The Rule of Law Without Separation of Powers

Legality in the Classical Tradition

Among theorists of legal liberalism, a common assumption is that the rule of law, rightly understood, entails some version of the separation of powers — especially the separation of adjudication from the making or enforcement of law. The Stanford Encyclopedia entry on the rule of law, written by Jeremy Waldron, says a procedural requirement of the rule of law is that “there must be courts and there must be judges whose independence of the other branches of government is guaranteed. This side of the Rule of Law is connected with the constitutional principle of the separation of powers. That principle is sometimes justified simply on the ground that it is unhealthy for power to be institutionally concentrated in society. But it also has a Rule of Law justification inasmuch as it assigns distinct significance to distinct stages in the making and application of laws.” (Here Waldron draws upon an earlier argument of his own). So too, in a recent paper on the essentials of the rule of law, Cass Sunstein says that “[m]any institutional designs can be compatible with the rule of law, but it is important to ensure that the people who make the law, and who enforce the law, are not the same as the people who interpret the law. Legislators ought not to be allowed to decide what their laws mean. Law enforcement officers are not supposed to decide what the law is…. We might have long debates over what an ‘independent judiciary’ is, exactly, but the basic point should be plain: The adjudicative task must be separate from the task of making or enforcing the law.” Where no such separation of powers obtains, especially the separation of adjudication from lawmaking and law-enforcement, the liberal legal theorist may or may not grudgingly concede that there is thin “rule by law,” but denies that the thick “rule of law” obtains in a substantial sense.

This view has strange consequences, however. It implies first that throughout millennia of the western legal tradition, there was no genuine rule of law. (The Stanford Encyclopedia entry says, imperturbably, that “the heritage of argument about the Rule of Law begins with Aristotle (c. 350 BC); it proceeds with medieval theorists like Sir John Fortescue (1471).” The poor student is left to wonder why no one discussed the subject during the eighteen centuries in between). The classical tradition of the ius commune, literally the “law common,” was shared and developed across polities of all shapes and sizes, from republics, to monarchies, to elected principates or empires, to the elected papacy, featuring a bewildering variety of institutional designs and arrangements. Many of these polities did not feature anything like the modern separation of powers, especially the separation of adjudicative from lawmaking powers, at least at the pinnacle of the legal order. But all of these polities took themselves to be firmly committed to law; in all of them, we observe thick cultures of legality, in which leading jurists of civil or canon law were honored by state, church and society, in which states competed to have the most prestigious law faculties, and indeed in which the cult of justitia, legal justice, was if anything even more pronounced than it is today.

In the many jurisdiction-states of the classical, medieval and early modern eras, the impartial decision of cases on petition or appeal, involving the application or at most marginal clarification of extant law, was a central function of government, specifically of imperial or royal or papal authority, supported by an array of jurists and offices charged with the writing of responses and rescripts. Cassius Dio recounts that when the Emperor Hadrian refused to halt a journey to hear the petition of an ordinary woman, the woman replied “Stop being Emperor then!” Hadrian turned around to hear the case — the dual point of the story being that the imperial administration of justice was seen as a defining feature of the office and as properly available even to the lowly. The exercise of this core function of adjudication was surrounded by countless legal, political and theological norms and practices emphasizing the duty of rulers to administer justice impartially, as in the famous promise of the coronation oath of the English monarchs.

Was all this just thin rule by law, or perhaps merely a sham, a flimsy cover for arbitrary rule? Or should we attend rather to historians and legal historians like A.J.B. Sirks, who observed that “[a]lthough the emperor had since the beginning of the second century been de facto the sole lawgiver, supreme judge, and administrator of the empire, that certainly did not imply that he had absolute power. It is repeatedly stated that a good emperor subordinates himself to the rule of law, and the emperor himself said so: ‘our authority depends on the authority of the law’ (de auctoritate iuris nostra pendet auctoritas). This was not mere propaganda and ideology. Practice and legislation were a development of the existing structures rather than brusque innovation, and mostly a response to litigation or problems which had arisen at the lower levels of government.” Making the same point in different terms, Ronald Syme observed that “[t]he Principate, though absolute, was not arbitrary. It derived from consent and delegation; it was founded upon the laws.” The liberal theorist takes it on constitutional faith that the fusion of powers equates to a-legal arbitrariness, but the first does not conceptually entail the second, nor does history allow the assumption that the second flows predictably from the first.

The liberal view further implies that, even today, the rule-of-law credentials of major legal orders are suspect. In strict constitutional orthodoxy, the constitutional order of the United Kingdom centers on a fusion of powers, rather than separation of powers, of the Crown-in-Parliament. So too, the United States federal Constitution contains no freestanding separation of powers principle, and instead enacts a complicated “bundle of compromises” that subjects the judiciary to a range of legislative and political checks. Even more importantly, the constitutional order of the United States has developed into an administrative state that features an array of sweeping delegations to the President and executive bodies or agencies; many of those agencies combine rulemaking, enforcement, and adjudicative powers, subject to (only) deferential judicial review. The structure of adjudication in such agencies is that enforcement and adjudication are separated at the level of initial adjudication, but combined at the top level of the agency itself, a structure resembling that of many classical adjudication-states, in which powers were fused at the apex of the structure but exercised by different officials at lower levels. And beyond the Anglosphere and Western Europe, we see legal systems drawn in part from European legal codes, in part from ancient domestic traditions, many of which do not feature anything like the modern liberal version of separation of powers. Are all these lacking in the rule of law?

The liberal view is thus parochial in the extreme — both intertemporally parochial and geographically or comparatively parochial as of the present. It entails that, at least before Montesquieu or perhaps Fortescue at the level of theory, and really before the 18th or 19th century at the level of (European) institutions, there was no genuine rule of law to be found anywhere in the world, and that even today, there is no genuine rule of law to be found across large stretches of the globe. At a minimum, such implications should encourage us to entertain the heretical thought that, perhaps, building the separation of powers into the conceptual essence of the rule of law is too demanding. As the Oxford philosopher John Tasioulas observed, “there is no a priori reason to suppose that there is a single best institutional arrangement for securing the Rule of Law. Failure to register these points can lead to the Rule of Law ideal being parochially equated with the detailed arrangements in place in a specific jurisdiction … which are then to be simply ‘exported’ to jurisdictions that currently lack those precise arrangements.” Tasioulas’ institutional agnosticism is echoed in judgments of the U.S. Supreme Court, which has repeatedly upheld combined-powers arrangements in agencies against challenges based on the constitutional principle that adjudication must be impartial. The Court has reasoned, in part, that “[n]o single answer has been reached” to the question of how powers should be distributed across offices or officers, and that “the growth, variety, and complexity of the administrative processes have made any one solution highly unlikely.”

Why would it be, how could it be, that the rule of law does not require the separation of powers? The classical legal tradition contains several answers, which combine to give us what current theory almost entirely lacks — a normatively appealing account of the public-regarding ends, the affirmative public benefits, that underlie the combination of powers. One is that the rule of law has to attend to several desiderata simultaneously, and perhaps even trade them off against one another to some degree. Separating lawmaker and law-enforcer from adjudicator promises impartiality, but impartiality is not all there is to the rule of law as an ideal. It is also desirable, for example, that the identification, declaration and implementation of law be both well-informed by accumulated experience and pragmatically effective. It is desirable that the law be not only announced in the abstract, but vigorously enforced as law, and that its application in particular cases be informed by experience and prudence. An arrangement that combines powers in one set of hands, thereby putting the interpretation and application of laws in the hands of the same official(s) who formulated the law in the first place, may reasonably be thought to offer more prudent and effective implementation of the law in these senses. As the Liber Augustalis, the great law code promulgated by Frederick II Hohenstaufen in 1231, put it, “when the sources of justice and of protection are both united in the same person, enforcement will not exist apart from justice, and justice will not exist without enforcement.” Here too, this venerable reasoning finds an echo in modern judgments of the United States Supreme Court, which has said that the combination of investigative, enforcement, and adjudicative powers in the same set of hands helps to ensure the vigorous and well-informed execution of the laws, and furthermore that allowing the holder of combined powers flexibility to choose between more or less general forms of lawmaking — between rulemaking and case-specific adjudication — conduces to more effective and coherent implementation of legal policy.

Put otherwise, when Federalist 78 explains that the federal judiciary “may truly be said to have neither FORCE nor WILL, but merely judgment; and must ultimately depend upon the aid of the executive arm even for the efficacy of its judgments,” it articulates a very specifically liberal conception that has both costs and benefits even from the standpoint of the rule of law itself. We may justly wonder whether the inability of the law-applier to guarantee the efficacy of its own judgments is, on net, a good or bad feature of a law-governed constitutional order — good or bad from the standpoint of the rule of law itself. In contrast to Federalist 78, Pascal in effect argued for a combination of powers along the lines of the Liber Augustalis, observing that “la justice sans la force est impuissante; la force sans la justice est tyrannique” (justice without force is powerless; force without justice is tyrannical). Whether or not Pascal’s view is correct, the question he poses is internal to the ideal of the rule of law.

Another point is that the liberal legal theorist assumes that only a particular institutional technology — focused on institutional separation — can guarantee or even promote impartiality in adjudication. This obsession with institutional technology is indeed highly characteristic of liberal constitutional theory generally, but it overlooks the role of internalized norms within a system of combined powers. Those norms are themselves norms of role, requiring a switch of roles and of attitudes when the holder of combined powers exercises different functions. When the lawmaker switches hats, as it were, and becomes the law-applier or adjudicator, it is an open question, ultimately an empirical question, whether the arrangement produces an intolerable risk of bias. The Court, for whatever it may be worth, holds that a “realistic appraisal of psychological tendencies and human weakness” does not suffice to show that the risk of bias arising from combined functions is generally intolerable, or that the risk suffices to overcome the law’s general “presumption of honesty and integrity in those serving as adjudicators.” Indeed, on one standard view of the relationship between formal legal institutions and norms, formal institutional separation may send precisely the wrong message, “crowding out” intrinsic public-regarding motivations and norms of role-switching that would otherwise obtain. That view may or may not hold in particular conditions or under particular legal arrangements, but the question is ultimately contingent and empirical, and then it is not true that the separation of powers, especially the separation of adjudicative from other powers, is a conceptual essential of the rule of law; it does not straightforwardly fall out of the very idea of either rule by law or rule of law. The classical law indeed embodied a norm of role-switching in the Code and thus in the law itself, through a famous public statement of imperial self-restraint, the Digna Vox, that subjected the imperial authority itself to the extant law, and which thus hovered somewhere between a self-binding instrument and a nonbinding statement of ideal principle: “It is a statement worthy of the majesty of a ruler to proclaim himself bound by the laws; so much does Our authority depend upon the authority of the law. And in truth, to subject the sovereign power to the laws is something greater than imperial rule. And by the oracular pronouncement of the present edict, We show what We do not allow Ourselves.” (Cod. 1.14.4; Bruce W. Frier trans).

Finally, the liberal theorist — as in both the papers I mentioned at the outset — tries to fold the separation of powers into other components of the rule of law, both formal components such as the prospectivity, intelligibility and clarity of laws, and procedural components such as the right to a hearing before the infliction of deprivations to liberty or property. In fact, however, the separation of powers is detachable from those other components. Modern legal theorists usually and quite rightly associate the formal and procedural components with a famous treatment by Lon Fuller, but the classical law itself emphatically embraced and implemented these components millennia beforehand, even in legal orders not featuring any apex separation of powers. (Fuller’s major work, incidentally, says very little about the separation of powers generally or the separation of adjudication from other powers in particular, and does not seem to regard these as essential to legality). The Code of Justinian, in its first book, contained a set of meta-rules for the interpretation of laws that expressly adopts the major formal criteria for the thick rule of law. As to the retroactivity of laws, for example — a central concern of Fuller’s — the Code establishes a strong presumption against retroactive application, saying that “laws and constitutions create a rule for future cases and cannot be applied retroactively to past acts, unless provision has expressly been made concerning the past for cases that are still pending” (Cod. I.14. 7; Bruce W. Frier trans.). This is an excellent summary of current law in the United States. So too, the Code mandates the clarity and intelligibility of laws (I.14.9) and provides that the law binds even imperial officials (I.14.10). And the duty to afford a hearing before a legal imposition could be inflicted is a central idea of the classical legal tradition; jurists sometimes traced it to the hearing that even an omniscient God willingly afforded to Adam and Eve in the Book of Genesis.

It is even plausible that, under certain conditions, the combination of powers might better promote the formal and procedural rule of law than does the separation of powers. Fuller, for example, was rightly concerned about lack of congruence between the law as declared or enacted and the law as applied. The combination of lawmaking, law-enforcement and law-applying power in one office plausibly helps to ensure such congruence, whereas separating these functions across different institutions with divergent powers, capacities, information, methods and agendas risks distortion of the lawmaker’s aims in application to particular cases. For this very reason, the classical law emphasized that interpretation of the laws was ultimately reserved to the apex authority (Cod. I.14.11), although initial application could be and often was delegated to subordinate tribunals. The point is not that the combination of powers always and invariably does better with respect to the formal and procedural aspects of legality; it is merely that there is no reason to assume, as a matter of constitutional faith, that the separation of powers is invariably necessary to promote effective legality.

What unites these points about the public-regarding ends underlying the combination of powers is that classical legal theory is premised on an account of the ends of constitutionalism generally that differs from the liberal account. Whereas liberal constitutional theory focuses, to the point of myopia, on negative freedom and the dangers to individual freedom from the abuse of public power — the “liberalism of fear” — the classical view holds that the master aim or end of constitutional government is to promote the general welfare or common good. On the latter view, although abuses of public power are an undoubted evil, so too are private abuses of legal entitlements, which themselves threaten to undermine the rule of law. The aim of classical constitutionalism is thus not to maximize constraints on state power or to strictly minimize the risk of arbitrary public action, but rather to secure overall peace, order, and good government, which requires ample scope for vigorous public action through law, in order to secure the conditions of political flourishing. As Michael Foran emphasizes, the collaboration or combination of powers can in principle contribute to this account of effective legality and the ends of constitutionalism; hence the classical account does not rule it out of legality by conceptual fiat, but is open to the thought that the combination of powers might itself help to promote the rule of law.

Overall, liberal legal theory defines itself, implicitly but unmistakably, as the negative or antonym of classical legal theory, which remains generally agnostic about the separation of powers, but which holds that a combination of powers is equally consistent in principle with a profound commitment to the rule of law and legal justice. Ironically, then, there is a kind of imperialism — conceptual rather than constitutional imperialism — at the heart of liberal legal theory itself; it stamps as uncivilized and beyond the pale of legality all constitutional arrangements that are not created exactly in its own image, down to the details of the distribution of powers across offices. The classical law is more tolerant, institutionally speaking, believing that no particular institutional technology, including the separation of powers, is defined into the rule of law, so long as the constitutional order as a whole is rationally and adequately ordered to the proper end of law, the general welfare or common good.