What Is a Classical Liberal Constitution?

• Author: Dove, John A.

Classical liberalism is a belief system that links prosperity, both material and immaterial, to respect for the individual. Such respect is embodied in rights to one's own self and property, safeguarded under the rule of law. Classical liberals favor an institutional environment of economic freedom, "[t]he cornerstones of [which] are personal choice, voluntary exchange, open markets, and clearly defined and enforced property rights" (Gwartney et al. 2019, 1). There is considerable evidence that links this sort of institutional environment to prosperity (Hall and Lawson 2014).

Further, a society's institutional environment is a set of political outcomes. Those outcomes take shape within a higher-order framework of constraints and prescriptions---call it the metainstitutional environment, or a society's constitution. All else equal, different constitutions will yield different outcomes at the level of ordinary politics: some will tend to yield economic freedom, and it stands to reason that classical liberals would favor these constitutions. However, what does a classical liberal constitution look like?

Richard Epstein notes that foundational to classical liberalism are the "twin pillars of private property and limited government" (2014, ix). The latter must have sufficient capacity to define and enforce the former but also restraint in not employing it toward predation that erodes those rights (e.g., Buchanan 1975; Weingast 1993, 1995; Acemoglu and Robinson 2019). What sort of framework of prescriptions and constraints will strike this balance, empowering a state's protective Jekyll while shackling its predatory Hyde? This question is of paramount importance for classical liberals. However, although scholars of that bent have devoted great effort to exploring the relationships between ordinary-level institutions and prosperity--citations to many of them are, again, to be found in Hall and Lawson 2014--they have produced next to nothing regarding the links between constitutional-level institutions and prosperity.

In this paper, we aim to highlight the scholarly neglect of the constitutional level by classical liberals. This may strike many readers as misguided, especially if they have a preconceived notion of what a classical liberal constitution looks like. (Perhaps the U.S. Constitution comes to mind?) However, there is actually very little evidence--either compelling or merely suggestive--about which aspects of constitutional design are associated with economic freedom and limited government. This is troubling. As Aziz Huq puts it, "Structural aspects ... are not valued for their own sake.... Second-order constitutional design instead succeeds only if it creates desirable first-order goods" (2014, 1006). Secure property rights and rule of law--these are relevant first-order goods. Yet there is scant evidence as to the sorts of constitutional design that yield them up. In an empirical illustration, we demonstrate that certain aspects of constitutional design, favored by classical liberals, are not at all clearly linked to greater economic freedom.

Before we proceed further, we should note that we are concerned here with neglect of de jure constitutional design. A distinction can be drawn between de jure (formal, codified) constitutions and those that are purely de facto (informal, based on unwritten conventions and norms). (1) This distinction is worth noting because some scholars are dismissive of what James Madison referred to as mere "parchment barriers." (2) Nikolai Wenzel, for example, is very clear on this: "I emphatically reject the notion that good constitutional parchment is sufficient for successful constitutionalism" (2010, 65). Alexander Salter and Glenn Furton echo and elaborate on the point: "A written constitution can specify which set of rules, from a much larger set of feasible rules, will operate.... But this is very different than the intended, and broadly agreed-upon, utility of de jure constitutions, which is as mechanisms for binding or constraining. De jure constitutions cannot perform this role" (2018, 38). Although these authors do not outright reject the relevance of de jure constitutions, they clearly perceive them to be of second-order importance.

However, the rise of the de jure constitution and its near ubiquity today are undeniable. Since the adoption of the U.S. Constitution in 1789, 95 percent of states adopted a constitution within five years of coming into existence (Elkins, Ginsburg, and Melton 2009, 42). Untold resources have been devoted to the drafting and subsequent amending of these documents. It is also fair to say that most people in almost all countries believe that de jure constitutions are--or at least can be--of great importance. Among scholars, "[e]very serious student of the topic recognizes, indeed insists, that each nation, especially each emerging nation, should adopt a written constitution to shape the organization of its government" (Epstein 2011, 290).

And there are good reasons to believe that de jure constitutions are--or, again, at least can be--of great importance. They can serve as coordination devices (e.g., Hardin 1989; Ordeshook 1992; Weingast 1997; Hadfield and Weingast 2014). By codifying feasible rules, they can provide "a focal solution ... so that citizens gain the ability to act in concert and police their government" (Weingast 2005, 105). Just as importantly, such a focal solution can provide political agents with means by which to police one another (Young 2020). Until those feasible rules are codified, they are merely conventions and norms--shared beliefs regarding permissible behavior. The extent to which those shared beliefs are actually binding is determined only by the extent to which they are shared. Codification, in contrast, provides citizens and political agents with a common reference point. (3)

Furthermore, de jure constitutions can determine which feasible rules are coordinated on. As Salter and Furton note, they can "specify which set of rules, from a much larger set of feasible rules, will operate" (2015, 38). (4) But that seems to be critically important because only the (smaller) set of feasible rules will actually operate--that is, help to determine the practices and patterns of governance. Constitutional designers, then, play a role in determining whether the operative set is a "good" one or a "bad" one.

So, again, what does a classical liberal constitution look like? In using this terminology--including in the paper's title, no less--we are making an obvious reference to Richard Epstein's masterful book The Classical Liberal Constitution (2014). However, Epstein's book is concerned largely with the U.S. Constitution in relation to the Founding Fathers' ideological backdrop of classical liberalism. Epstein is not specifically concerned with the structural elements of constitutional design and their relationships to classical liberalism. Alternatively, the lack of scholarship exploring those relationships is of primary concern here. (5)

In the next section, we review the few empirical studies of constitutional design in relation to a classical liberal institutional environment. The rest of the paper is devoted to arguing that this paucity of applied study should be addressed. To do so, we begin in the second section by overviewing an ongoing debate regarding two alternative models of constitutional design: the entrenched/spare one versus the nnentrenched/detailed one. These models focus narrowly on two broad dimensions of constitutional design. We concentrate on them as a specific illustration because classical liberals have a preference for the former but without empirical support to back up that preference. In the third section, then, we describe the wealth of data available from the Comparative Constitutions Project (Elkins, Ginsburg, and Melton 2009). These data can provide invaluable insights into what a classical liberal constitution might look like. We employ the project data to take a preliminary look at whether entrenched/spare constitutions are associated with a classical liberal environment. The results are decidedly mixed, which should be a clarion call for systematic and serious study of the data, which we argue for in our concluding discussion.

(The Few) Existing Studies

The most straightforward studies of constitutional design in relation to an institutional environment of economic freedom are papers by Zane Spindler and Xavier de Vanssay (2002); de Vanssay, Vincent Hildebrand, and Spindler (2005); and Ram Mudambi, Pietro Navarra, and Christ Paul (2002). All of these studies focus on the Fraser Institute's Economic Freedom of the World (EFW) index, which can be interpreted as a measure of classical liberal institutions.

Spindler and de Vanssay (2002) explore the relationship between economic freedom and whether a country's constitution provides for (1) a federalist structure and/or (2) a bicameral legislature. They estimate cross-sectional OLS regressions with EFW index values as the dependent variable, finding that when the lagged value of EFW is controlled for, only the bicameral legislature value enters significantly. In that case, it enters positively, but the point estimate of the coefficient is very small (less than 0.5, whereas the EFW index runs from 0 to 10 with a standard deviation of about 1.7). De Vanssay, Hildebrand, and Spindler (2005) then consider panel EFW regressions that include a mix of de jure and de facto constitutional factors front the Database of Political Institutions (DPI) (Beck et al. 2001; Keefer and Stasavage 2003). Among the former, in cross-sectional and panel EFW only dummy variables for a finite executive term and a parliamentary system enter significantly (though not the latter when Organization for Economic Cooperation and Development countries are excluded). The point estimate on the latter is again fairly small (around 0.4); more importantly, despite what...