January 29-30, 2016
Matthew L. Gentry, Tatiana Komarova, and Pasquale Schiraldi, London School of Economics
Motivated by the empirical prevalence of simultaneous bidding across a wide range of auction markets, Gentry, Komarova, and Schiraldi develop and estimate a structural model of strategic interaction in simultaneous first-price auctions when objects are heterogeneous and bidders have preferences over combinations. In this model, bidders have stochastic private valuations for each object and stable incremental preferences over combinations, nesting the standard separable model as the special case when incremental preferences over combinations are zero. The researchers establish non-parametric identification under standard exclusion restrictions, providing a basis for both testing on and estimation of preferences over combinations. They then apply their model to data on Michigan Department of Transportation highway procurement auctions, quantify the magnitude of cost synergies, and assess possible efficiency losses arising from simultaneous bidding in this market.
Marie-Laure Allain, CNRS and Ecole Polytechnique; Claire Chambolle, INRA and Ecole Polytechnique; Stephane Turolla, INRA; and Sofia Villas-Boas, University of California at Berkeley
This paper analyzes the impact of a merger in the French supermarket industry on food prices. Allain, Chambolle, Turolla, and Villas-Boas stress the importance of considering whether price decisions are taken at the local or at the national level. Using consumer panel data, the researchers perform a difference-in-differences analysis. They provide a novel approach to define local areas affected by a merger when merging firms set prices nationally while other firms price locally. On average, they find that the merging firms significantly raised their prices after the merger but rather nationally than locally. The merger caused a significant increase of the competitor's prices, which is stronger in local markets in which more merging firms operate, and in which differentiation changed after the merger.
BLP Turns 21
Steven Berry and Philip Haile, Yale University and NBER
Empirical models of demand for – and, often, supply of – differentiated products are widely used in practice, typically employing parametric functional forms and distributions of consumer heterogeneity. Berry and Haile review some recent work studying identification in a broad class of such models. This work shows that parametric functional forms and distributional assumptions are not essential for identification. Rather, identification relies primarily on the standard requirement that instruments be available for the endogenous variables - here, typically, prices and quantities. The researchers discuss the kinds of instruments needed for identification and how the reliance on instruments can be reduced by nonparametric functional form restrictions or better data. They also discuss results on discrimination between alternative models of oligopoly competition.
Amit Gandhi, University of Wisconsin, and Jean-François Houde, University of Pennsylvania and NBER
Bradley Larsen, Stanford University and NBER, and Dominic Coey and Kane Sweeney, eBay Research Labs
Coey, Larsen, and Sweeney introduce a simple and robust approach to address several key questions in empirical auction analysis: discriminating between models of entry and quantifying the revenue effects of improved auction design or from bidder mergers or collusion. The approach applies in a broad range of information settings, including common values or asymmetric correlated private values, and auction formats, such as first price or ascending auctions. Furthermore, the approach does not require instruments, exogeneous variation in the number of bidders, or complex estimation techniques. The researchers demonstrate the approach using U.S. timber and used-car auction data.
Kerem A. Cosar, Stockholm School of Economics; Paul Grieco, Pennsylvania State University; Shengyu Li; and Felix Tintelnot, University of Chicago and NBER
In the automobile industry, as in many tradable goods markets, firms earn their highest market share within their domestic market. This home market advantage persists despite substantial integration of international markets during the past several decades. The goal of this paper is to quantify the supply- and demand-driven sources of the home market advantage and to understand their implications for international trade and investment. Building on the random coefficients demand model developed by Berry, Levinsohn, and Pakes (1995), Cosar, Grieco, Li, and Tintelnot estimate demand and supply in the automobile industry for nine countries across three continents, allowing for unobserved taste and cost variation at the car model and market levels. While trade and foreign production costs as well as taste heterogeneity matter for market outcomes, the researchers find that preference for domestic brands is the single most important driver of home market advantage — even after controlling for brand histories and dealer networks.
Anita Rao, University of Chicago, and Emily Y. Wang, University of Massachusetts
Firms often make selective or deceptive claims in their advertising. Such claims can have negative consequences for consumers, especially if consumers are not fully informed and the claims are hard to verify. This paper aims to measure the impact of such false claims on consumer demand, and to understand which type of consumer these claims primarily affect. Using a panel dataset of consumer purchases and firm advertising, Rao and Wang exploit the fact that four popular products settled charges raised by the Federal Trade Commission, leading to an exogenous discontinuation of the false advertising campaigns, to measure this impact. The researchers further control for and document firm responses in terms of price and advertisement changes around the date of the settlement. Their findings indicate a significant decline in demand following the termination of the claims resulting in a 12%-67% monthly loss in revenue across the four products, which amounts to a $0.40m-$3.82m loss in monthly revenue. The researchers further find that these claims primarily affect consumers who are least loyal.
Anita Rao, University of Chicago
Do pharmaceutical firms respond to the actions of their competitors in R&D, and how much? Answering this has implications on the impact of a faster FDA approval process — something pharmaceutical companies are pushing for. While a faster approval process leads to quicker realization of profits and more remaining time on the firm's patent, it also intensifies competition reducing per-firm profits. Which effect dominates depends on the degree of competition. To this end, Rao estimates a dynamic investment model using Phase-3 data. Solving the new equilibrium, she finds an expedited process is beneficial only when competitors are far from launching.
Fernando Luco, Texas A&M University
This paper studies whether mandatory disclosure of information increases competition or facilitates coordination among firms using the sequential implementation of regulation requiring gas stations in Chile to post prices on a government website. Luco uses a difference-in-difference approach to show that, on average, disclosure increased margins by 10%, while margin dispersion did not change. Luco also uses data on the exact location of consumers when they searched for price information using a smartphone app to show that margins increased in low-income areas, where consumer search was low or nonexistent, and decreased in high-income areas, where consumer search was significant. This shows that consumer search is able to overcome coordination among firms only when the newly disclosed information is easily accessible for consumers.
Jean-François Houde, University of Pennsylvania, and Yuya Takahashi and Ricard Gil, John Hopkins University
Mitsukuni Nishida, Johns Hopkins University, and Nathan Yang, McGill University
Nishida and Yang develop a tractable dynamic oligopoly model for retailing, in which forward-looking firms strategically choose expansion plans taking into account of the costs and revenues of franchising and corporate outlets. The researchers estimate this model in the market for convenience-store industry in Japan. First, they demonstrate noticeable differences in expansion strategies across ownership types. Second, they confirm that franchisee-run outlets generate higher revenues (all else held equal) than their corporate-run counterparts. Finally, the researchers sunk cost estimates reveal that it is more costly to open (and close) corporate-run outlets than franchisee-run outlets. The results suggest that both revenue and cost considerations are important drivers behind franchising decisions. Despite such benefits of expansion via franchisee-run outlets, the results also show that corporate-based expansion can still be rationalized, as franchisee-run outlets are more sensitive to cannibalization. Furthermore, the researchers' counterfactual analysis provides a salient connection between preemptive motives and expansion via corporate-run outlets, despite the inferred revenue and cost-based benefits of franchisee-run expansion. Finally, the authors show that a sudden increase in the share of corporate-run outlets may precede a threat of entry.
Mitsuru Igami and Kosuke Uetake, Yale University
Igami and Uetake study the process of industry consolidation with endogenous mergers, innovation, and entry-exit. They develop an empirical model of a dynamic game with a random proposer of merger in each period, and estimate it using data from the hard disk drive industry. They find mergers became a dominant mode of exit and sometimes generated productivity improvement (i.e., synergies). The researchers counterfactual simulations feature antitrust policy regimes with alternative tolerance levels of mergers, and highlight a dynamic welfare tradeoff between the ex-post pro-competitive effects of blocking mergers and its negative side effects due to the destruction of ex-ante option values. The results suggest approximately four firms as the optimal regulatory threshold.
Tiago Pires, Northwestern University, and Alberto Salvo, National University of Singapore
Storability constrains firms' ability to implement price discrimination, as it enables consumers to separate the timing of purchase from the timing of consumption. Scanner data, however, indicate that significant price differences exist for the same storable good and brand sold in containers of different sizes, with small sizes accounting for a larger share of purchases of low-income households than for their more affluent counterparts. Low-income households therefore tend to be less willing or less able to purchase large storable-good containers typically offered at "bulk prices."
Viplav Saini, Oberlin College
Using the framework of computable Markov-perfect dynamics, Saini endogenizes the heterogeneity among bidders in a procurement auction market. Bidding firms can repeatedly and simultaneously make costly investments in cost reduction. Saini finds that the industry market structure changes dramatically with the chosen auction format. The first-price auction (FPA) tends to exacerbate bidder asymmetries, while the second-price auction (SPA) tends to reduce them. As a result, the static revenue ranking of asymmetric auctions (Maskin and Riley, 2000) reverses once the asymmetry is endogenized. The SPA can also be the more dynamically efficient format. Accounting for entry and exit, however, produces examples where the SPA increases industry concentration relative to the FPA, which then becomes the dynamically superior format.
Bradley Shapiro, University of Chicago
Promotional strategies that pharmaceutical firms employ to convince physicians to prescribe their products are the subject of considerable regulatory scrutiny. In particular, regulators worry firms may use sales reps to try to convince physicians to prescribe drugs for uses that the Food and Drug Administration has not approved. Since 2004, 31 federal cases alleging off-label promotional practices have settled, totaling over $12 billion. In this paper, Shapiro studies the effects of detailing on physician prescribing in the anti-psychotic category, which was the category most heavily targeted for off-label promotion. Using physician level panel detailing data combined with patient chart information, the researcher explores how detailing causes physicians to prescribe for on-label versus off-label uses. This question is important for considering whether regulators should spend more or less of their scarce resources pursuing such cases. Additionally, with the risk of huge fines, this question is relevant to managers in deciding the exact role of detailing in their marketing mix. Shapiro studies this question using the case of Seroquel, a prominent branded anti-psychotic marketed by AstraZeneca. During the sample, Seroquel received two informational shocks in the form of good news about its side effect profile relative to other treatments. These shocks were each immediately followed by large increases in detailing to primary care physicians, providing a significant amount of within-physician detailing that is useful in estimating detailing effects. Shapiro takes advantage of the fact that not every physician was detailed at exactly the moment of the informational shock to separate the direct effect of the information from the incremental effect of the detailing. He finds that while detailing did raise off-label prescriptions, the effect is small both in absolute terms and in relative terms. Over the course of the sample, detailing shifts the prescribing distribution more towards on-label uses.
Yufeng Huang, University of Rochester
This paper shows that using a product builds up specific human capital that is difficult to transfer to another product, and this creates a form of consumer switching cost. In the context of digital cameras, Huang uses novel data to directly characterize the cost of switching products, by measuring the changes in consumer picture quality. He estimates a structural model of demand with learning by doing, in order to quantify the effect of learning-driven switching cost on the demand for new products and on consumer welfare. Huang finds that 10% of consumer human capital is lost to every brand switching, for which a consumer is willing to pay $40 to avoid. This explains a quarter of the consumer's persistence in brand choice, and lowers her cross price elasticities by a factor of 3.
Amil Petrin and Boyoung Seo, University of Minnesota
The identification of discrete choice demand models since Berry (1994) and Berry, Levinsohn and Pakes (1995) (BLP) has relied on the assumption that observed product characteristics (excluding price) are uncorrelated with the unobserved product characteristics. Questions have been raised as to whether the assumption is reasonable and what the intuition is for identification of the heterogeneity in taste parameters. Noisy demand estimates are often the outcome with market level data. Since Spence (1975) economists have understood that firms' decisions about prices and characteristics are driven by their beliefs about the distribution of consumer preferences as both marginal and infra-marginal consumers enter the first-order condition for profit maximization. In this paper, Petrin and Seo show how to use these first-order conditions to estimate demand and supply parameters in a setting with product-specific unobserved demand and cost shocks. The researchers allow firms information sets at the time they choose characteristics to potentially include other firms' product characteristics, demand, and cost shocks, signals on all of these, or no information at all on them. Ex-post firms may wish they had made different decisions and the author' identification is based on the assumption that firms are correct in their choices on average (Hansen and Singleton (1982)). Using the same automobile data from BLP, they find that some of the slightly puzzling parameter estimates of BLP go away as all of their parameter estimates are of the correct sign. The researchers find significantly more precise estimates given the same exact data. They strongly reject the standard identification assumption; the conditional correlations of the demand and cost unobservables with observed characteristics equal 0.85 and 0.7 respectively.
Frank Wolak, Stanford University
Mar Reguant, Northwestern University and NBER
In many economic settings, counterfactual analysis can be difficult for two reasons: (i) we do not know how to compute the equilibrium of the game, or (ii) even if we know how to compute one equilibrium, the game might feature multiple equilibria, which are difficult to exhaustively characterize. Reguant proposes a new methodology to allow for counterfactual analysis even when these problems might arise. The method relies on determining valid (conservative) bounds to counterfactual outcomes that contain any outcome that could be sustained in equilibrium, i.e., any outcome that can be supported by a set of equilibrium constraints. To ensure that all potential solutions are considered, the researcher proposes to reframe equilibrium constraints as a relaxed mixed-integer linear program. Reguant shows that the framework can also be used to narrow down equilibria, by imposing additional equilibrium constraints. She provides examples related to discrete choice models, dynamic games and multi-unit auctions, which exemplify how the method can be used in a practical context.
Michael Dinerstein, University of Chicago, and Troy D. Smith, Stanford University
School policies that cause a large demand shift between public and private schooling may cause some private schools to enter or exit the market. Dinerstein and Smith study how the policy effects differ under a fixed versus changing market structure in the context of a public school funding reform in New York City. They find evidence of increased private school exit and reduced entry in response to the reform. Using a model of demand for and supply of private schooling, the researchers estimate that 32% of the reform's effect on school enrollments came from increased private school exit and reduced private school entry.
Mark L. Egan, University of Minnesota Carlson School of Management, and Ali Hortaçsu and Gregor Matvos, University of Chicago and NBER
Deposit Competition and Financial Fragility: Evidence from the US Banking Sector Egan, Hortaçsu, and Matvos develop a structural empirical model of the U.S. banking sector. Insured depositors and run-prone uninsured depositors choose between differentiated banks. Banks compete for deposits and endogenously default. The estimated demand for uninsured deposits declines with banks' financial distress, which is not the case for insured deposits. The researchers calibrate the supply side of the model. The calibrated model possesses multiple equilibria with bank-run features, suggesting that banks can be very fragile. They use their model to analyze proposed bank regulations. For example, their results suggest that a capital requirement below 18% can lead to significant instability in the banking system.