The last time the Department of Justice took on a big tech firm for antitrust was in the 1990s with Microsoft. Now, there are a number of high-profile lawsuits alleging unfair and anti-competitive practices: Google is under scrutiny, with the DOJ alleging that the tech giant has used anti-competitive practices to dominate and maintain monopoly power in the search engine market, deploying tactics they say have harmed competition and stifled innovation. Yesterday, the FTC announced that it is suing Amazon for also illegally maintaining a monopoly.

Studying the effects of antitrust intervention on competition and innovation is Riitta Katila, the W. M. Keck Sr. Professor in the Department of Management Science and Engineering in the Stanford School of Engineering.

Here, Katila discusses some of her research, particularly her work examining market effects in the aftermath of the DOJ case against Microsoft. She and her co-author, Sruthi Thatchenkery PhD ’17, found that while antitrust intervention did lead to more inventions, these new products failed to gain traction on the market and make a profit – hardly the type of innovation regulators hoped for. Katila also offers lessons about the role of regulation on innovation and competition in technology and business.

What is the case for antitrust regulation?

Antitrust work is based on the underlying belief that competition and innovation go hand in hand. The idea is that when there’s more competition, there’s often more new ideas too, such as more innovative products for consumers to choose from.

In the world of digital platforms, antitrust regulators often worry that powerful tech platforms “block” competition in their ecosystems. This is called self-preferencing – the situation where a platform gives its own apps unfair advantage and thus reduces the chances that rival apps would want or be able to innovate in that space. Simply put, if big companies make it hard for others to join in and compete, there will be fewer new ideas and innovations in that field in the long run. So one could argue that regulators need to step in to ensure healthy competition in a marketplace.

What is the case against antitrust regulation?

Regulation often has unexpected consequences. This is true of antitrust as well. In our research that looked at the aftermath of the Microsoft antitrust settlement in the early 2000s – a case that is often referenced in the current antitrust action – we found the effects of the settlement to be a mixed bag. Regulating Microsoft did trigger a substantial increase in patent activity, especially among firms that had lower market shares before the regulatory intervention. So, antitrust action did help technical invention. However, firms struggled to turn technical inventions into product innovations that the market would value, so we didn’t really see the innovation benefits that the regulators probably were hoping for. Another unexpected consequence was that the real beneficiaries seemed to have been the second- and third-place players just behind Microsoft in market share. They seem to have benefited most by becoming more efficient, not more innovative.

Is there anything about antitrust regulation that is misunderstood? What does your research reveal about market interventions?

Big tech platforms often get a bad rap for killing innovation, but our findings show that it’s more nuanced than that. The surprising finding is that the “apps” we traced in the Microsoft case – enterprise software complementors – seem to have been beneficiaries of some of the platform’s complementary assets before the antitrust intervention. One example – Microsoft’s well-liked and widely used proprietary implementation of Java was used by rival complementor developers but was phased out following the antitrust intervention, forcing companies to develop their own alternatives. By eliminating some of Microsoft’s “complementary assets” offerings and shifting these costs onto rival complementors, the settlement may have inadvertently eliminated rival complementors’ paths to commercialization.

Another surprise was that the Microsoft case did not enable new firms to enter the marketplace. If regulators hoped to level the playing field by helping new firms with innovative approaches to enter the enterprise software market, that didn’t happen. Rather, the changes in the ecosystem that we traced, i.e., the technical invention increases, were driven by incumbent firms, not the new entrants. The payoff for most firms – particularly those that did not already have high market share themselves – thus appears limited. Overall, while policymakers may be satisfied with a boost in ecosystem invention, it is not clear that antitrust delivered for firms in the ecosystem.

What are some of the differences with the DOJ’s case against Google today relative to the one against Microsoft 20 years ago?

Much of the arguments around competition and innovation overlap with those used in the Microsoft case 20 years ago. Unfair visibility of a “preferred app” and calls to bar exclusionary contracting with industry partners are examples. One big difference now is the importance of data. In contrast to software licensing fees for Microsoft, access to user data and getting access to data at scale that underlies digital business models is important today.

What are some other lessons from the Microsoft case that can be applied to the Google antitrust trial?

Concerns about self-preferencing, i.e., platforms favoring certain apps over others, are again spotlighted in the current high-profile antitrust cases. Our findings illustrate the potential effectiveness of “behavioral conduct” remedies in these cases. When regulators want to boost competition, they have two options: a structural remedy, in which a dominant firm is broken up or forced to divest from specific markets, or a behavioral conduct remedy, in which a dominant firm is barred from engaging in specific anti-competitive behaviors. While behavioral conduct remedies can work, regulators must be careful with the specific toolkit, such as making sure not to have platforms prune assets that the ecosystem firms in fact need to flourish.

Altogether, adding more competition is not an automatic lever for more innovation. In digital ecosystems, there are two sides of the coin. So even though we saw that Microsoft’s in-house complementors’ market share decreased post antitrust, leveling the playing field didn’t yield all the innovation that regulators probably expected, offering some lessons for current regulatory interventions.

How can antitrust regulation be used to spur innovation?

To maximize the effects of antitrust action, regulators need to understand the technology and the marketplace closely. Diluting a platform’s market power may in extreme cases backfire for the rival firms. Knowing ecosystem dependencies and how the ecosystem is dependent on the platform is thus important.