Sometimes asking the right question makes all of the difference. In July of this year, my friends and former students at the National Communications Authority of Ghana (0) asked me to speak (1) on the question: “Has digitization redefined the boundaries of market definition?” They didn’t ask if market boundaries change. That would have been too easy. Rather they asked if the boundaries of market definition change. My answer (2) was, “Yes they have! And thus we should rethink market power.”
In a nutshell, the traditional approach to assessing market power is to first define “market” and then define “power.” But what if markets are rapidly changing so that defining one is illusive? And what if what appears to be power is fleeting or is actually earned by a company simply doing a great job for customers? It’s time to look for a new approach.
How is market power analyzed?
The US Department of Justice’s (DOJ) merger guidelines (3) describe the current approach to assessing competition in a market. The first step is to define the market. In simple terms, this is done by analyzing historical patterns in customer purchasing to determine which products customers deem to be effective substitutes. Products that are found to be substitutes are considered to be in the same market, such as would be the case with Sprint LTE and Verizon LTE. Non-substitute products are in different markets, such as would be the case with The Wall Street Journal online and Netflix.
Once the market is defined, the analysis looks at various factors to determine the amount of competition. Factors considered include potential for market entry, the presence or absence of powerful buyers, and profits. Many analysts include market structure—the number of direct competitors and their market shares—in this analysis of competition, but others are skeptical of doing so. Still, the DOJ emphasized market shares (4) when seeking to block the proposed DirectTV/DISH Network merger.
What’s wrong with this approach?
There are several problems with the traditional approach. As Bronwyn Howell recently explained (5), the method is based on an outdated understanding of the relationship between market structure and market performance, and generally omits consideration of multisided platforms.
In addition the approach cannot keep up with fast changing circumstances. This inadequacy leads to some embarrassing outcomes:
- Remember that thing called “long distance telephone service”? In the year 2000, the DOJ stopped (6) the proposed merger of MCI and Sprint, claiming it “would result in higher prices for millions of consumers and businesses” and would threaten “to undermine the competitive gains achieved” by past DOJ actions. Within a few years, long distance as a distinct service was effectively no more. (Full disclosure: I was employed at Sprint from 1993 to 1996.)
- Remember AOL Instant Messenger? Using traditional analyses (7), regulators approved the 2001 AOL/Time Warner merger only with numerous conditions (8), including requirements to open AOL’s Instant Messenger to rivals before AOL/Time Warner could offer advanced instant messaging. The regulators believed that AOL was dominant in messaging and that this dominance would continue through future generations of the service. Not only did AOL completely lose (9) instant messaging, the merged company proved so weak (10) that it broke up ten years later.
These problems result in part from the traditional method’s need to define a market. In today’s fast moving tech industry, data and analyses decay quickly, making yesterday’s market data largely irrelevant for making decisions about tomorrow.
What is an alternative?
It is time for economics to return to its fundamentals. Market boundaries and market outcomes result from basic conditions and people’s decisions. In tech, by the time we observe the boundaries and outcomes, they are no longer relevant, and we should therefore rely instead on analyses of the basic conditions and decisions
What basic conditions lead to market power? This is an under-researched area of economics. When I ask this question, most people say economies of scale and barriers to entry. But scale economies and barriers to entry relate to specific products, which quickly decay in tech, so economies of scale and barriers are just as problematic as market definition.
Fortunately, two of the founders of modern economics—Adam Smith and John Stuart Mill—addressed what creates monopolies. They point to collusive agreements, control of essential resources, and government-imposed restrictions on competition as the chief culprits. So competition analysis could focus on those. More recently, William Baumol addressed what creates economies of scale. Professor Baumol explained that these economies result from the presence of a production input that is both costly to acquire (in relative terms) and necessary for the service and that, once acquired, is nearly costless to use for multiple services and quantities of sales. So for economies of scale to matter for market power in a dynamic industry, such an input must exist and must be essential for multiple generations of services. Investigating whether such an input is present in a market is therefore another way to assess the likelihood of detrimental market power.
What about large market shares? Unless they result from one of the factors described above, they are likely earned and shouldn’t be considered a market failure. For example, Google’s Android operating system is on nearly 90 percent of smartphones (11) worldwide because of the high value and low cost it offers device manufacturers and consumers. Should the presence of such value be considered a market failure? Also, providing the system makes sense for Google in part because it allows Google to provide more apps. Some might view this vertical relationship with concern, but would customers really be better off if tech companies viewed each product in isolation and, in doing so, took a pass on investing to create synergies across products? (Full disclosure: I consulted for Google in 2012.)
The bottom line is that our traditional approaches to markets are too static for tech. Going back to the fundamentals seems like the right approach, but it will require us to battle our conditioning and accept that what appears to be detrimental market power may be a quickly-passing phase or evidence of great products.