The Actuarial State
Why the Tannehill–Hoppe theory of private law produces plutocracy, not liberty
The seduction of private law
The most dangerous arguments are not the foolish ones. Foolish arguments announce themselves and are dismissed. The dangerous ones are the arguments that have been polished until only the attractive parts remain, so that what is left looks like a proof when it is in fact an incentive story with half the incentives deleted.
The private-law thesis associated with Morris and Linda Tannehill and with Hans-Hermann Hoppe is exactly such an argument. It does not say, in the adolescent way, that there should be no rules, no courts, no enforcement, and no protection. It says something more sophisticated, and therefore more seductive: that law can be supplied as a market service; that insurers and private defence agencies, because they bear the costs of violence, will be economically driven to prevent it; and that property will be honoured more reliably by firms that profit from order than by a state that profits from disorder. The state, on this account, is a territorial monopoly over final judgement, and like all monopolies it overcharges and underperforms. Replace it with competition and the abuses dissolve. Hoppe places “free and unregulated” insurance agencies at the centre of this competitive order; the Tannehills build it from private protection firms and private arbitration. Both rest the system on a single load-bearing claim: insurers dislike loss, violence is loss, therefore insurers will buy peace.
The claim has a genuine surface plausibility, and the honest way to meet it is not to mock it but to take its strongest form seriously and then ask whether the conclusion actually follows from the premise. It does not. The premise is about incentives. The conclusion is about institutions. Between them lies a gap that the thesis never crosses, and the things that fall into that gap — the definition of property, the finality of judgement, the standing of the poor, the concentration of force — are not minor details. They are the whole of what a legal order is.
This essay argues that the Tannehill–Hoppe thesis mistakes a partial incentive for a complete institutional order, and that its most probable failure mode is not the chaos its critics usually predict. Chaos is the easy objection and the wrong one. The harder and more damaging point is that the thesis is most likely to succeed at producing order, and that the order it produces is plutocracy: a stratified regime in which protection tracks wealth, law becomes a tiered subscription product, and coercive capacity concentrates into an insurance-security-arbitration cartel that is, in everything but name, a state — only a state with no public accountability, no equal standing, and no appeal beyond the price list. The thesis promises the abolition of sovereignty. What it actually offers is the sale of sovereignty as an insurance product. It does not abolish coercion. It changes the billing department.
The thesis at its strongest
To attack a weak version of an argument is a waste of everyone’s time, so the thesis deserves to be stated in the form its ablest defenders would recognise.
Begin with the diagnosis of the state, which is the strongest part of the case and largely correct. The state is a compulsory territorial monopolist of final decision-making. Because it faces no competitor in the supply of ultimate adjudication and coercion within its territory, it has the standard monopolist’s incentives: to raise its price (taxation), to lower its quality (bureaucracy, delay, capture), to expand its jurisdiction, and, most corrosively, to act as judge in its own cause whenever the citizen and the state are themselves the disputing parties. This is not a caricature. It is a serious public-choice observation, and the libertarian tradition from Rothbard through Hoppe has pressed it with real force. A monopoly on justice is still a monopoly, and monopolies are not improved by being called governments.
From this diagnosis the thesis derives its prescription. If monopoly is the disease, competition is the cure. In a private-law order, protection agencies compete to supply defence; insurers compete to underwrite the risks of life, property, and conflict; and arbitration firms compete to resolve disputes. No single agency holds a territorial monopoly, so each is disciplined by the possibility that dissatisfied clients will leave.
The insurer then enters as the mechanism that converts competition into peace. Violence destroys insured property, raises claims, lowers asset values, and threatens portfolios. An insurer therefore has, on this account, a direct financial interest in suppressing violence among and against its clients, in coordinating with other insurers to standardise rules and contain conflict, and in disciplining any protection agency whose aggression generates claims. David Friedman’s version sharpens the economic logic: rights-enforcement and dispute-resolution are services with prices, and a competitive market in such services will, he argues, tend to generate efficient legal rules because the parties to a dispute share an interest in cheap resolution and will gravitate to the arbitration arrangements that minimise the joint cost of conflict.
Reputation and contract are then offered as the safeguard against predation. An aggressive agency that preys on non-clients invites retaliation, loses the custom of those who fear it, finds itself denied access to the arbitration networks every agency needs, and becomes uninsurable. The market, in this telling, ostracises the predator more reliably than any regulator could, because the discipline is automatic and self-interested rather than political and corruptible.
The conclusion is that law survives without the state. It becomes private, contractual, polycentric, and competitive. The courthouse does not vanish; it is unbundled, priced, and sold by firms that want repeat business.
This is the claim in its best clothes. It is not “no law.” It is the claim that market incentives can do the work that public sovereignty now does badly, and do it better. Everything that follows is an attempt to show why the work does not in fact get done — why each of the load-bearing moves smuggles in a premise it has not earned.
Incentive is not institution
The first and most fundamental defect is a non sequitur sitting at the centre of the argument. Strip the rhetoric and the inference reads:
Violence creates losses. Insurers dislike losses. Therefore insurers will produce a just legal order.
The conclusion does not follow, because the operative verb in the premise — “dislike losses” — does not mean what the conclusion needs it to mean. An insurer does not dislike losses in the way a moralist dislikes injustice. An insurer dislikes expected payouts in excess of premium income. Its objective function is the management of a balance sheet: to classify risk, price it, exclude it where it cannot be priced, cap liability, deny weak or expensive claims, settle cheaply, monitor insured behaviour, and protect capital reserves against correlated shocks. These are not vices. They are the legitimate operations of a competently run insurer. But they are operations of portfolio optimisation, and portfolio optimisation is not justice. The two coincide in some cases and diverge sharply in others.
They coincide when suppressing a given act of violence is cheaper than paying for its consequences and the victim is a valued client. They diverge precisely where justice is most needed and least profitable: where the claimant is poor and commercially marginal, where the facts are complex and expensive to establish, where the defendant is wealthy and able to impose litigation costs, where the harm is real but the expected recovery is low, and where doing right by the victim would cost more than abandoning him. In all these cases the insurer’s rational course and the just course point in opposite directions, and the thesis gives us no reason to expect the insurer to choose justice over the balance sheet. It gives us, in fact, every reason to expect the opposite, because the firm that consistently chose justice over its accounts would be outcompeted by the firm that did not.
This is the cleanest way to put the error. The thesis proves, at most, that insurers have an incentive to reduce those forms of violence that threaten insured portfolios by more than the cost of suppressing them. That is a real incentive and it will do real work. It is also a far weaker thing than the thesis needs. It is not an order of justice. It is actuarial housekeeping — the management of loss within a book of business — and it stops exactly where the book of business stops. An insurance company is not a guardian of rights. It is a balance sheet with lawyers, and a balance sheet has no opinion about the dignity of those who cannot afford its premiums.
Property must be defined before it can be honoured
The thesis says that protection agencies and insurers will “honour property.” Tucked inside that verb is the assumption that does the most quiet damage to the entire argument: that property already exists, in determinate and settled form, prior to and independent of the institutions that are supposed merely to guard it. On this picture, titles arrive from somewhere already labelled, and all the enforcement apparatus has to do is stand watch over labels that nature or reason has already affixed.
This is false to the point of being the reverse of the truth. Property is not an object with a tag. It is a structured relation among persons with respect to things — a relation specifying who may use, exclude, transfer, encumber, and bequeath, against whom, on what conditions, and with what priority when claims collide. The hard cases that consume real legal systems are almost never the storybook case of A simply stealing B’s cow. They are disputes about which claim counts: contests over title and the validity of the acquisition chain, over priority between competing creditors, over adverse possession, inheritance, fraud, duress, insolvency, easement, nuisance, trespass, intellectual property, corporate control, fiduciary duty, and the externalities one person’s use imposes on another’s enjoyment. Before any of these can be “honoured,” someone must decide what the property right is.
Ronald Coase’s analysis of social cost is decisive here, and it cuts against the thesis from within economics rather than from sentiment. Coase showed that harm is reciprocal: the factory’s smoke harms the laundry, but forbidding the smoke harms the factory, and there is no natural fact of the matter about whose use should yield. With zero transaction costs the initial assignment of rights would not affect the efficient outcome, but transaction costs are never zero, and where they are positive the assignment of rights determines both the outcome and its efficiency. The implication is fatal to the idea that enforcement is logically downstream of an already-given property order: the content of the rights, and the transaction-cost environment in which they are exercised, are doing the real work, and someone must set them. Harold Demsetz’s account of how property rights emerge points the same way. Rights are not primitives; they are institutional responses that develop to internalise externalities when the gains from doing so exceed the costs of definition and enforcement. Property is historically contingent, evolved, and constructed — which is to say that defining it is not a preliminary to the legal order but the substance of it.
The decision about which property theory to enforce is therefore unavoidable, and it is political in the broad and exact sense that it allocates rights, settles priorities, and determines whose expectations the system will protect and whose it will sacrifice. A private agency that proposes to “honour property” cannot do so without first choosing a theory of property, and that choice is a legislative act however it is dressed. Calling the agency private does not make the act non-political. It merely relocates the politics into a firm that answers to its paying clients rather than to a public. The thesis assumes the settled property order it is supposed to derive, and then uses the assumed order to claim that agencies merely enforce property rather than make law. That is question-begging of the most consequential kind.
The circularity: markets presuppose the law they are said to supply
The deepest problem is structural, and it is best stated plainly because its very simplicity is what makes it hard to evade. Markets do not run on air. They run on a dense substrate of legal infrastructure: rules of property, contract, agency, fraud, evidence, title, priority, insolvency, liability, and corporate form, together with procedures for adjudicating disputes about all of these and mechanisms for enforcing the results. Strip that substrate away and there is no market — only a collection of people making assertions at one another.
The fully private-law thesis claims that the market will supply this infrastructure. But the market requires the infrastructure in order to function at all. The result is a circle the thesis never escapes: the market is said to produce law, yet the market presupposes law in order to produce anything. This is not a quibble about sequence. It goes to whether the system can get started.
Consider what each of the thesis’s own mechanisms requires. A contract is not self-enforcing; it is a promise that becomes binding only because some authority will compel performance or its equivalent, and “the market will compel it” only pushes the question back, because the compelling itself requires rules of formation, interpretation, breach, fraud, duress, and remedy that someone must have established. An arbitration award is not a judgement; it is an opinion, advice with no force, unless there exists an enforcement structure standing behind it that can make the loser comply when the loser would rather not. An insurance obligation is a string of words unless there is law determining when the policy is formed, what it covers, when it is breached, and what happens when the insurer alleges fraud or the insured alleges bad faith. A title record is ink unless there is a recognised rule — public authority or accepted network protocol — that gives the record legal effect against third parties who never agreed to it.
The thesis gestures at reputation, arbitration, and inter-insurer coordination as the substrate’s substitutes, and these mechanisms do real work in the easy cases: among repeat players, with visible facts, low stakes, and rough equality of power, private ordering can and does function without a state standing over every transaction. The error is the leap from “private ordering works in these conditions” to “private ordering can constitute the entire legal order for all persons, all disputes, and all stakes.” That is a fallacy of composition — the inference that because a part has a property the whole must have it too — and it fails for the same reason it always fails: the conditions that make private ordering work in the easy cases are precisely the conditions that the hard cases violate. Reputation disciplines those who need to deal again tomorrow; it does not bind the party for whom this dispute is the last move, or the party powerful enough that no one can afford to refuse him tomorrow regardless of his conduct today. The thesis cannot conjure a legal order out of mechanisms that presuppose one, and waving the word “competition” over the gap does not fill it.
The final-authority problem: feud or sovereignty
Every legal order must eventually answer one question, and it is the question on which the private-law thesis breaks. What happens when two parties disagree, each is genuinely convinced of his right, each is backed by a protection agency, and neither will submit?
The standard answer is arbitration. But arbitration presupposes either prior agreement on a forum or an enforcement structure able to impose the forum’s decision, and in the hard case at least one party has every incentive to refuse both. When the stakes are high enough, the facts contested enough, and the parties unequal enough, the losing side — or the side that anticipates losing — does not consent to the process that will rule against it. At that point the order arrives at a fork from which there is no third path.
Either there exists a final authority with the capacity to impose the outcome on the recalcitrant party whether or not he agrees — in which case sovereignty has reappeared under a new logo, and the entire promise to abolish the territorial monopolist of final decision has been quietly broken. Or there is no such authority — in which case the dispute is resolved not by law but by bargaining under threat, by the relative coercive capacity of the two agencies, by escalation, cartel negotiation, or force. The first outcome is the state, rebranded. The second is the feud.
This is not an incidental weakness; it is structural, and it has been recognised even by thinkers sympathetic to the libertarian project. The private-law model oscillates permanently between the two outcomes and can rest at neither. Centralise enforcement enough to bind the recalcitrant, and you have recreated the sovereign you set out to abolish, now without the constitutional constraints, the public accountability, and the formal equality before the law that even a flawed state at least professes. Decline to centralise it, and you have no answer to the party who simply will not comply, which means you have no law in the hard cases — only negotiated coexistence among armed firms, stable when their interests align and violent when they do not. Private law either becomes law, in which case it owes us an account of the sovereignty it claimed to have escaped, or it remains genuinely private, in which case it cannot bind the one who refuses to be bound. There is no version that is both fully private and fully binding, and a legal order that cannot bind in the hard case is not a legal order. It is a market with weapons.
What insurance actually does
Because the thesis rests the whole apparatus on the insurer, the actual economics of insurance are not a side issue; they are the heart of the matter, and they are unkind to the argument. Insurance is not a benign solvent that dissolves conflict into priced and managed risk. It is a specific institution with well-documented structural features, and each of those features, applied to the supply of law, points toward exclusion rather than universality.
The first is adverse selection. Those most likely to suffer a loss are most likely to seek cover, which means the pool an insurer attracts is systematically worse than the population unless the insurer screens it. So insurers screen. They classify applicants, price by risk, cap exposure, attach deductibles and exclusions, and refuse the risks they cannot price profitably. A market in legal protection subject to adverse selection does not extend equal cover to all; it sorts people into risk classes and prices them accordingly, and some classes it declines outright.
The second is moral hazard. The insured party, shielded from the consequences of his conduct, has weaker incentives to avoid loss, so the insurer must monitor and constrain behaviour, attaching conditions, surveillance, and controls to the policy. Translate this into a protection market and the insurer’s rational response to moral hazard is the surveillance and behavioural regulation of its own clients — a private disciplinary apparatus operated not for justice but for loss control.
The third, and the most damaging to the thesis, is correlated risk. Insurance functions on the law of large numbers, which requires that losses be substantially independent or at least diversifiable, so that the many who do not suffer loss can fund the few who do. But the losses that a legal order exists to address are precisely the ones that are not independent. Civil disorder, organised crime, insurrection, war, mass unrest, territorial conflict, financial collapse, and the breakdown of the legal order itself are correlated risks: they strike many policyholders at once, and they are the systemic events against which insurance is structurally weakest. An insurer pricing household burglary is doing ordinary actuarial work. An insurer asked to underwrite the stability of the legal order is being asked to price sovereignty failure, and against correlated catastrophe an insurer has only three options: withdraw cover, raise premiums beyond the reach of ordinary people, or build coercive infrastructure large enough to control the systemic risk directly rather than merely insure against it. The third option is the road to private government, and it is the option a sufficiently large and rational insurer will take, because controlling the risk is cheaper than paying for it.
The fourth is the claims-denial incentive, which the thesis politely ignores. Insurers do not simply pay. They investigate, contest, delay, narrow coverage, dispute causation, allege misrepresentation, and settle for less than the claim where they can. These are not abuses; they are the normal management of a book, and they are sometimes the difference between solvency and ruin. But “the body that decides whether your claim is honoured has a direct financial interest in denying it” is a description of a profound structural conflict, and it is the conflict the thesis proposes to install at the centre of justice. Underwriting is not standing. Claims management is not due process. Risk exclusion is not universal protection. Settlement pressure is not legal equality. The thesis treats the insurer as a neutral keeper of the peace; the institution is in fact a sorter, a denier, and a manager of correlated catastrophe, and none of those functions is justice.
Rights at the price of coverage
Here the plutocracy argument becomes explicit, and it is an institutional argument rather than a sentimental one. The objection is not that rich people will become villains. It is that, when enforcement is a commodity, the enforceable value of a right comes to track the holder’s ability to pay for it — and that this is not an abuse of the system but its design.
In a public legal order, however imperfectly it lives up to the claim, a person’s right is not supposed to depend entirely on his current purchasing power. The principle is that the right inheres in the person and the law stands behind it regardless of his bank balance. The practice falls far short: the rich already obtain better lawyers, faster service, and more favourable outcomes, and the gap between formal and effective access to justice is one of the genuine indictments of existing legal systems. But the private-law thesis does not solve this defect. It constitutionalises it. It takes the contingent, regrettable, fought-over fact that wealth buys better law and makes it the explicit organising principle of the entire order.
Trace the mechanism. In a fully private market for protection, the wealthy buy comprehensive cover, elite representation, rapid intervention, privileged access to the best arbitration, and strong enforcement. Those of moderate means buy thinner cover, slower service, weaker representation, and access to whatever forums their budget reaches. The poor buy little. The destitute buy nothing, and a right that no one can be made to enforce is not, in any operative sense, a right at all — it is a sentiment with no remedy. The poor do not lose their rights by any declaration or decree. They lose them by under-insurance, which is quieter and harder to protest and therefore more durable.
The point sharpens further once we see that in this model the poor are not merely poorer consumers of law; they are worse legal risks, and rational firms treat them accordingly. The poor have fewer assets to seize, which makes them low-value claimants; they have less capacity to fund a counterclaim or endure a long dispute, which makes them weak defendants; they have little reputational leverage and less ability to take their custom elsewhere, which makes them captive. A rational insurer or protection agency therefore has reason to treat the poor as commercially marginal — high-cost, low-recovery, easily abandoned — which means the poor are not simply served worse but rendered progressively invisible to the enforcement system, present in it only as risks to be priced out rather than persons to be protected. When enforceable right is a function of purchasing power, inequality of wealth becomes inequality of law, and inequality of law with the full backing of force is the precise institutional meaning of plutocracy. Legal standing ceases to be civic and becomes actuarial. That is the substitution the thesis performs while presenting itself as a charter of liberty.
The economics of coercion: scale, concentration, and the dominant agency
The thesis depends on a market that stays competitive — on the persistence of many agencies, none dominant, each disciplined by the others. But coercion is not an ordinary good, and a market whose product is force does not behave like a market in sandwiches. It has scale economies that drive it toward concentration, and concentration in the supply of force is not a large firm. It is a sovereign.
The advantages of size in the production of coercion compound in a way they do not for ordinary goods. A larger protection-insurance firm commands better intelligence, more armed personnel, deeper databases, superior surveillance, stronger legal drafting, larger capital reserves against correlated shocks, and greater bargaining leverage over everyone it deals with. Crucially, it can threaten a smaller agency far more credibly than the smaller agency can threaten it, and it can impose its terms on the arbitration networks that every agency needs, refusing to deal with firms that will not accept its rules. Each of these advantages raises the return to being larger, and the firm that is larger is better placed to become larger still. The competitive equilibrium the thesis assumes is not stable, because the product itself rewards the accumulation of the capacity to coerce.
This is not a hostile invention; it is the conclusion reached by the libertarian tradition’s own most careful philosopher. Robert Nozick, setting out to defend the minimal state to anarchists on their own terms, argued that competing protection agencies would not remain plural. Through an invisible-hand process driven by nothing more sinister than each client’s preference for the agency most likely to prevail in a conflict with clients of other agencies, one agency becomes dominant within a territory; it then suppresses, by force or by buying out, the risky private enforcement of independents, and arrives at a de facto monopoly on the legitimate use of force — which is to say, at a state. The significance for the present argument is that even a thinker determined to vindicate liberty against the anarchist concluded that a market in protection does not stay a market. It collapses into a monopoly. Nozick called the result a minimal state and thought it justified; whether or not one accepts his normative conclusion, his structural prediction is the one that matters here, and it is the opposite of the Tannehill–Hoppe assumption that competition persists.
Tyler Cowen pressed the point from inside the economics of the question and reached a conclusion still more uncomfortable for the thesis. The polycentric system requires a network: agencies must recognise one another’s arbitration, honour one another’s judgements, and coordinate enforcement, because without such a network there is no system, only a scatter of firms unable to resolve cross-agency disputes. But a network with the power to admit and exclude is a network with the power to collude. The very coordination that makes private law function — the shared protocols, the mutual recognition, the interlocking arbitration — is the coordination that allows the established agencies to act as a cartel: to fix terms, to exclude entrants, and to discipline mavericks, exactly as the dominant firm would in any other industry, except that here the industry’s product is force. The mechanism that the thesis needs for the system to work is the mechanism that turns the system into a coercive monopoly. Hoppe himself anticipates that private law would tend toward a unification of law through inter-insurer agreement; he presents this as a benign convergence on good rules, but a unified, coordinated, network-enforced body of law backed by concentrated force is not the abolition of the state. It is the state’s functional re-creation by a cartel, with the added defect that the cartel was never even nominally accountable to anyone but its members. The state does not disappear in this story. It is reconstituted as a premium-funded enforcement conglomerate, and the only thing genuinely abolished is the public’s claim on it.
The natural equilibrium: limited-access order and elite rent-sharing
If the supply of coercion concentrates, the next question is what kind of order the concentration produces, and the political-economy literature gives an answer that is more precise and more damning than the usual prediction of collapse. The likely end-state of a private-coercion order is not Hobbesian chaos. It is a stable, ordered hierarchy in which violence is contained by sharing its proceeds among those capable of deploying it — and that hierarchy has a name.
Douglass North, John Wallis, and Barry Weingast argue that the overwhelmingly dominant way human societies have controlled violence is not the rule of law but the limited-access order, or “natural state,” in which powerful actors with the capacity for violence are given privileged access to economic rents — to land, trade, offices, and organisations — on the condition that they refrain from fighting. The logic is that open violence destroys the rents, so the elites who hold them have a shared interest in keeping the peace among themselves, and the order is stabilised precisely by restricting access: by limiting who may form organisations, who may compete, who may enter the privileged circle. The result is genuine order, often durable, but it is not equal, not open, and not liberal. The few coordinate to preserve a profitable peace; the many receive whatever standing the coordination finds convenient to grant them. The rare alternative — the open-access order, with impersonal rule of law, open economic and political competition, and entry on equal terms — is historically exceptional and rests on demanding conditions that do not arise automatically and are not the default toward which societies drift.
This framework maps onto the private-law failure mode with uncomfortable exactness. The actors a private-coercion order would empower — large insurers, dominant security firms, major creditors, landlords, infrastructure owners, and the arbitral institutions that serve them — are exactly the actors who, in the North–Wallis–Weingast account, form the coalition of a natural state. They need no conspiracy and no smoke-filled room; shared incentives are sufficient. Each prefers predictable titles, enforceable debts, low disruption, stable returns, and barriers against the low-cost entrants, defaulting debtors, squatters, strikers, and inconvenient outsiders who threaten the value of what they hold. A coalition of such actors, coordinating to preserve a profitable peace and to restrict access to those who might disturb it, is the textbook structure of a limited-access order. The private-law society does not collapse; it congeals — into precisely the rent-sharing elite settlement that most of human history has produced whenever the capacity for violence was privately held.
Mancur Olson’s analysis explains why concentration is the attractor rather than an accident. Under genuine anarchy, a “roving bandit” who can plunder and move on has no reason to leave his victims enough to prosper, so roving predation is ruinous for everyone. But a “stationary bandit” who monopolises theft within a territory acquires an encompassing interest in the productivity of those he robs: because he will be there to rob them again next year, he has reason to leave them enough to invest, to provide the order that lets them produce, and to rationalise his extraction as regular and predictable taxation rather than random seizure. Olson’s startling implication is that monopolised coercion can be Pareto-superior to competitive predation — which is exactly why a dominant private-protection firm has every incentive to become the stationary bandit: to stabilise its extraction by monopolising violence, to provide order because order is what makes its clients worth protecting, and to call the regular fees premiums and the protection a service. The economic logic that built states out of anarchy in the first place does not pause for the word “private.” It runs in a market for protection exactly as it ran everywhere else, and it runs toward monopoly.
Protection and the racket
There is an empirical version of all this, and it is the version the thesis most needs to avoid, because the historical cases in which protection was genuinely supplied by private firms in the absence of an effective state are not flattering. The clearest record we have of privately produced protection is the record of the mafia, and that is not an accident of vocabulary.
Charles Tilly placed the point at the level of theory. War-making, state-making, protection, and extraction, he argued, sit on a single continuum, and the difference between a government and a protection racket is far less principled than the government would have us believe. Both sell protection; both sometimes manufacture the very threats they then charge to defend against; both extract payment under the implicit understanding that refusal carries a cost. Legitimacy, in Tilly’s deflationary account, is largely retrospective — a story told after the fact by those who won the competition to control violence in a territory. The relevance to the private-law thesis is direct: the thesis proposes to launder the protection–extraction continuum by inserting the word “insurance,” but inserting a word does not change the structure. A firm that says “pay us and we will protect you” and a racket that says “pay us and we will protect you, including from ourselves” differ not in their offer but in the constraints upon them — in whether there is legality, accountability, contestability, exit, and due process limiting the coercion. And in a fully private order, those constraints are supplied, if at all, by the same enforcement ecosystem that profits from coercion. The distinction between protection and racket becomes endogenous to the firms that have the most interest in erasing it.
Diego Gambetta’s study of the Sicilian mafia supplies the empirical flesh and removes any comfort the thesis might draw from the abstractness of the objection. The mafia, Gambetta showed, is best understood not as a criminal aberration but as an industry: an industry supplying private protection and the guarantee of transactions in a society where the state failed to provide reliable enforcement and trust. Where public protection of property and contract was weak, a market for private protection emerged to fill the gap — and what filled it was not a competitive ecology of mutually disciplining firms converging on liberal law. It was a violent, territorial, rent-extracting hierarchy that guarded the property of those who paid, settled disputes on its own terms, suppressed rivals, and treated outsiders as objects rather than clients. Federico Varese’s parallel study of the Russian mafia in the chaotic aftermath of the Soviet collapse found the same pattern: when state enforcement disintegrated and a sudden demand for the protection of new property rights went unmet by any public authority, private protection markets arose — and again they took the form of mafias, not of the benign insurer-arbitrators the thesis imagines. The point must be made carefully, because the careless version is false: not every private insurer is a mafia, and the claim is not that private protection is always criminal. The claim is narrower and harder to dismiss. The thesis owes us a mechanism that prevents a market in protection from becoming a market in protection rackets, and the only mechanism it offers is competition — but coercive competition, the competition to supply force, is exactly the thing that the historical cases show produces rackets rather than restrains them. The one large-scale natural experiment we have in privately produced protection returned the mafia. That is the evidence, and the thesis has no answer to it beyond the assertion that this time the firms would behave differently.
Why reputation and ostracism fail in the hard cases
The thesis’s last line of defence is reputation. Aggressive agencies, the Tannehills argue, would be ostracised — boycotted by clients, denied arbitration, made uninsurable — and so the market would discipline predators without any public authority. The mechanism is real, and the question is not whether it ever works but whether it works in the cases that matter, and there the answer is no.
Reputation disciplines a firm under a specific and demanding set of conditions: many competitors, low switching costs, transparent facts, rough equality of bargaining power, reliable information about conduct, and the absence of any agency dominant enough to ignore the others’ disapproval. These are the conditions of the easy case, and in the easy case the firm did not need much disciplining to begin with. The hard case — the case the legal order exists for — violates every one of them. There the facts are contested and expensive to establish, so the predation may never become common knowledge; switching costs are high, so dissatisfied clients are captive rather than mobile; bargaining power is grossly unequal, so the victims are exactly the parties least able to organise a boycott; and, decisively, the predatory agency is often the dominant one, serving the wealthy clients, controlling the critical infrastructure, and commanding the superior force.
Against a dominant agency, ostracism is theatre. The parties with the capacity to discipline a powerful protection firm — the other large firms, the major asset holders, the creditors — are precisely the parties who benefit from its protection and have no interest in disciplining it, while the parties with the interest — the poor, the marginal, the preyed-upon — are precisely the parties without the leverage. Reputation works against the small and dependent firm that needs everyone’s goodwill to survive; it does not work against the firm large enough that others need its goodwill more than it needs theirs. Exit disciplines a seller only when exit is real, and for the captive client of a dominant coercive firm exit is not real. Under those conditions reputation ceases to be a constraint and becomes a brand — a thing the powerful firm cultivates precisely so that its coercion may be called service. The ostracism argument assumes away the concentration that the economics of coercion produces, and an argument that depends on assuming away the central problem is not an answer to it.
Peace is not liberty
It is worth pausing on a conceptual confusion that runs beneath the whole thesis, because once it is named, much of the argument’s appeal dissolves. The thesis treats the production of peace as though it were the production of liberty, as though a demonstration that insurers prefer order were a demonstration that they would supply a free society. But peace and liberty are not the same thing, and the gap between them is where every illiberal order in history has made its home.
A cartel can produce peace. A feudal manor can produce peace. A dictatorship can produce peace, often a very orderly one. A prison produces peace. A plantation produced peace. A mafia, where it is unchallenged, produces a notably effective peace. Peace — the absence of open violence — is compatible with almost any distribution of power and almost any degree of subordination, and the more total the domination, frequently the deeper the peace. So a proof that a private-law order would tend toward peace, even if it were sound, would establish far less than the thesis supposes. It would establish that the order is quiet, not that it is free.
The questions liberty actually turns on are the ones the thesis does not ask: peace under what rules, by whose consent, with what right of appeal, with what possibility of exit, with what equality of standing before the law, with what protection for those too weak to defend themselves, and with what limits on the very firms that enforce the rules. A plutocratic order answers all of these in a particular and unfree way. It supplies a very specific peace: stable titles for incumbents, low disruption for capital, predictable enforcement for creditors, and quiet subordination for those without bargaining power. That is exactly the peace an insurance-security-arbitration cartel would have every reason to supply, and exactly the peace the historical limited-access order has supplied for most of human history. To prefer this peace is not to prefer liberty. The thesis proves, at the very most, an incentive toward a certain kind of order, and then helps itself to the word “freedom” as though order and freedom were synonyms. They are not, and the difference is the entire subject.
The honest counterargument: non-state governance can work
The critique made so far would be cheap if it amounted to the claim that nothing works without the state, because that claim is false, and the strongest non-statist scholarship demolishes it. Intellectual honesty requires meeting that scholarship squarely rather than hiding behind a strawman, and the most serious version of it is Elinor Ostrom’s.
Ostrom’s work on common-pool resources is the most rigorous demonstration available that communities can govern themselves, resolve disputes, and sustain cooperation over long periods without either privatisation or an external state. Studying irrigation systems, fisheries, forests, and grazing commons that endured for centuries, she identified a set of design principles that long-surviving self-governing institutions tend to share: clearly defined boundaries around the resource and the community; rules matched to local conditions; collective-choice arrangements that let those affected by the rules participate in making them; effective monitoring, often by the users themselves; graduated sanctions that escalate with the severity and repetition of violations; accessible and low-cost conflict-resolution mechanisms; at least minimal recognition by external authorities of the community’s right to organise; and, for larger systems, nested layers of governance. Her achievement was to refute decisively the lazy dichotomy that the only choices are state control or privatisation, and to show that polycentric, self-organised governance is a real and sometimes superior third path. Any honest critic of the private-law thesis must concede this, and it is conceded here without reservation: the state is not the only alternative to the market, and non-state governance is not a fantasy.
But the concession does not rescue the thesis; it locates precisely where the thesis fails, and Ostrom’s own findings are what locate it. The institutions she documents work because of their conditions, and those conditions are the inverse of the conditions a society-wide private-coercion order would face. Ostrom’s commons are bounded: a defined community of identifiable members who interact repeatedly, can observe one another’s conduct, share a stake in the resource’s survival, and possess local legitimacy and the means to monitor and sanction at low cost. They govern a common resource among known neighbours, not the totality of high-stakes disputes among strangers across a whole society. They rely on graduated, communally administered sanctions, not on the deployment of organised coercive force against the recalcitrant powerful. And, tellingly, several of her principles presuppose exactly the thing the private-law thesis denies — minimal recognition by an external authority, and nested governance that includes higher tiers — which is to say that Ostrom’s successful self-governance is typically self-governance within a broader legal order, not a replacement for one. The Tannehill–Hoppe thesis takes the genuine fact that bounded communities can govern shared resources among repeat-dealing neighbours and inflates it into the entirely different claim that a market can supply the complete coercive legal order for an unbounded society of strangers, across every dispute, every wealth level, and every act of violence. That is the fallacy of composition again, in its most consequential form. Private ordering can supplement a legal order; Ostrom shows as much. It does not follow that private ordering can constitute one, and Ostrom’s conditions are precisely the evidence that it cannot scale to do so. The honest reading of the strongest non-statist authority is therefore not a vindication of the thesis but a specification of its limits.
The state is bad; private plutocracy is not the cure
None of this is a defence of the state as it exists, and it is important to say so plainly, because the thesis draws much of its energy from the real and serious failures of public legal orders. Those failures are not in dispute. State legal systems are slow, costly, and bureaucratic; they are captured by organised interests; they are unequal in practice however equal in principle; they are sometimes corrupt and sometimes abusive; police forces can become predatory; regulation can ossify into a shield for incumbents; and the public-choice critique of the state’s incentives is largely correct. The libertarian diagnosis with which this essay began is not a setup to be knocked down. It is true.
But the truth of the diagnosis does not establish the soundness of the cure, and here the thesis commits its most basic methodological error: it compares the real state, with all its documented defects, against an idealised market, with its defects assumed away. That comparison is rigged. The honest comparison is between real public authority, defects and all, and real private coercion, with its defects — the concentration, the cartelisation, the exclusion of the poor, the conflict of interest at the heart of claims, the historical tendency to produce mafias and limited-access orders — fully in view. Set the real against the real, and the private-law order does not emerge as the freer of the two. It emerges as the older and uglier one: private sovereignty for those who can afford it, which is the arrangement the rule of law was painfully constructed to replace.
There is, moreover, a coherent alternative to both horns of the false dichotomy the thesis offers, and it is the tradition of constitutional political economy the thesis ignores. James Buchanan and Gordon Tullock distinguished the constitutional level, at which the rules of the game are chosen, from the operational level, at which actors play within them, and argued that the remedy for the abuse of collective power is not the abolition of public authority but its constraint — the binding of the sovereign by rules agreed at the constitutional level, designed to hold whoever happens to wield power later. The same impulse runs through Acemoglu and Robinson’s distinction between inclusive institutions, which distribute power broadly and open access to opportunity, and extractive institutions, which concentrate power and channel resources to a narrow elite — with the historical record showing that prosperity and freedom track the inclusive form and that concentrated, unaccountable power, whether public or private, tends toward extraction. The remedy for a coercive monopoly that overcharges and abuses is to constrain it: to subject it to constitutional limits, transparency, division of power, independent appeal, due process, equal standing, legal aid, and competition where competition is genuinely possible. The remedy is emphatically not to auction the coercive function to whoever can pay the most for it, because an unconstrained coercive power sold to the highest bidder is not less dangerous than an unconstrained coercive power held by the state. It is more dangerous, because it has shed even the pretence of public accountability and the formal commitment to equality before the law that gives the citizen a standpoint from which to demand better. The cure for bad public power is better-constrained public power. It is not private power, unconstrained and for sale.
Conclusion: the actuarial state
The Tannehill–Hoppe thesis begins from a genuine insight and ends in an illusion. The insight is that incentives matter, that a monopoly on justice is still a monopoly, and that the state’s incentives are often as bad as its defenders are reluctant to admit. The illusion is the inference that because insurers dislike loss, insurers will supply justice — that a single incentive, isolated and idealised, can stand in for an entire civilisation’s worth of institutions.
It cannot, and the reasons compound. Insurers minimise expected liabilities, not injustice, and the two part company exactly where justice costs more than abandonment. Property cannot merely be honoured, because it must first be defined, and defining it is a political act that allocates rights and cannot be performed by a firm without that firm legislating for its clients. The market cannot supply the legal substrate it presupposes, on pain of a circularity it never escapes. Arbitration cannot bind the recalcitrant in the hard case without an enforcement structure that is sovereignty under another name, leaving the order to oscillate forever between feud and a rebranded state. Coercion is not an ordinary good; it has scale economies that drive it toward concentration, as the libertarian tradition’s own philosopher and its own economists were honest enough to conclude. Concentration plus elite coordination is not chaos but a limited-access order — the rent-sharing settlement that most of human history has produced whenever violence was privately held. The one large natural experiment in privately produced protection returned the mafia. And reputation, the thesis’s last defence, disciplines the small and dependent firm while leaving the dominant one untouched, which is the only firm that needed disciplining.
The more plausible end-state of the private-law project is therefore not a society without sovereignty. It is sovereignty without accountability: a private order in which coercion is priced, law is bundled into coverage tiers, and justice is delivered according to the level of protection a person can afford. The poor are not stripped of rights by any decree; they are priced out of them, which is quieter and lasts longer. The wealthy are not granted privileges by any statute; they buy them, which sounds like freedom and functions like aristocracy. This is the actuarial state — a regime that has all the coercive substance of the thing it claimed to abolish and none of the public claims that even a flawed republic concedes to its citizens.
The state is not abolished when the courthouse is moved into the insurance office. The monopoly on final judgement is not dissolved when it is reconstituted as a cartel of firms that answer only to their shareholders. The thesis does not abolish coercion; it changes the billing department. And the order it most likely produces is not liberty against the state. It is the state behind a paywall.
Works Cited
Acemoglu, D., & Robinson, J. A. (2012). Why Nations Fail: The Origins of Power, Prosperity, and Poverty. Crown.
Buchanan, J. M., & Tullock, G. (1962). The Calculus of Consent: Logical Foundations of Constitutional Democracy. University of Michigan Press.
Coase, R. H. (1960). The Problem of Social Cost. Journal of Law and Economics, 3, 1–44.
Cowen, T. (1992). Law as a Public Good: The Economics of Anarchy. Economics and Philosophy, 8(2), 249–267.
Demsetz, H. (1967). Toward a Theory of Property Rights. American Economic Review, 57(2) (Papers and Proceedings), 347–359.
Friedman, D. D. (1973). The Machinery of Freedom: Guide to a Radical Capitalism. Harper & Row.
Gambetta, D. (1993). The Sicilian Mafia: The Business of Private Protection. Harvard University Press.
Hoppe, H.-H. (1999). The Private Production of Defense. Journal of Libertarian Studies, 14(1), 27–52. (Also issued as a monograph by the Ludwig von Mises Institute.)
North, D. C., Wallis, J. J., & Weingast, B. R. (2009). Violence and Social Orders: A Conceptual Framework for Interpreting Recorded Human History. Cambridge University Press.
Nozick, R. (1974). Anarchy, State, and Utopia. Basic Books.
Olson, M. (1993). Dictatorship, Democracy, and Development. American Political Science Review, 87(3), 567–576.
Olson, M. (2000). Power and Prosperity: Outgrowing Communist and Capitalist Dictatorships. Basic Books.
Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action. Cambridge University Press.
Rothbard, M. N. (1970). Power and Market: Government and the Economy. Institute for Humane Studies.
Rothbard, M. N. (1973). For a New Liberty: The Libertarian Manifesto. Macmillan.
Tannehill, M., & Tannehill, L. (1970). The Market for Liberty. Self-published (later reprinted, Laissez Faire Books / Ludwig von Mises Institute).
Tilly, C. (1985). War Making and State Making as Organized Crime. In P. Evans, D. Rueschemeyer, & T. Skocpol (Eds.), Bringing the State Back In (pp. 169–191). Cambridge University Press.
Varese, F. (2001). The Russian Mafia: Private Protection in a New Market Economy. Oxford University Press.
Williamson, O. E. (1985). The Economic Institutions of Capitalism: Firms, Markets, Relational Contracting. Free Press.