I outline some general critical comments on the recent commissioned reports for ICANN’s proposed introduction of new top-level domain names (TLDs).

1. The reports cite seminal papers in economics, but the papers’ applicability here is dubious. For example, for economists a “good” is a product intended for consumption, which is a different sort of animal than a financial investment. Moreover, the good must have an implicit or explicit quality.

Despite a sunrise period for trademark owners to register domain names before a new TLD is made open to the public, most of the initial public registrations will be speculative investments. Thus, traditional consumer economics is not applicable.

The new registrants face tremendous risk, which can result in a TLD fiasco similar to Wall Street’s recent sales of derivates on derivatives (a practice that, at best, amounts to legalized gambling). When analyzing risky investments, you need to look at their risk-reward tradeoffs. For a new investment instrument to be welfare adding, it must provide a new class of risk-return pattern. The pattern for the new TLDs may be a new legalized gambling instrument whose social benefits are, at best, dubious.

When the quality of a consumer good is not obvious, the producer may provide warranties. In addition, a secondary market with asymmetric information between buyers and sellers about product quality, such as the market for used cars, can end up being a market for lemons, as formally demonstrated by Nobel laureate George Akerlof. Because it is not clear what kind of warranties and revenue information the registrants can provide, the new TLDs may be lemons.

Nevertheless, one can convincingly make welfare-increasing arguments for domain name registrations that are signaling and search experience—enhancing (sometimes referred to as “defensive”). Hence, I am not suggesting that there can be no welfare gain, but that the cited academic references fail to convincingly demonstrate social value creation.

2. The reports recommend a prelaunch screening process of viable registry applicants. But such a command-and-control allocation method typically favors incumbents. For example, the Federal Communications Commission’s (FCC’s) spectrum allocation principally benefits the incumbent license holders, who typically favor regulatory barriers to entry. FCC’s protection has come under the umbrella of “public interest” concerns. Moreover, the FCC has been accused of, for example, side deals to appease industry factions in the short term while putting off analytically sound but politically difficult policy choices needed to promote long-term economic benefits. For this reason, such a qualification process is sometimes called “a beauty contest.” On the other hand, without some kind of due diligence, the auction winners can later face tremendous problems, as we have learned from past spectrum auctions.

3. It is not obvious that the new TLDs will compete with .com. In addition to the common argument that .com is the brand of choice, registrants can have anticompetitive economic incentives. For example, a registrant that chooses a short domain name under a new TLD may also register a longer name under .com. This creates benefits to the .com registry (complementary externalities), not the desired increase in competition.

4. Under the argument of monopoly incentives, the reports suggest that incumbent platforms should be banned from bidding on new TLDs. However, the one monopoly profit principle under no-price regulation for the new registry entrants implies that each registry monopolist has no incentive to discriminate against other TLDs. Under regulated prices, the rule does not hold (an exception sometimes called “Baxter’s Law” in honor of William Baxter, the Justice Department official who cited it in the early 1980s as a reason for breaking up the AT&T monopoly). Thus, the impact of the proposed restrictions on incumbents needs to be reassessed.

5. Allowing bids for an introductory low registration price with the option to raise renewal fees in the future complicates the auction mechanism design. The complication can restrict participation and thus competition. Moreover, such a pricing structure complicates brand owners’ registration decisions.

6. Based on the observed limited brand registrations done under alternative TLDs, the reports make heuristic statements about the costs to brand owners of registrations under the new TLDs. The reports should have analyzed the merits of the new TLDs using, say, a signaling model, which justifies the brand owners’ observed behavior. (The other plausible explanation for the lack of brand registrations under all existing TLDs is that the cybersquatter’s ask price may be too high to justify the domain’s acquisition.) Moreover, the reports could have examined the market premiums among TLDs [PDF] before and after previous introductions of TLDs. This would provide a measure of the impact of new TLDs on competition to those TLDs then available.

7. Comments by others.