My recent post on private auctions for new gTLDs under contention generated many responses from applicants. Here I follow-up with answers to the questions raised.

Q: ICANN recently announced that it is waiting until March 1 to announce the results on “Confusingly Similar” strings. Does this impact your schedule for the First Applicant Auction.

A: Yes. It is essential for the uncertainty about confusingly similar strings to be resolved before the first auction. We won’t learn what can be auctioned until this announcement. As such it makes sense to postpone the first commitment date until early March, which would mean that the first auction cannot take place until the end of March. We are hoping that there will not be further delays in the process.

Q: My understanding of your “simultaneous ascending clock auction” is that as the auction progresses and the buyer of a string is identified that the seller’s share of the buyer’s payment is added to the seller’s deposit and therefore expands the seller’s bid capability for other strings. Why would a bidder without budget constraints like Google participate in this auction? Shouldn’t Google wait until the Last Resort Auction? Why would Google fund competitors?

A: This is a subtle question involving difficult tradeoffs. Some of these tradeoffs are carefully addressed in our research paper. What we find is that Google and other well-capitalized bidders are better off participating in the applicant auction for the same reason as others. Compared with the Last Resort Auction, the Applicant Auction provides: 1) early resolution of string contention, so that investments in the domain can begin earlier and value created sooner, 2) a lower payment for the buyer, and 3) a positive payment for a seller. These results are strong and robust to a variety of assumptions about bidder values.

One of the standard economic assumptions is that all bidders have access to capital markets to fund purchases. Economists don’t assume that all companies have equal access to capital markets, rather we assume that the differential access is a proper reflection of the risk inherent in the investment. Capital is less expensive when the investment is less risky.

Capital markets take into account the balance sheet of the company. Thus, a company with a rich portfolio of TLD applications is valuable in the capital market. These assets facilitate the company’s ability to acquire further cash as either debt or equity. That is, capital markets value the company’s portfolio of applications even before any private auctions. What the private auctions do is resolve some of the uncertainty about value. Thus, as time goes by, capital markets will have better information with which to value the company’s risky asset. This reduction of uncertainty makes the asset (slightly) more valuable.

Thus, the only benefit to Google or other well-capitalized company of waiting for the Last Resort Auction is the possibility that by doing so uncertainty about the value of applications will be increased so much that the competitive advantage from this higher uncertainty creates more value than the direct and obvious value losses coming from late resolution, higher buyer payments, and no seller profits. To me, it is highly implausible that this would be the case. The direct impacts are first order, whereas the competitive advantage coming from greater uncertainty is highly speculative. Further, Google or any other company cannot unilaterally prevent the revelation of information about TLD values. Google has no ability to prevent the vast majority of strings from resolving early. This early resolution will reduce uncertainty not just about the non-Google strings, but all strings, since string values likely are positively correlated.
Bottom line: The “companies won’t fund competitors” critique of the Applicant Auction has been grossly overstated. The critique essentially ignores how actual capital markets work.

Q: This sure sounds like a lot of economic theory. Maybe its works in the classroom, but it won’t work on the street.

A: You should be quite hesitant to dismiss science and knowledge simply because it sounds academic. The field of auctions has been extensively studied in theory, in the lab, and on the street. The same goes for finance and our understanding of capital markets. Our analysis of these issues is based on some of the most-tested models in economics.

The theory has been shown to be highly relevant to practice. This is why my colleagues and I are frequently asked to advise the boards and top management of companies participating in high-stake auctions. Yes, much academic work has little relevance for practice. But the areas of auctions and finance are highly applicable to practice. Ignore this science at your peril.

Q: In the mock auction I participated in Santa Monica in December, I noticed that each participant got to see the pricing and demand information on all the strings in the auction, not just those strings the participant was able to bid for. It seems a better approach would be to limit demand and pricing information about a string to only those bidding on the particular string.

A: Yes. The information policy is for all bidders to see the demand and pricing information for all strings in the auction. This was a conscious choice of the design team. We certainly recognize that some participants would like less price transparency. However, at this point we believe that the benefits of greater transparency dominate any downside. Note that this level of transparency is quite common in other transparent markets, such as those for financial securities, electricity, radio spectrum, and most commodities. Open auctions such as Sotheby’s and Christies’ also provide this pricing information to all participants. The information policy is not limited to fungible assets like financial securities, but is used in many long-term investment markets, such as those conducted for spectrum or long-term electricity products. These latter auctions are more similar to the Applicant Auction. Also it appears likely that the Last Resort Auction will adopt the policy of revealing demand and price information on all strings, as this is the standard approach with simultaneous ascending clock auctions. There are benefits to maintaining a parallel structure with the Last Resort Auction.

It is natural that some applicants would prefer the less transparent information policy. In particular, an applicant with a large portfolio of strings can reap the benefits of excellent price discovery even when the bidder’s information is limited to only the strings the bidder holds. In sharp contrast, a bidder with only one or two strings is certainly harmed by not having the pricing information of the strings the bidder does not hold.

Thus, our proposed information policy serves to level the playing field between large and small bidders. This level playing field improves competition and efficiency. It also supports the transparency objective.
We have been unable to identify compelling reasons for not disclosing full price information. One that we have heard is that it may lead to strategic behavior with respect to budgets, such as a bidder attempting to exploit budget constraints revealed from the pricing information. However, this potential problem is readily addressed by not disclosing deposits or winner identities until the conclusion of the auction. Even if a bidder is suspected of being a buyer in early rounds and therefore perhaps capital constrained, the bidder can always relax the constraint with an additional deposit to the escrow account. This extra deposit is not observed by other bidders.

For these reasons, we continue to recommend full price information. However, we are open to arguments in favor of alternative information policies provided the arguments are framed in why the alternative supports the objectives of efficiency, transparency, simplicity, and fairness.

Q: I’m a community applicant. I don’t want money. I want this domain for my community, and this community use creates the most value to society.

A: You may be right. ICANN recognizes that there are factors other than highest willingness to pay that should resolve string contention in some circumstances. This is the motivation for special types of applications, such as community and geography. For these ICANN conducts a separate process to determine what is best for society. Our private auction has nothing to do with that.
The private auction is focused on those domains for which the community/geographic/trademark elements are not dominant. In those cases, the market economy addresses scarcity with prices. A well-designed auction has been shown over the centuries to be an effective way of establishing prices to resolve scarcity. You may not like prices and markets but they are a reality of our global economy and are just as relevant in China as in the US.

If ICANN decides that community/geographic/trademark factors do fully resolve contention for a string. Then the contention must be addressed in another manner. In these cases the Applicant Guidebook states that the Last Resort Auction will be used to resolve the remaining contention. The private auction is an alternative to the Last Resort Auction. If you or any other applicant does not want to participate in a private auction you don’t have to. Each applicant has veto power to take the string to the Last Resort Auction. Thus, the relevant question is: In the event that string contention is not fully resolved by community, geographic, or other factors, do you prefer a private auction (or other negotiation process) or the Last Resort Auction? And if private auction, which one? That is the choice.

To me, applicants have much to gain by making use of a private auction like the Applicant Auction that has been designed by experts to achieve the objectives of efficiency, transparency, simplicity, and fairness. Then the contention can be resolved more quickly and at no additional expense to the applicants, since collectively the applicants pay no more to ICANN, and potentially receive a refund depending on the timing.

I wish you success in your community application. If you think that you have a strong community case, then you certainly should wait until the community issue is resolved by ICANN and not participate in an early private auction. The private auction is not meant to resolve contentions in this case, but rather the cases where community/geographic/trademark factors are not dominant.

Conclusion

Many, but certainly not all, applicants have a strong motivation to participate in a private auction. This incentive is especially strong for domains for which the community, geographic, and other factors are not compelling. Among the private auction alternatives, the Applicant Auction stands as the most studied and tested with rigorous scientific methods. The approach builds on the desirable properties of the last resort auction, but allows contentions to be resolved more quickly and without additional cost to applicants.