Semi‐nonparametric Estimation of Consumer Search Costs

Abstract

This paper studies the estimation of the cost of non-sequential search. We provide a new method based on semi-nonparametric (SNP) estimation that allows us to pool price data from different consumer markets with the same underlying search cost distribution but different valuations or selling costs. We show that pooling data from different markets increases the number of estimated critical search cost cutoffs at all quantiles of the search cost distribution, which increases the precision of the estimates. A Monte Carlo study shows that the method works well in small samples. We apply our method to a data set of online prices for memory chips and find that the search cost density is essentially bimodal such that a large fraction of consumers searches very little, while a smaller fraction of consumers samples a relatively large number of stores.

Price dispersion, chain heterogeneity, and search in online grocery markets

“…If we now use the same change of variable in the equilibrium profit equation y yields 17 Moraga-Gonzalez et al (2013) show that the model described here is nonparametrically identified if the number of searches does not tend to infinity.…”

Dynamic selection and distributional bounds on search costs in dynamic unit‐demand models

“…With only data on prices but not on individual consumers, a necessary limitation of our model, and others in the literature on which we build, is that consumer heterogeneity is one‐dimensional. Importantly, as Hong and Shum (), Moraga‐González and Wildenbeest (), and Moraga‐González, Sandor, and Wildenbeest () have shown, even with only the price distribution at hand, identifying features of the search cost distribution is still feasible. This heterogeneity leads consumers to search different numbers of firms to obtain prices, which in turn supports the mixed strategy pricing equilibrium among firms.…”

“…In this paper, we propose a dynamic model with forward‐looking consumers to estimate the population consumer search cost distribution. Like many other recent studies, such as Hong and Shum (), Moraga‐González and Wildenbeest () (henceforth MGW), Wildenbeest (), Moraga‐González, Sandor, and Wildenbeest (), and Sanches, Silva Junior, and Srisuma (), we focus on the case where only the distribution of prices in each period is observed, but no individual‐level data on consumers is available. We show that if consumers have the option to delay purchase until a later period, a dynamic selection problem exists that will cause estimates obtained from a static model to be biased.…”

“…() and Moraga‐González and Wildenbeest () used the equilibrium conditions on supply and demand to consistently estimate the search cost distribution within one period at a finite number of points. Moraga‐González, Sandor, and Wildenbeest () estimated the search distribution more completely on an interval by pooling multiple markets with the same search technology and exploiting the variation in valuations and marginal costs across markets. Sanches, Silva Junior, and Srisuma () used a minimum distance approach, modifying Hong and Shum () to create a consistent and asymptotically normal estimator.…”

A Generalized Model of Sales

“…Sorensen (2000), Lach (2002), Brown and Goolsbee (2002), Barron et al (2004), Lewis (2008), Chandra andTappata (2011), Pennerstorfer et al (2015), and Sherman and Weiss (2015), or ii) structural estimations, e.g. Wildenbeest (2011), Moraga-González et al (2013), Giulietti et al (2014), Allen et al (2014), and An et al (2015). or 'surplus' (Simester (1997), Hosken and Reiffen (2007), Wildenbeest (2011), Dubovik andJanssen (2012), Anderson et al (2015)). However, these papers only use it to compute sales equilibria in specific market settings, and do not use the associated profit function, π(u), to explore any general results or implications.…”