Peter Cane: A General Theory of the Executive? (Book Symposium: Margit Cohn’s A Theory of the Executive Branch – Tension and Legality (OUP, 2021))

Margit Cohn’s A Theory of the Executive Branch: Tension & Legality is a rich, complex and stimulating book. In this blog I will briefly address three questions: the first concerns the ideas of ‘complementarity’ and ‘tension’ that are central to Margit’s theoretical approach; the second is whether Margit’s theory is a theory of the executive? and the third concerns Margit’s goal to produce a general theory.

A. Margit’s Theoretical Approach

1. Value-Pluralism

Margit aims to offer a ‘general theory of the executive branch’ (9). ‘I submit’, she says ‘that the executive is best viewed as straddling the line between subjection to law and dominance beyond law’ (42-3). By ‘the executive’, Margit apparently refers primarily to the chief executive. She presents the theory as ‘explanatory’ or ‘descriptive’ (57), based on data from the US and the UK systems. In both systems, she observes, the chief executive is neither fully subject to, nor entirely dominant over, the legislature. Rather, fluctuating tension between subjection and dominance makes the relationship between the executive and the legislature fluid and variable. Margit distinguishes her theoretical approach from those she characterises as ‘hierarchical and dichotomous’: executive subjection to and dominance over the legislature are not dichotomous alternatives but ‘complementary and simultaneous’ (24) (though ‘self-contradicting’ (25) and ‘irresolvable’ (43)) tensions. Personally, I find such an approach attractive. In analysing and understanding human institutions (such as the executive) it is typically helpful, I believe, to model their activities in terms of the promotion of various incommensurable values that are held in creative tension with one another in response to particular policy objectives and factual circumstances: not either/or but both/and. Value-pluralism characterises the human condition.

2. The Concept of Fuzzy Law

In Margit’s account, the tension between executive subjection to law and dominance over law finds its practical expression in the concept of ‘fuzzy law’. Fuzzy law ‘presents a façade of legality that derives from its binding formal status, while its content or application offer broad options for action (and possible abuse) which conceals a reality of a-legality’ (59). In other words, the concept of fuzzy law rests on a distinction between action within the law and action outside the law. From the claim that fuzzy law authorises a-legal action by virtue of its formal pedigree as law, it would seem to follow that a-legal action is not, as such, illegal. At the same time, Margit distinguishes between acceptable and excessive fuzziness (270, 272). Excessive fuzziness results when law fails adequately to cabin executive action because it does not convey ‘a general understanding of the purposes of the grant of power and discretion’ (270). Excessively fuzzy laws confer ‘virtually unlimited, or highly unrestricted, choice’ (270). Excessive fuzziness ‘renders an arrangement unintelligible [and] offers a mere façade of legality’ (72). According to Margit, excessive fuzziness is unacceptable for two reasons. First, it threatens the rule of law, cashed out in terms of Lon Fuller’s inner morality of law (260); and secondly, it undermines principles of participatory and deliberative democracy (262-5; see also 287). The threats to the rule of law and participatory democracy presented by excessive fuzziness ground Margit’s normative case, made in Part IV of the book, for giving the judiciary a central role in regulating the relationship between the executive and the legislature.

In my mind, all this raises a question of whether the distinction between acceptable and excessive fuzziness may not actually mark a transition from (legally authorised) a-legality to (legally unauthorised) illegality? If so, does the distinction reintroduce into Margit’s theory, in the form of the distinction between the legal and the illegal, the very dichotomizing that she seeks to replace with a concept of complementarity? Margit might say that the location of the line between acceptable and excessive fuzziness will depend on context and circumstances; but by itself, this would not erase the dichotomy which, in Margit’s normative theory, it is the role of the courts to implement. It is perhaps this issue that Margit attempts to address in her charming but opaque idea that ‘fuzziness is itself fuzzy’. The point may be approached differently by asking whether the theoretical approach that, to my mind, Margit adopts successfully in the descriptive and analytical part of the book can be as successfully applied to the normative aspect of the project. Does the role Margit assigns to courts perhaps presuppose a dichotomous distinction between the acceptable and the unacceptable?

B. Is Margit’s Theory a Theory of the Executive?

1. Fuzzy Law and the Bureaucracy

As already noted, Margit’s account of the executive focuses on the chief executive. She briefly brackets the issue of the relationship between the chief executive and the administrative bureaucracy by including the administration within ‘the executive branch in its broad sense’ (31-39) on the ground of ‘functional similarity between the core executive and administrative agencies’. This is surprising given that massive growth of the non-political, non-elected bureaucracy is the most tangible manifestation of the modern state. Margit focuses on the relationship between the executive and the legislature. However, much debate about the nature and control of ‘executive power’, especially in the US system, is as much concerned with the relationships between the executive and the bureaucracy on the one hand and the legislature and the bureaucracy on the other, as with that between the (chief) executive and the legislature.

Another puzzling feature of Margit’s account is her claim that fuzzy law may be ‘executive-generated’ as well as ‘constitution-generated’ and ‘legislature-generated’ (see esp pp 70-71). By ‘executive-generated’ fuzziness, she does not seem to refer to law made by the executive under delegated law-making powers but rather to exploitation by the executive of fuzzy statutory (and, maybe, constitutional) law. Nevertheless, mass transfer of law-making power from the legislature to the executive and the bureaucracy is characteristic of the modern state. In her theory, Margit seems to model the executive as an applier of law made by other institutions and to underplay the importance of administrative rulemaking as a distinctive feature of constitutional arrangements and, itself, a generator of fuzzy legality.

2. Fuzzy Law and The Legislature

Whereas Margit models the executive (predominantly) as an exploiter of fuzzy legality, she models the legislature as a (prime) generator of fuzzy law. However, at least in a system (such as the US) with a canonical documentary constitution, the legislature might equally be modelled as an exploiter of (constitution-generated) fuzzy law.

3. Fuzzy Law and The Judiciary

By contrast to the legislature, in Margit’s account the judiciary ‘rarely develops novel forms of fuzziness’ (290). This view underpins her normative claim that the courts are well placed for ‘setting constraints on fuzzy executive power’ (290). It seems to rest on an image of courts as law-appliers and not also as lawmakers. Two issues arise in my mind. First, superior courts (in the UK system, at least) can and do make, as well as apply, public law; and the common-law process by which they make law might arguably be seen as itself a distinctive source of fuzzy law. Secondly, modelling courts as external constraints on power may be problematic. On the one hand, the construction of the Supreme Court, under the US Constitution, as an integral component of a system of checks and balances, might suggest that the US judiciary should itself figure as a player in a theory of executive power rather than as an external regulator of such power. On the other hand, modelling the courts as law-appliers suggests that they, too, might be understood as exploiters of fuzzy (constitutional and statutory) law. In short, the judiciary, like the legislature and the executive may, under appropriate circumstances, be modelled as both generators and exploiters of fuzzy legality.

C. Is Margit’s Theory General?

As already noted, Margit bases her explanatory theory mainly on data from the US and UK systems. However, her aspiration is to produce a ‘general’ theory. By ‘general’ Margit means something like: yielding ‘some insights that are valid beyond specific domestic borders’ (40).

Applying George Tsebelis’s veto-points concept (28), Margit argues that despite the differences between presidential and parliamentary regimes, the US and UK systems have enough in common to provide a firm foundation for a general theory. The easy response would be to say that Margit’s dataset is far too small to accommodate her aspiration to generality. However, my query focuses on her more specific claim that there is sufficient similarity between the US and UK systems to generate a general theory. I find this unconvincing. For one thing, the relationship between the executive and the bureaucracy is very different in the US system than in the UK system. For another, I would suggest that the US concept of (federal) judicial power is quite different from the UK concept.

However, here I want to focus on what seems to me to be a much more fundamental difference between the US and UK systems relevant to Margit’s project: that between ‘written’ and ‘unwritten’ constitutions – the US constitution being an exemplar of the former and the UK constitution of the latter. Like many commentators, Margit treats unwritten constitutionalism as an archaic aberration. However, for her purposes, she seems happy to treat both as equally capable of generating fuzziness in the forms of fuzzy law and fuzzy legality. It is this approach that I want to destabilize.

Typically, deployment of the written/unwritten distinction is qualified by the concession that the UK constitution has numerous written elements. Less commonly, until recently at least, it is (conversely) qualified by a concession that the US constitution has significant unwritten, as well as written, elements. Sometimes, such qualification is followed by a nod towards James Bryce’s alternative distinction between rigid and flexible constitutions. However, this move is of little help because ‘flexibility’ and ‘rigidity’ are both relative concepts and both are important constitutional values.

I want to argue that the UK and the US constitutions are fundamentally different, not in either of these respects but rather in terms of the relationship between law and the constitution. The US Constitution is understood to be a legal document which, in theory, gives effect to ‘the sovereignty of the people’. Under the Constitution, Congress, the President and the Supreme Court are all understood equally to exercise not ‘original’ power but rather power delegated to them by the people. The Constitution is supreme law, and the whole governmental system is understood to be directly or indirectly derived from and subject to that supreme law.

By contrast, although law plays (and for many centuries has played) an important part in the UK constitution, it has never enjoyed supreme status: unlike the US constitution, the UK constitution is not a creature of law. For one thing, the basic organs of government – Parliament, the Executive and the central courts – are all understood to have ‘original’ or ‘inherent’ power as opposed to power delegated by ‘the people’ or by some other organ of government. (Of course, many (perhaps all) other governmental organs possess only delegated power). One way of putting this is to say that the source of the power of these organs is not legal but ‘political’ – it is, if you like, ‘organic’ rather than ‘synthetic’ (to adopt Hayekian imagery). And just as the US Constitution is recognised as both a source of power and a source of limitations on power (most conspicuously in the Bill of Rights), so we may say that in the UK constitution, law and politics are sources of both power and limitations on power. Put differently again, we may say that in the UK constitution, both law and politics can create constitutional norms conferring and limiting power. AV Dicey famously gave the name ‘constitutional conventions’ to political norms. In his view, the force of conventions was ultimately dependent on law. I disagree. In my way of thinking, the UK constitution consists of political norms made and enforced by political institutions, and legal norms made and enforced by legal institutions. The UK constitution is both legal and political. In Margit’s terms: in the UK constitution, law and politics are truly complementary sources of power and its limitation.

But how does any of this relate to Margit’s theory? I would say, first, that whatever the correct analysis for the US system, in its UK counterpart the ‘tension’ to which Margit refers is between the legal and the political, not between the strictly legal and the ‘a-legal’ (or ‘illegal’). In the English way of thinking, politics precedes law. One implication of this is that courts might play a less central role in controlling the exercise of public power than in a system founded on a canonical, legal document, and less ‘central’ (289) than they play in Margit’s theory. Secondly, I would argue that the difference, in this regard, between the two systems is important for understanding the different constitutional roles and powers of their respective legislatures. More particularly, I would argue that the principle of Parliamentary sovereignty is essentially political, not legal. At the apex of the UK system we find not a legal document but a political institution with inherent power – the Queen in Parliament. Thirdly, I would speculate that the difference between the two systems is also likely to have significant implications, for theorisation of the role of the executive, that cast doubt on the feasibility or value of an entirely general theory. One may reply, of course, that the UK system is an anachronistic aberration that will (and, for some, should) sooner or later fall into line with the US model. To the contrary, I would argue that the two models embody quite different ways of imagining and constructing constitutionalism that both demand and deserve theoretical attention because each has distinctive practical strengths and weaknesses.

Peter Cane is Senior Research Fellow at Christ’s College, Cambridge; Emeritus Distinguished Professor of the Australian National University; and a Corresponding Fellow of the British Academy.