Consumer Behavior, Monopolistic Competition, and International Trade: CES Redux? Paolo Bertoletti y Pavia University and IEFE Paolo Epifani z Bocconi University and IGIER February 2012 Abstract We argue that monopolistic competition models featuring pro-competitive e/ects, such as the model in Krugman (1979) or models assuming linear demand systems, build on dubious assumptions about consumer behavior and are therefore potentially misleading. In any case, departing from CES preferences in an otherwise standard Dixit-Stiglitz monopolistic competition setting seems to lead to controversial results. We also consider alternative settings featuring, i.e., consumers preference for an ideal variety, strategic interactions and heterogenous rms, and nd that a pro-competitive e/ect may plausibly arise only in the presence of a small number of rms, an assump- tion clearly at odds with monopolistic competition. Our results suggest that existing monopolistic competition models are not suitable to study the e/ects of competition on markups, thus making a point in support of CES preferences, which overlook the latter e/ects. Moreover, we show that trade-induced selection e/ects  la Melitz (2003) are generally robust to the assumptions about preferences. This suggests that mo- nopolistic competition is better suited to analyze frameworks in which larger markets select more aggressively on productivity rather than forcing rms to move down their average cost curves. JEL Classication : F1 Keywords : Monopolistic Competition; Consumer Behavior; Elasticity of Substitu- tion; Markups; Anti-Competitive and Anti-Selection E/ects of International Trade. A previous version of the paper circulated under the title: "Price-Increasing Monopolistic Competition? The Case of IES Preferences". We are grateful to Eileen Fumagalli and Clara Poletti, who were initially part of the project, and to Kristian Behrens for his comments on a previous draft. We also thank Luca Salvatici, Bob Staiger, Piero Tedeschi and seminar participants at many venues for helpful comments and discussions. All errors are our own. y Department of Economics and Business, Universit di Pavia, Via San Felice 5, 27100 Pavia (Italy), and IEFE, Bocconi University. E-mail: paolo.bertoletti@unipv.it z Department of Economics, IGIER and Ba¢ Centre, Universit Bocconi, Via Rntgen 1, 20136 Milan (Italy). E-mail: paolo.epifani@unibocconi.it. 1