The Law of Constitutional Capture

3.1 Legal Systems are Prediction Systems

In his landmark The Concept of Law, H.L.A. Hart^[54] observed that at the heart of every legal system is a rule of recognition that tells what counts as law.^[55] In the European civil law tradition, Hans Kelsen^[56] proposed the comparable concept of a Grundnorm.^[57] When people come together in community, their success in living together involves a shared understanding of how sayings can become the law that guides their life together, including the law that says who leads, who governs, who can make more law. Hart argued that this shared understanding was needed at least among key people whom we could circularly call ‘officials.’^[58] Of course, the status of those people as officials depends on the very law that they together recognize.^[59]

What makes the rule of recognition real? The fact that enough key people are treating it as real. Its reality is self-generating, self-fulfilling. People treat as law the sayings that are pedigreed by the emerging system as law. People treat as leaders those who are pedigreed by the emerging system as government. Why do people take the system seriously in this way? Because they expect that other people will do so and will share such expectations in turn. Law and government self-generate through converging expectations that produce converging conduct that feeds back to reinforce shared expectations. People follow because they expect others will follow, and that following by others creates strong reasons to follow too.

What are the strong reasons for following law? First and foremost, that we cannot live together in large groups without a way to share understanding about what others are likely to do and to expect when our paths cross. Jürgen Habermas^[60] observed that ‘the very function of law’ is ‘to stabilize expectations.’^[61] To help achieve this, legal systems often signal how others will respond to failures to conform to law’s signals, and how still others will respond if the designated first responders fail to act as signaled. As the game theorist Ken Binmore^[62] puts it:

If someone deviates and we go to one of the subgames off the equilibrium path, it remains optimal in a perfect equilibrium to stick to your original strategy provided everybody else does. In particular, if your strategy tells you to punish the deviant at some cost to yourself, it will be optimal for you to carry through on the punishment. If you deviate by trying to escape your duty to punish, you will take us to yet another subgame where it is optimal for some other player to punish you for your dereliction of duty. If he fails to discharge his duty, we go to yet another subgame – and so on forever. With only a finite number of players, these chains of responsibility are necessarily closed in the manner that Kant failed to consider.^[63]

Oliver Wendell Holmes^[64] called law ‘systematized prediction.’^[65] Law is a signaling system^[66] for what people in law’s community are likely to do and to expect.^[67] A legal system can emerge through customs of following leaders that put their community on a fast track to many more customs. Customs of following leaders can sprout branches of further following as leaders delegate some leadership roles to others, expanding law’s network of effective signalers and signals.^[68] Each participant looks from side to side and sees strong reasons to act as expected. All this mutual expecting of actions and expectations creates a shared internal point of view^[69] among participants toward the say-so of those who lead.

In his original book, Hart articulated a ‘fundamental objection’ to seeing law as all about prediction. He observed that ‘where rules exist, deviations from them are not merely grounds for a prediction that hostile reactions will follow or that a court will apply sanctions to those who break them, but are also a reason or justification for such reaction and for applying the sanctions.’^[70] When we see that law’s self-fulfilling, custom-creating character is what makes it real and valuable to our lives, we have a complete answer to Hart’s objection. We have a strong moral reason to keep law real, so that we can live together in large groups. That moral reason informs our legal systems’ characterizing conformity as a duty. Law is real only if it succeeds in signaling what people in law’s community are likely to do and to expect. Keeping law real means keeping up its predictive effectiveness, which depends on most of us most of the time doing what it predicts we will do. That strong moral reason to do as law says we will do is pro tanto, not preemptive, and can of course sometimes be outweighed by stronger moral reasons to do differently. But as Binmore observes,^[71] it is bolstered by the way the system self-generates incentives to conform and deterrents against departure.

In his postscript to The Concept of Law’s second edition, Hart indirectly seemed to concede law’s essentially predictive character by acknowledging that the rule of recognition is a convention.^[72]

Rules are conventional social practices if the general conformity of a group to them is part of the reasons which its individual members have for acceptance. … [C]onventional social rules … include, besides ordinary social customs (which may or may not be recognized as having legal force), certain important legal rules including the rule of recognition, which is in effect a form of judicial customary rule existing only if it is accepted and practised in the law-identifying and law-applying operations of the courts.^[73]

If at the heart of the legal system there is just shared expectations leading to shared practices, just predictable patterns of customary behavior, then all of the law and leadership that branches out from that heart grows ultimately from shared expectations. To reassign David Hume’s^[74] notorious claim about authority, we have law and leadership ‘because every one thinks so.’^[75] What everyone thinks does not create preemptive moral duties to obey, as Hume surely understood. He elsewhere famously observed that no ‘ought’ follows merely from an ‘is.’^[76] What everyone thinks does not in itself determine conclusively whether someone ought to be obeyed.^[77] But what everyone thinks does in itself determine whether there is a shared expectation about what and who will be followed, and it is from such shared expectation that we have law and government.

Hart memorably distinguished law from the threatening demands of a bank-robbing gunman.^[78] That contrast is made particularly resonant by the ad hoc, arbitrary character of the robber’s behavior. Law makes our relations with others less ad hoc, more predictable, and so lets us live together in large groups. But those who make our laws are not necessarily more morally justified in leading as they do than the bank robber is in brandishing his weapon. Some leaders of nations really are robbers. That does not stop their words from being law and their gangs from being government. Using the language of norms and normativity in accounts of law and government can risk conflating the concept of being normal with the concept of being moral. Law helps establish what is normal, what is customary, in the life of a society, and we have strong moral reason to want there to be a known truth about what is normal, so that we can live together.^[79] What is normal may, however, be bad. That’s the point of legal positivism – what’s legal isn’t necessarily what’s moral, all things considered. Of course, we would rather have good leaders than bad ones, but good or bad, people are lawgivers if their society in fact treats their words as law. Whether lawgivers came to power in good ways or bad, and whether their acts of purported lawgiving are morally justified or not, their words are law for us if and only if enough people in our community understand and expect them to be law. Our ever-present reason to care about law is not that we owe something to lawgivers, but that we need law’s help to live together.

3.2 Legal Systems Tell Stabilizing Stories

Feeding the feedback loop that lets us have law, that lets us live together in a large society, is a weighty moral reason for following the law. It is not difficult to see how much we benefit from the expectations that law helps us have, the confidence we acquire from law about what strangers will likely do at an intersection or in a commercial transaction. Yet throughout our history, lawgivers have wanted more moral support than that.^[80] Throughout our history, lawgivers have sought stories that tell not only why having law is good, but why having them as the lawgivers is good, and better than alternatives. At a minimum, these stories need to explain why the leaders are morally justified in leading. Beyond that, leaders have sought to say that their people are morally required to follow. To secure the duty to follow against claims that sometimes we have stronger reasons for not following, our leaders historically sought to claim that they had moral rights to be followed, rights that preempted, overrode, or outweighed any reasons their people might ever have had for not following. This was the old idea of authority.

Throughout the ancient world, and widespread until recent centuries, leaders advanced variations on the same story to support their rule: the creator or creators of humans had a right to be obeyed by humans and had delegated that right to human leaders.^[81] That argument helped sustain leaders staying leaders longer than people wanted to follow them. The argument pointed to the source of the very qualities that first let the leader or their ancestor lead – the creator or creators of the leader and everyone else.^[82] The leader had a right to lead their community, it was said, a right delegated by the divine.^[83] We see that claim made by leadership in many times and places. We see it still in the motto on the British monarch’s coat of arms – Dieu et mon droit.^[84] In Rousseau’s^[85] words: ‘The strongest is never strong enough to be always the master, unless he transforms strength into right, and obedience into duty.’^[86] The old claims of a divine right of kings had a conceptual coherence and a moral reach that no substitute theory of authority has ever attained. As exemplified by the biblical story of a divine command to Abraham to sacrifice his son Isaac,^[87] the divine right to be obeyed was paramount and preemptive. It could not be morally outweighed by any competing considerations.^[88]

When the Enlightenment demolished claims of delegated divine rights to lead, human leaders’ quest for stabilizing stories to support their rule shifted from claiming their leadership was the will of God to claiming their leadership was the will of ‘the People.’ Yet no one then or since has ever come up with a credible way to claim that popular support for leaders is reason enough for us to shut down our own moral reasoning and defer blindly to those leaders. Social contract^[89] notoriously does not supply such a reason, both empirically^[90] and because even when promise-making does happen, promise-keeping is a good of obviously only finite and relative weight,^[91] certainly not always the right thing to do.^[92] And Razian epistemic authority^[93] is neither about having rights to be followed nor a source of the reliably continuous moral guidance that the old preemptive idea was supposed to supply.^[94] What’s the point of an ‘authority’ that is not coextensive with the claimant’s power and not clear ex ante about its own extent? ‘Piecemeal’ authority^[95] would not really help us know when we should follow. We follow law not because we owe moral duties of ‘allegiance’ to lawgivers;^[96] we follow law, when we do, both because we can see that doing so will make our own lives better in many ways and because we owe moral duties to one another as people living together in community.^[97] There is no symmetry between when would-be leaders are morally justified in doing acts that lead and when other people are morally required to do acts that follow.^[98] The old idea of authority tried to bind the two together, and abjectly failed to do so.^[99] Yet that failure did not stop law and government from achieving what human communities have always needed them to achieve. Law and government let us live together in large groups. The notion of a preemptive right to be followed is a fifth wheel, as Holmes put it.^[100] Legal systems can work just fine without it. The value of an effective prediction system is obvious and morally weighty, without being preemptive.

Fortunately for leaders, the value to them and their legal systems of a stabilizing story does not depend on that story claiming for leaders a preemptive moral right to be obeyed. It is enough that the story helps morally justify the leaders in leading. The substitute stabilizing story of ‘will of the People’ can provide that moral support for leadership. And once up and running, legal systems can self-sustain regardless of the leadership’s bona fides.^[101] By systematically and credibly predicting good consequences from conformity and bad consequences for non-conformity, prediction systems can keep themselves going^[102] regardless of whether the way leaders reached power and what leaders command is good or bad. Yet even the most morally bankrupt leader sees value in having a stabilizing story to tell as an aid to perpetuating their power.

Stabilizing stories are especially valuable at moments of discontinuity, when revolutionary change visibly breaks with the past. New governance becomes real, new sayings become law, when enough people in a community expect that enough other people in that community will treat those new sayings as law, that is, will treat those sayings as expressing what people in the group are likely to do and to expect. What mattered crucially about the circumstances of the United States Constitution’s adoption was not that the process of adoption through elected ratifying conventions was really the most morally worthy one. That ratifying process was a revolutionary departure from the amendment rules of the existing legal system, and leading opponents condemned the process for that reason.^[103] What mattered crucially was that having the document endorsed by all those specially-elected people in all those specially-called conventions helped turn the United States Constitution into what the American people were thereafter likely to treat as the law. Initially fragile expectations about the Constitution’s status strengthened as, month by month, people continued to act as if it was the law by doing the acts that it contemplated – holding elections, convening, governing. We could call the converging thoughts behind these converging actions the internal point of view.^[104]

3.3 Changing the Rule of Recognition

When we see how crucial a real legal system is to enable living and acting together with large numbers of other people, we can see why people living inside legal systems so rarely succeed in rising up to overthrow those systems. Even when a system of law and government is a tyranny that makes most of its people miserable by governing in disregard of their interests and offering them no easy way to change leaders, people only rarely rise up to force revolutionary change. Overthrowing an existing legal system by force is hard because bringing people together against the existing legal system is hard. It is the existing legal system that reliably coordinates our actions with large numbers of people we do not personally know. How can we know what others will likely do and expect when the law is not on our side, pointing the way? If we rebel against existing law and government, we risk making the sound of one hand clapping. And we know that even if many people do come out on the streets, we will have law against us, coordinating riot police and soldiers and all manner of other officials of the existing system. They will likely act as the system predicts they will, not because they love it much more than we do, but because a giant collective action problem traps them and us in cells of ignorance to which the existing legal system holds the key.

When extralegal overthrow does happen, it is often because a sub-system of the existing legal system with disproportionate access to means of force detaches itself from the rest of the legal system and uses its internal legal system to attack its parent. We call that a military coup. Those who lead the coup rely on military law to ensure their subordinates’ compliance and coordinate their conquest of the mothership of state.

Military coups depend on part of the preexisting legal system that sets up chains of command. That dependence often defaults ultimate post-coup power to a group of senior military officers. The usual fate of such military juntas tells us something very valuable about which distributions of governing power best support democracy and which distributions of governing power best support tyranny. In their analysis of the empirical evidence, political scientists Barbara Geddes, Joseph Wright, and Erica Frantz^[105] identified studies showing ‘that dictatorships led by somewhat collegial groups of officers (juntas) end sooner than other kinds of dictatorship.’^[106] There are two ways that such tyrannies by committee can end – by moving toward democracy or by consolidating tyranny in one pair of hands. It is concentrating ultimate power in one person that lets tyranny last long term.^[107]

Why does tyranny by committee tend to be transitional? A government unaccountable to its people becomes estranged from their interests and intent on the interests of leadership, but having no true purpose higher than self-interest is destabilizing to a group. A group focused on self-interest will end up preying on one another.^[108] Violent overthrowers vary in how ideologically fervent or transparently self-serving they are, but ostensible ideological fervor absent actual accountability to the people has a poor track record of maintaining shared governance. Members of a tyranny by committee therefore face a choice – move toward democracy so government, disciplined by accountability to its people, can truly be for the common good, or collapse into a tyranny of one, in which government becomes about serving that person’s interests and preoccupations. As Milan Svolik^[109] puts it:

[T]he reasons for the emergence of personal autocracy are structural. … [A]uthoritarian elites operate under distinctly hazardous conditions. They cannot rely on an independent authority to enforce mutual agreements, and violence is the ever-present, ultimate arbiter of their conflicts. These dismal circumstances ensure that any dictator’s aspiration to become the next Stalin is matched by the opportunity to do so. … Rather than an accident of history, the emergence of personal autocracy is a systematic phenomenon.^[110]

Tyrannies by committee become tyrannies of one because singular leadership turns out to be the ideal model for building a system of law and government that is unaccountable to the governed and focused on the self-interest of those who govern. By being clear about whose interests ultimately matter,^[111] whose interests those in government will be rewarded for serving and punished for not serving, government under a singular leader can organize people into a governing system in which it is plain to each participant what they need to do to be rewarded and to avoid being punished. That can create disciplined conformity with a minimum of communication, indeed such a system can reliably serve the interests it exists to serve even if that service contradicts much of what the government formally says.^[112] Singular leadership allows for underarticulated signaling of expectations.^[113] Governing is about what’s best for the leader, which well situated participants in the system are rewarded for being adept at figuring out. It is the leader who supplies the focal point for shared expectation. The rule of recognition for officials in such a system becomes assimilated to the will of the leader, which can be discerned, absent explicit instructions, from asking what would serve the self-interest of the leader – their glory, their power, their permanence.

One true tyrant can and often does create an incentive structure that lets tyranny last. Government by one person produces the stable equilibrium of tyranny: a power pyramid with one bright point at the top that is the focus of shared expectation, converging participants’ understanding of how to behave to succeed in sharing in the spoils of the system without falling out with one another.^[114] An exception that helps prove the rule is government in bifurcated societies, such as in the American South under Jim Crow.^[115] Shared governance in service of a common good can sustain itself in such societies through its actual accountability to the segment of society whose common good it serves, even as simultaneously it may tyrannize another segment of society for generations.

Tyrannies that share ultimate power are playing on democracy’s turf. Democracies with one true leader are playing on tyranny’s turf. And that brings us to constitutional capture, a situation where there never was a violent overthrow, there was just an election that brought someone to power inside a democracy who wants to change the system. When we see that the rule of recognition comes from converging expectations, we can see that the rule is vulnerable to shifts in expectations. If someone becomes situated to shift expectations and converge them elsewhere, then the old rule’s handiwork, the existing formal legal system, may be hollowed out and occupied by a shadow system that grows from expectations converging elsewhere. A power configuration that creates the salience to grab collective attention and shift the focal point for shared expectation may change the rule of recognition.^[116]

For would-be tyrants, capture may be better than overthrow in several big ways. First, capture can modulate its speed to maximize its chances. Second, capture can turn the existing system against itself. Third, capture can coopt the existing system’s stabilizing story.

By maintaining continuity of institutional façades rather than just bulldozing existing institutions and setting up new ones, capture can happen slowly over time in ways that avoid spontaneous mobilizing and coalescing of opposition. Opponents can be isolated and picked off one at a time. Stealth and subterfuge allow the new system to emerge in a more stable way that improves its prospects of success. Normally it is necessary to go slow because without overt threats of overwhelming violence there is likely to be resistance, and victory requires lots of small steps that wear away the points of resistance. That makes the democratic ‘erosion’^[117] metaphor especially apt.

Attempts at extralegal overthrow may face formidable multifaceted resistance from an existing legal system. In contrast, capture uses levers of the existing legal system to change formal structures and official practices from within in ways that stop existing institutions from fighting an emerging shadow system and subject those institutions to that emerging system. A military coup detaches and deploys one subsystem against the parent legal system, but capture uses much more of the existing legal system to self-subvert in slow motion. Capture also uses the levers of the existing legal system as sources of favors and fear that help staff a shadow system of personally loyal and dependent officials. Those participants come to understand that future plum jobs and lucrative other opportunities for themselves and their families depend on conforming not to signals of the formal legal system, but to the signals of the shadow system. Sometimes capturers have the added asset of an existing party or movement structure that can slide seamlessly into staffing the government in service of the shadow system. But incumbency can enable capture even without such readymade sources of placemen, as Kais Saied is demonstrating in Tunisia now.^[118] What can sufficiently guide and persuade and coax and even inspire disparate people to come together to pull off a constitutional capture? One salient person’s will.

Overthrowing an existing system can be a group effort, perhaps by military coup, perhaps coming up spontaneously from the street with no single salient face behind or in front of it. But capturing a system of government requires a process of attaching strings to people who were formerly free, and almost imperceptibly turning them into puppets. That takes a puppet master, someone who is situated to grow the relationships that attach the strings. Capturing the constitution of a free people starts by capturing particular people and in small steps turning them unfree.

Personal loyalty is a powerful tool to swap out one prediction system for another. Once we see that law is all about signaling what people in community are likely to do and to expect, we can see that a shadow system can grow by supplying better guidance than the preexisting formal system about what people are likely to do and to expect. Singular leadership readily enables expectations to converge anew. The salience of the one true leader, the relational dynamic between leader and followers, the personalizing of the emerging system’s raison d’etre, all help shift shared expectations to what the leader wants or is likely to want.

Personal loyalty can build a new pyramid-shaped prediction system inside an existing system of constitutional checks and balances and gradually capture it. People can be bullied or bought into personal loyalty. Our singular executive models supply one person with just the tools they need to end up atop a personalized power pyramid. A leader with the will and the skill can spin webs of dependent relationships through favors and fear to concentrate power in fewer and fewer hands and ultimately all in their own. They can embrace the forms of democracy en route to a government of, by, and for the leader. Through this creative process, the leader can circumvent and even repurpose formally countervailing power centers by filling them with persons who will follow the leader because they depend on the leader. Those institutions will no longer check and balance, they will just acquiesce according to the agenda of the shadow system. They can then be used to help sideline any remaining institutional holdouts. This is what Vladimir Putin did in Russia.^[119] This is what Xi Jinping is almost finished doing in China.^[120] This is what many other leaders discussed in the literature on constitutional capture seem to be trying to do too. And as Hart explained, the shared expectation that supports the new rule of recognition needs only to inform key insiders, the so-called officials.^[121] Everyone else can be kept in the dark about the true shape of the shadow system that really governs; everyone else can continue to be presented with the forms of the old system that has been superseded and that now operates as a passthrough for rules of behavior dictated from the shadows. And continuing the forms of that old system lets the shadow system coopt the old system’s stabilizing story.

When the menu of stabilizing stories shrank from including the will of God to including only the will of the People, would-be tyrants wanted a way to tell that story. Today’s tyrants sense that they need the forms of constitutional democracy. The less the reality of democracy – the more of a sham its forms are, the emptier its institutions are going to be – the more valuable its institutions’ familiarity is. The more the institutions look the same and formally act the same – use the same procedures – the easier it is to obscure the fact that those institutions are acting under instruction from a puppeteering shadow system. The stabilizing story the institutions are pantomiming can be wrapped around actions dictated by a shadow system. This stabilizing strategy has a long pedigree. Edward Gibbon^[122] memorably recounted the care with which actually omnipotent Roman emperors deployed it, masking their power and contriving to keep the Senate looking like it mattered. Rome became ‘an absolute monarchy disguised by the forms of a commonwealth. The masters of the Roman world surrounded their throne with darkness, concealed their irresistible strength, and humbly professed themselves the accountable ministers of the senate, whose supreme decrees they dictated and obeyed.’^[123]

Today’s leaders have no coherent way to justify their leadership other than the people’s support. So they choose to pretend that the support exists. They choose to claim that government comes from the people, and they choose to maintain the forms of institutions that look responsive to what the people want. This they can do, as we see in nation after nation, while turning the forms of democratic government into puppets controlled by a shadow system that keeps the incumbent leadership securely in place through its control over who gets to occupy the formal institutions of government. Government can continue in name to be all about the will of the people, just as it was formerly in name all about the will of God.^[124] The new claim can continue to be made without being any truer than the old claim. Where the old claim just needed the formal blessing of religious institutions, the new claim needs to be played out through the institutions of constitutional democracy – elections, assemblies, courts. To capture these as securely as past monarchs established temples and churches requires some agility. It requires nurturing a set of shared understandings and expectations that become a shadow system. Through those shared understandings and expectations, the shadow system can gut each of the great constitutional guardians of democracy, and then operate inside their shells.

3.4 The Role of the One True Leader

Without one true leader, it is hard to derail a democracy. Without one true leader, it is hard to stay a tyranny. In 2020, Russia’s puppet master, Vladimir Putin, orchestrated a sweeping rewrite of the Russian constitution. Along with a range of nationalistic and social welfare crowd pleasers, the changes formally concentrate more power in the president over all the other institutions of government, and blow past the old presidential term limits.^[125] Under the new dispensation, President Putin’s actual hold on power has an official blessing for almost as far as the eye can see.^[126] In 2022, it was Xi Jinping’s turn to burst through old expectations. He ignored the precedents set by China’s recent leaders and secured an extension of his leadership beyond a decade, perhaps far beyond.^[127]

What enables Vladimir Putin and Xi Jinping to perpetuate themselves in power? Leadership of shadow systems that make other participants so dependent on them for advancement, so deeply in their debt, or so afraid of harm, that those other participants will do whatever best serves the leader. Those participants are convinced that other participants in the system will do whatever best serves the leader, so they had better do so too. When we see how legal systems are built on converging expectations, we can see how a leader can build their own master system in the shadows and deploy it to control the formal outcomes of the formal legal system. From legislators to judges to party bureaucrats, those who have a formal say in whether the leader stays leader will formally say what the leader wants them to say. Vladimir Putin and Xi Jinping will continue to concentrate power in their own hands because they understand what makes tyranny work. From atop a pyramid, they can harness patronage and fear to ensure that everyone positioned to vote on their leadership will vote the right way.

These late-stage stories exemplify the role of singular leadership in a capture process that can start small and grow over time even inside real democracies, which Russia barely was, and China never was. Singular executive leadership is the pivotal structural opening for this insidious transformation. The office of chief executive gives one person powers to hire and fire and spend and command. Those formidable powers can create focal status and secure personal loyalty. Ultimacy lets the one true leader shift the spotlight and become the focus of future expectations. And inside real democracies, having one chief executive has a range of other, complementary subversive effects on elections and political parties that show up sooner and set the stage for capture.

Having one chief executive dangerously raises the stakes of elections. Electoral contests to be president or prime minister visibly pave the way for a transiently popular person to parlay their transient popularity into apex stature and power. That stature and power then open up a well-worn, well-understood path to turning temporary power permanent. In winning election and then in initiating capture, a candidate may be aided by group psychological aspects of their popularity, in particular the dynamic that Max Weber^[128] called charismatic authority.^[129] The sorts of personalities drawn to compete in such contests, the loyal followings they often attract, and the frequency with which those circumstances have led to tyranny in the past can fuel suspicions and fears on all sides that undermine confidence in the democratic process and scuttle expectations of a democratic future.

Having one chief executive also directly and deeply affects why and how political parties exist and function. Where the top of the electoral ticket is a contest for chief executive, political parties readily become preoccupied with winning that contest and scoring the rewards that flow from its outcome to those on the winner’s team. As someone has to be chief executive, parties do not necessarily need broad and diverse support to get their candidate over the finish line first. Then they have their foot in the door to an awesome array of powers and privileges. The spoils of executive power – the jobs, the contracts, the regulatory accommodations for allies -- can become what really hold parties together, rather than actually shared beliefs in and commitments to policy ideas for serving the common good. This debasing of political parties’ raison d’etre predisposes and prepares their members to participate in constitutional capture, so that the rewards of incumbency can be theirs in perpetuity. And in the meantime, their fixation on winning the executive can give them a reason when out of power to want the incumbent government to fail at being a good government. That can cause them to use the checks and balances of constitutional democracy in obstructive and destructive ways, so that what is best for the people cannot happen on someone else’s watch. Having one chief executive in a democracy is a source of gridlock, which in turn can be cited in support of steps that weaken constitutional checks and balances and begin the process of capture. And when capture has happened, the truth of the system ceases to be democracy and becomes something else, something even dispassionate observers may call a tyranny, a despotism, or more often now in the clinical prose of the academy, an authoritarian regime.

The existing literature on so-called authoritarian regimes is puzzlingly unclear about why that description is apt^[130]– what distinctive relation to authority do authoritarian regimes possess? When we speak of authority, we sometimes mean to invoke an idea about the source of law and government. On other occasions, and more often now, we use the word merely as a synonym for law and government. When a judge asks counsel do you have any authority for that proposition? she means do you have any case law? When we speak of the local authority we mean the local government. Synonymous usage will likely always be with us, but so used, the word does not distinguish one kind of government from another – all real governments are authorities. On the other hand, authority as a source of law and government was the idea that particular persons had rights to lead and everyone else had duties to follow them. The old claims were top-down: authoritative leaders were delegates of divine rights to be obeyed.^[131] Since the Enlightenment, most of humanity has moved emphatically to bottom-up visions of how law and government come to be. We have moral reason to choose and follow leaders because law and government let us live together in large groups. Our duties to follow are owed not to leaders, but to one another as people living in community.^[132] This difference offers a clue to what lies behind the intuition to call certain kinds of leadership authoritarian. That leadership shares with the monarchies of old an aspiration to hold more power for longer than dependence on popular support would reliably allow. Unlike the old monarchies, however, most modern authoritarians cannot overtly assert a coherent theory of authority in support of their long-term leadership.

Authoritarians are not so called because they are distinctively sticklers for rule following. As Juan Linz^[133] recounted, one was more likely to be reprimanded for minor infractions in democratic Switzerland than in Franco’s Spain.^[134] Authoritarians are so called not really because of their attitude to rules at all, but because of their attitude to themselves. They want to make government all about them, to claim like the monarchs of old a right to lead long term. But in our post-Enlightenment world, We the People think government is all about us. For would-be monarchs who feel the pressure of the Enlightenment’s call to ground government in popular support, there is shadow law and constitutional capture. Moving real government into the shadows lets tyrants make government all about them, while pretending it is all about us.

The public relations of shadow systems call for maintaining the forms of their captured constitutions. Those forms may be shams^[135] in two formally distinguishable but substantively indistinguishable ways. First, the formal text may fail to signal accurately who leads and how. In that case, the officeholders it describes are to varying degrees puppets, whose actions, if any, may be micromanaged from the shadows. Stalin’s 1936 Constitution of the Soviet Union exemplifies this phenomenon. Second, and more commonly, the formal text may accurately describe who leads but not how and why. The official president will lead, but not really in virtue of the processes described in the formal text, because those processes have been captured by the shadow system. The president will be president because he leads the shadow system, and all actors involved in his formal selection know from that system whom they must choose. When a shadow system is in charge, we often see that formal voting for leadership is improbably lopsided in favor of the predetermined leader.^[136]

Formal constitutions that have been contrived or captured by shadow systems may accomplish more than public relations for those systems – a substantial body of scholarship explores further ways in which such documents and the institutions they describe help shadow systems govern.^[137] An uninitiated lay person who came across a book on authoritarian constitutions might expect that the book would be devoted to distinguishing between authoritarian and democratic constitutions by pointing out what authoritarian constitutions distinctively say. But in fact what the reader will find is explanation and argument about why undemocratic systems sometimes say and seem to do democratic things. The scholars who write about constitutions in authoritarian regimes are deriving their reasons for calling those regimes authoritarian mostly not from the distinctive content of those regimes’ purported constitutions, but from other real-world evidence. Their reasons for calling regimes authoritarian come mostly not from the kind of text-bound research that legal scholarship normally involves,^[138] but from what we learn through the methods of investigative journalism, whether by listening to the journalists or by joining them.

We can distinguish unwritten power dynamics that serve to subvert democracy from unwritten power dynamics that serve to strengthen it. The architecture of real constitutional democracy includes constitutional conventions, established practices that didn’t make it into the original drawings but which complement written constitutionalism by helping to make government more of, by, and for the people, not less. Consider, for example, the convention of responsible government, by which the British monarch acts on the advice of ministers chosen for the support they enjoy in an elected Parliament and who resign if they lose that support.^[139] When we notice unwritten power dynamics in any system of government, the crucial question we need to ask is: do these dynamics help keep the promise of democracy? Do they help government serve the common good, the public interest? Or are they complicit in breaking the promise of democracy and in reorienting the focus of government from the public interest to the narrow self-interest and preoccupations of the leadership?^[140]

5.1 Plural Executives

Billionaire businessman Christoph Blocher’s calls for Switzerland to cut back on immigration and keep its distance in international relations helped his Swiss People’s Party win more votes than any other party in Swiss national elections.^[181] That success gave them many seats in the Swiss Federal Assembly and saw him elected by the Federal Assembly to Switzerland’s national executive, the Federal Council.^[182] He and his party gained greater influence over Swiss governing policy. But his appointment gave him no way to stand out and become a focal point for following, no way to start changing the system.

Blocher was only one of seven. The Swiss Constitution requires the seven members of the Federal Council to act collegially in administering the nation together. Though each heads a government department, all major decisions are made collectively by the seven under conditions that keep them individually out of the limelight.^[183]

Why couldn’t Blocher’s party get majority control on the Federal Council, and then bend it to his will? Swiss democracy insists on proportional representation in the first chamber of the Swiss national legislature.^[184] Every party that attracts significant support gets seats, even if they do not have the most support in any particular geographic area. That makes for many parties in the legislature and no party winning a majority of the seats.^[185] As no party can gain control of the legislature, no party can gain control of the executive.^[186] To be elected to the Federal Council, candidates must enjoy multipartisan support in the Federal Assembly. Given the need for consensus candidates, the major parties agreed, many decades ago, on a ‘magic formula’ under which representatives from each major party are elected to the national executive.^[187]

Sharing power in the national executive frees political parties to focus on having good policy ideas for serving the Swiss people. The legislature, not the executive, is the dynamic center of major decision making.^[188] Sharing power in the national executive frees the national legislature from gridlock because the system gives no legislator a reason to want Swiss government to fail; legislators are genuinely focused on getting to good policy outcomes, not on competing to win the executive. James Bryce, who memorably coined the ‘laboratories of experiment’ metaphor for policy initiatives in the American states,^[189] observed that ‘strife for office and the sweets of office felt as always present in the background of debates in the assemblies of England, France, and other parliamentary countries, finds little place in the Swiss legislature.’^[190]

After one awkward term on the Federal Council,^[191] Christoph Blocher was not reelected. He declared himself ‘torn between relief, disappointment and outrage’^[192] by the Federal Assembly’s decision and took his party into ‘opposition.’ But that move proved ineffectual.^[193] The system offered him no realistic route to replacing the government. Eventually his party agreed to nominate two people to fill a later vacancy on the Federal Council. One of them was Christoph Blocher. The Federal Assembly chose the other person. The system had seen him off.^[194]

Switzerland’s plural executive protects Swiss constitutional democracy from capture because Switzerland’s plural executive cannot itself be captured. Almost two centuries of Swiss experience prove its resilience.^[195] Proportional representation in the legislature has proved adequate to prevent any one person pulling the executive’s strings. Were we designing a plural executive from scratch, we could augment this safeguard by conditioning parties’ participation in elections for government on having internal power dynamics that mirror those required of the government. We can ensure the no-one-true-leader principle permeates a system of government if we insist that this principle permeate the parties that compete for election to government.

But what of energy, responsibility, and dispatch?^[196] Christoph Blocher was held responsible for his performance on the Swiss Federal Council when the Federal Assembly decided not to reelect him. Swiss experience shows that members of group executives can be held just as individually responsible for their contributions to decisions as are legislators and judges. And when compared with actual governance in European nations that have one person as chief executive, Swiss governance has not been especially inefficient.^[197] Even faced with violent separatist agitation in one region, the Federal Council proved up to the task of managing the situation through democratic processes.^[198]

Government by committee provides a more reliable level of energy and speed. Singular leadership is far more vulnerable to vagaries of character and health that may compromise an executive’s capacity for timely decision making. Committee members can help hold one another accountable in ways unavailable to presidents’ and prime ministers’ cabinets and staff. Committee procedures can be calibrated to needs for speed, with clear protocols for proceeding when particular members cannot participate.

Committee deliberation may lead to better decisions.^[199] In Thomas Paine’s words, a singular leader ‘will have no person to consult with of a standing equal with himself, and consequently be deprived of the advantages arising from equal discussion. Those whom he admits in consultation will be ministers of his own appointment, who, if they displease by their advice, must expect to be dismissed.’^[200] True group deliberation pulls itself toward the common good, partly because it impels committee members to express themselves in those terms and partly because when we leave behind the ego trip of singular leadership, we find a different kind and caliber of leader.^[201] In Socrates’ words, ‘[t]hat city in which the destined rulers are least eager to rule, will inevitably be governed in the best and least factious manner.’^[202]

5.2 Divided Executives

Plural executives are not the only way to disrupt efforts to capture executive government. An existing literature draws on American state government experience to scrutinize the merits of dividing the executive.^[203] Dividing the executive differs from pluralizing it in that ultimate decision making may still be assigned to individuals rather than committees. For example, within a divided executive, decisions to prosecute may ultimately rest with an attorney general who is truly independent of the chief executive.^[204] The chief executive is not chief for all executive purposes – other individuals are chiefs for particular purposes. In contrast, a singular executive model makes other senior executive decision makers, such as the attorney general, subject to removal and therefore control by the chief executive.^[205] If the chief executive is intent on constitutional capture, having a legal right to control the whole executive fills his arsenal with weapons, such as the power to prosecute, that may be deployed to overcome resistance to capture.

American state government experience shows that divided executives can meet the criteria of energy, responsibility, and dispatch, though like singular executives, they vary in how well they do.^[206] When it comes to preventing capture, however, we have reason to rank divided executives as second-best behind plural executives. Divided executives, like singular executives, set up individuals as focal points for following, with personal apex control over their own hierarchies of officials – effectively a series of power pyramids rather than just one. Divided executives raise individual profiles much as singular executives do, allowing magnetic personalities to grow public popularity that may turn into a mass following through the psychological dynamic that Max Weber called charismatic authority.^[207] Those in command of enforcers, such as police or soldiers, may nurture within their commands a culture of personal loyalty that speeds shadow system construction.^[208] Competing personal fiefdoms may end up in a subversion race, as each recognizes first-mover advantage.^[209] Only plural executives can reliably deny salience to individuals.^[210] Among divided executives is one frequent division that has been notoriously ineffectual at preventing capture – the division between head of state and head of government. All three Axis powers had separate heads of state when their heads of government turned them into tyrannies.