1 Introduction Automobile emissions

1 Introduction Automobile emissions are partly responsible for air pollution and global warming. If car owners care about the emissions of their cars, we expect the used-car market to react to the arrival of new information about the emissions of cars. One possible reaction is sorting based on preferences for environmental quality (see, e.g., Hendel and Lizzeri, 1999, for a comprehensive discussion of sorting in markets for durable goods under heterogeneous preferences). Car owners with high preferences for environmental quality could sell their cars to individuals with low preferences for environmental quality. In this study, we exploit the decline in the observed environmental quality of Volkswagen (VW) diesel cars after the disclosure of the stunning VW emissions scandal to investigate reactions to information disclosure. On September 18, 2015, the US Environmental Protection Agency (EPA) disclosed the installation of defeat device software in the emission control systems of VW diesel engines (EPA, 2015). This disclosure reduced the observable environmental quality of VW diesel cars because the defeat device negatively affects the environmental and engine performance of the manipulated cars. Furthermore, the durability of several parts of the exhaust system (e.g., the catalytic converter) could significantly decline after retrofitting, thus increasing maintenance costs. Since market participants did not expect the information revealed by this disclosure, the resulting decline in observable environmental quality is a quasi-experimental exogenous shock to the used-car market. Several studies show that car buyers’ willingness to pay is sensitive to fuel prices (e.g., Allcott and Wosny, 2013; Busse et al., 2013; Sallee et al., 2016). These studies argue that car buyers take the future fuel costs of vehicles with different fuel economies into account. In contrast to these studies, we investigate the reactions of potential car sellers to information disclosure about the environmental quality of cars, which is related to fuel economy. Tadelis and Zettelmeyer (2015) investigate the random disclosure of information (for potential bidders) about the quality of used cars during online auctions. The unexpected disclosure of new information increases sales probability and revenue