ABSTRACT
In this paper, we apply a recently popularized methodology called ‘matched listings’ to eBay API data to study the effect of enabling eBay’s bargaining feature on auction outcomes. Our results indicate that enabling the bargaining feature increases sale probability by 3.17%. In addition, hypothesis testing indicates this estimate is consistent with eBay’s claim where enabling bargaining increases sale probability by 5%. Finally, we find that enabling bargaining has no impact on transaction price. We provide an innovative evaluation of whether the matched listings methodology works well. We find that this methodology is prone to endogeneity with longer time windows, causing a downward bias when estimating the treatment effect of enabling bargaining.
Acknowledgments
The authors would like to acknowledge the helpful comments from Xiaoyong Zheng and Robert Hammond.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1 We also use to index seller since one of conditions for matched sets is the group of listings need to be listed by the same seller.
2 We run robustness checks using 30-day, 14-day, and 7-day windows. displays the distribution of time windows for the full sample.
3 While we do not observe bargaining history, this is unnecessary information for our research questions.
4 Based on the statistic is −1.21, we fail to reject the null hypothesis of .