James W. Roberts
 Assistant Professor, Department of Economics, Duke University and Faculty Research Fellow of NBER

  photo

Education:

Ph.D., Economics, Northwestern University
BA, Economics, Davidson College

Fields of Specialization:

Industrial Organization, Applied Microeconomics

Contact Information:

227B Social Sciences
Box 90097, Durham, NC 27708-0097
919-660-1822 (office)
919-684-8974 (fax)
j.roberts"at"duke"dot"edu


Curriculum Vitae

Curriculum Vitae (Updated Spring 2013)

Published and Forthcoming Papers

When Should Sellers Use Auctions? (with Andrew Sweeting)
American Economic Review, Forthcoming.

Robust Firm Pricing with Panel Data (with Ben Handel and Kanishka Misra)
Journal of Econometrics, 174(2), 2013.
Copyright 2013: Elsevier B.V.

Can Warranties Substitute for Reputations?
AEJ Microeconomics, 3(3), 2011.
Copyright 2011: American Economic Association

Network Structure of Production (with Enghin Atalay, Ali Hortacsu and Chad Syverson)
Proceedings of the National Academy of Sciences, 108(13), 2011.
Supporting material here.
Copyright 2011: Proceedings of the National Academy of Sciences.

Entry into Auctions: An Experimental Analysis (with Seda Ertac and Ali Hortacsu)
International Journal of Industrial Organization, 29(2), 2011.
Copyright 2010: Elsevier B.V.

Working Papers

Regulating Entry Through Indicative Bidding (with Vivek Bhattacharya and Andrew Sweeting)
Submitted.

Bailouts and the Preservation of Competition (with Andrew Sweeting)
Submitted.
Replaces: Competition versus Auction Design
Which Replaced: Entry and Selection in Auctions (NBER Working Paper 16650).

Unobserved Heterogeneity and Reserve Prices in Auctions (Updated Version Coming Soon)
Revision Requested.

Gender Disparity in Urology: Preferences, Competition, and Quality of Care (with Ryan McDevitt - Updated Version Coming Soon)
Revision Requested.

Speculators and Middlemen: The Role of Intermediaries in the Housing Market (with Patrick Bayer and Christopher Geissler)
NBER Working Paper 16784.

Research in Progress

Airline Mergers and the Potential Entry Defense (with Andrew Sweeting)

Focusing on airlines, we estimate an empirical model that allows for selection along cost or quality dimensions and we find it to be quantitatively important. We show how a selective entry model helps us to explain the stylized fact in the data that airline mergers have tended to lead to higher prices on the most affected routes with only very limited being induced. We also use our estimated model to perform counterfactual experiments on mergers.

An Empirical Model of Dynamic Limit Pricing: The Airline Industry (with Chris Gedge and Andrew Sweeting)

Theoretical models of strategic investment often assume that information is incomplete, creating incentives for firms to signal information to deter entry or encourage exit. However, the very simple one-shot nature of these models has limited the scope for testing whether these models can quantitatively or even qualitatively fit the data. We develop a fully dynamic model with persistent asymmetric information, where an incumbent has incentives to repeatedly signal information about its costs to potential entrants. The model is well-suited for empirical work in that it has a unique Perfect Bayesian Equilibrium under a standard form of refinement, with strategies that can be computed quite easily. We are in the process of using our model to test whether dynamic limit pricing can explain why a dominant incumbent airline drops its price when Southwest becomes a potential entrant on a route. The current version of the paper uses some existing tests to show that there is strong evidence of some form of strategic pricing behavior on the routes in our sample.

Competition and Dynamic Revenue Management in a Perishable Goods Market (with Andrew Sweeting)

Most revenue management models assume that the seller is a monopolist. In this project we investigate the optimal strategy of a large seller who knows that his pricing and listing decisions can move the future distribution of market prices, using a dynamic model in continuous time.