thomas m. lenard is a senior fellow and president emeritus of the Washington-based Technology Policy Institute.
Illustrations by Jim Frazier
Published October 28, 2019
The emergence of gigantic tech platforms like Google, Facebook and Amazon has led to increasingly heated debates about whether antitrust policy needs to be revised to cope. Critics claim these platforms have amassed too much market power, undermining the potential of the digital economy to drive productivity growth.
Responding to this concern, Britain created a blue-ribbon Digital Competition Expert Panel headed by Jason Furman, chairman of the United States’ Council of Economic Advisers under President Obama, to recommend ways to change the country’s policies to reflect new realities. And the panel’s analysis is likely to influence the debate in both the United States and the European Union, which, like Britain, are struggling to adapt.
The panel’s report does not endorse populist prescriptions, such as breaking up large tech platforms — or, for that matter, doing away with treating consumer welfare as the primary goal of antitrust, which has been the cornerstone of U.S. policy for half a century. However, its recommendations for a new regulatory regime for large digital platforms to supplement traditional antitrust is arguably as radical: It would create what amounts to an “internet commerce commission” with broad powers to override market incentives.
The Expert Panel report can’t be rejected as representing the views of a fringe. Indeed, it reflects a nascent center-left consensus on competition policy for digital industries that has been brewing for some time. As the aphorism goes, however, those who fail to learn from history … . Similar regulatory regimes in the U.S. — the original Interstate Commerce Commission for surface transportation and the Civil Aeronautics Board for airlines, to name two — were eliminated when centrist reformers convinced a bipartisan majority in Congress that they largely served to protect incumbents at consumers’ expense.
Is There a Competition Problem?
The British report’s proposals are rooted in the view that the inherent characteristics of digital markets make them incompatible with sustainable competition. Digital markets do often exhibit strong economies of scale and scope along with network effects that make them subject to “tipping,” where the winner takes all (or most, anyway) of the market. Moreover, the inherent incompatibility is exacerbated by the advantage that incumbents enjoy in amassing data from customers, which the report argues is a major barrier to entry.
To buttress its argument, the panel cites high concentration levels for various platform “markets.” But a close look suggests these concentration figures are misleading. Google is presented as having a 90 percent share of the “general” online search market without demonstrating that general online search is a well-defined market in terms of antitrust. In fact, Google faces more competition than the 90 percent figure suggests. For example, consumers use Facebook as well as Google to search for news, and nearly half of U.S. internet users go to Amazon first for product searches, compared with about 35 percent for Google. Moreover, consumers use many other platforms for searches in specific areas that generate substantial advertising revenues such as real estate (Zillow, Trulia) and travel (everything from TripAdvisor to Expedia to Airbnb).
For that matter, Google faces more competition in general search than the 90 percent number suggests. While Bing, Yahoo and DuckDuckGo have single-digit shares, they still force Google to stay on its proverbial toes because users can easily switch and because at least two are backed by deep-pocketed corporations. Thus, the fact that Google remains dominant suggests it earned (and keeps on earning) the dominance.
Or consider social media. The report points out that British users spend more than 70 percent of online social media time on Facebook and Instagram (down from 92 percent a few years ago). But it doesn’t demonstrate that the market is no longer contestable. Nor does it engage on the question of what constitutes the relevant antitrust market in a world in which many other platforms provide niche social media services for groups ranging from antique car buffs to travel hackers to strident knitters.
The report notes that Google and Facebook combined garner 54 percent of digital advertising revenue, but does not show that digital advertising is a distinct antitrust market. In the U.S., digital ad spending is just over half of total ad spending, suggesting that digital platforms must compete with both traditional media and upstart marketing approaches like social influencers. Moreover, we are seeing ongoing entry into the digital advertising space — Amazon, in particular, represents a formidable threat — implying there are serious checks on market power.
The digital market power report also points out that Google (with Android) and Apple (with iOS) together account for almost all mobile phone operating systems. Close to 90 percent of smartphones operate on the Android platform, which is free to any device maker. Google makes money on Android only when Android devices lure consumers into using Android-powered apps. And there is no evidence that consumers have been adversely affected by higher prices or less innovation because Android is everywhere.
There is little evidence of economies of scale and scope in amassing digital data. The ability to know what to do with data — to develop smart predictive algorithms — is much more important. Data can strengthen network effects only if pooling data provides visible benefits to consumers.
Amazon is described as dominant in a “meaningful distinct sector of online retail,” with eBay as its closest competitor. The analysts acknowledge that other online sellers, not to mention brick-and-mortar stores, are still in business, but suggest they don’t provide meaningful competition. Still, retailers like Walmart and Target do compete with Amazon offline and are also investing heavily in their own online platforms.
The report notes that Amazon has become a gateway to consumers for many smaller retailers in the Amazon Marketplace, implying an added source of market power. Having that gateway available, however, is a major benefit to those sellers — not just an acknowledgement of Amazon’s marketing clout — by obviating the need for their own marketing efforts. Consumers also benefit from the ability to search for products from many sellers in one place rather than having to search multiple platforms.
In building its case for the need for regulation, the panel claims that access to capital for entrants into digital businesses is limited by the fact that a large portion of investment for digital companies is in intangible capital (and therefore not easily accessible through debt) and that digital companies may incur losses for many years before becoming profitable. But this reality is offset by the existence of a very active venture capital market (at least in the U.S.), as well as by equity valuations for online platforms that incorporate wildly optimistic expectations of growth that allow them to raise capital at very low cost. The most prominent example, ironically, is Amazon, which focused on growth at the expense of profits for many years. By the same token, the ride-sharing companies Uber and Lyft have yet to earn profits, but enjoy market valuations that would be the envy of many long-profitable service businesses.
Barriers Built From Data?
A fundamental theme of the report is that large digital platform markets are not contestable. This has plainly not always been the case — Facebook, for example, pushed MySpace into oblivion. But the panel argues that absent policy changes, the potential for this happening in the future is limited. A major reason: accumulated consumer data are a more formidable barrier to entry today than when now-established platforms were startups.
The economist Catherine Tucker of MIT addressed this issue in a recent paper. In particular, she examined whether possessing large amounts of data enhances market power. Tucker found little evidence of economies of scale and scope in amassing digital data. The ability to know what to do with data — to develop smart predictive algorithms — is much more important. And she argues that data can strengthen network effects only if pooling data provides visible benefits to consumers.
The Regulation Fix
The panel recommends the creation of a new British government agency with authority to offset what the authors view as inherent limitations of after-the-fact antitrust enforcement in digital markets. This new “digital markets unit” could either be a stand-alone agency or part of an existing competition enforcement agency. It would identify digital platforms with “strategic market status” (SMS) — companies in a position to exercise market power over a gateway or bottleneck in a digital market — and focus its enforcement on them.
But identifying platforms with SMS would not be simple. Firms are likely to lobby aggressively to be excluded — and to have other firms included. The definition of a “digital platform” itself is unclear. Does it, for example, include Walmart and Goldman Sachs, which have extensive digital consumer operations? The difficulty of drawing bright lines seems to be inherent in the technology.
Forcing Good Behavior
The panel recommends imposing a binding code of conduct for companies with SMS, which would be developed by the regulators in collaboration with stakeholders. This code would require access to the platform (including rankings and reviews) on a consistent and transparent — that is, nondiscriminatory — basis, much the way common carriers like pipelines and monopoly utilities like electricity distribution networks are regulated.
Note, though, that such a code would preclude business conduct that is common (and legal) in competitive markets — conduct that is often discriminatory, but that nonetheless benefits consumers. The typical supermarket, for example, does not have market power, but engages in practices that, from a regulatory perspective, would be considered discriminatory. Examples include the price discrimination that accompanies “frequent shopper” cards and the slotting fees charged to wholesalers for shelf space and shelf placement. These practices are key to the sometimes-messy system by which the market allocates scarcity (shelf space, consumer attention, etc.). We see analogous messy competitive processes on digital platforms — e.g., preferential rankings.
Consider, too, that for producers with large fixed and low marginal costs, which includes most digital platforms, charging different amounts to different customers depending on their elasticity of demand is almost always efficient — and in some instances is the only way to cover fully allocated costs. To the extent that open-access requirements such as those proposed in the report preclude such discrimination, it almost goes without saying that investment incentives and consumer welfare would suffer.
The report recommends measures for data mobility and open standards that would help make digital systems interoperable. The idea is to lower the costs of switching among platforms and make it easier for entrants to compete with incumbent firms that have amassed large amounts of consumer data.
Tucker observes, however, that “for most consumer-facing applications, it is difficult to see there would be real switching costs due to the storage of historic data,” and gives Amazon as an example. Amazon has a consumer’s historic purchase records, which may make it a bit easier to remember which brand of kibble your dog prefers, but hardly seems likely to prevent the consumer from buying from another platform.
On the other hand, the value of a health analytics app might depend on having the consumer’s health history, which could give rise to very real switching costs. In these circumstances, open standards and interoperability could benefit consumers and providers. Whether the government should mandate such standards and participate in their development is, however, another question.
The creation of open standards and standards for data mobility involves complicated technical issues and presents opportunities for strategic behavior on the part of market participants. The British report does not discuss the potential for this process to be used for anticompetitive rent seeking — in this case, by fashioning standards that favor incumbent enterprises. Nor does it consider whether having government at the center of the standard-setting process (as envisioned in the report) would help or hinder innovation.
Is Data a Barrier?
“The evidence suggests that large data holdings are at the heart of the potential for some platform markets to be dominated by single players,” the report concludes, “and for that dominance to be entrenched in a way that lessens the potential for competition.” As previously noted, however, there are many alternative sources of data. I would argue that the analytical capacities possessed by a large platform are more important than access to its database, and more difficult to replicate.
Mandatory sharing could have significant adverse effects on investment incentives, both for incumbents and entrants. The costs of assembling and organizing data can be substantial, including setting lower prices or offering higher quality to attract more users who add to the data.
Nonetheless, the panel recommends that, to increase competition, the new regulator should have the authority to mandate access by other firms to data held by the strategic market status companies. Mandating access to an allegedly critical asset is a signal feature of public utility regulation. For example, under the Telecommunications Act of 1996, incumbents were required to lease their facilities to entrants at administratively determined prices.
But how would the price of the data be set? A fair price would be difficult to determine both in theory and in practice. Note, moreover, the wheels within wheels: to be useful, data need to be continuously updated.
Consider, too, that mandatory sharing could have significant adverse effects on investment incentives, both for incumbents and entrants. The costs of assembling and organizing data can be substantial, including setting lower prices or offering higher quality to attract more users who add to the data. Requiring a firm to share these assets with competitors would diminish the incentive to undertake these investments. Similarly, the incentive for entrants to invest in data collection and organization themselves would be lower if they could obtain it below their own cost.
The report argues that competition authorities should set a higher bar for acquisitions of potential competitors by large digital platforms. Indeed, the panel suggests (as others have) that if authorities had blocked the Facebook-Instagram merger in 2012, Instagram might have emerged as a competitor to Facebook. The only catch: the merger couldn’t realistically have been blocked at the time because the two entities were not a significant part of any relevant market, thus precluding an analysis of the anticompetitive effect.
The panel suggests being tougher on acquisitions in adjacent market niches, under the assumption that they are more likely potential competitors. This would produce some false positives — instances in which a benign merger was blocked — which authorities should be willing to tolerate. However, determining which acquisitions might be competitors would probably become a highly subjective exercise and the focus of strategic advocacy by various interested parties.
Consider, too, that acquisition by a large platform, such as Facebook or Google, is a frequent exit strategy for successful tech startups. So making that strategy more difficult would be a disincentive to investing in those startups in the first place.
First, Do No Harm
In my view, implementing the British report’s key recommendations would be a risky experiment in public utility-style regulation for the digital sector, a radical departure from current policy. In contrast to after-the-fact antitrust enforcement, this new regime would consist of a set of before-the-fact regulatory requirements that would apply to large digital platforms, regardless of whether they had yet exhibited behavior harmful to consumers. But the report does not weigh the costs and benefits of its proposals.
The panel proposes that competition rules be developed collaboratively by the Digital Markets Unit’s regulatory authority and the commercial platforms as well as other affected parties. I find it ironic to argue that a pro-competition agenda is needed because of the market power of the largest firms, and then to invite these same firms to join in rule-making. Incumbents can be expected to use the collaborative processes envisioned by the report to erect barriers to entry.
Equally to the point, regulation of the sort proposed by the panel has historically favored large incumbents. And although the report wants enforcement focused on large platforms with what the regulators decide is strategic market status, many of the regulations would have the potential for collateral damage. The issue with digital platforms is not whether they ever reduce consumer welfare and inhibit innovation, but whether the proposed intervention will make it better or worse. By this test, the report’s approach seems a nonstarter.