SEMI‐NONPARAMETRIC ESTIMATION OF CONSUMER SEARCH COSTS
研究了如何利用多个市场的价格数据来估计消费者搜索成本的分布，提出半非参数估计方法，并应用于内存芯片在线价格数据，发现搜索成本密度呈双峰分布。
SUMMARY This paper studies the estimation of the distribution of non‐sequential search costs. We show that the search cost distribution is identified by combining data from multiple markets with common search technology but varying consumer valuations, firms' costs, and numbers of competitors. To exploit such data optimally, we provide a new method based on semi‐nonparametric estimation. We apply our method to a dataset 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, whereas a smaller fraction searches a relatively large number of stores. Copyright © 2012 John Wiley & Sons, Ltd.