Price Discontinuities in an online used Car Market∗

We examine empirically whether individuals evaluating used cars efficiently aggregate all relevant information on its constituent characteristics. Based on detailed field data on almost 90,000 used car offers in a large online market-place, we provide evidence for biased information processing. While the pre-cise date of first registration, i.e., its “age”, is publicly and prominently stated for each car, we identify an amplified value adjustment for otherwise identi-cal cars at year-count changes. These discontinuities indicate that individuals over-react to the figure displayed in the latter, while underrating the finer in-formation on a car’s age as conveyed through the month of first registration. Moreover, we are able to replicate the findings from Lacetera et al. (2012) and find discontinuous drops in prices at 10,000km odometer thresholds. While the latter finding, as suggested by Lacetera et al. (2012), is consistent with a left-digit bias in the processing of numerical information, the first finding can-not be explained by this. Our findings underline that information-processing heuristics might also matter in markets with large stakes and easily observed information