The Data Boom Is Here — It’s Just Not Evenly Distributed

Joseph Schumpeter was deeply worried about innovation. The renowned economist, who coined the term creative destruction, championed entrepreneurship as the engine of economic growth but feared that small players lacked the key resource needed to implement their pathbreaking ideas: capital. Fortunately, he turned out to be wrong. Since the 1950s, a thriving ecosystem of angel investors and venture capitalists has supplied enough money to startups for their ideas to change the world.

But the data age has revived Schumpeter’s concern that innovators could be blocked from accessing the resources they need. As Big Tech becomes evermore powerful thanks to the vast troves of data that the major platforms have collected, and innovation becomes increasingly data-driven, entrepreneurs and enterprises may find it difficult to seize new opportunities. Keeping the engine of innovation running will require access not only to capital but to data as well.

For many innovators exploiting digital technologies — especially AI and machine learning — great ideas need to be paired with relevant data to create a viable product. Without mountains of suitable training data, safe and reliable autonomous vehicles will remain a pipe dream. The same is true for AI-based medical diagnostics or predictive maintenance systems. We require great volumes of data to develop voice and image recognition, as well as for use cases such as fraud detection, product recommendation, and protein folding. Whether a brilliant idea can make “a little dent in the universe” (to quote Steve Jobs) or at least be turned into a successful product increasingly depends on access to data.

While great ideas can spring up anywhere, access to data is not evenly distributed. Today, the collective we refer to as Big Tech — Google/Alphabet, Amazon, Facebook/Meta, Apple, and Microsoft — has a far greater capacity to collect data than startups and smaller competitors do. Even large corporations in other sectors often have only limited ability to access relevant data.

Lacking access to data, small would-be competitors can only hope to be bought up by Big Tech — which is precisely the dynamic witnessed in Silicon Valley over the past 15 years. An analysis by Stanford Law School’s Mark Lemley and Andrew McCreary showed that at the turn of the century, most successful Silicon Valley startups went for an initial public offering, whereas by the 2010s, most were bought up by incumbents.

Access to data is turning into a key strategic advantage. For many enterprises, it will be the single most important resource bottleneck they face in the coming years. Addressing data access at the company level is as crucial as it is challenging.

But far more is at stake: Lack of data access not only leads to more concentrated markets; a concentration of data in the hands of a few makes it harder for society to meet the substantial challenges we face. We need a policy response.

So far, debates on how to regulate Big Tech have examined market concentration and monopoly power. Remedies include breaking up large platform companies (as White House technology and competition adviser Tim Wu suggests) and treating them as common carriers (as Federal Trade Commission chair Lina Khan has argued). These tough measures are based on previous cases of dealing with powerful monopolies, such as Standard Oil and AT&T. But they are a bad fit for what we face today.

Such traditional antitrust remedies don’t address the source of Big Tech’s competitive power. Unlike Microsoft during the browser wars a quarter century ago, in most cases the platform giants aren’t dominant because they are illegally tying multiple services. They are successful because their exclusive access to huge piles of valuable data translates into competitive advantage.

Breaking them up would wipe out billions of dollars in value while not helping smaller startups and competitors gain access to the resources they need. And as the case of AT&T illustrates, breaking up monopolies without addressing the underlying concentration dynamic means the problem will arise again.

Designating the Big Tech companies as common carriers and turning them into de facto public utilities doesn’t facilitate innovation either. It may force the large platform providers to accept any customer and even prescribe the fees they can charge. But it leaves the dominant companies with exclusive control of their data treasures while limiting the ability of others to be innovative. Worse, treating them as common carriers cements their powerful market position and gives them little incentive to advance innovation.

A far better approach is to ensure that many players can innovate, by granting them access to platform companies’ data. Unlike a tax, a data access mandate does not expropriate property. The Big Tech companies can continue to use the data they have. But it ensures that others have access to data to transform their ideas into innovative new products and services. This is economically efficient because no company can extract all value from the data it has: Others will have different ideas on how to use data to create additional value.

Access to big troves of data disproportionally helps smaller competitors. But unlike a common-carrier regime, it does not redistribute funds to them by prescribing what fees platforms can charge; it incentivizes them to innovate — precisely what’s needed, from a macroeconomic perspective.

Of course, a data access mandate would not cover trade secrets. It would not force Big Tech to share results of data analysis, only the “raw” data those companies collected. And it would not extend to personal information that is subject to privacy regulations; such data would be depersonalized first.

Numerous studies have found that in North America and Europe, most of the data collected is never used. This suggests that we don’t need incentives to gather data, but rather to utilize more of it. And because a data access mandate does not take any data away from Big Tech, platform companies still have an incentive to collect data — even if they must share it with others.

Policy makers around the world, but especially those in Europe, are working on such data access mandates. Germany’s ruling Social Democrats have a comprehensive approach in their party manifesto, envisioning an economy in which data is plentiful and easily accessible to all businesses. Their mandate would cover not just dominant online platforms but all businesses above a certain size threshold and holding a certain amount of data. The mandate would also be progressive in that bigger companies would have to provide access to more of their data. More narrowly focused, sector-specific versions of access mandates are also being considered in the U.K. (for platform giants), France (for public transport and insurance), and the Netherlands.

At the European Union level, a set of proposed measures includes data access mandates in specific situations (highly concentrated markets, such as platform e-commerce) and sectors. The European Commission is also working on a comprehensive legislative framework to facilitate data sharing through the creation of what it calls European data spaces. While it is too early to predict which measures will be enacted, it is clear that improved access to data is in the policy makers’ crosshairs.

Access mandates aren’t unique to Europe, nor are they new or unprecedented. Almost 70 years ago, the U.S. government forced Bell Labs to give competitors and startups access to its patented technologies — including those for semiconductors. It was the intellectual seed that enabled Silicon Valley’s amazing innovation journey. And more than a decade ago, the U.S. Department of Justice forced Google to open up access to data as it acquired travel back-office behemoth ITA Software.

Access to data is the antidote to Schumpeter’s nightmare of a world deprived of innovation through the concentration of resources in the hands of a powerful minority. It enhances competition, enables great ideas, and ensures sustainable economic and societal progress. Mandating it will deliver an unprecedented data dividend.

Editor’s note: An adapted version of this article appears in the Spring 2022 print edition under the title “The Data Oligopoly That Could Choke Innovation.”