Constitutions Un-entrenched: Toward an Alternative Theory of Constitutional Design

Rosalind Dixon

In their recent paper in the American Political Science Review, Versteeg and Zackin offer an important contribution to evolving debates on constitutional design, convergence and diffusion. They suggest that, far from being the only model in circulation in global constitutional thinking, the US constitutional model of highly abstract and entrenched constitutionalism is in fact no longer even the dominant model: at a US state level, and globally, a quite different model of very specific and flexible constitutionalism is in the ascendancy. This model blurs the line between constitutions and ordinary legislation. It also reflects a quite different kind of thinking about the relationship between constitutions, democracy, and the people: rather than empowering courts to interpret vague or abstract constitutional guarantees, and entrenching those decisions against repeal by ordinary democratic majorities, Versteeg and Zackin suggest that this model seeks to constrain courts, legislators and executive actors to act in line with the preferences of a majority of citizens.

In this sense, it represents a quite different take on traditional understandings of democracy and distrust: it is the expression of a form of popular distrust of elite institutions generally, rather than more particularized distrust of legislators of the kinds such as John Hart Ely envisaged.¹ Versteeg and Zackin further argue that there is a close logical relationship in this context between a preference for constitutional specificity and flexibility: specific constitutions may help popular majorities control elite actors, but they are also more likely to require active updating by citizens themselves, rather than elite actors. As I have also suggested in prior work, whatever the scope for courts and legislators to update of a constitutional standard by way of ‘common law interpretation’,² or polycentric forms of interpretation,³ there is far less scope to apply such approaches to more specific rule-like constitutional provisions.⁴

In addition to identifying this new model of flexi-specific democratic constitutionalism, and its logical inter-dependence, Versteeg and Zackin offer compelling evidence of its increasing diffusion or predominance both at a US state constitutional level and globally. Indeed, this empirical dimension to their project is particularly wide impressive, and no doubt one of the reasons the piece found a home in the APSR. One of the things I also particularly liked about the piece in this context is that it combines both state-of-the-art large and empirical approaches with a more historical, qualitative analysis of the historical origins of various flexi-specific constitutional provisions: drawing on Zackin’s earlier archival work for her excellent book, Looking for Rights in All the Wrong Places, the piece quotes from a range of legislative, newspaper and scholarly sources from the late 18^th and early 19^th century as to the origins of many specific constitutional provisions.

While I am very sympathetic to their project, their work did, however, raise two broad questions for me, in part connected with my own prior work on constitutional specificity and amendment. First, would it be useful to further disaggregate the model of flexi-specific constitutionalism into two broad categories: one aimed at the control of legislative and executive decision-making, and another at the control of courts? The two models potentially have somewhat different origins, or logics: in most cases, it seems likely that attempts to control legislatures or executive officials will arise out of the kind of principal-agent situation Versteeg and Zackin identify. I was reminded of this only last week visiting Chicago, when I came across a letter to Illinois residents outlining the scope of the proposed amendment to section 11 Art IX of the Illinois State Constitution: this amendment is very long, detailed, and if successful, could still be relatively easily repealed – and is all about the relationship between voters and their elected representatives. It attempts to limit the ability of certain local units to use funds appropriated for transportation for other purposes.

But for provisions speaking more directly to courts, in some cases at least, the origins of such provisions may be somewhat different: they might originate in a belief that it should be legislatures rather than courts that decide certain questions, or in a concern to provide ‘insurance’ for certain legislative policies against the risk of judicial invalidation.⁵ In some cases, provisions that purport to authorize certain forms of legislative action, and thereby prevent courts from invalidating such action, might thus have quite different origins or underpinnings to provisions that require forms of legislative action.⁶ A good example concerns the kinds of labour rights provisions found in various state constitutions in the US and discussed by Versteeg and Zackin at some length: some of these provisions were clearly directed at requiring state legislatures to limit working hours (mandatory by-law clauses), but others to insulating legislative action from judicial invalidation (permissive by-law clauses).

Second, I wonder whether flexi-specific constitutions are in fact a true alternative model of constitutional design, or something closer to an important sub-type or sub-form of constitutional design. As an empirical matter, it is not clear that the increasing specificity—and amendment—of many US state and global constitutions is in fact distributed evenly across all areas of constitutional law. Rather, it may well be that flexi-specificity is concentrated in certain areas – dealing with federalism, various legislative and executive policies, and not the fundamental democratic, separation of powers rule of law, or rights-based provisions of a constitutions. In Brazil, for example, as an important example of the trend Versteeg and Zackin identify, there have been large numbers of constitutional amendment since the adoption of the 1988 Constitution, but almost none of these amendments have touched what one might call ‘the basic structure’ of the Brazilian Constitution. In the US, state constitutions have also become increasingly specific against the backdrop of a quite abstract, and entrenched, set of federal constitutional guarantees that clearly modify or set limits on the scope of the flexi-specific state constitution. Think of the recent decision of the US Supreme Court in Obergefell: the Court held that the attempt to create increasingly specific prohibitions on same-sex marriage under state constitutions was inconsistent with the more abstract and entrenched guarantee of equal protection under the Fourteenth Amendment, and thus invalid and ineffective as a form of constitutional change.

Theoretically, we might have two broad reasons to doubt whether flexi-specific constitutions can succeed as a standalone constitutional model. First, if constitutions are extremely long and detailed, citizens may have difficulty understanding them, and thus be relatively unlikely to identify and defend the constitution. (Indeed in other important work, Versteeg has shown with Stephanopoulos that the ability to understand a constitution is a key determinant of constitutional identification.) This may itself mean that flexi-specific constitutions are far more likely to be replaced than more general, abstract ones, so that there is a kind of Darwinian pressure toward more abstract rather than specific constitutions.

Second, if constitutions are too specific, they may ultimately end up creating forms of inconsistency or overlap that necessarily require courts themselves to resolve a range of interpretive ambiguities. Tom Ginsburg and I make this point in our work on insurance swaps: we suggest that if drafters adopt highly specific but conflicting provisions, this effectively delegates to courts the task of resolving relevant conflicts. I have also made a similar argument about attempts to codify certain constitutional exceptions or carve-outs, and both the ambiguities this creates, and the degree to which ambiguity can be empowering rather than constraining of constitutional courts.

Ultimately, this may mean that for flexi-specific constitutions to endure, and/or achieve their aims in terms of judicial constraint, they cannot be entirely standalone models of constitutional design. Instead, they must be nested within a broader, more abstract and entrenched constitutional framework.

Constitutional ‘tiering’ is already a relatively familiar idea in the context of the design of constitutional amendment rules: as Richard Albert has pointed out, there are a range of constitutions worldwide that adopt different levels of entrenchment or flexibility, based on the nature of the question at stake. David Landau and I have also recently suggested that constitutions often adopt a tiered approach to constitutional language, as well as entrenchment: they adopt quite specific provisions for a range of routine or ordinary constitutional matters, but more abstract, parsimonious language for core constitutional values designed to bind a community together, or protect the basic constitutional or democratic order.⁷

Tiering of this kind may be done better or worse in different constitutions, but it seems a relatively common feature of democratic constitutional design: just think of the role of preambles or constitutional value statements. As Landau and I have pointed out in the context of constitutional entrenchment debates, tiering of this kind can also often occur ex post – by judges selecting certain provisions for special reinforcement and validation – as well as ex ante.

In sum, I am highly persuaded by both the empirical evidence Versteeg and Zackin present, and by their claim that the field of comparative constitutional studies needs to do more to recognise and grapple with the increasing centrality of flexi-specific forms of constitutionalism. But I also wonder if we shouldn’t divide this trend into two types (legislative and judicial-facing), and do more to recognise it as a form of nested democratic constitutional development, which is best understood as part of a tiered approach to both constitutional amendment and language.

1. Compare John Hart Ely, Democracy and Distrust: A Theory of Judicial Review (1980); Rosalind Dixon, Constitutional Drafting and Distrust, 13 Int’l J. Const. L. 819 (2015).
2. David Strauss, Common Law Constitutional Interpretation, 63 U. Chi. L. Rev. 877 (1996).
3. Robert C. Post & Reva B. Siegel, Legislative Constitutionalism and Section Five Power: Policentric Interpretation of the Family and Medical Leave Act, 112 Yale L.J. 1943 (2003).
4. Rosalind Dixon, Updating Constitutional Rules, Sup. Ct. Rev. 319 (2009); Rosalind Dixon, Amending Constitutional Identity, 33 Cardozo L. Rev. 1847 (2012).
5. Tom Ginsburg, Judicial Review in New Democracies: Constitutional Courts in Asian Cases (2003); Tom Ginsburg & Rosalind Dixon, Constitutions as Political Insurance (Unpublished manuscript, 2016).
6. Compare Rosalind Dixon & Tom Ginsburg, Deciding Not to Decide: Deferral in Constitutional Design, 9 Int’l J. Const. L. 636 (2011) (identifying mandatory versus permissive by-law clauses).
7. Rosalind Dixon & David Landau, Exporting Art V? Tiering Constitutional Amendment (unpublished manuscript, 2016).