While Professor Wylie provides substantial insight into the efficiencies of the Internet auction, I'd like to suggest that a countervailing influence may be at work: the cost of asymmetric information.

Unlike a brick and mortar auction where the problems, if any, relate to the authenticity of the good sold, Internet auctions compel neither a legitimate seller nor a legitimate buyer.

Case in point: I tried to sell an old laptop on eBay. My reserve price set at $1k, I watched with excitement as the bids climbed to $1,240. A few hours later, the auction closed and I dispatched the customary email to the winner. No response. I emailed twice more per eBay policy. No response--from the winner, that is. I DID receive an email from another eBaysian, informing me that the person who had bid up my laptop had bid fraudulently.

Certainly this is only one example but examples such as these number nearly one-third of my total eBay experiences (about 100 in total--I maintained a small web-design business on eBay). The reality is that in the amorphous cyber-world of the online auction, at least one transactor has less information than the other. And we know that in cases of asymmetric information, risk abounds and where risk abounds, costs rise.

It could be argued that eBay's "user feedback profile"--a record of users' feedback on a buyer or seller that any other buyer or seller can access at any time--helps to resolve asymmetric information. In fact, I won't buy from any eBay seller who doesn't have a clean and comprehensive feedback profile. However, if there's just one of me in every eBay category, that is, if there's just one selective eBay auctioneer requiring better information on whom s/he's doing business with, then the Walrasian auctioneer is in jeopardy. If I'm selling and observe that a bidder of dubious feedback has bid on my item, then I might close the auction down early, preferring to swallow the slotting cost but avoiding the revenue-share. If I'm buying, then I won't bid on goods offered by profile-less auctioneers or auctioneers with complaints in their profiles. In either case, my actions as the selective auctioneer ripple through the market causing problems for that poor Walrasian auctioneer.

Protracting the argument a step further, we have to wonder if we wouldn't encounter adverse selection and moral hazard problems after a while. Presuming that veteran eBaysians do business strictly with other veterans, then what do newbies have to look forward to? Even if we all grant the basic good of humankind, we would still have to agree that newbies draw from a more tainted pool, one relatively more infected with deadbeats, frauds, and cheaters. The problem of adverse selection here becomes even more noteworthy given the fact that newbies constitute the clear majority of eBay users.

One of eBay's solutions is to provide insurance by way, incidentally, of Lloyd's! (Lloyd's was the subject of a Tuck first-year case in Managerial Economics). Covering bad purchases up to $200, less a $25 deductible, eBay provides coverage for all eBay users in good standing at no extra charge. Here we see the moral hazard problem: eBay provides the cheater an opportunity to commit yet another fraud. After all, who better to seek out other cheaters than the cheater himself? He knows all the red flags to look for because he raises them himself. And he knows how to disappear if the auctioneer does turn out to be legitimate. It's an easy $200 for the cheater and yet another problem for the Walrasian auctioneer interested in a good market-clearing price.

On the positive side, eBay has gone beyond feedback profiling and insurance protection to combat information asymmetry. Introducing regional auctioning at no extra cost, sellers can now specify the region in which their products are available for sale and pick-up by the buyer. One might suppose that such a service makes the most sense where vehicles and fragile antiques are on the auction block but few Porsches and antique Ming vases grace the local listings. What we find instead are an abundance of higher-priced, heavily bid-upon items. This suggests that when the stakes are high and the risks of a bad transaction are even higher, buyer and seller have a greater interest in meeting each other in the non-virtual world. Even if buyer and seller never actually connect, the mere potential for this meeting and the accountability it implies, compel better information from more transactors more of the time. The result is transactions with all the advantages of an Internet auction and none of the complications of information asymmetry. Of course, the limitation of regional auctioning is its strength: not all auctioneers can or should take advantage of this more physical isotope of the virtual auction experience.

Regards
Mike Reynolds T'01