Differentiated Products Demand Systems from a Combination of Micro and Macro Data: The New Car Market

By Steven T. Berry, J. Levinsohn, Ariel Pakes. 1998

In this paper, we consider how rich sources of information on consumer choice can help to identify demand parameters in a widely used class of differentiated products demand models. Most important, we show how to use “second‐choice” data on automotive purchases to obtain good estimates of substitution patterns in the automobile industry. We use our estimates to make out‐of‐sample predictions about important recent changes in industry structure.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Counterfactual inference for consumer choice across many product categories (2019, background): “…common to include many latent variables describing user preferences; for example, in a mixed logit model, it is common to include a user-item random effect, as well as individual-specific preference parameters for price and other observed item attributes (e.g. Berry et al., 2004; Train, 2009).”
Testing the Production Approach to Markup Estimation (2020, background): “The demand approach to markups, exemplified by Berry et al. (1995) and Berry et al. (2004), cannot do so because models for firm competition and demand vary substantially across industries.”
Dynamics of Consumer Demand for New Durable Goods (2012, unknown): “27 See Petrin ð2002Þ and Berry, Levinsohn, and Pakes ð2004Þ for details on calculating weighting matrices when combining micro moments with aggregate moments. consumer demand for new durable goods 1199 the set of consumers who care very little about price.”
Beyond Truth-Telling: Preference Estimation with Centralized School Choice and College Admissions (2019, methodology): “may expect that the stability-based approach leads to a loss of information; in particular, we may lose some precision in estimating the substitution patterns when we allow for more flexible random utility models (Berry et al., 2004; Abdulkadiroğlu et al., 2017).”
Best practices for differentiated products demand estimation with PyBLP (2020, methodology): “Other common modifications to the BLP model that are beyond the scope of this article include 4 the pure characteristics model of Berry and Pakes (2007) and the inclusion of micro moments, as in Petrin (2002) and Berry et al. (2004a).6”
Discrete Choice Models with Multiple Unobserved Choice Characteristics (2007, unknown): “The random coecients or mixed logit approach was generalized in an influential paper by Berry, Levinsohn and Pakes (1995, 2004) (BLP from hereon) and applied to settings with a large number of choices.”
Ownership of intellectual property and corporate taxation (2014, background): “This is an open access article under heterogeneity in firms' location choices (see inter alia, Berry et al. (1995, 2004), Nevo (2001) and Train (2003)).”
The Welfare Effects of Coordinated Assignment: Evidence from the New York City High School Match (2017, background): “Berry, Levinsohn, and Pakes (2004), for instance, argue that data on top and second choices can be used to estimate these parameters by exploiting common characteristics between subsequent rankings for a given student.”