The rallying cry against monopoly power has long been, until relatively recently, one of the most compelling forces in U.S. history. Antitrust was the most prominent issue that united the working class and urban middle classes with substantial portions of the rural populists in the early 20th century Progressive Movement. That movement ultimately captured both major parties and left a legacy of substantial antitrust legislation on the books, especially through the creation of the Sherman Antitrust Act and Clayton Antitrust Act and the accompanying bureaucratic bodies charged with preventing the formation of monopolies.

The need for such bodies was obvious at the time—trusts had consolidated market power across a wide variety of products, from steel and oil to tobacco and matches. To maintain market position, these trusts aggressively bought out or undersold competitors, bribed regulators, suppliers, and merchants, and engaged in all manner of collusion to secure price fixing cartels where a single monopolist was unable to corner the market. Enterprising journalists—known as the muckrakers—made their names exposing the depth of these trusts and the flagrant corruption that allowed them to build their empires. The response was relatively rapid. From the assassination of President William McKinley, and rise of Teddy Roosevelt, to the outbreak of World War I, ‘trust busting’ was seen as a noble pursuit for the federal government. Every president in the era engaged at least in anti-monopolistic theater, even if sometimes their efforts generated more smoke than fire.

A key tool for busting these monopolists was criminal law. Legislators defined a set of anticompetitive practices, and companies accused of this behavior could be prosecuted. Notable examples included exclusive dealing, tying the sale of products, and predatory underpricing. Companies found guilty were broken up or fined heavily, both to deter monopolistic behavior and to give an opening to potential competitors. The U.S. Department of Justice’s Antitrust Division retains jurisdiction over criminal enforcement.

America’s other national antitrust body is the Federal Trade Commission (FTC), charged with preventing market consolidations. The FTC and the Justice Department share federal responsibility for civil enforcement, jointly overseeing issues like merger review, litigation, and conduct review.

The FTC is quite clear that simply being a monopoly is not illegal. What is illegal, however, is engaging in anticompetitive practices to seize or maintain a monopolistic position. Over the 20th century, the logic of antitrust jurisprudence and enforcement went from understanding monopoly power as harmful in-and-of-itself, to seeing monopoly as mainly a problem when it can be shown to “harm the consumer.” The logic rests on the assumption that while some monopolies are a natural market outcome that is good for the consumer, the worst monopolies are the result of specific efforts to subvert markets, in order to win prices or levels of market share that would not be achievable in a competitive market.

This is as true for matches or cigarettes as it is for steel and oil, all of which were prominent 20th century monopolies. If you run a match factory, your selling price will be above your manufacturing cost so that you can make money. This leaves an opening for other firms to win market share by having a smaller profit margin. As such, you can’t get a monopoly on matches. Upfront cost of manufacturing equipment complicates this picture, but doesn’t fundamentally change it unless it becomes a dominant part of the cost structure.

In such a market, companies engage in anticompetitive practices to maintain their position—pressuring potential suppliers or buyers into exclusive contracts, for example. The FTC and the Antitrust Division of the Department of Justice have tools to identify and intervene in these cases, even if it may not do so with adequate gusto. However, it is the nature of many products to produce monopolistic markets, regardless of the behavior of the specific firms.

The existence of natural monopolies is hardly a novel realization. Regulators are well aware that in many cases, there is no good economic reason for two competitive firms to emerge. This is often because the good they are selling is ‘non-rival’—meaning that the purchase of the good doesn’t substantially impact supply. Matches are a rival good—if I buy a pack of matches, you can’t buy the same pack, and the overall quantity available for sale in the short-term has diminished. If the match factory has to increase costs to produce more matches in the short-term, other factories can step in and compete if their marginal costs are lower. So a monopoly in matches will generally require some manner of illegal anticompetitive behavior to maintain.

A train ticket on the other hand, is a rival good only when the train starts to fill up. On an uncongested line, train tickets are essentially non-rival goods. A single rail line between two medium-sized cities may be perfectly capable of handling all the passengers likely to want to move between those cities. The upfront costs of building a competing rail link are so high, and the marginal costs of adding a few more cars or locomotives to the existing line so low, that there is no reason another firm would be able to spring up between the two towns. In other words, we have a natural monopoly.

Recognizing this problem, Congress established the Interstate Commerce Commission in 1887, primarily to regulate railroads. The commission was intended to establish political control over some of the most powerful monopolistic forces of the day. The power which these monopolies exercise when unchecked are why transportation, utilities, and other apparent natural monopolies are generally heavily regulated or outright controlled by local or national governments. Attempting to build a functioning market for these commodities would be nearly impossible, as having two firms competing would require the creation of immense amounts of redundant infrastructure and would thus likely raise, not lower, overall prices faced by consumers.

However, in the most quickly growing sectors of the modern economy, natural monopolies are rising in prominence. Google is much more like a railroad or electrical company than the old monopolies centered on the production of physical products. Almost all of Google’s costs are connected to the creation of the initial research, development, and design of its search and advertising systems. If a good search algorithm, infrastructure, and dataset are difficult to create, but once created can be used by a billion individuals for nearly the same cost as it can be used by a hundred million, there is little economic logic for having multiple firms creating search algorithms. Other search engines like DuckDuckGo have to compete in niches that Google doesn’t fill, like privacy, different features, and not being Google.

Similarly, any firm engaged in the creation of digital media faces a similar cost curve. The cost of producing an operating system, video game, or even a film is nearly entirely in the initial production and marketing, and in the built-up productive infrastructure and intellectual property rights, again meaning that competing with an established firm on their turf is exceptionally difficult.

The results are easily visible in the markets. Apple and Google control almost the entire market for smartphone operating systems, three firms control a majority of console gaming revenue, and Disney earned nearly 40% of total film industry revenues in the United States since 2000. Even firms producing physical products—like automobile companies or device firms—have concentrated their efforts on the parts of their business that are the most amenable to natural monopoly formation. As such, fewer corporations are earning most U.S. profits. In 2016, 28 corporations together earned half the profits of the entire S&P 500, and nearly every sector is more consolidated than it was a few decades ago.

Of course, there are many small and medium-sized enterprises still operating in the U.S. Indeed, many of these larger firms depend on these smaller firms—smaller manufacturers that create the component parts of many of these products, or even the products themselves, as well as contractors, merchants, franchisees, etc. These firms in many cases compete with one another. Coca-Cola benefits from the low prices that result from competing bottling companies (producing a rival service), which allows even greater margins for their core business: branding and developing new formulas—the non-rival components of the business. Apple can take bids for various components of their iPhone and keep manufacturing costs to about a third of their overall sales price, thereby deriving their major profits from branding and design.

There is nothing illegal about this. It’s simply making savvy use of different market structures to maximize profit. Even when a crime may have been committed in forming these monopolies, or in using it to force entry into other markets, the burden of proving it may be far too much to handle. The FTC has an annual budget of around $300 million. By contrast, these large corporations report profits in the billions, per quarter. Taking such a corporation to court on an issue that strikes at the very heart of their profit engine is an incredibly expensive undertaking.

Regulating mergers to prevent the creation of these megacorporations runs into obstacles quite quickly. It doesn’t prevent companies from organically growing to immense size, or from pre-emptively swallowing up competitors while those competitors are still small enough to slip under the FTC’s authority. And in practice, it doesn’t even prevent some anticompetitive mergers like Facebook’s purchase of Instagram.

But a more fundamental problem cuts to the heart of monopoly power’s advantages: some markets tend so strongly towards consolidation that firms that can’t grow to immense size simply perish.

Google never had to purchase the myriad search engines they were competing with. They simply had to outperform them, and the nature of its business was such that it could rapidly grow to take on customers without increasing the marginal cost. At this point, another firm sinking billions into coming up with an algorithm and network effects to match Google may require an enormously inefficient replication of economic resources. This is the weakest point in the American state’s current legal and political frameworks for governing monopolies, and the one which requires the most systematic rethink of the technological power of these monopolies is going to be properly harnessed.

What are the implications of this trend for governance? Simply put, the current antitrust model may not address the whole problem. When firms can replicate their ‘product’ a million or a billion times for minimal extra cost, firms can easily establish monopolistic or oligopolistic positions without participating in any anticompetitive practices, at least not on a scale that is provable in court.

The large economic and financial power that comes with the natural monopoly position can then be used to subsidize entry into other markets. For example, Google’s smartphone market share could not have been achieved without the search and advertising monopoly cash cow to back the project financially. These activities could be considered anticompetitive, but under the current regime they are not.

Monopolies in these sectors of the economy are a natural tendency that should be treated as a relatively normal part of the economy, albeit one that requires special regulation, like the sort we already apply to other natural monopolies.

What are the options? One, of course, is nationalization. Outside of the U.S., this has been frequently applied to sectors of the economy lending themselves to natural monopoly. Just as rail, electricity, and telecoms have frequently been nationalized in other countries, functions like internet searching engines and social media could also be. However, this may require more state capacity than we have. Currently, the institutional expertise for governing large nationalized industries is low. The extreme factionalism of U.S. politics may also provide too great an incentive for politicians to treat nationalized entities as vehicles for short-term patronage.

Another, more feasible option is to treat such monopolies like utilities. The logic for this decision is simple: having judged the benefits of monopolistic power to be advantageous, the state nonetheless acts to ensure that these benefits are not simply privatized. Abuses of market power must be regulated against, and the freedom to take risks and innovate properly used. If high market power translates into nothing more than lazy stagnation and executive bonuses, then it is useless. If it translates into throwing enormous financial power around to expand the monopoly into other markets, in some cases it may be harmful.

Effectively regulating technology infrastructure monopolies as utilities, however, requires an expanded understanding of who is impacted by market consolidation. Current monopoly policy has a sole focus on the costs to consumers alone, ignoring other metrics like wages or innovation. By this logic, a monopolistic firm with in-house supply chains located in China, which produces little by way of groundbreaking research, would be an American economic success story—provided only that its products had the lowest possible cost.

Highly concentrated markets tend to reduce not only the number of sellers of a good, but the number of buyers for goods it uses. This is particularly true for labor: as few tech companies dominate the market, there are also fewer employers for programmers and similar careers. Silicon Valley companies have already been found illegally colluding to lower wages, but any standard economic theory holds that even with free competition between firms for employees, wages are going to be inefficiently low in an economy with high levels of monopsony power.

The distinction between consumers and producers, moreover, becomes increasingly difficult to maintain with a platform business, like YouTube. In this case, the same individual may partner with YouTube to produce content, consume other people’s videos, and advertise their own content on the platform. As Chris Gillett has pointed out, one way Amazon has avoided creating too much political resentment is by focusing on providing a stellar consumer experience. Gillett notes that Amazon does this not only with its prices and service, but by a similar confusion of roles: “many of those adversely affected by Amazon in one capacity benefit from it in another, [and] this throws a major wrench into the state’s task of deciding how to deal with its growing power.” Widening the definition of who has a legitimate grievance to bring against a monopolist is critical to adequately regulating this problem.

So, states which seek to govern monopolies must refine the definition of who is a stakeholder.  Not only consumers, but also employees, competing firms, and vendors who deal with monopolies need to be considered when assessing if the size of a firm necessitates a specialized response. Moreover, the state itself is a stakeholder—if very large firms threaten the freedom of state power, action must be taken. Amazon and other tech giants which take on work contracting for the government are major examples, especially in Amazon’s HQ2 campaign, where it extracted concessions from multiple municipalities to house its new office. Having too few and too big firms dealing with the government gives them too much negotiating leverage against the public interest. Firms in a position to wield editorial power over the public conversation, like Facebook and Twitter, threaten the ability of the state to commit to freedom of speech. The appropriate action will depend on determining what the sources of a firm’s monopoly position are.

One of the most challenging components of these enormous firms is the control of property-producing monopolies. Apple, Disney, and Google all control specific intellectual property—both copyrights and patents—that allow them to exert monopoly power by preventing competition against future products as well as present ones. A major advantage of this is the tax advantage to holding international property. Not only is it untaxed by local authorities, unlike basically any physical property, it is easy to move around and register in different jurisdictions and with subsidiary companies. Apple came under fire for doing exactly this with its Irish subsidiary until recently, to avoid income taxes.

The effect of property itself, especially intellectual property, as a type of monopoly is particularly salient when looking at these firms. If Apple holds a key smartphone patent, for example, it can set its price to license it out. That price will likely be higher than is socially efficient, because there are no competitors offering the same license; as a result, fewer firms will pay to use it, and the overall efficiency of the economy is diminished due to this ‘hold out’ problem.

Economist Glen Weyl’s proposal for a tax based on a self-evaluated sales price for such monopolies, at which the owner is thereafter compelled to sell the monopoly right, seems particularly appropriate for these cases. This tax would deal with the holdout problem by requiring that registered property be self-assessed in value. Firms would then pay a percentage of that value as a tax. Market discipline forces them to honestly disclose their true valuation of ownership rights: the obligation to sell if offered their stated price prevents undervaluation, while the tax on the same value penalizes them for overvaluation.

If firms faced the possibility of a mandatory sale of their intellectual property, they would likely be much more open to licensing it. This would open the possibility of greater competition while also spreading technological advances more rapidly. Both of these mechanisms allow for greater economic efficiency and less wasted effort in the creation of redundant technologies and in litigation over patents. But insofar as this ownership grants legitimate value, particularly in terms of gains from research and innovation, companies would be able to take advantage of it—and be properly priced for the privilege of doing so. By employing a legal framework based on understanding the value of property as a mechanism and the incentives of large firms, the state can both harness the power of intellectual property and pre-empt their stagnation. If another firm were able to use the intellectual property rights to a certain technology more profitably, they could gain control of it.

Apart from the benefits of property as such, monopolies are abetted simply by the massive amounts of cash that certain firms can accumulate. Temporary natural monopolies—like those gained by inventing a key technology or marketing a revolutionary product—can be converted into something like corporate dynasties, where the cash flow from previous products is used to quickly buy up any competing products and either incorporate or smother them. The top 10 cash-rich firms in the U.S. currently hold liquid assets worth $721 billion dollars. By comparison, total annual U.S. non-defense discretionary spending stands at $700 billion. At some point, this sheer amount of cash becomes a threat to the proper functioning of the market and state alike.

Simply put, that amount of disposable cash makes firms difficult to regulate and prosecute, as they can retain legal resources far in excess of what the government can pit against them. And it’s not clear that their ability to buy start-ups is always economically efficient. While they may scale up new technologies at times, in many cases truly disruptive technologies or products are underdeveloped when they are owned by the exact firms they would disrupt. Reducing firms’ ability to accumulate such an enormous amount of liquidity may require a direct tax on the revenues of the largest firms in monopolistic or oligopolistic markets. Natural monopolies are effectively private rights to extract rents from the economy. Taxing these rents by some means shifts the tax burden off of more efficiently productive work, diminishes the incentives for cornering the market, and reduces the ability to accumulate too much cash. Of course this must be balanced with the value of private earned wealth for investment, and the financial incentive for innovation, which often creates new monopolies.

Finally, there are some firms, like Amazon, that are near unassailable because of their sheer scale. This scale is part of their unique value as well: it allows them to provide lower prices and greater selection to their customers.  Breaking up Amazon into several competing firms, therefore, may help its vendors and competitors, but would almost certainly raise prices, as the smaller resulting companies would lack that critical scale and have more redundancy. But also, Amazon’s division structure preempts the effectiveness of a breakup if it should occur. Amazon’s Internet retail empire, it would seem, is a natural monopoly. The advantages of scale are too great to replicate at a smaller scale.

As such, large platform companies are becoming political actors in which states have significant interest. Weakening their capacity seems to be a doomed endeavor, undermining not only the company but even the interests of the state which pursues this strategy. In such a case, the FTC’s common approaches—fines and break ups—can do little good. Instead, the firm needs to be governed more like a utility, common carrier, or some new category of natural monopoly understood to be semi-public infrastructure. Just as a power company is regulated in how much profit it can earn, how it must treat its customers, the kind of environmental impact it must strive for, and what additional businesses it can participate in, such a regulatory system is needed for Amazon and other large firms that compare to it in size. A competent regulatory structure must be able to referee the claims of consumers, employees, vendors, and the state, and ensure that the inevitable scale of these natural monopolies is not used to unfairly disadvantage other stakeholders.

Our current FTC, Antitrust Division, Department of Commerce, and other government departments are of course inadequate to the task of applying this paradigm. They don’t currently have the competence to assess various claims by stakeholders, and how those interests can be coordinated with each other, with national interests more generally, and with interests of state. Nor is there any politically realistic path by which they could gain these capabilities in the short term.

They would need to operate on a strategic vision for the economy. Just as a committee regulating a utility needs to set the metrics for success by how they contribute to a long-term plan, those charged with regulating monopolies must direct these megafirms to ends that are beneficial to the U.S. as a whole.

Gaining the ability to strategically direct the largest firms and the economy overall will require a systematic study and application of how countries with more robust forms of state capacity have managed monopolistic firms. To quote Ben Landau-Taylor and Oberon Dixon-Luinenburg’s article on lessons from the machine tool industry: “coordination among the economic power centers of American society must be reestablished to rebuild the engine of creation.”

In the current period, no less than the industrial revolution, the state has a responsibility to direct the economy so as to ensure that we retain access to important technologies and industrial capacity. As AI, energy technology, and other sectors enter into a new era, it will be more important than ever to ensure that the massive companies that dominate the U.S. economy, which may be at the cutting edge of these tech revolutions, are accountable to the interests of the country as a whole.

Increasing the institutional and legal competence of the Commerce Department and other relevant bodies, or their successors,will not come overnight. It would require funding, staffing, and research in order to create a body that can start to understand and quantify the immense and complex firms it is charged with regulating. This would require statutory changes, of course—expansion of the FTC’s authority to implement monopoly governance even in cases where no wrongdoing has been proven, and more political will to enforce current legal tools against monopoly empires.

But first we need a new consensus for how monopolies are governed. As long as governments are limited to regulating mergers or to criminal prosecution of the worst abuses of monopoly power, they’ll be several steps behind the most powerful and profitable corporations, and the economies they are supposed to be managing will suffer as a result. Moreover, they will risk clumsy and reactive approaches to regulating monopoly power which could end up gutting American technological power altogether.

The biggest monopolies in need of regulation are in many cases household names enjoying considerable consumer goodwill. They have access to enormous material, legal, and rhetorical resources. No casual or half-baked approach can be taken to regulating them. For an agency or regime to gain the political capital to properly govern these corporations, it would need to carefully build broad public and elite understanding of the issue of monopoly power. It would also need careful negotiation with the firms in question to ensure their relative cooperation.

Rather than idealized heroes of humanity or caricatured villains, monopolies and the people who run them must be understood as a natural consequence of certain markets. Their regulation is a natural matter of state.