changes on supply- and demand-side outcomes, ride production and platform outcomes. We show that the design change led to substantial changes in driver engagement, with separate effects from the guarantee and the transparency. Drivers increased their working hours and utilization, leading to more completed trips and higher per-hour and per-trip earnings. These effects were strongest for drivers with lower pre-policy earnings and greater income uncertainty. We unpack the economic mechanism by which these changes led to a positive spillover on demand. We also provide some evidence that points to platform transparency potentially leading to unintended strategic driver behavior. In ongoing work, we outline a counterfactual simulation framework that models ride production as a function of driver supply hours and rider intents, examining how small changes in driver choices might have amplified the positive effects of the intervention even further, and develop a self-supervised machine learning model that leverages driver trajectory embeddings to predict multihoming behavior and examine whether supply increases came from both the expansion of driver activity as well as substitution from competing platforms. Our study shows how platform-led interventions present an intriguing alternative to government-led minimum pay regulation, and provide new strategic insights into managing platform change.
By Rubing Li; New York University    Xiao Liu; New York University    Arun Sundararajan; New York University
Presented by: Rubing Li, New York University
Discussant:   Chung-Ying Lee, National Taiwan University
Sharing is (not) caring: strategic information disclosure and platform’s business model Abstract This study explores how data and data-sharing affect the strategic dynamics between a monopolistic platform and the third-party sellers active on its digital marketplace. We analyze how data-sharing alters the strategic choice of the business model operated by the platform in each market category. Moreover, we investigate the conditions under which a platform chooses to share user data with the sellers, particularly when such data reveal consumer preferences and can be leveraged for price discrimination. First, we show that data-sharing induces the platform to operate more often in agency mode, due to the seller increased ability to extract surplus from the consumers, which is shared via the revenue-sharing fee. Second, we show that the platform would strategically share data only with those sellers it does not want to compete with (i.e., the efficient ones) as it can exploit their efficiency through the fee. Third, when considering the possibility for the platform to charge a separate fee to those sellers who want to access the data, the platform earns an even higher profit. Finally, in the presence of perfect downstream competition among third-party sellers, price discrimination is redundant, and so is sharing data that allow for that practice.
By Federico Navarra; Charles River Associates    Flavio Pino; Politecnico di Torino    Luca Sandrini; ZEW - Leibniz Centre for European Economic Research
Presented by: Federico Navarra, Charles River Associates
Discussant:   Shiyun Xia, Tianjin University
Platform Design in Dynamic Differentiated Goods Markets - The Case of Airbnb Abstract This paper studies the allocation of goods in two-sided, one-to-one matching markets with finite time horizons. In these markets, the timing of seller participation distorts the allocation of goods. This is due to dynamic externalities inherent in one-to-one matching markets -- sellers fail to fully consider how their participation impacts the evolution of the market. Using data on Airbnb, I estimate a dynamic structural model of a two-sided matching market in order to quantify externalities related to the timing of seller entry. I evaluate policies that a platform can implement to improve total surplus on its market. Entry subsidies that encourage later entry reduce total welfare by $3 for every $1 spent, while subsidies that encourage earlier entry increase total welfare by $2 for every $1 spent.
By Wenxuan Xu; National University of Singapore Business School
Presented by: Wenxuan Xu, National University of Singapore Business School
Discussant:   Sebastian Valet, ZEW Mannheim
Commission Fee Structure and Innovation in Digital Platforms Abstract This paper quantifies the welfare effects of regulating commission fees in digital platforms, focusing on third-party app developers' innovation and pricing decisions. I employ a comprehensive dataset of music apps within Apple's App Store in the United States from October 2018 to February 2024 to estimate app users' demand and app developers' cost parameters. The paper reveals key findings with three policy counterfactual simulations where I sequentially solve for optimal innovation and pricing decisions. First, capping commission fees stimulates third-party developers’ innovative efforts and improves social welfare. Second, when the platform adds a unit fee scheme under the fee cap, developers partially pass unit fees onto app users by increasing in-app purchase prices. Third, a hypothetical buy-out of a streaming app by the platform leads to a significant decrease in the innovative efforts and market share of the acquired app. Notably, welfare analysis without quality adjustment is predicted to underestimate the impact of fee cap on social welfare by 0.91% - 2.06% points compared to the full-stage model estimates. This research highlights the importance of considering quality changes along with price effects when evaluating regulatory intervention in digital platforms.
By Byoungmin Yu; University of Florida
Presented by: Byoungmin Yu, University of Florida
Discussant:   Tianli Xia, University of Rochester
Session 68: Technology, Spatial Access, and Public Program Design April 12, 2026 10:15 to 12:15 Location: Georges (4th Floor)
Session Chair: Lucy Xiaolu Wang, University of Massachusetts Amherst; Max Planck Institute for Innovation and Competition
The Cost of Cost Savings: Procurement Auctions and Moral Hazard in Drug Market Abstract Procurement contracts are central to big buyers like governments, yet their design can generate a moral hazard problem: winning firms have incentives to cut costs by reducing quality. We examine this issue in China’s volume-based procurement (VBP) of generic drugs. Drawing on regulatory filings and novel measures of patient retention and switching, we find that procured drugs exhibit more supplier changes, lower consumer retention, and more switches to branded alternatives, consistent with reduced quality. To assess welfare implications, we estimate a structural demand model with unobserved heterogeneity and state dependence in drug choice. Our results indicate that while procurement achieves large cost savings, the decline of drug quality offsets 87.7\% of consumer welfare gains. Scoring auctions that account for quality could improve consumer surplus. Together, these findings provide empirical evidence of supplier moral hazard in the healthcare procurement setting and underscore the importance of contract design to regulate quality.
By Tianli Xia; University of Rochester    Liyu Zhao; University of Rochester
Presented by: Liyu Zhao, University of Rochester
Discussant:   Chenyuan Liu, Tsinghua University
Spatial Disparities, Selection, and Segmentation in Health Insurance Abstract Most U.S. means-tested programs use uniform eligibility thresholds that ignore regional cost-of-living differences. I study whether indexing eligibility to local costs improves coverage and targeting in Medicaid and health insurance marketplaces: two adjacent programs serving low-income individuals without employer insurance. I develop and estimate a model where the regulator sets eligibility rules, insurers compete on price and quality, and consumers choose or are assigned to plans. I find that adjusting eligibility thresholds for cost-of-living creates competing effects: it improves coverage by crowding low-income families into insurance in high-cost areas, but also reshuffles risk pools across both programs in ways that change premiums and plan quality. Counterfactual simulations using administrative data from California show that partial cost-of-living indexing yields higher welfare per public dollar than either uniform thresholds or full indexing. These results demonstrate how eligibility design interacts with market structure and insurer incentives in segmented health insurance systems.
By Salpy Kanimian; Rice University
Presented by: Salpy Kanimian, Rice University
Discussant:   Stephan Sagl, Indiana University
From Free Rider to Innovator: The Rise of China’s Drug Development Abstract This paper examines the transition of developing economies from pharmaceutical “free riders” to innovators, using China’s recent ascent as a case study. Since 2020, China has surpassed the US in its annual volume of registered clinical trials. By linking global clinical trial registries with granular drug sales data and scholarly publication data, we provide evidence that this shift was primarily driven by the National Reimbursement Drug List (NRDL) reform. This policy shock increased the effective market size for innovative drugs by combining price negotiations with massive insurance expansion. We document a sharp rise in both the quantity (94% increase post-reform) and novelty of drug trials, with growth concentrated in reform-exposed disease categories, in novel drug development, and among domestic firms. A decomposition exercise shows that the NRDL reform accounts for 40% of the growth in trial activity in the Oncology sector. Our results suggest that strategic public purchasing can simultaneously improve patient access and provide robust incentives for indigenous innovation.
By Panle Jia Barwick; UW Madison and NBER    Hongyuan Xia; Cornell University    Tianli Xia; University of Rochester
Presented by: Hongyuan Xia, Cornell University
Discussant:   Anwita Mahajan, University of California, San Diego
Health IT Diffusion and Physician Density Abstract This paper examines how the diffusion of advanced health information technology (HIT) affects the density of hospital-based (HB) physicians. Leveraging sharp county-level increases in HIT adoption driven by federal incentives, we compare physician density per 100k population in counties with rapid diffusion (treatment group) to those with slower or no uptake during our sample period (control group). Using an event-study framework, we find that HIT diffusion led to a 11.1% increase in the HB physician rate in treated counties relative to control counties, and medical and surgical specialties account for most of the increase. Moreover, the growth in HB physician density is evident, absent major consolidation activities, and contributes to overall growth in total physicians. This growth is concentrated among early-career physicians and in physician shortage areas. Mechanism tests suggest that physicians benefit financially from practicing in treated counties, with higher Medicare reimbursement and hospital profits. Outpatient surgeries rise most in counties with moderate pre-period care utilization. Various robustness checks support our results. Our findings suggest that strategic HIT investments can attract physicians, expand care capacity, and reduce geographic disparities in health care access.
By Jason Huh; Rensselaer Polytechnic Institute    Jianjing Lin; University of Massachusetts Amherst    Lucy Xiaolu Wang; University of Massachusetts Amherst; Max Planck Institute for Innovation and Competition
Presented by: Lucy Xiaolu Wang, University of Massachusetts Amherst; Max Planck Institute for Innovation and Competition
Discussant:   Ching-to Albert Ma, Boston University
Session 69: Merger Retrospectives April 12, 2026 10:15 to 12:15 Location: Aegean (3rd Floor)
Session Chair: Flavia Roldán, Universidad ORT Uruguay
Coordinated Pricing After a Blocked Merger: Evidence from the JetBlue-Spirit Case Abstract We examine pricing behavior following the blocked merger between JetBlue and Spirit Airlines. We particularly focus on whether firms engaged in coordinated pricing. Using a two-way fixed effects model with instrumental variables, we compare fare patterns before and after both the merger announcement and the regulatory block. Despite the merger’s termination, we document patterns consistent with coordinated pricing, greater price alignment and lower dispersion on overlapping routes, possibly reflecting information gained during merger due diligence. To deepen our understanding of firm conduct, we also tested whether JetBlue’s post-block decisions reflect strategic responses to cost information revealed during merger negotiations. The findings support the possibility that information driven behavior, along with tacit coordination, shaped post-block market outcomes. In line with Adam Smith’s caution in The Wealth of Nations, even unrealized mergers may create channels for coordination and challenge the assumption that market competition naturally reasserts itself once consolidation is blocked.
By Minhae Kim; Oklahoma State University    Myongjin Kim; U Oklahoma    Junyeol Ryu; University of Oklahoma
Presented by: Junyeol Ryu, University of Oklahoma
Discussant:   Joe Mazur, Purdue University
Multimarket Contact and Prices: Evidence From an Airline Merger Wave Abstract We study the US airline merger wave from 2008 through 2013, which included mergers between Delta/Northwest, United/Continental, Southwest/AirTran, and American/USAir. We first show these mergers occurred between airlines with complementary networks and very little head-to-head competition on overlap, nonstop routes. Consequently, each merger led to minimal changes, on average, in route-level HHI but large increases in multimarket contact. We analyze the causal impact of the mergers on prices using synthetic difference-in-differences and the synthetic control method. We find that merger-induced increases in multimarket contact led to higher prices, especially in the latter two legacy mergers. We therefore find that these mergers led to coordinated price effects. In contrast to the previous literature, we implement econometric methods that match on pre-merger price trends, and we do not find a significant impact on overlap routes in legacy airline mergers, suggesting that a primary channel through which mergers affect prices is an increase in multimarket contact.
By Marc Remer; Swarthmore College    Reed Orchinik; MIT
Presented by: Marc Remer, Swarthmore College
Discussant:   Federico Ciliberto, U Virginia
Airline Merger Effects Along the Exposure Distribution Abstract I evaluate four major U.S. airline mergers (2008–2013) using a continuous difference-in-differences model with route-level exposure measures. Standard merger retrospectives typically rely on binary treatment classifications that collapse important variation in competitive exposure. I construct three continuous exposure measures—simulated ∆HHI (direct competitive overlap), merger share (the merging carriers’ combined presence), and non-overlap merger share (merger share on routes where only one merging carrier operated)—and estimate dose-response curves for each. Results reveal substantial heterogeneity both across mergers and within each merger across the exposure distribution, with different exposure measures yielding different conclusions, as each captures a distinct competitive channel.
By Benjamin Leyden; Cornell University
Presented by: Benjamin Leyden, Cornell University
Discussant:   Chenyu Yang, University of Maryland, College Park
Price, Variety, and Quality Effects of Retail Consolidation: Evidence from a Supermarket Merger Abstract We study the impact of a supermarket merger on consumer well-being, examining prices, product variety, and quality using georeferenced supermarket data from Uruguay. We employ a difference-in-differences approach with a robust control group of non-large supermarkets isolated from larger competitors to ensure immunity from merger-driven competition. We measure variety through intensive (within-category) and extensive (portfolio breadth) margins, and develop a quality measure based on relative price positioning within categories, and use Mahalanobis distance to assess overall strategic repositioning. The merger harms consumers through multiple channels: increasing prices by 2.6 percentage points, reducing variety by 0.16-0.19 varieties per category and 4-10 percentage points in portfolio breadth, while shifting toward higher-quality segments with 4.2-5.3 percentage points increase in quality positioning. These findings reveal complex strategic repositioning that traditional price-focused merger analysis would miss, highlighting the importance of multidimensional evaluation frameworks for competition policy.
By Flavia Roldán; Universidad ORT Uruguay
Presented by: Flavia Roldán, Universidad ORT Uruguay
Discussant:   Farasat Bokhari, Loughborough University
Session 70: Firm Relationships and Policy April 12, 2026 10:15 to 12:15 Location: Thompson (4th Floor)
Session Chair: Zhonglin Li, National University of Singapore
Common Ownership and Risk Selection in Medicare Part D Abstract This paper shows how common ownership among Medicare Part D insurers affects premium setting and interacts with risk selection. Using the 13F filing, we document substantial overlap in insurers' institutional owners. Reduced form evidence shows that greater exposure to common ownership is associated with higher premiums. We then estimate a structural model of demand and supply for prescription drug plans (PDP) that quantifies the effects of common ownership and risk selection. Non-nested tests and conduct parameter estimates favor common ownership conduct over own profit maximization. We find that common ownership magnifies the effect of risk selection on premiums, and removing both would reduce average premiums by 20 percent and increase plan enrollment by 14.7 percentage points, improving allocative efficiency.
By Pinka Chatterji; The University at Albany    Chun-Yu Ho; University at Albany, SUNY    Jaehak Lee; University at Albany, SUNY
Presented by: Jaehak Lee, University at Albany, SUNY
Discussant:   Mario Leccese, Boston University
The Welfare Effects of Government Intervention into the Licensing of Standard-Essential Patents: An Analysis of the Chinese Smartphone and SoC Markets Abstract The licensing of standard-essential patents ("SEPs") in the cellular communications field has been a contentious issue worldwide. In particular, the licensing policy of the U.S. technology firm Qualcomm for third-generation and fourth-generation cellular communication standards has been the subject of lawsuits and government intervention in several countries. Given Qualcomm's dual role as a technology licensor as well as a supplier of baseband processors and System-on-Chips ("SoCs") to smartphone manufacturers, there has been a concern that Qualcomm's royalties on smartphone sales could have an exclusionary effect on rival chipmakers and cause consumers to pay higher prices for smartphones. We evaluate the impact of the most drastic intervention to date: the Chinese government's 2015 decision to forcibly lower Qualcomm royalty rates by 1.23 - 1.75 percentage points. Using a simple theoretical model, we argue that such an intervention could have ambiguous effects on consumers; it could lead to higher or lower smartphone prices. To quantify the policy's impact, we construct a structural econometric model of the Chinese smartphone and SoC markets which allows for strategic pricing in the two vertically related markets. Counterfactual analysis using the estimated model allows us to quantify the intervention’s impact on smartphone manufacturers' marginal costs and product prices. Our simulation results indicate that the intervention tended to cause an increase in smartphone manufacturers' marginal costs (around 1.1 percent on average). However, this was more than offset by smartphone manufacturers' incentive to lower their prices under the reduced royalty rate, leading to a slight reduction in smartphone prices (around 0.6 percent on average). Taken together, these results suggest that the Chinese government's intervention had the intended effect on social welfare, although its magnitude was fairly limited.
By Kensuke Kubo; Keio University    Mariko Watanabe; Gakushuin University
Presented by: Kensuke Kubo, Keio University
Discussant:   Jin-Hyuk Kim, University of Colorado at Boulder
Competition Enforcement Effects across the Value Chain Abstract How do shocks to competition in one sector propagate through the value chain? Using firm-level panel data from 25 countries, we study this question by estimating the value chain effects of the formation and breakups of collusive cartels. We match the names of cartel firms to a large firm-level panel and generate time-varying measures of exposure to cartelized sectors’ linkages in the production network for firms in upstream, downstream, and complementary sectors. We then leverage quasi-random changes in exposure to cartelized sectors triggered by the dates of cartel formation and their dissolution by competition authorities to study the value chain effects. We find that the negative shocks to competition generated by cartel conduct have significant negative effects on the output, investment, employment, and profits of downstream firms and firms in complementary sectors, while upstream firms may capture part of the cartel rents. We document significant sectoral heterogeneity in these effects. Our results provide new cross-country firm-level evidence on value-chain spillovers of de facto competition enforcement and underscore the importance of strong antitrust institutions.
By Chiara Criscuolo; IFC and CEP    Tomaso Duso; DIW Berlin, TU Berlin, CEPR, CESifo    Leonard Le Roux; International Finance Corporation - World Bank Group    Antonis Tsiflis; World Bank
Presented by: Leonard Le Roux, International Finance Corporation - World Bank Group
Discussant:   David Benson, Federal Reserve Board
Vertical Control and Retail Competition: Evidence from Consumer Battery Industry Abstract This paper examines the impact of a vertical agreement between a downstream retailer and an upstream manufacturer in the context of the U.S. consumer battery industry, using detailed consumer receipt panel data. Following the agreement between Walmart (the retailer) and Energizer (the manufacturer), Walmart increases the retail prices of Duracell, Energizer’s main competitor, while Energizer increases the wholesale prices charged to other retailers competing with Walmart. We document a shift in Walmart’s sales from Duracell toward Energizer and retail price increases of about 6–7 percent for both brands across major retailers. To interpret these patterns, we estimate a model of consumer demand and retailer–manufacturer bargaining that treats Walmart and Energizer as an integrated party in the post-agreement period. Counterfactual analyses show that Duracell, though not part of the agreement, benefits from higher prices. Moreover, when similar agreements involve smaller retailers, they induce even larger wholesale and retail price increases elsewhere, resulting in greater consumer harm.
By Zhonglin Li; National University of Singapore    Hsin-Tien Tiffany Tsai; National University of Singapore
Presented by: Zhonglin Li, National University of Singapore
Discussant:   Kosuke Shimamoto, Duke University
Session 71: Panel Discussion: Doing IO at Antitrust Consulting Firms April 12, 2026 10:15 to 12:15 Location: Atlantic 1 (3rd Floor)
Session Chair: Lawrence White, Stern School of Business, New York Unive
Discussants:      Mark Ponder, NERA      Prerna Rakheja, Bates White      Matias Escudero, Compass Lexecon      Evangelos Constantinou, Keystone      Riccardo Marchingiglio, Analysis Group
Session 72: The Impact of Generative AI on Content, Information, and Society April 12, 2026 10:15 to 12:15 Location: Atlantic 2 (3rd Floor)
Session Chair: Eungik Lee, FRB NY
Censorship by Perception: LLM Predictions of Social Media Behavior in the Face of Government Scrutiny Abstract We investigate how knowledge of news regarding government surveillance of social media affects behavior. Prior research documents user self-censorship through surveys and case studies, which make generalizing hard. We address this gap by using large language models (LLMs)—including ChatGPT, Claude, Llama, and Gemini—to simulate how diverse U.S. user profiles respond to knowledge of government surveillance. We design a multistage game in which LLMs act as social media users. We give them news on which they can comment and share, sometimes with and sometimes without knowledge from other news sources that the government is monitoring social media. Preliminary results show users modify their postings and the degree of change depends on each user’s political leanings and the topic.
By Isabelle Hansen; University of Florida    Mark Jamison; University of Florida
Presented by: Mark Jamison, University of Florida
Discussant:   Renjie Bao, Princeton University
AI Summaries Overrepresent Fake Reviews: Evidence from Amazon Abstract AI summarization of reviews has been widely deployed to distill information from large volumes of consumer reviews. In this research, we argue that AI summarization produces an unintended consequence: AI summaries overrepresent fake reviews. Our key insights are that (i) AI summarization is designed to extract common themes, and that (ii) fake reviews are more linguistically similar to one another, creating more common themes. To provide empirical support for our argument, we study AI summaries on Amazon. Using varying assumptions and proxies for products more likely to contain fake reviews, we consistently find evidence that AI-generated summaries on Amazon overrepresent fake reviews. We also study the influence of the overrepresentation of fake reviews on market outcomes. We find that review manipulators receive more positive AI summaries than other sellers, even after we control average ratings and other observable characteristics. We also find that review manipulators experience a greater increase in sales following the introduction of AI summarization relative to their competitors.
By Sihan Zhai; Harvard Business School    Andrew Ching; Johns Hopkins University
Presented by: Sihan Zhai, Harvard Business School
Discussant:   Cole Williams, University of Nebraska-Lincoln
Does Generative AI Crowd Out Human Creators? Evidence from Pixiv Abstract Using a comprehensive dataset of posts from a major platform for anime- and manga-style artwork, we study the impact of the launch of a prominent text-to-image generative AI. Focusing on the majority of incumbent creators who do not adopt AI as a primary tool, we show that the AI launch led to a significant decline in post uploads by illustrators, relative to comic artists who are less affected by the launch due to comics' need for tight stylistic alignment across sequential images. We present empirical evidence for two underlying mechanisms: (1) illustration posts experience a loss of viewer attention---measured by bookmarks---following the AI launch, which can significantly harm creators' business models; (2) direct competition from AI-generated content plays a role: illustrators who work on intellectual properties (IPs; e.g., Pokémon) that are more heavily invaded by AI reduce their uploads disproportionately more. We further examine creators' responses and show that illustrators who are highly exposed to AI avoid using tags favored by AI-generated content after the AI launch and broaden the range of IPs they work on, consistent with a risk-hedging response to AI invasion.
By Ginger Jin; University of Maryland    Sueyoul Kim; KDI    Eungik Lee; FRB NY
Presented by: Eungik Lee, FRB NY
Discussant:   Sihan Zhai, Harvard Business School
“Just One More Clip”: Short Videos, Big Self-Control Problems Abstract I study self-control problems in media consumption and their amplification by short-form content. Using microdata from a U.S. short drama series, I show viewers watch 23 episodes (82%) more than intended and overspend by $5.51 (23%). The structural estimation indicates that temptation takes effect over a short horizon, averaging 6.6 minutes per decision, making minute-long videos repeatedly trigger self-control problems. Policy interventions such as default time limits and mandatory breaks can meaningfully improve consumer welfare. Extrapolating the analysis to TikTok underscores the broader relevance of these findings.
By Renjie Bao; Princeton University
Presented by: Renjie Bao, Princeton University
Discussant:   Eungik Lee, FRB NY