Discussion: (1 comment)
Comments are closed.
A public policy blog from AEI
The latest on technology policy from AEI, published daily.
View related content: Technology and Innovation
Sometimes pundits and journalists worrying about tech companies’ supposed market power base their fretting on little more than bad analyses and name calling. Some just get their numbers wrong, and others fail to understand what the numbers mean. And the companies are called “cash-rich internet moguls” and “data barons” as if the labels justify regulations or breakups.
In part 1 of this series, I explained that understanding competition in tech requires answering three questions: Who is really in the tech space? Where do these companies experience competitive pressure? What is the nature of rivalry in digital markets? I addressed the first two questions in that blog. I now address the third question.
What is the standard method for assessing market competition?
The traditional view of competition is that it occurs in markets. For example, merger cases often hinge on identifying whether customers would change their buying practices to avoid a 5–10 percent price increase. If customers appear unwilling to shift purchases to another product, the product in question is considered to be in its own market. If customers are unwilling to go to a different physical location to obtain a substitute product, then these customers’ geographic area is also considered its own market. This is called the hypothetical monopolist test.
The hypothetical monopolist test is the state of the art in economics, but it has some problems. One is the cellophane fallacy, which derives its name from a mistake made in the 1950s: Economists and courts analyzing competition for food wrapping materials failed to understand that du Pont was a real functioning monopolist, so the test gave meaningless results.
How well does this standard approach work in tech?
The standard approach has numerous problems when applied to digital markets (using the term “markets” loosely). One is that companies have many complex strategies available in which price is only one of the moving parts. This was one of the snags in the Department of Justice’s case against AT&T–Time Warner: The argument against the merger relied on a specific game theory model, and it was hard to prove that the industry participants actually behave according to it.
The rapidly changing nature of digital technologies also challenges the traditional approach. Constant change makes it hard, if not impossible, to gather enough valid data to measure product substitutability and other market features. Also, the winner-takes-(almost)-all nature of products with network effects means that static views of market shares are deceiving. Furthermore, in instances where companies believe that network effects and other synergies are intertemporal, rivalries don’t exist just over what customers do today but primarily over what customers will do tomorrow. And much of the competition is for consumers’ time and attention, meaning that services that appear different — such as social media and internet search — are actually in direct competition.
How should rivalry be understood?
In addition to the standard product and geographic dimensions, rivalry should be examined along aspects of time, resources, and interrelations. Time matters because a company has no market power for the next generation of products — which are often emerging quickly and with great uncertainty — unless it controls an important resource that its rivals for the future will need. And the company has no power for today’s products if they are largely a prologue for tomorrow.
Related to the time issue is the rivalry in resources. Consumer time and attention are sought-after resources. So are information, knowledge, and understanding, which today are being augmented with artificial intelligence. Companies accumulate these resources to launch what happens next. The companies that accumulate the most have an earned advantage over rivals. The advantage is earned and benefits customers because the prospect of gaining the advantages gives companies a strong incentive to compete for the future.
Value chain is the traditional one-dimensional view of interrelations. Today’s firms in the tech space interrelate along threads that link with multiple companies and customer groups and evolve over time. Consider the multilateral competitive linkages illustrated in Figure 1 in part 1 of this blog series, as well as the intertemporal rivalry between network and edge providers, as illustrated in a previous blog.
How should governments react to this form of rivalry?
Given the complex nature of rivalry in the tech space, heavy burdens of proof should be placed on those proposing government fixes to perceived problems. It is unlikely that the advocates of intervention know enough about the rivalry system as it stands to clearly define a problem and possible solutions, and it is impossible to know the intertemporal aspects with any certainty.
The focus should be on unearned advantages. Generally, these are government barriers to competition, such as favors done for particular companies or regulations that favor incumbents, including some of the emerging privacy regulations. The goal would be to undo the unearned advantages so that competition for today and tomorrow can flourish.
Comments are closed.
1789 Massachusetts Avenue, NW, Washington, DC 20036
© 2018 American Enterprise Institute