Demand Heterogeneity: Implications for Welfare Estimates and Policy

Thumbnail Image

Persistent link to this item

View Statistics

Journal Title

Journal ISSN

Volume Title


Demand Heterogeneity: Implications for Welfare Estimates and Policy

Published Date




Thesis or Dissertation


This dissertation is comprised of three essays, the first two of which are coauthored with Kevin R. Williams. In the first chapter, we develop new empirical methodology to estimate consumer demand and across-market consumer demand heterogeneity when faced with products that have zero sales at the local market level, but have positive sales at the aggregate (national) market level. The absence of sales is a common problem among data sets with a large number of products and this problem is exacerbated when markets are narrowly defined. Observing zero sales creates two major empirical issues. First, standard estimation techniques rely on the Law of Large Numbers in sales. When the Law of Large Numbers does not apply, mechanically forcing the estimation will result in biased demand estimates. Second, failing to account for small sample sizes may create spurious heterogeneity. That is, the randomness generated by small sample sizes may overstate the degree to which demand differs across locations. Alternatively, aggregating over markets to obtain a large sample size will obscure the heterogeneity across local markets. We propose a modification to Berry (1994) and Berry, Levinsohn, Pakes (1995), where both local and aggregate level sales information are used to recover geographically varying demand. This new estimation approach is easy to implement and we will show, using Monte Carlos exercises, that it fits the data well. In the second chapter, we apply this new methodology to examine the welfare benefit of the increased access to variety from online retail. The proliferation of online retail has greatly increased consumers' access to variety. The value of this additional variety depends crucially on the extent to which local demand can be captured by local retailers. The existing literature has found huge welfare benefits from online variety, but these studies have been limited by national level data. As a result, they are unable to speak to the differences in demand across locations. Using an original data set of online shoe sales, we show that failing to account for across-market heterogeneity can greatly overstate the consumer welfare gain from increases in product variety. In the third chapter, I develop a theory of non-collusive basing-point pricing under Bertrand competition. Basing-point pricing occurs when the delivered prices faced by a consumer is determined by the consumer's distance from a common location, known as the basing point. That is, the price faced by the consumer is equal to the price charged by firms at the basing point plus the cost of transportation from the basing point to the consumer's location regardless of the actual location of the selling firm. The existing non-collusive theories of basing-point pricing rely on the exploitation of market power by non-basing point firms to obtain the basing-point result. I show that basing-point pricing is consistent with Bertrand competition by non-basing point firms when these firms face increasing marginal costs of production or capacity constraints.



University of Minnesota Ph.D. dissertation. July 2015. Major: Economics. Advisors: Thomas Holmes, Amil Petrin. 1 computer file (PDF); vii, 100 pages.

Related to




Series/Report Number

Funding information

Isbn identifier

Doi identifier

Previously Published Citation

Suggested citation

Quan, Thomas. (2015). Demand Heterogeneity: Implications for Welfare Estimates and Policy. Retrieved from the University Digital Conservancy,

Content distributed via the University Digital Conservancy may be subject to additional license and use restrictions applied by the depositor. By using these files, users agree to the Terms of Use. Materials in the UDC may contain content that is disturbing and/or harmful. For more information, please see our statement on harmful content in digital repositories.