Thomas Hobbes is the name that is most often invoked, at least in the context of political philosophy, in conversations focused on justifications of government. The story Hobbes tells should be familiar. In the absence of government, the ‘state of nature’ obtains, in which each individual, endowed with roughly equal faculties as everyone else, is engaged in perpetual war with every other individual. We kill each other for resources (competition), we kill each other pre-emptively (diffidence), and we kill each other for egoistic superiority (glory). What is Hobbes’s solution to this miserable state of affairs? For everyone to hand over to a government absolute power and authority (presumably, only over a limited territory; otherwise, there would only be one government in the world). This government, or State (I make no distinction here, although one might), protects persons from one another, and it does so by holding a monopoly on force.

Clearly, such a solution – supposing it could work otherwise – would not work if everybody were as well-armed as the government itself is. That would defeat the whole purpose. The State is an absolute ruler, and this means that it may use force in ways, and to degrees, that no other person or non-governmental agency may. So, there are three features of the Hobbesian State: force, monopoly, and territory. This lines up neatly with Max Weber’s famous definition of the State, which is the most frequently cited: the State is a human community that claims the monopoly of the legitimate use of physical force within a given territory. Even if this definition proves slightly inadequate in some particulars, it captures well enough the essence of what it means to be a State.

One understandable concern which one may have is as follows: what if the State itself, monopolistic wielder of coercion that it is, is corrupt and seeks to use its coercive power for ill rather than for the social good? What if, in other words, the State is functionally a gang which plunders and dominates us, much as an aggressor would do in a state of nature? It is by no means obvious that a State would not do so; and it is therefore not obvious that a Statist society would be an improvement upon one that is Stateless.

Hobbes insists that such is not particularly problematic. He provides a number of reasons for this, but the most compelling reason is that, in order to hold such a State in check, there would have to be another, even more powerful State, which would only serve to create the same problem anew.
So, Hobbes is making two claims here, both of which may be understood (anachronistically) in game-theoretic terms.

1. In the state of nature, where power is divided roughly equally among all persons, it is most rational for the individual to engage with her fellow human being violently rather than peacefully.

2. Where power is distributed unequally, with enormous coercive power concentrated around a single person or agency, it is most rational for the individual to engage with her fellow human being peacefully rather than violently (this includes the persons comprising the State as well).

Unfortunately for Hobbes, both of these theses are false. I wish to spend the most time discussing the second thesis, although I would like to devote some attention to the first claim too, since this is one on which many seem to agree with Hobbes. Thesis 1 is essentially a description of market failure, in which each individual acts correctly in her own narrow self-interest, and yet the net result is to make virtually everybody worse off than if they had all acted otherwise (for a fuller exploration of this, see my discussion ‘Anarchy, State, and Market Failure’). Specifically, everybody would be better off if we all engaged with one another peacefully rather than violently; our lives would not be as ‘solitary, poor, nasty, brutish, and short’ as they are in the state of nature. But, any individual who takes it upon herself to be peaceful rather than violent is making herself worse off: she will have fewer resources (because she does not kill for ‘competition’), she will be in greater danger from others (because she does not kill for ‘diffidence’), and she will have less glory too, although this strikes me as less urgent.

Typically, cases of ‘market failure’, where individually rational decisions add up to sub-optimal outcomes for everyone, are expressed in game-theory by means of the ‘prisoner’s dilemma’ game. However, the best scholarship on the prisoner’s dilemma suggests that there is far more to the ‘game’ of violence than Hobbes’s analysis allows. Robert Axelrod’s The Evolution of Cooperation, published in 1984 and still a subject of discussion in recent literature, explores the ‘iterated prisoner’s dilemma’, which is simply a repetition, indefinitely many times, of the basic prisoner’s dilemma game. Interestingly, the correct strategy for approaching the game alters fundamentally given iteration. Although there is still some dispute about precisely which strategy is superior, there is no question about the fact that the kinds of strategies that perform best are those which are ‘cooperative’; i.e. those which work well with other strategies, and allow them to do well too. Axelrod leverages this insight as a means of explaining the evolution of peaceful interaction between persons, making use of historical examples. His work has proved influential in game theory, economics, political science and evolutionary biology.

Axelrod’s book: http://www.eleutera.org/wp-content/uploads/2015/07/The-Evolution-of-Cooperation.pdf

A recent article, indicating the present state of scholarship on the iterated prisoner’s dilemma, and offering some constructive developments: http://jasss.soc.surrey.ac.uk/20/4/12.html

So, the best contemporary treatments of Hobbes’s (unintentionally) game-theoretic claims indicate that he is very probably wrong. In fact, peaceful cooperation can, does and did occur, even in the absence of a centralised government, because, when repeat dealings are taken into account, the strategies which work best are those which allow others to benefit as well, which implies a tendency towards peaceful cooperation rather than opportunistic predation (Hobbes’s mistake was to consider the prisoner’s dilemma only as a one-off game, but the logic of the game alters entirely once iteration is accounted for).

So much for Hobbes’s first thesis. What of the second? Thesis 2 is also false, but much less controversially. Thesis 2 is a defence of totalitarianism, which few reasonable people would defend. An ‘absolute ruler’ is not a particularly attractive notion, and it is doubtful that it would make the world a better place, relative to its alternatives. What alternatives, exactly? One alternative would be anarchy. My own view is that Axelrod’s game-theoretic insights vindicate libertarian anarchy, although arguing for this directly is not something I shall be doing here (for more reflections on how we might bargain our way up to civilisation out of a Hobbesian state of nature without a government, see David Friedman, ‘A Positive Account of Property Rights’: http://www.daviddfriedman.com/Academic/Property/Property.html).

Another alternative is the view which is most prevalent today among social democrats, conservatives, classical liberals, and practically everyone who is not an anarchist: a limited government. In answer to the question, ‘What is preventing a State from acting as it pleases, dominating and exploiting us endlessly?’, the answer is that there are certain ‘controls’ which act as ‘checks and balances’ on the State’s power (note that, when I speak of the State ‘acting’, I refer to the actions of those persons who comprise it, for collectives do not of themselves have the capacity to act).

So, the success of a defence of the State must provide a satisfactory account of these checks and balances. It must explain convincingly how it is that the State’s power is restrained, so as to prevent a tendency towards totalitarianism. What I plan to do is to examine the three most commonly-cited checks and balances – separation of powers, written constitution, democratic process – and to provide some reasons for thinking that none of them serves effectively to restrain the power of the State. My claim is not that all States are totalitarian, but that there is a tendency towards totalitarianism, such that the checks and balances, to which such confident appeal is typically made, are shown to be unfit for purpose. To put it another way, both the anarchist and the non-totalitarian Statist agree that Hobbes’s thesis 2 is false, but for different reasons. I want to show that the latter’s reasons are inadequate.

The Separation of Powers

First, we look to the separation of powers. Simply, this is the idea that the State is divided into distinct, co-equal branches, each of which has a substantial measure of regulatory influence over the others, such that no single branch may exceed its constitutional mandate and abuse its power. Ideally, it is impossible for any given branch of government to become totalitarian, because the other two branches will not allow it to do so. If the executive becomes ‘too big for its boots’, then the legislature and the judiciary responsibly ‘step in’ and exercise their regulatory influence over it, so as to keep it in check. But is this, in fact, the way that it works? To address this question, we must not ask, ‘What are the three co-equal branches of government notionally intended to do?’ Rather, we must ask, ‘What do the three co-equal branches of government have a realistic incentive to do?’ Do the three co-equal branches of government have a realistic incentive to hold one another in check, and to prevent each other from growing in power?

The correct answer is ‘For the most part, no’. It is not inconceivable that one branch of government exercise its influence responsibly and nobly over another, but this is untypical. In general, there is little or no incentive for this to happen, and in some cases there is an incentive for it not to happen. In order to explain why, let us consider a market illustration. Suppose there are two firms which sell a similar product, and are therefore market competitors. Let us say that they are both vendors of tea. Let us also imagine that each of these firms has some kind of regulatory influence over the other: perhaps they each own land that crosses the other’s import route, so that they can each disrupt the other’s supply of tea leaves. The two firms have a long-standing agreement that they will not exercise their disruptive influence, provided that their opponent’s market share not increase past a certain point (say, 50% of the market). Now, it is possible that one’s market share increase in such a way that does not threaten the other; perhaps the demand for tea increases, and both firms enjoy more sales. But, assuming that the demand for tea remain constant, it seems that any increase in the market share of one firm results in a decrease in the market share of their competitor. So, if one firm starts to increase in ‘power’ (market share), the competitor assuredly does have an incentive to make good on its threats and implement whatever regulatory controls it has at its disposal to hold it ‘in check’.

But what if the two firms are not competitors at all, but sell complementary products? Suppose that one firm is a tea vendor and the other is a vendor of chocolate biscuits. Tea and chocolate biscuits are complements: ceteris paribus, as sales of one increase so do the sales of the other. If, therefore, one firm in this scenario attempts to boost its sales and increase its market share, this is actually to the benefit of the other firm. Each does well when the other does well.

What does this tell us? It suggests that, when two or more agencies are in a position to exercise regulatory influence over the other(s), they have a realistic incentive to do so if the agencies are competitive, and they do not have a realistic incentive to do so if they are complementary. Are the executive, legislative and judicial branches competitive or complementary? The answer is implied in the question itself: they are each providing a different and complementary service. The legislature passes laws, the judiciary interprets laws, and the executive enforces laws. The more laws are passed, the more laws there are to interpret and enforce. What is more, all three branches of government benefit from perpetuating the following (untrue) commitment in the minds of their citizens: that the services which each branch provides can only be provided effectively by them. Quite simply, it is in the interests of the State in toto to maintain and grow the power of the State in toto; it is plausible that the growth of the power and influence of the State involves the growth of all three branches of government. In this sense, they have a ‘common cause’. Suppose that the majority of people could be persuaded that the State is not necessary in order to have a system of rights-enforcement and dispute-resolution. All three branches would be equally threatened by such a phenomenon.

This would imply that, typically, the three co-equal branches of government do not have a realistic incentive to exercise regulatory control over the others. What, then, would a true ‘separation of powers’ look like? We have already seen the beginnings of an answer in the analysis we have given. Competition separates powers. Suppose that, instead of power being separated between three different and complementary service-providers (i.e. law-producers, law-interpreters, law-enforcers), power was separated across n many competitors. One may suggest that this already obtains, since, if one does not like one’s own State, one may simply leave and go to another. But this, I believe, is to miss the point. States are territorial sovereigns. When we speak of checks and balances, we are looking for controls on the State’s power over the territory it governs. If the ‘You can always leave’ argument is taken seriously, then there really is no reason why the government should not be totalitarian after all. What we are looking for, then, is a separation of powers within the territory, and not between territories. And, as we have seen, it is markets, and not monopolies, which separate powers.

It is beyond my purpose here to provide an account of rights-enforcement and dispute-resolution in a Stateless society, but for explorations of these issues, see David Friedman, The Machinery of Freedom: Guide to a Radical Capitalism.

The Written Constitution

Next, let us consider the limits established by a written constitution. This provides a set of formal boundaries, beyond which, notionally, the government may not act. Clearly, the mere existence of a written constitution serves as no protection in itself. Minimally, three further things are required. First, it must be created in a way that is fair (consider, for instance, if a constitution did exist, but it afforded the State totally unacceptable powers). Second, it must be interpreted in a manner that is fair and competent. Written constitutions are produced in specific historical contexts, and cannot possibly anticipate the myriad ways in which the world may change, let alone the ways in which its own language may change. Third, and perhaps most importantly, it must be enforced. At the risk of glibness, a piece of paper does not stop the bullet from a gun. It must be accompanied by the commitment, on behalf of those who are equipped to do so, to enforce its stipulations.

A written constitution, then, must be accompanied by certain services: law-production, law-interpretation, and law-enforcement (a constitution, after all, is essentially a kind of law). These services are familiar: they are the services provided by the State itself. But this raises a problem, that the State is, in effect, made the judge in its own case. Since the State is responsible for establishing its own limits, interpreting the sense of these limits as time moves on, and seeing them enforced, there is a clear conflict of interests. It is not at all difficult to appreciate the nature of this conflict. Suppose that the State consisted of precisely one supreme ruler, who is responsible for establishing, interpreting and enforcing the limits on his own power. Clearly, he cannot be trusted to do so. If he could, it would only be because we have judged him to be a trustworthy ruler, in which case the constitution would be unnecessary in any case. Suppose, instead, that the State consisted of a group of rulers, all of whom were equally responsible for establishing, interpreting and enforcing their own limits. Would this be an improvement? Hardly; if they are all co-rulers, then they all share a common interest, that of growing the power of the State, and interpreting its constitutional limits as liberally and permissively as possible.

It is at this point that the State’s apologist will draw attention to the fact that, while the State is indeed responsible for the production, interpretation and enforcement of its own limits, these services are provided by different branches of government, and therefore it is not a straightforward case of a group of individuals serving as their own judges. Appeal is being made, in other words, to the separation of powers. For this reason, I suggest that the ‘checks and balances’ provided by a written constitution do not serve as an extra defence, over and above the separation of powers. Rather, the defence against totalitarian tyranny which the constitution is intended to provide can only be trusted if we already have some independent reason for trusting the separation of powers. However, I have already provided reasons for thinking that the way in which power is separated in the States of our acquaintance is not adequate to protect against totalitarianism. This, in turn, would seem to render ‘constitutional limits’ largely impotent.

As John C. Calhoun observed, scathingly yet lucidly, in his A Disquisition on Government, it will always be in the interests of the larger and more influential political party to exercise the powers granted by the constitution, and to ignore its limitations as far as is possible. Meanwhile, the weaker and less influential party will have the opposite incentives, to hold the larger party to obey its restrictions, and to resist and question every exercise of power, whether legitimate or not. The larger party will seek to ignore and flout its restrictions, and the weaker party will seek to uphold them. What will the outcome be? ‘In a contest so unequal,’ says Calhoun, ‘the result would not be doubtful.’ The larger and more powerful party will tend to be successful in enforcing its own interpretation of the limits placed upon it, and so, over time, there is an inevitable tendency towards a ‘liberalising’ (in a pejorative sense) of the constitution’s interpretation, and thence to totalitarianism.

Democracy

We look, finally, at the democratic process itself. The common-sense or ‘naïve’ model of democracy is that, if those in power start to exceed their constitutionally defined mandate, we, the people, simply vote them out. A minimal requirement for such an arrangement to succeed is that voters be, in general, politically well-informed. Unfortunately, from an individual perspective, there is very little reason to become well-informed. Suppose you live in a society, virtually all the members of which are politically well-informed already. If such is the case, there is little reason for you to take it upon yourself to become well-informed as well, for the rest of society can probably already be relied upon to reach the ‘right result’. Suppose instead that you live in a society, virtually nobody in which is well-informed. Here, again, it profits you little to become suitably well-informed, as here, too, your vote is highly unlikely to impact the outcome of any decision-making process. This, sadly, is an unavoidable ingredient in any collective decision-making process (except on an extremely small scale): because the individual’s contribution has mathematically negligible odds of swaying the final decision, it simply does not pay the individual to take it upon herself to become well-informed, given the significant costs in time and intellectual power that is required in order to do so.

The proof is in the pudding. From experience, we know that the average voter is woefully ignorant. Countless studies have shown that a majority of voters cannot even name their constituency’s representative, let alone enumerate their representative’s voting decisions on recent bills. Even the rudiments of a country’s political system are unknown to an astonishing number of people. We can therefore understand the voter’s incentive structure as giving rise to a kind of market-failure: sure enough, if every voter becomes politically well-informed, then this will probably yield the best result possible, but any individual who takes it upon herself to do so is taking a course of action which costs her more than it benefits her.

But, even supposing that a substantial number of voters were suitably well-informed, this of itself is not sufficient to keep a State from exceeding its constitutional limits. Voters, understandably, will tend to vote in their own immediate interests and in favour of the politicians who appear to represent their own values. If the predominant party happens to be the one which suits a given voter, that voter has little reason to vote against it. Let us remind ourselves that the State is an agency of coercion, and it is in the nature of coercion that it tends to be predatory in nature, benefitting one party at another’s expense. It follows that a State can only ever be a weapon to be wielded by some (to their benefit) against others (at their expense). In a representative democracy, the majority are positioned to benefit at the expense of those who are outvoted, regardless of whether or not the State exceeds its prescribed mandate.

Let us lay these considerations aside. Let us assume, counterfactually, that a substantial majority of voters are both well-informed and care suitably about the State acting appropriately. What kind of effect will this have on the actions of the State? A private firm must continually satisfy its consumers in order to survive; the profit-and-loss system encourages the firm to be hyper-sensitive to the degree to which it effectively satisfies consumer preferences. If a large firm has a commercially poor week or month, it must ask itself what it must do immediately so as to reclaim its customers, who are at liberty to shop elsewhere upon the turn of a dime. The State, on the other hand, hardly faces such constraints. Elections are typically held once every four or five years, and so a voter whose preference alters must wait as long as this before they can select a new political preference. Even then, unless the voter can persuade suitably many of her neighbours, she may continue to be governed by a person or party whom she does not endorse. Contrast this with the freedom she has as a consumer: she may decide, as she lies in bed at night, to patronise a new supermarket in the morning, effective immediately, and her changed preferences are not dependent upon her successfully persuading her neighbours to shop where she shops.

Finally, and perhaps most crucially, the representatives for whom we vote do not have an enforceable contract with us. If your elected representative violates every promise he makes on the campaign trail, you have no legal recourse. All you may do is wait for however many years it takes, cross your fingers, and hope in vain that your next choice has more integrity.

How Optimistic Should We Be?

In the above treatment, I have tried to show that there are good reasons for seriously doubting the effectiveness of the ‘checks and balances’ which are notionally intended to restrain State power. Recognising the inadequacy of these mechanisms should be cause for serious thought. The question with which we began was: 'Given that the State holds a monopoly on force, what is preventing it from acting as an aggressor in a state of nature would act?' Nothing less than the philosophical justification of the State may hang on this question. If, it transpires, there is little reason for thinking that the State is going to obey the limits placed upon it, then it may be little improvement at all. For clarity’s sake, I am doubtful that Hobbes’s analysis of the state of nature is the correct one, so what I am not arguing is that the ‘war of all against all’ is preferable to the present Statist situation. As Axelrod, Friedman and others have shown, there are good game-theoretic reasons for thinking that a State is not a necessary precondition for the emergence of peaceful cooperation; though, admittedly, this is a thesis to which I have not done justice in this piece. But if, as I argue, we have good reason for thinking, not only that Hobbes’s totalitarianism is mistaken, but also that the defences that are set in place to protect us from totalitarianism are not fit for purpose, we might have reason to look again at anarchy as the least-bad alternative. A positive defence of anarchy will, however, have to wait for another occasion.

I will finish with this thought. It is often assumed that the anarchist must be terribly optimistic about human nature; she must take the view that human beings are essentially peaceful and selfless creatures who can be trusted to do the right thing, in the absence of a restraining influence. Meanwhile, those who are suitably pessimistic (or perhaps simply 'realistic') about human nature ought to want a State, so as to protect people from each other. Statism, so it goes, is the more grounded, sober, realistic position, which takes human nature as it is, and which takes human incentives seriously. I hope to have given pause to someone who might take such a view uncritically.