Romain Espinosa, CREM, CNRS, Université de Rennes 1
In this paper I investigate the role of e-reputation mechanisms on illegal platforms that specialize in drug sales. In detail, I ask whether online reputation systems can limit the risk of scamming (i.e. fraud) by dishonest sellers, and thus prevent Akerlof-like market destruction. I do so by analyzing all published offers covering four types of drugs collected from the second-largest platform operating in March 2017 (the Hansa marketplace). I first show that the distribution of seller reputation is very similar to that observed on clearnet platforms such as eBay. I then carry out multivariate analyses to estimate the relationship between reputation and pricing, and show that the more-reputed sellers charge significantly higher prices for three types of drugs (Weed, Hash and Ecstasy). I then develop a theoretical model of reputation-building on Darknet Markets that I estimate for the Hansa marketplace. I show that the three types of drugs have relatively low scamming risks, with the average probability that a random seller effectively sends the ordered good of over 83%. With respect to consumer welfare, my results indicate that the Weed and Ecstasy markets leave consumers with about 25% of their surplus, while the Hash market extracts almost all of the consumer surplus (with only 1.4% left to consumers). Finally, I discuss the potential impact of the shutdowns of the two leading platforms in 2017 on the scamming risks on future platforms. The probability of good shipment is likely to rise from 2.7 to 9.7%, depending on the type of drug and the scenario considered, and average price per sale falls from 2.9 to 14.2%.