The digital economy has transformed the way we communicate with each other; the way we consume information, products, and services; the way we entertain ourselves. It’s revolutionized seemingly non-digital industries—think of how different financial services, for instance, are today from what they were two decades ago—and investors expect it to soon transform others, which is why Tesla Motors is worth more than General Motors despite making a tiny fraction as many cars as GM makes and earning a tiny fraction of the revenue.

This phenomenon explains why the so-called Big Five of the digital economy—Apple, Alphabet, Microsoft, Amazon, and Facebook—have, at various points over the last year, been the five most valuable companies in the world.

So you might say that the digital economy has lived up to the expectations people had for it 20 years ago, in the early days of the Web. In other important ways, however, its consequences have been smaller than you might think. U.S. GDP growth has, by historical standards, been disappointingly slow since the arrival of the Internet. Productivity growth, which many assumed would be invigorated by the impact of digital technologies, has been dismal for much of this century. In the heyday of the Internet boom during the late 1990s, productivity growth accelerated for the first time since the 1970s, and it appeared for a short while as if technological innovation had solved a central problem with the U.S. economy. But the productivity boom ended in the early 2000s, and it never resumed. Some observers initially suspected mismeasurement, arguing that GDP was not capturing the true value of the many free goods the digital economy offers. But there’s little doubt that the productivity revolution we hoped digitization would usher in has yet to materialize.

The digital economy also has not transformed the job market as much as one might have expected. To be sure, we now have entirely new categories of workers: fleets of Uber drivers, and the TaskRabbiters who haunt Whole Foods stores in major cities. But Americans aren’t job-hopping more than they used to—in fact, by some measures they’re changing jobs less than at any point in the past two decades. And digitization has also eliminated swaths of workers—not just because of automation, but also because of things like online shopping, which has put hundreds of thousands of retail workers out of a job. More important, the digital economy has not been the source of a huge number of good, well-paying jobs. In fact, the rise and consolidation of the digital economy has coincided with an extraordinarily weak labor market. Wages for American workers have just recently started to grow at a clip faster than inflation, but for most of this century they’ve been close to stagnant. (The same is true in most developed countries.) That’s not the fault of digitization. But digitization has not been the driver of job and economic growth that many hoped it would be. Information and communication technologies—which include software and IT companies and Internet firms, along with entertainment and publishing—have seen their share of GDP rise just 1 percent since 2000. And while that’s almost certainly an understatement—and doesn’t capture the impact of digitization on other industries—the number is striking. So, too, is the fact that only a small percentage of private-sector workers work in what you would think of as digital companies.

Still, the most surprising—and potentially troubling—thing about today’s digital economy is how remarkably stable it has become. The buzzword that was always associated with digitization was “disruption.” The Internet and other digital technologies, it was assumed, would accelerate competitive pressures and make it harder for incumbents to hold onto power. If the old industrial order was characterized by companies that stayed at the top for decades, the digital economy, with its supposedly low barriers to entry and low switching costs, was going to be characterized by constant turnover at the top. Instead, the opposite is true. Today’s digital economy, at least on the consumer side, is dominated by the same five giants that have dominated it for at least the past decade and that almost everyone seems to anticipate will dominate it for the foreseeable future (at least if you go by their market capitalizations, which anticipate many more years of enormous profits for all of them). The digital economy is an economy in which platforms are the biggest source of value, and the Big Five’s platforms are the most lucrative ever invented. The result is that this economy is governed, in effect, by an oligopoly. The Big Five sometimes compete and sometimes coöperate, but ultimately each has solid control over its core markets.

“Oligopoly” sounds sinister, but this one was not primarily created by overtly anticompetitive or monopolistic behavior. Rather, digital markets are what economists call winner-take-all markets, in which success tends to breed nearly insuperable advantages. The rule that seems to govern today’s digital economy, in fact, was well stated in Matthew 13:12: “For whoever has, to him more shall be given, and he will have an abundance.” Which is great for those who have, and not so great for everyone trying to compete with them.

Power of numbers

How did we end up with a digital economy dominated by a few big players? The simplest explanation focuses on what are called network effects, whereby a product or service becomes more valuable the more people use it. In the classic example of a network effect, a telephone is worthless if only one person has one, since there’s no one to call. If two people have telephones, they now have some value. And if a million people have telephones, the phone network suddenly becomes enormously valuable.

The implication is that the more users a network has, the easier it becomes to add more users. And direct network effects in this sense are important in understanding the success of a company like Facebook. Facebook’s single biggest advantage over any would-be competitors, at this point, is simply that it’s such an immense network that if you want to connect with people, it’s the logical place to start. The same is true of services like Instagram and China’s WeChat. For digital companies like Snap and Twitter, which are struggling to become profitable, direct network effects are just about the only value they have at all.

The Big Five also benefit from what are sometimes called indirect network effects, including the fact that sellers want to be where buyers are, and vice versa. Because Google has such an enormous user base, companies want to advertise with it. And that makes Google a natural place to go if you’re looking to buy something. Similarly, because Amazon has such a critical mass of customers, it’s the natural place for third-party sellers to gravitate. When Amazon made the decision to allow third-party sellers on its site, competing with its own wares, the decision seemed crazy to many at the time. But it positioned the company to benefit from the network effect: having third-party sellers made Amazon more appealing to customers, which in turn made it more appealing to sellers, creating a virtuous cycle for the company.

Beyond the network effects is another, related way that the sheer scale of the Big Five helps them stay on top: through the access they have to enormous amounts of user data. That data, which is far more detailed and granular than anything companies have been able to access in the past, helps these companies improve their products and services, which in turn helps them add more users, which gives them access to more data, and so on. This data flywheel effect did not get as much attention as network effects in the early days of the digital economy, but it’s become clear that user data is an enormous competitive advantage for the powerhouses of the digital economy and a key reason why it’s difficult to imagine them being toppled anytime soon. By tracking people’s clicks, Google is continually improving its search results and ad servicing. Amazon and Netflix and Apple mine their data to improve their recommendation algorithms, making it more likely they’ll offer you things you want to buy or watch. This process isn’t automatic—you need to have lots of smart data scientists and be willing to invest the resources to ceaselessly upgrade your product. But if you do that, and all of the Big Five do, the rewards can be immense—far greater than what comes merely from the traditional online business model of packaging data and selling it to advertisers.

The ability to gather enormous amounts of data and analyze it efficiently is also at least part of the reason investors think Tesla is more valuable than General Motors. After a traditional car company sells a car to a customer, its relationship with that customer typically is limited (except for maintenance and servicing). Tesla, by contrast, collects terabytes of driving data—including, in some cases, video data—from its customers. That data is then put to use in improving the self-driving features of its cars. According to Adam Jonas, an analyst at Morgan Stanley, Tesla cars are now logging five million miles a day. Since making self-driving cars work depends on machine learning, which in turn requires reams of the data that AI learns from, Tesla’s advantage in data is likely to translate into a huge advantage in making safe and effective self-driving cars. Indeed, Jonas has argued that Tesla’s new mass-market Model 3 sedan could be up to 10 times safer than the average car.

Finally, the Big Five have entrenched themselves in a more traditional fashion, by using their highly valued stock and their enormous amounts of cash to buy other companies, something they’ve done far more aggressively in recent years. Together, Google, Apple, and Microsoft have roughly a quarter of the cash reserves in the entire S&P 500. Google, the most active buyer, has averaged one acquisition a month. Acquisitions have become increasingly important as a way to gain new technology and new engineering talent, expand into new markets or new product areas, and in some cases squelch potential competition. And since no competitor has the resources to outbid the Big Five, it’s another way in which simply being big makes it easier to keep getting bigger.

Digital monopolies

So on the one hand, we have a digital economy that, for all the value it’s created, has not dramatically improved economic growth or wage growth for ordinary workers; on the other, a big chunk of that economy is dominated by a very small group of players. And what’s interesting is that there’s some reason to think these two things are, in fact, connected to each other.

To begin with, the most important fact about platform companies is that they scale, meaning that they can create enormous amounts of value while employing a relatively small number of workers. This is a good thing from the perspective of efficiency. But it also helps explain why today’s digital giants have a smaller impact on the economy than dominant companies had in the past. Together, the Big Five employ around 400,000 full-time workers in the United States. That might sound like a lot. But roughly half of those workers are Amazon employees, many in relatively low-skill, low-paying warehouse jobs. And it’s actually fewer employees than General Motors alone had in 1979, when the U.S. workforce was a lot smaller. What’s more, where GM’s production led to eight jobs in its supply chain for every one person it employed directly, the ripple effects of the Big Five’s businesses, with the exception of Apple, are much smaller. The result is that the rewards of the digital economy are more concentrated among a small number of workers than the rewards of the industrial economy were.

This is aggravated by the fact that the Silicon Valley dream of starting a company in your garage and making it into a huge business has become less realistic than ever. Even though billions continue to be poured into venture funding (more than $200 billion between 2011 and 2016), and even though the number of so-called high-growth startups has not declined in recent years, work by the MIT economists Scott Stern and Jorge Guzman shows that fewer of those startups are succeeding than did in the past. Of course, we still have the Teslas and Ubers (or Lyfts) of the world. But they are rarer than they once were. And one plausible reason is the sheer scale and scope of the Big Five, which can meet competitive challenges either by copying the innovations of others (as Facebook arguably did with Snapchat), and thereby making them seem superfluous, or simply by buying potential competitors at an early stage. Regardless of why it’s happening, though, the result is less dynamism in the economy, and less spreading of the wealth.

One obvious solution to the problems caused by the concentration of power in just a few companies is to break up the Big Five or regulate them as, in effect, public utilities. And of late, there have been more and more calls for dramatic action. But this is difficult for a variety of reasons. First, these companies don’t, for the most part, fit the stereotype of the monopolist. They aren’t “natural monopolies,” like power companies, in markets where it would be practically impossible for a competitor to arise. Anyone who wants to build a new search engine, or a new online retailer, can do so. These companies also, with some exceptions, didn’t achieve their dominance by engaging in conventionally anticompetitive behavior so much as they exploited the nature of the digital economy to build and maintain their empires.

Nor do you hear too many complaints from consumers, although issues of privacy obviously still matter. Indeed, relative to industries like cable television or airlines, digital companies tend to do well on customer satisfaction, and the digital economy as a whole has become a cornucopia of “free” stuff (paid for with consumers’ attention but not their cash). And even though, in practice, consumers often get locked in to the technologies these companies offer (if only because once your data is in the cloud, it’s so much easier to stay than leave), the businesses have no real leverage over consumers. Most also continue to pour billions of dollars into research and development, and they are constantly upgrading their products and services. So it’s hard to argue that these giants have been anything but beneficial to consumer welfare, which since the 1970s has been the standard that antitrust regulators apply.

Indeed, when we look at what the digital economy has done over the past two decades, what becomes clear is that it has created an enormous amount of value for consumers and for a small group of big companies, even as it has diminished competition, centralized power, and made life much more difficult for businesses that produce content or try to compete with the economy’s dominant players. (In one way or another, if you want to make money in the digital economy, you will almost certainly find yourself working with, rather than against, one of the Big Five.) In the industrial economy, the benefits were spread widely among companies, employees, and consumers. The digital economy is giving us a world in which the benefits are concentrated among consumers and the Big Five who serve them. Everyone else just lives in it.

James Surowiecki is the author of The Wisdom of Crowds and a senior story producer at Vice News Tonight.