In the landmark antitrust suit against Microsoft, Judge Thomas Penfield Jackson singled out one aspect of the company’s dominant power as particularly injurious.
“Most harmful of all is the message that Microsoft’s actions have conveyed to every enterprise with the potential to innovate in the computer industry,” Jackson wrote in his 1999 decision, adding that “Microsoft has demonstrated that it will use its prodigious market power and immense profits to harm any firm that insists on pursuing initiatives that could intensify competition against one of Microsoft’s core products.”
Two decades on, regulators continue to cite boosting innovation as an important justification for market intervention. But does antitrust regulation of a behemoth like Microsoft actually promote innovation as it intends to? Scholars have long debated this question, but it’s not easy to answer, partly because isolating the effects of antitrust action in the real world is so difficult.
New research by Riitta Katila of Stanford’s Department of Management Science and Engineering and Sruthi Thatchenkery, PhD’17, of the Vanderbilt University Owen Graduate School of Management sheds light on the matter by leveraging a quasi-natural experiment in the software industry.
The outcome of antitrust intervention on innovation is a mixed bag, they discovered.
“Antitrust can be a lever for innovation, but we find it’s difficult for firms to create value after regulators step in. For instance, profits go down—especially for the most innovative firms,” says Katila, who is also Faculty Director of STVP, the Stanford Engineering Entrepreneurship Center.
A natural antitrust experiment
In their paper “Innovation and profitability following antitrust intervention against a dominant platform: The wild, wild west?” Katila and Thatchenkery focused on what are called platform ecosystems and complementors. A simple example is Apple’s iOS operating system. iOS is the platform; apps that run on it are the complementors. In some markets, the platform itself offers an in-house complementor service. Think of Apple’s own podcast app, which sits in the App Store alongside rival podcast players like Stitcher and Overcast.
Markets where a platform’s in-house complementor competes with rival complementors are of particular interest to antitrust regulators because platforms can give their own offerings strong advantages—like default installation, high visibility, and interoperability benefits. Figuring out which advantages are fair and which are anticompetitive is the crux of many antitrust matters.
In the Microsoft case, the “browser wars” were the best-known example of in-house complementors scrapping with rival complementors, but the same dynamic was playing out more quietly with enterprise infrastructure software—a type of software that underlies companies’ complex back-end functions like database management, email servers, and IT security.
In the late 1990s and early 2000s, Microsoft’s Windows Server operating system was the dominant platform for enterprise infrastructure software—in other words, companies had to make sure their software worked on and “played nice” with it. At the time, there were five major enterprise infrastructure software markets: application integration tools, developer tools, database management, network and system management, and IT security.
Katila and Thatchenkery noticed that Microsoft had popular in-house complementor offerings in two of these markets—developer tools and database management—but not the other three. This created a quasi-natural experiment to test the effects of antitrust regulation. The final 2001 settlement against Microsoft was unlikely to have much effect on the three markets where the computer giant wasn’t an important player, creating a “control group”—but it should impact the two where Microsoft had significant market share, making them the “treated group.”
It was a hugely lucky break for the researchers. “When we collected the data, we saw that Microsoft is majorly present in two segments, which helps our design, but not the other three,” Katila says. “It’s perfect, and nobody noticed it before!”
After antitrust action, little guys innovate but fail to profit
The researchers looked at innovation, as measured by patent activity, and at profitability, as measured by return on sales, among all the complementors in these five markets between 1998 and 2004, the three years before and after the 2001 settlement. They also looked at firm entry—how many new complementors entered the markets.
As expected, little changed for the control group—innovation and profitability remained about the same before and after the settlement. But it was not business as usual for the treated group. In these markets, patenting activity across the market increased dramatically.
When Katila and Thatchenkery looked more closely, they realized that the bulk of the new patents were coming from the “little guys”—firms with limited market share. In contrast, the “second-best” complementors, the ones just behind Microsoft in market share, didn’t innovate much more after the settlement than they had before.
Profitability decreased most markedly among the firms with low market share. In other words, despite regulation ostensibly aimed at helping them, the most innovative companies ultimately failed to thrive. Nor did new competitors come on the scene—contrary to what one might expect in a newly open marketplace, firm entry in the treated markets did not increase.
“Antitrust can be a lever for innovation, but we find it’s difficult for firms to create value after regulators step in.”
– Riitta Katila
“When the markets opened up, there was more innovation—and this is what regulators want. But all that innovation was by the little guys, who ended up spending the money on innovation but not reaping the benefits,” Katila says. “In contrast, it’s the number-two and number-three competitors who benefit. They got more efficient in their operations and at the same time didn’t spend on innovation. If the regulators were hoping to push down entry barriers and see new firms enter these IT segments, that didn’t happen either.”
In their analysis, Katila and Thatchenkery speculate that several forces contributed to the uneven innovation and profitability they observed. Antitrust regulation created what they call a “Wild West” in the two treated markets, where complementors were newly willing to try anything and everything but may have lacked the discipline to focus or identify the right opportunities that customers would have been willing to pay for.
Complementors, especially the smaller ones, may also have underestimated some of the infrastructural benefits of a strong platform and in-house complementor. For instance, Microsoft offered a proprietary implementation of Java that many complementors liked and relied on; after the settlement, which had deemed the proprietary implementation anticompetitive, the company stopped offering it, forcing complementors to use other versions of Java that they found less functional.
Lessons for regulators
To Katila, the research is important because it shows that antitrust action does indeed increase innovation—but also suggests that increasing innovation, on its own, doesn’t magically create a truly competitive marketplace.
“There is this assumption, especially among regulators, that competition will lead to more innovation in the form of products that customers will pay for, and that will lead to profits for those who innovate. But it turns out it’s much more complex than that,” says Katila. And because innovation, profitability, and firm entry don’t move in lockstep, regulators need to consider which of those specific goals they want antitrust action to achieve and what behavior they want to incentivize.