Xingjie Wei

Thematic seminars
big data and econometrics seminar

Xingjie Wei

Leeds University
Analysing automobile recalls using text mining
online
Date(s)
Tuesday, February 28 2023| 2:00pm to 3:30pm
Contact(s)

Michel Lubrano: michel.lubrano[at]univ-amu.fr
Pierre Michel: pierre.michel[at]univ-amu.fr

Abstract

The decision to recall is costly for firms. Despite its importance to customer choice, the value of consumer complaints in product recall decisions has not been investigated. We develop a novel measure of complaint relevance that denotes similarity between the description of a consumer complaint and the defect summary. We analyse the effect of complaint relevance on time to recall, and the role of product characteristics in moderating this relationship. Using a sample of 1161 recalls in the automobile industry in US between 1995 and 2019, we find that consumer complaints with high relevance to defect summary are more likely to be the attentional focus of decision-makers, thus leading to a faster recall. Specifically, the results show that complaint relevance decreases time to investigation and increases discovery to recall, while a recency effect exists. An increase in complaint relevance will result in a much faster recall when the level of component sharing is higher. Likewise, the impact of complaint relevance on time to recall is greater for car models that are at least three years old. Our findings are robust to alternative explanations when using vehicle recall, component variety, and industry crisis as moderating variables. By uncovering a new predecessor to product recall decisions, this study provides valuable insights to both manufacturers and regulators.

More information