Product Differentiation and Mergers in the Carbonated Soft Drink Industry

Abstract

Standard-Nutzungsbedingungen:Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden.Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen.Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in AbstractEconomists increasingly rely on structural methods to assess mergers in di¤erentiated products industries. They typically estimate aggregate demand models using market level data, requiring a trade-o¤ between the scope of products considered and the implied underlying consumer behavior. Instead, I use a unique micro data set containing detailed transaction-level information for a panel of household purchases of carbonated soft drinks. I estimate household demand for soft drinks, allowing for more realistic multi-item purchase behavior for a large set of product alternatives. Integrating across the predicted household purchases, I combine aggregate demand with a model of static oligopoly. Finally, I simulate the competitive impact of several soft drink mergers on equilibrium prices and quantities. I compare my results with the popular aggregate mixed logit model. (JEL L15, C5, D4)

Multi-Category Competition and Market Power: A Model of Supermarket Pricing

“…The K × 1 vectors µ d and p d collect the tastes and prices (respectively) that are relevant for each category given store-category choice d. Λ is a symmetric K × K matrix of parameters, common across consumers. 18 The first two terms in (1) are variable utility (in terms of q): u(q, d) is gross utility from the categories bought, and p d q is the consumer's total payment for them. The price sensitivity scalar α corresponds to the marginal utility of expenditure on non-supermarket consumption (in which utility (1) is quasi-linear).…”

“…We develop a model that is multiple-discrete-continuous, in that the consumer can choose one or more (discrete) stores and makes a (continuous) non-negative choice of quantity for every category. We build on the existing literature on multiple-discrete choice (see Hendel (1999), Dubé (2005), Gentzkow (2007)), and discrete-continuous choice (see Dubin and McFadden (1984), Haneman (1984), Smith (2004)). 8 Our multi-category multi-store model brings together the empirical literature that measures market power for a single supermarket category (e.g.…”

Household willingness to pay for organic products

“…Individual food products (de…ned by a unique bar code) are grouped into categories such as "fresh lamb", "tea", "olives". 6 In total there are 146 such groups. We restrict our analysis to those food categories where more than one food product is available as organic and where we observe more than ten purchases of these goods over the year -this includes 98 food products.…”

Estimating households' willingness to pay

“…Methods have been developed to analyse these cases. Scotchmer (1984), Dubin and McFadden (1984), Haneman (1984), Smith (2004), Beckert, Gri¢ th and Nesheim (2009) analyse demand with both discrete and continuous margins and Hendel (1999) and Dube (2005) analyse multiple dis-crete choices. However, all of these methods remain intractable in cases where the consumer basket contains a large number (i.e.…”

“…Work using these data has looked either at the aggregate basket of groceries (e.g. Smith, 2004) or at single product categories, for example, breakfast cereals (Nevo, 2001), ketchup (Pesendorfer, 2002), yoghurt (Ackerberg, 2001) or carbonated soft drinks (Dube, 2005). and across households.…”