Industrial organization
Applied econometrics
Development economics

Industrial Organization
Applied Econometrics
Microeconomic Theory
International Trade

May, 1998 (Oral) Industrial Organization, International Economics
May, 1997 (Written) Microeconomic and Macroeconomic Theory

Essays on Spatial Product Differentiation

Professor Steven Berry
Professor Christopher Timmins
Professor Patrick Bayer

Summer 2001

M.Phil. (1998), M.A. (1997), Department of Economics, Yale University
B.A. (1995), Magna Cum Laude, Economics and Mathematics, Franklin & Marshall College

Ryoichi Sasakawa Young Leaders Fellowship, 1998
Yale University Graduate Fellowship, 1996–1998
Phi Beta Kappa Honor Society, Minnie Zeid (Economics) and Kershner (Mathematics) Prizes, Franklin & Marshall College, 1995

Teaching Assistant, Public Responses to Market Imperfections, Yale University, 1999
Teaching Assistant, Portfolio Choice, Yale University, 1998

Research Assistant to C. Timmins (Economics), Yale University, 1999–2000
Developed estimation algorithm for discrete choice model of crop variety adoption behavior. Presented results at 24^th International Conference of Agricultural Economists (Berlin, 2000).

Analyst, Lexecon Limited, London, United Kingdom, 1999
Contributed to economic report submitted to the European Union’s Competition Directorate regarding a pending merger between two postal services companies.

Research Assistant to L. Wantchekon (Political Science), Yale University, 1998–1999
Pursued research on the role of policy makers in promoting credible privatization programs with a focus on ownership transfers in Eastern Europe and Latin America.

Research Fellow, Centre for European Economic Research, Mannheim, Germany, 1998
Prepared econometric study analyzing the role of information technology in changing labor demand and productivity of West German service firms.

Research Assistant, National Economic Research Associates, White Plains, NY, 1995–1996
Employed in the Securities Practice responsible for performing economic research and analysis in connection with securities litigation cases.

Publications and Working Papers:

"The Impact of Information Technology on High-Skilled Labor in Services: Evidence from Firm-Level Panel Data," forthcoming, Economics of Innovation and New Technology 10, 2000, with Martin Falk.

"Geographic Differentiation and Market Structure: The Video Retail Industry," manuscript, Yale University, 2000.

Work in Progress:

"A Model of Sequential Entry and Location Choice in Retail Markets."

"The Role of Spatial Correlation in Measuring the Impact of Global Climate Change on Brazilian Agriculture," with Christopher Timmins.

"Spatial Differentiation as a Source for Price Discrimination in the Wholesale Market for Steel," with Rupa Athreya.

While competition in differentiated product markets has been extensively studied in the empirical industrial organization literature, questions of product introduction and optimal positioning in product characteristic space have received considerably less attention. Using location in geographic space as an easily quantifiable, observable measure of product differentiation, my dissertation attempts to fill this gap by analyzing the impact of spatial differentiation on firms’ entry and location decisions.

The first chapter develops a static equilibrium model of market structure and location choice under imperfect information, which circumvents some of the difficulties encountered in earlier studies of entry into markets with a large set of product types. Further, the model is solved numerically using a fixed-point algorithm that computes the equilibrium and nests it in a maximum likelihood estimation routine. The second chapter adopts a more realistic sequential equilibrium perspective to incorporate firms’ expectations of their rivals’ future actions into the entry process. These two models are estimated and contrasted using a unique data set, which combines detailed demographic data with geo-coded firm-level data on realized entry and location choices from the video retail industry.

Earlier empirical studies of equilibrium market structure, such as Bresnahan and Reiss (1991) and Mazzeo (1999), develop perfect information models, in which firms compare profits to be earned in every possible market configuration to make entry and possibly product-type decisions. In contrast, the static model presented in Chapter 1 formalizes firms’ endogenous location choices in a one-shot Bayesian game of asymmetric information about idiosyncratic sources of profitability, such as managerial talent. In equilibrium, each firm’s conjecture of its competitors’ behavior is matched to their conjecture of the firm’s own behavior. The Bayesian framework entails notable advantages over its complete information counterpart in terms of its ease of verifying and computing equilibria. By introducing a solution concept based only on equilibrium beliefs, instead of pure location strategies, the framework allows firms' conjectures to be drawn on a realistic, large-dimensional set of product-type choices. Incorporating such a large-dimensional choice set is infeasible in complete information games.

The empirical results demonstrate the endogeneity of market structure and firm profitability: while favorable demand characteristics provide an incentive for stores to cluster together, the intensified competition associated with increased clustering drives profits down. The results furthermore imply large payoffs to geographic differentiation since only the closest rivals exert strong competitive pressure on store profitability. The paper contrasts these results with those from a conventional model of entry without product differentiation and compares predictions of location patterns and total number of entrants. The results offer strong statistical support for the joint model of market structure and product differentiation. In extensions to this paper, I hope to incorporate systematic differences between firms, such as chain affiliation, which cause location specific profitability to be correlated with observed firm characteristics.

The second chapter addresses some of the limitations that arise in a static model of market structure and location choice. While entry or introduction of new products involves significant fixed costs, the entry decision is by no means irreversible. Repeated waves of the retail data used in the first chapter show a trend towards entry by chain affiliated video retailers at the expense of single-outlet stores. These facts motivate a model of sequential entry in which firms’ location choices are conditioned on the history of the entry-generating process and firms’ expectations of its rivals’ future actions. Contrasting the results from such a model to the corresponding results from the static equilibrium model provides insights into the importance of firms’ forward-looking behavior in driving industry conduct. In particular, the results allow for an evaluation of motives such as entry-deterrence and product space foreclosure in driving location choice.