Package-size level demand for carbonated soft drinks and its implications on market power

Date

2021-05

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

This study analyzes manufacturer-level carbonated soft drinks (CSDs) demand and market structure at two aggregation levels: brand level and bottle-size level by comparing their findings and implications. We use the BLP (1995) approach to estimate the demand and use the resulting estimates to evaluate market power for CSDs brands. Two pricing conducts, namely Bertrand Nash and joint-profit maximization, are assessed under the two aggregation levels. Aggregation level changes the implications of CSDs demand estimates. Demand-side empirical results for the bottle-size level do not imply the same directional effects as the brand level for the price, calorie, and caffeine contents when decomposed by income and the number of children in the household. For example, low-income households are less sensitive to CSDs prices at the brand level than middle-income and high-income households. In contrast, at the bottle-size level, high-income households are less price-sensitive than other income categories. In terms of elasticities, the bottle-size level's own-price elasticities are more elastic than those at the brand level. Furthermore, at both levels, cross-price elasticities show consumers have some brand loyalty and are more responsive to the leading brands' changes. For the pricing conducts, the results suggest the Lerner index, a measure of the market power, is quite sensitive to the price levels. Hence, when we use the demand estimates at the aggregate brand level, the percentage markups are higher than when we use the disaggregated bottle-size level, which is the more realistic one. Hence, the bottle-size level can be more accurate for the estimation because of having more accurate prices. Therefore, the market power assessment might be more accurate when the brands' prices are at the bottle-size level. Vuong’s (1989) test results indicate that the Coca-Cola Co., PepsiCo, and Dr Pepper Snapple Group compete in Bertrand-Nash's pricing conduct, rejecting the collusive price behavior. The Bertrand-Nash game yields Lerner indices that vary between 40% and 62% at the brand level and 20% and 39% at the bottle-size level. Our results support other authors' conclusion that the high Lerner index is not due to prices' collusion, but rather an implication of non-price competition, such as product differentiation and advertising.


Embargo status: Restricted until 06/2026. To request the author grant access, click on the PDF link to the left.

Description

Keywords

Random Coefficient Logit Model, New Empirical Industrial Organization, Demand, Differentiated Demand, Differentiated Good, Market Power

Citation