EconCS Seminar

 

Lecturer: 

Dr. Amir Ban (Hebrew University)

Title: 

Unending Sequential Auctions, with Applications to Bitcoin

Abstract: 

Sequential auctions for identical items by unit-demand, private-value buyers are common, and mostly occur periodically, without end, as randomly arriving bidders at least replace bidders who leave. We model bidder uncertainty by a probability of having to leave the auction in each period. Modeling the sequential auction as a Markov process, we show that it reaches a steady state. When uncertainty is zero, the steady state emulates a posted-price mechanism: Bidders with value above a threshold will almost surely win an item, by bidding the threshold price repeatedly, while bidders with a value below it will almost surely never win an item. The item price in equilibrium is the threshold value that balances “supply” (bidders with value above the threshold) and “demand” (auction winners). We show that Bitcoin transaction fees are a good fit for our model with a slight uncertainty.

When uncertainty is not zero, the same threshold value holds, but is less sharp, becoming fuzzier as uncertainty grows. The uncertainty benefits low-value bidders, those whose value is below the threshold, giving them a significant chance of winning an item. Surprisingly, up to some value limit, higher-value bidders also benefit from uncertainty, with a lower equilibrium bid and higher expected utility. Often, the consequence is that bidder uncertainty harms the auctioneer’s utility.

Location: 

Room 130, Feldman Building, Edmond J. Safra Campus.

 

Click here to add the EconCS Seminar to your Google Calendar