The recent House Committee on the Judiciary Hearing on Competition in the Evolving Digital Marketplace raises interesting questions on regulation, on the frameworks needed to regulate emerging digital businesses, and on the applicability of existing regulatory frameworks.
Does the high tech sector require additional regulatory attention, or is the Sherman Antitrust Act, at a geriatric 120 years and still counting, sufficiently general and sufficiently frisky to deal with modern online business models as adequately as it dealt with the emergence of smokestack industry monopolies and cartels? Or are there companies, even entire categories of business, that require new regulatory models? If so, of course, it would not be the first time that the intersection of new technology, new economic forces, and new corporate strategies required a significant extension to existing antitrust law.
Intent and capability
Military analyses focus on both intent (or what you think a potential opponent might want to do to you) and capability (or what you think that opponent would be capable of doing to you if it so chose). Intent changes over time, perhaps quite rapidly and quite without warning. Arming Iran under the Shah was not in America's best interests, as we discovered when Iran's intent changed dramatically and instantaneously after the installation of the Islamic Republic of Iran. Likewise, arming the Taliban in Afghanistan so that they could defeat the Russians turned out to have surprising consequences when we, not the Russians, were the opponent in the Taliban's cross-hairs. If the stakes are high enough, and capability is great enough, "trust me" is not an adequate way to assess intent or to balance benign present intent against possible future actions.
Should we be interested at all?
Is current antitrust regulation working or not? I will argue below that it demonstrably is not. Do regulators fully understand how the oft-touted new economics of the internet, beloved by internet and dot-com entrepreneurs, may be impossible to control with regulation designed to enhance competition in the old economy? I will argue, again, demonstrably not.
FTD Bureau of Competition Director Richard Feinstein, testifying on behalf of the Federal Trade Commission, said "Some have argued that there should be different rules for markets characterized by rapid technological development, but Congress drafted the antitrust laws in general terms to accommodate changing markets and new products, and the laws are flexible enough to meet the challenges of the high-tech era." This was part of his testimony before the House Committee on the Judiciary, Subcommittee on Courts and Competition Policy on September 16, 2010. Not only is this statement about the Sherman Antitrust Act not true today, but it was not true almost from the Sherman's inception.
The regulatory regime of the Sherman Act was found wanting almost immediately after its drafting, and almost immediately needed to be extended to deal with the emerging telecommunications industry. Telecommunications technology, network economics, and AT&T strategy interacted in a way not anticipated when Sherman was drafted. In particular, telecommunications offers positive participation externalities or network effects, where a service gains value as the number of users increases, leading to the idea of a natural monopoly. The Government had to go outside Sherman to regulate AT&T, leading to the Kingsbury Commitment of 1913, which created AT&T as America's first privately owned but regulated monopoly company.
Sherman needed a "manual override" because it could not accommodate the market power of AT&T in the new network economy. America wanted the clear benefits of interconnectivity, with everyone able to communicate with everyone else, which given the technology of the time required a single telecommunications company. Although network effects argued in favor of a single monopoly telecommunications service provider, America did not want monopoly power in this critical new industry concentrated in one firm. For the first time, consumer welfare and technology interacted in a way that demanded that telecommunications be provided by a monopoly, and that likewise demanded regulation of that monopoly in a way not anticipated when Sherman was drafted; clearly not one of the drafters of Sherman anticipated that there would be benefits if we all used steel or oil or detergent or canned soup from the same manufacturer.
So even in 1913 Sherman was not sufficiently flexible and required a "patch". Today it is not at all clear that Sherman is up to the challenge presented by third-party payment business models, where users have free access and third parties are billed. Since users pay nothing for service, and indeed are largely ignorant of the cost of service, prices may be effectively decoupled from the discipline of the marketplace. (At present this has been covered only in academic conferences or on YouTube; we will post on the regulatory challenges of third party payment business models shortly).
The Internet economy will almost certainly pose new regulatory challenges to Sherman, and may require solutions not envisioned in the drafting of Sherman. It is easy to dismiss the need for any regulation if you fail to understand the presence or absence of barriers to entry on the Internet. Although entering online appears easy, it is one thing to create a website, and it is quite another to ensure that it has traffic.
Ed Black, President of the Computer and Communications Industry Association, spoke at the House hearings and rejected claims that Google has reached a point where its practices are anti-competitive, arguing instead that there are "few barriers to entry" in the Internet marketplace. This sounds so plausible, and will be repeated so many times at so many future hearings, that it demands a response.
Ease of Entry?
Instead of assuming away barriers to entry and therefore assuming away the threat of monopolists, let's look at the data. One of the tell-tale signs of monopoly power is monopoly profits; a first course in economics assures us that there should be no long-term super-normal risk adjusted profits; high profits not only invite competition, they more or less guarantee it. And Google demonstrably has extraordinary profits in search. Another tell-tale sign of monopoly profits is the presence of cross-subsidies; if a company is earning enough in one market to cross-subsidize other lines of business, then the business providing the cross subsidies enjoys monopoly power in the absence of contestability; once again Google demonstrably is earning extraordinary profits in search, and once again it is clearly using its monopoly power in search to subsidize everything from high speed internet access and mobile phone service to online video (YouTube) and photo sharing (Picasa). Monopoly profits are not inherently evil nor are they inherently illegal; market power legitimately obtained, without intent to monopolize or to restrain competition, is not legally actionable. But it does demand rethinking statements about the absence of barriers to entry.
This suggests that we should at least look for the possibility of serious barriers to entry in search, and it's not hard to see that they exist. Just look at how much Microsoft has spent to develop and promote Bing (hundreds of millions of dollars on advertising alone) and how limited its current market share is, and you get a sense of how severe the barriers to entry can be.
How Can There Be Barriers to Entry Online?
But how can anyone have monopoly power online? Anyone can build an online website; this is not like building a huge industrial facility, right? And customers are only a click away from switching, right? How can there be barriers to entry online? Build a better and cheaper mousetrap, or a better one, or a cheaper one, or even just a newer one with a catchy name, and customers will flock to you, right?
The first and perhaps most important barrier to new competitors is the third party payer model. You search for a hotel in Arlington, Marriott and ArlingtonHotels.com bid for keywords, and if ArlingtonHotels.com bids enough, your search for Marriott Hotels Arlington takes you to ArlingtonHotels, one of dozens of websites operated by Otels.com. You can still book your Marriott, so you don't care, but the surcharge Otels.com charges Marriott is 30%, so they certainly care! They might prefer to have you go to Bing, to Yahoo!, or to almost anyone else, but since you are not paying for search, the third party is paying, you have no reason to switch search engines.
The second barrier is the geometry of the net. Mostly consumers can't find anything without going through some search engine. That may make search an Essential Facility, and one that enables search engine operators to control what we find, or what we do not find.
Even the "one click away" argument is misleading; as long as search is free to consumers, and as long as a search engine's results are good enough for consumers, then consumers have no reason to invest in even that single click. If consumers stay with their existing search engine choices, then corporations have to go where the consumers are. Once again, the geometry of the net and the third party payer model come into play, making it almost impossible for a new entrant to compete, and almost impossible for a corporate customer to drop out of a keyword auction on an existing search engine.
Why should we care?
There are lots of reasons why we should care about online monopolists. In general, monopolists can and do extend their reach, subsidizing new markets until they are able to obtain monopoly power there, and in general when a "practicing monopolist" obtains new monopoly powers, we can expect these markets to be exploited also. That is, abusive monopolists grow, extend their reach, and abuse their new monopolies. The Sherman Act does not exist to protect poorly run competitors but the Sherman Act does exist to protect competition and to protect consumers from abusive monopolists.
But we should worry even more about an online search monopolist, even if it could be shown that the search monopolist had not charged monopoly prices and had no intention of abusing new monopolies. Search has become the principal way in which most of us learn about everything online, from how to book a new hotel or the quality of a new movie, to how China is handling its dissidents or how the Tea Party candidates are preparing for the November elections. The net is a diverse place, and somewhere one can find exactly the right product, or the story that exactly supports or counters any opinion.
But what if "monopoly of search" can trump "diversity of source"?
If monopoly of search can trump diversity of source, then a dominant search engine has enormous power to promote its own offerings and to stifle innovation in a range of industries. Other problems likewise follow automatically. At present, a dominant search engine can charge almost whatever it wants for keywords, effortlessly switching the balance of power between compliant and uncooperative companies, or compliant and uncooperative politicians. It can promote or stifle points of view by shifting what news stories we find, or don't find, hiding stories it wants to hide, or allowing companies to pay to alter what we learn about them.
I am, of course, not saying that any of this has happened or will happen. But if online monopolies are possible, they are also particularly dangerous.
It's time for Washington to wake up and smell the antitrust
If it walks, talks, acts, and smells like a monopolist, odds are it's a monopolist. And if this monopolist is earning extraordinary profits, and if there is even the possibility that this monopolist might be using those profits to restrain trade, then perhaps the Sherman Antitrust Act is not working. The possibility of online monopolists demands better theory than "there are no barriers to entry online" and purported monopolists need better defense than "trust me."