The Enclosure

Between the fifteenth century and the nineteenth, the common lands of England were fenced. The open fields and shared pastures on which villagers had held customary rights — to graze a cow, to gather firewood, to glean the stubble after the harvest — were progressively enclosed: hedged and walled into private property, the common rights extinguished, the land consolidated into the hands of those who held the title. Enclosure had two faces, and the argument over it has never ended. To its defenders it was improvement: private property in land, unlike the muddled commons, gave its owner the incentive to invest and to farm it well, and agricultural productivity rose. To its critics it was theft — the conversion of a shared resource into a private rent, and the dispossession of the commoners who had depended on it, driven off the land and into the towns. The rhyme that outlived them put it most sharply: the law will hang the man or woman who steals the goose from off the common, but lets the greater felon loose who steals the common from the goose.¹

This plate is about the new enclosures — for the act did not end with the fields. In our own age the same fence is going up around things that were never land at all: around ideas, around networks, around the very record of human attention. The patent and the copyright fence an idea; the platform fences a network and charges a toll on every exchange that crosses it; the data empire fences our behaviour and sells what it has gathered. Each is an enclosure in the oldest sense — the conversion of a commons into a private rent — and each wears the same two faces the enclosure of the fields wore: the face of improvement, which promises that the fence is the necessary spur to creation, and the face of extraction, which collects a toll on what the fence has captured. The whole difficulty of the plate, and of the age, is to tell the two faces apart.

To see what is peculiar about the new enclosures, begin with what is peculiar about the things they fence. Land is rivalrous and excludable: if I graze my cow on an acre, you cannot graze yours there, and a fence can keep you out. An idea is neither. If I teach you what I know, I do not thereby forget it; if you copy my design, mine remains. Thomas Jefferson put it exactly: he who receives an idea from me receives instruction without lessening mine, as he who lights his taper at mine receives light without darkening me. Ideas, and information generally, are by nature non-rivalrous — any number can hold the same idea at once — and naturally hard to fence, for they spread by the mere telling. In their natural state they are not scarce but abundant: the opposite of land.²

And here is the paradox on which the enclosure of the immaterial turns. To fence an idea is not, as with land, to put a wall around an already-scarce thing; it is to manufacture scarcity out of abundance — to take something that could be everywhere at once and make it artificially rare, by forbidding the copying that its nature invites. This is a more radical act than the enclosure of the fields, and it would seem indefensible, but for the argument the previous plate already met in another form. The argument is Schumpeter’s: that the artificial scarcity is the deliberate and necessary incentive. An idea may be costly to create and almost free to copy, so that if anyone could copy it the moment it appeared, its creator could never recoup the cost of making it, and — so the argument runs — fewer ideas would be made. The fence around the idea is the temporary monopoly that lets the creator profit, and so calls the idea forth in the first place. The scarcity is at once the incentive and the cost: the very thing that rewards the creation is the thing that withholds it from everyone else.³

This is the bargain of intellectual property, and it is a real and defensible one — within limits. A patent grants its holder a monopoly on an invention for twenty years; a copyright grants its holder a monopoly on a work of expression for the life of the author and seventy years beyond. In exchange for the temporary monopoly, the invention is disclosed and the work is made, and in time — when the term expires — both pass into the public domain, the commons to which, by their nature, they always belonged. Where the term and the scope are well-judged, the bargain is sound: the fence buys creation that would not otherwise have happened, and then it comes down. The trouble is that the fence has a way of growing past any incentive it could possibly provide, and where it does, it ceases to buy any creation at all and becomes the pure rent this whole volume has been tracking.⁴

The clearest case is the retroactive extension of copyright. When a legislature extends the term of works that already exist — as the United States did in 1998, pushing the term to life-plus-seventy in an act its critics named for the cartoon mouse it kept enclosed — it cannot possibly be buying any new creation, for the works in question were already written; their authors, many of them long dead, will not write more for the promise of a longer monopoly. The extension is an incentive for nothing; it is pure rent, a windfall transferred to the holders of old copyrights at the public’s expense. The same logic exposes the other abuses of the fence: the pharmaceutical company that patents a trivial variation of an expiring drug to prolong its monopoly, a practice called evergreening; the thicket of overlapping patents so dense that no one can build without infringing someone, each patent a tollbooth and the thicket a maze of tolls; the patent troll who manufactures nothing and acquires patents only to sue those who do. In each, the fence has outgrown its purpose. William Nordhaus framed the question precisely, more than half a century ago: there is an optimal length and breadth to a patent, the point that best trades the incentive to create against the cost of the monopoly, and a fence built larger than that optimum adds cost without adding incentive. Much of our intellectual property, and copyright above all, is built well past it.⁵

The second of the new enclosures fences not the idea but the network, and it is the most powerful and the most novel. A platform — a marketplace, a search engine, a social network, an app store — is what economists call a two-sided market: it sits between two groups who wish to find each other, buyers and sellers, riders and drivers, the makers of applications and the users of them, and it makes their meeting possible. What gives the platform its power is the network effect: the more users gather on one side, the more valuable it becomes to the other, so that a platform which gets ahead tends to get further ahead, and the market tips toward a single winner that everyone must use because everyone else already does. And once it has tipped — once the platform is the one place the two sides can reliably meet — the platform is a bottleneck, the bridge that every transaction must cross, and it can charge a toll on the crossing.⁶

The tolls are not hypothetical. Apple’s App Store takes a commission of thirty per cent on the sales that pass through it, and long forbade the developers who depend on it from so much as telling their users that a cheaper option existed elsewhere — a prohibition that has cost the company a five-hundred-million-euro fine in Europe and years of litigation in America, where a court found it had defied an order to stop, in bad faith. Google pays Apple some twenty billion dollars a year to remain the default search engine on the iPhone, buying the bottleneck’s exclusivity; and an American court has found Google an unlawful monopolist twice over, in search and in the advertising technology that sits astride the open web — its conduct likened by one prosecutor to a bank that owned the stock exchange. The European Union has given the pattern a name and a law: its Digital Markets Act designates the largest platforms as gatekeepers — its own word for the bottleneck — and lays upon them a set of obligations whose entire purpose is to keep the bridge open: to let users be steered to other sellers, to let rival applications be installed, to forbid the platform from preferring its own goods. The platform is the new landlord, and the toll on the bridge is the new rent.

The third enclosure is subtler than the others, for it fences a thing we do not usually think of as property at all: our attention, and the record of our behaviour. Attention is genuinely scarce — there are only so many hours in a day, and only so much that any of us can attend to — and in a world drowning in information, the scarce thing, as Herbert Simon observed, is no longer the information but the attention to consume it. A wealth of information creates a poverty of attention; and where attention is scarce, its capture is a rent. The platforms of what is rightly called the attention economy compete to seize and to hold our hours, and to sell them, in the form of advertising, to those who would reach us.⁷

Beneath the enclosure of attention lies the enclosure of data — the record of what we do and click and buy and say, gathered as we use the platforms and held by them as a private asset. Data has a peculiar and consequential economics. Like an idea, it is non-rivalrous; and unlike most things, it improves with scale in a self-reinforcing way, for the more data a platform holds, the better its service becomes, which draws more users, who generate more data still — a loop that turns an early advantage into an entrenched one. It was exactly this loop that an American court identified as the engine of Google’s search monopoly, whose index of the world’s behaviour is more than twice its nearest rival’s; and the court’s central remedy was to order that the hoard be shared. The data is gathered from us, and from the commons of our collective behaviour, and enclosed; the scholar Shoshana Zuboff named the system that does it surveillance capitalism — the conversion of human experience into a private asset, and a private rent.⁸

And now the hard question, the sharpest form of the question the whole volume has circled: where does the earned reward for creation end, and where does the unearned rent of an entrenched position begin? In the new enclosures the two are harder to separate than anywhere else, because — and this is the crux — they are the same thing at different stages. The temporary monopoly that rewards the inventor is the same monopoly that, extended too far, becomes pure rent. The network effect that is the just reward for building a platform people want is the same network effect that, once the market has tipped and the users are locked in, becomes the barrier entrenching the platform against all comers and letting it charge a toll beyond any value it adds. The data advantage earned by serving users well is the same advantage that, compounded, forecloses the competition that would discipline it. Innovation rent and monopoly rent are not two different things; they are one thing, early and late. This is why the policy problem is so genuinely hard, and it is the lesson of the previous plate’s defence carried to its limit: one cannot abolish the monopoly rent without threatening the innovation incentive, because they are the same fence — and the whole art is to tell which stage one is looking at.⁹

The saving grace, where it exists, is the discipline the previous plate also named: creative destruction, the gale that topples yesterday’s monopoly with tomorrow’s innovation. Sometimes it blows. In the very year these lines were written, an American court declined to find the largest social network an unlawful monopolist, on the ground that it now faced fierce competition from newer platforms that had risen to challenge it — the river, the judge wrote, quoting Heraclitus, is never the same twice. But sometimes the gale does not blow, or blows too late: the same year, the same courts found the dominant search engine a monopolist precisely because its position had proved impregnable, its data moat too deep for any rival to cross. And the platforms know the difference, which is why the most entrenched of them buy the disruption before it can grow — acquiring the nascent competitor while it is still small, in the strategy a regulator called buy or bury. The dynamic defence is real but unreliable: the gale disciplines some rents and not others, and the entrenched incumbent does what it can to still the wind.¹⁰

One difference from the old enclosure cuts in the public’s favour, and should be marked. The enclosure of the fields had, for all its cruelty, a genuine efficiency argument on its side beyond the incentive to improve: a common pasture could be overgrazed, the tragedy of a resource that no one owns, and private property solved it. But an idea cannot be overgrazed; a non-rivalrous thing has no tragedy of the commons, for my use of it subtracts nothing from yours. The congestion that justified fencing the pasture has no analogue in the fencing of ideas and information. So the case for enclosing the immaterial rests on the incentive argument alone, with no help from the tragedy of the commons — which means that wherever the fence exceeds what the incentive requires, there is no second justification to fall back upon, and the over-enclosure is pure loss.¹¹

The remedies follow the diagnosis, and they are versions of the volume’s recurring prescription: keep the fence where it buys creation, and open it where it merely collects. For the enclosure of ideas, the cure is to cut the fence to its purpose — to shorten and narrow the terms that have outgrown any incentive, above all to stop extending the monopoly on works already made; to raise the bar against trivial patents and curb the evergreening and the thickets; to strip the patent troll of the power to tax those who actually build. The fence should be no larger than the creation it must buy.¹²

For the enclosure of networks, the cure is to open the bottleneck. If the platform’s power is that users cannot leave, then let them leave: mandate interoperability and the portability of data, so that a user may take her network and her history elsewhere, and the bridge loses its captive toll. Where a platform is a genuine and durable bottleneck, regulate it as one, as we have always regulated the essential bridges and railroads and utilities that everyone must cross. Train the antitrust laws on rent rather than on size, for it is not bigness that is the harm but the toll that bigness extracts; and block the buy-or-bury acquisition that forecloses the disruption before it can discipline the rent. Here two roads diverge — the American, which waits for the harm and then litigates it, case by patient case, and the European, which writes the obligations in advance and imposes them on the gatekeeper from the start — and between the slowness of the one and the bluntness of the other lies the unfinished work of governing the platform age. For the enclosure of attention and data, the cure is of a piece: give people the ownership and the portability of the data drawn from them, and capture for the public the rent extracted from a commons that we all, merely by living, produce.

What unites the remedies is the principle the volume has carried from its first plate to this one. The fence is justified only by what it creates, and only for as long as it creates it; beyond that it is rent, and the commons it encloses should be restored. The enclosure of the fields could not be undone, and the commoners never got their land back. But the immaterial commons is not gone — an idea released to the public domain is as abundant as it ever was, a network opened is as useful, a hoard of data shared loses nothing by the sharing. What was fenced from abundance can be returned to abundance, if we choose. That is the one mercy the new enclosures hold that the old did not.