Past as Prologue

In both Washington and Brussels, the debate over whether or how to reign in Big Tech has become increasingly philosophical in nature. The long-standing view that antitrust interventions are only justified when real consumer harm has been demonstrated is being challenged by those who believe market dominance is inherently a problem. These Neo-Brandeisians argue that overly powerful companies will inevitably treat their competitors and employees unfairly, eventually slowing innovation, and thus proving to be bad for consumers too. They conclude that it’s best to intervene early to preempt more serious problems later.

But no matter which side of this debate you are on, the practical questions are the same. Will the marketplace eventually address today’s concerns about “Big Tech”? If not, what specific policy interventions will do the most good and the least harm? Should these interventions be led by the courts, Congress, or a federal agency? Perhaps most importantly, will government-imposed changes improve American innovation and global competitiveness over the longer term?

Although history doesn’t always repeat itself, 50 years of high-tech antitrust experience suggests a cautious approach. The four biggest information technology (IT) antitrust cases thus far have all either proved to be unnecessary or have resulted in serious adverse consequences. In contrast, requiring dominant tech companies to modify certain carefully selected business practices has consistently helped the digital world move forward.

The IT business has a particularly rich body of antitrust experience because IT markets have always had unusually strong winner-take-all tendencies. Whether we are assessing IBM’s dominance of the mainframe business, AT&T’s control of America’s telecom infrastructure, Microsoft’s and Intel’s influence over personal computers, or the triumphs of Alphabet, Amazon, Apple, Meta, and Microsoft in today’s mobility and Internet eras, the pattern is the same. The virtually infinite scale economies of software and data, the increasing returns that stem from network effects, and customers’ desire to make safe choices in complex, risky, and fast-moving markets have helped the strong get stronger—at least until the technology paradigm changes. During each era, complaints about unfair competition and abuse of market power have echoed across Europe and the United States. So, let’s consider the key lessons from this history.

IBM and AT&T in the Mainframe Era

During the long history of high tech antitrust, no date was more momentous than January 8, 1982. On that fateful Friday, the Reagan administration’s Justice Department dropped the 13-year-long antitrust suit against IBM—deeming it as “without merit.” DOJ also ended its eight-year antitrust battle with AT&T. In a landmark consent decree, the telecom giant agreed to divest its local telephone service business into seven new Regional Bell Operating Companies.

DOJ had sought to break up IBM into separate mainframe and small business system companies because it believed IBM was too powerful for others to compete against. Throughout the 1970s, it was widely believed that IBM would soon extend its dominance into satellite communications, telecommunications, financial services, robotics, and other domains. But by the mid-1980s, IBM had already lost its control of the IT industry. A new paradigm of commodity PCs connected to local area networks was rapidly making many proprietary mainframe and minicomputer systems obsolete. By the late 1980s, the idea that IBM needed to be broken up was laughable. Today, the company is a shadow of its once-gigantic self.

Unlike IBM, AT&T was broken up. However, as with IBM, it was changing technology—not antitrust intervention—that led to the biggest industry restructuring. Over the last 30 years, traditional telephone services all around the world have been largely replaced by cable/broadband offerings, cellular operators, and Internet service providers. While consumers have benefited greatly from these changes, the breakup of AT&T had one major downside: It led directly to the decline of Bell Labs (once the world’s leading research laboratory) and allowed European and Asian firms to dominate the global telecom equipment market, as they do to this day.[1] The demise of Lucent, formerly AT&T’s Western Electric, is another sad chapter in America’s manufacturing decline, but it’s also a powerful example of adverse unintended consequences. What remains of Bell Labs and Lucent is now owned by the Finnish firm Nokia.

Microsoft and Intel in the PC Era

As with IBM, it was widely believed that Microsoft was unstoppable, and in May of 1998 DOJ filed antitrust suit. Once again, the goal was to break up the market leader, this time by separating Microsoft’s applications from its operating system business. But even as the suit was being filed, Microsoft was already on the defensive, as the company missed the early years of both the Internet and mobile phone industries. Microsoft and DOJ settled in 2001, with minor, but still helpful, actions such as requiring Microsoft to provide information about its application programming interface to third-party providers. More importantly, Microsoft’s steady profits from its applications and operating system businesses enabled it to survive its strategic mistakes and eventually become the essential supplier and competitor it is today. A breakup might well have made Microsoft’s extraordinary revitalization impossible.

With Intel, there was no talk of breakups, as Intel had only one dominant business: microprocessors. In 1981, IBM chose to use Intel chips in its PCs, giving Intel a temporary monopoly in this highly strategic market. In 1991, Advanced Micro Devices—a maker of Intel-compatible semiconductors—filed an antitrust suit alleging that Intel “engaged in unlawful acts designed to secure and maintain a monopoly.” After much acrimony and numerous related investigations in the United States, Europe, and Japan, the two companies settled in 2009, ending their patent and licensing disputes, with Intel paying AMD $1.25 billion. But although a viable and healthy AMD clearly helped consumers by lowering PC prices, it weakened Intel’s global position. AMD’s success made it harder for Intel to compete in capital-intensive semiconductor manufacturing, let alone build factories in the United States. Unlike Lucent, Intel remains a rich and powerful firm, but it’s more strategically vulnerable than it’s been in decades. Like AT&T, Intel provides a cautionary tale for policymakers who can see how to increase domestic competition in the short run but can’t possibly foresee the unintended consequences over the longer, global term.

History Points to the Need for Humility

The IBM and Microsoft stories show that previous high-profile efforts to break up Big Tech were unnecessary. The AT&T and Intel cases highlight the worrisome unintended consequences that major government interventions can have. All four histories suggest the need for policymakers to pursue more narrow remedies.

Here, the evidence is on the interventionist’s side. IBM, AT&T, Microsoft, and Intel did make life difficult for their competitors, often in very controversial ways. Rulings and pressures from governments in Europe and the United States effectively addressed some of these concerns. For example, over the years IBM agreed to stay out of the computer services business (temporarily), unbundle its hardware and software pricing, and make it easier for firms to build and sell IBM-compatible products. AT&T was forced to allow third-party telephones and new long-distance carriers such as MCI to connect to its telephony services. Microsoft and Intel both modified their business practices to treat PC suppliers more even-handedly. These were all significant improvements, with few serious downsides.

Taken together, the failures and successes of the past point the way forward for today. Once again, the case for sweeping antitrust intervention is weak. In both the mainframe and PC eras, there were just two powerful firms—IBM/AT&T and then Microsoft/Intel—neither of which competed directly with the other. Today’s claims that five companies—Alphabet, Amazon, Apple, Meta, and Microsoft—are unassailable monopolies are almost self-refuting, especially as these firms increasingly compete with each other and face stiffening competition from China and others, with numerous disruptive scenarios on the horizon.[2] While it might seem justified and tempting to separate Amazon’s retail and cloud computing businesses, breaking up Alphabet into search, email, Android, YouTube, and other businesses, or demand that Meta divest Instagram and/or WhatsApp, this type of armchair chess playing has historically either proved to be unnecessary or had serious unintended consequences.

We should also be skeptical of efforts to restrict these firms’ future acquisitions or unwind previous ones. Acquiring small companies is an important form of R&D in the tech sector, and an essential part of the start-up and venture capital ecosystems. To spur innovation, new companies need the freedom to sell. Ask yourself: Was it right in 2013 for the FTC to allow the Israeli firm Waze—and its 100 employees—to sell itself to Google for $1.1 billion, or should Waze have been forced into a long battle with Google Maps that this tiny start-up knew it would likely lose? If you were one of Instagram’s 13 employees in 2012, would you turn down Facebook’s $1 billion offer? Ditto for the 55 WhatsApp employees who were no doubt pleased by Facebook’s $19 billion offer in 2014. Large acquisitions should still be subject to traditional M&A scrutiny, but the only other companies willing to spend so much cash on such risky new ventures were other tech giants such as Google. How different would that have been?

Rather than sweeping breakups and/or divestments, government oversight should focus on particular business practices and complaints, and as in the past, plaintiffs will sometimes have a compelling case. Companies are not saints, and they inevitably push hard to maximize their interests. Without judging the current efforts by Congress, the courts, or regulatory agencies, it’s certainly fair game for society to scrutinize Amazon’s retail pricing strategies, Apple’s app store policies, Google’s ad auction dynamics, Facebook’s interoperability practices, or similar areas. Such inquiries are directly analogous to the targeted actions that proved so helpful in the mainframe, telecom, and PC eras.

The bottom line is that whether you subscribe to the consumer welfare standard or the Neo-Brandeisian school of high tech antitrust, targeted remedial actions are sometimes justified, and if done well they can result in increased competition. However, as in the past, the biggest transformations will come, not from antitrust interventions, but from shifts in technology and the marketplace. Policymakers need to be humble about how much they should do and very smart about how to do it. This has never been an easy task, and it won’t get any easier going forward. Faced with a rising China and the massive investments needed for the future, America’s digital economy needs to be nudged forward, not dismantled. History clearly shows that doing too much is riskier than doing too little.

Endnotes