Emory Law Journal

A Comment on The Rule of Law Unplugged
Jeffrey K. Staton Assistant Professor, Department of Political Science, Emory University.

Introduction

There is a familiar rhetorical template around which rule of law papers are framed. It involves a dizzying number of legal scholars, both ancient and modern, who remind us of the subject’s gravity and intellectual pedigree. Although we may disagree over precisely what defining the concept entails, 1 See, e.g., Jeremy Waldron, Is the Rule of Law an Essentially Contested Concept (in Florida)?, 21 L. & Phil. 137 (2002) (providing a useful summary of the multiple definitions of the rule of law). the contours of the conceptual debate are clearly demarcated, and in any sense, the rule of law is normatively appealing. From economic development 2 See Daron Acemoglu et al., The Colonial Origins of Comparative Development: An Empirical Investigation, 91 Am. Econ. Rev. 1369 (2001) (using an instrumental variables approach to present evidence that the quality of a state’s formal legal institutions is an important cause of economic development). to political order 3 See, e.g., Douglass C. North et al., Order, Disorder, and Economic Change: Latin America vs. North America, in Governing for Prosperity 17 (Bruce Bueno de Mesquita & Hilton L. Root eds., 2000) (suggesting that legal institutions can lower the stakes of holding power, and in that way, reduce incentives to capture or maintain control of the state). and the expansion and protection of human rights, 4 See, e.g., Charles R. Epp, The Rights Revolution: Lawyers, Activists, and Supreme Courts in Comparative Perspective (1998) (observing that an independent judiciary is a necessary, if not sufficient, condition for a rights revolution). aspects of the rule of law seem critical to the maintenance of core elements of human welfare. In light of its normative and positive standing, we are informed of the dense international network of rule-of-law-reformers, who spend considerable resources tracking and advocating institutional change around the globe. 5 Thomas Carothers, The Rule of Law Revival, in Promoting the Rule of Law Abroad: In Search of Knowledge 3, 4, 10–11 (Thomas Carothers ed., 2006).

Insofar as this template accurately reflects rule-of-law scholarship and advocacy, and I believe that it does, it captures succinctly how intimidating the rule-of-law project is to anyone who has ever considered whether it just might be flawed in some fundamental ways. To be blunt about it, taking on the rule of law means taking on a cadre of academic luminaries, an impressive collection of empirical results, and a dense, well-funded, and ideologically committed network of policy advocates. Building a critical argument in the context of the template, then, is quite a challenge.

It is precisely this challenge that McCubbins, Rodriguez, and Weingast (MRW) take up in The Rule of Law Unplugged6 Mathew D. McCubbins, Daniel B. Rodriguez & Barry R. Weingast, The Rule of Law Unplugged, 59 Emory L.J. 1455 (2010). In so doing, they raise important questions about our ability to guide rule-of-law reform sensibly. Their critical goal alone makes the Article worth reading, but the authors also promise to construct a clear pathway toward effective rule-of-law advocacy. 7 See id. at 1458–59. As they write, the rule of law is not a “vacuous Rorschach test upon which legal scholars and reformers simply project their own views about the content and purpose of law. Rather, [they] see the rule of law as expressing a worthy aspiration that rightly finds voice in the hard work of good-intentioned activists.” 8 Id. at 1459. Yet, reflecting on the critical element of their article, I am left questioning whether the constructive enterprise is really what we want to consider constructive. Perhaps, it would be better to think about their second goal as reconstructive. If the argument is sound and if other scholarship, which they do not invoke but which suggests similar conclusions, is sound, then simply put, we have reason to question how well we can currently advocate reform.

The main problem they raise is a disconnect between conceptual and advocacy work on one hand and institutional research on the other. 9 See id. at 1456–58. My goal in this Essay is to ensure that the implications of their argument are placed clearly on the table. Institutional research raises questions about the way that we theorize about and measure reform and about the way we hope to learn about our models from data. By calling our attention to the ways in which this is so, MRW raise questions about core elements of the science underlying reformism.

The argument suggests a number of alarming implications. If we are clear about what we mean by the rule of law, and if we are faithful to that concept, then it is not clear whether we have developed a valid indicator. It also follows that we may not be able to develop one that can be applied equally to all communities around the globe. If we were forced to pay attention to all the possible processes underlying the construction of the rule of law, we would likely then have to question a key element of reform orthodoxy: that we know which institutions to change and how. We would want to question how much we can advance the rule of law via institutional design, especially in particular contexts. We might even question whether it is possible to design the rule of law at all. What is more, assuming that we can give instructions for how to build the rule of law, some strategies for doing so might undermine other core societal values. It is even possible that we may have to advocate the violation of some rule-of-law principles in order to construct it. 10 The authors largely put aside another fundamental concern—that we have yet to identify the right causal effect of the rule of law on the elements of human welfare that we believe it advances (for example, development and human rights). It may be that the effect is small or non-existent, and we have yet to develop an identification strategy that is fully satisfying. The authors do cite Stephen Haggard, et al., who summarizes this debate nicely. Id. at 1457 n.12 (citing Stephan Haggard et al., The Rule of Law and Economic Development, 11 Ann. Rev. Pol. Sci. 205 (2008)). To sum up then, if we take MRW seriously, we have reason to question the theories around which we wrap our reform advice, the measures we use to track it, and the inferences we draw from our empirical tests.

The gap between institutional research and advocacy is in part a failure of institutionalists. If advocates want clear strategies for reform and scholars cannot offer them, the two worlds may understandably grow apart. But it does seem that our desire for reform has gotten out in front of the science that should support it. This does not mean that we should scrap the rule of law concept or give up on trying to reform states that systematically violate rule of law values. On the contrary, like MRW, I believe strongly that the rule of law is a worthy aspiration, and I am optimistic enough to believe that the science can catch up to our reform interests. That said, The Rule of Law Unplugged demands that we carefully reconsider the ways that we research the subject and ultimately advocate for change.

The remainder of this Essay is divided in two parts. Part I identifies the target of the MRW critique. Part II restates their claims, expands on some of them, and develops others, which are not stressed in their Article, but which I believe advance the conversation. I conclude by underlining the implications of the argument for rule-of-law research and reform and suggest, if only briefly, how we might proceed.

I. The Target

To appreciate the implications of the argument set forth by MRW, it is useful to stress four assumptions on which they believe current reform practices are predicated. First, the rule of law is a universal value, which is applicable equally across and within state boundaries and over time. 11 McCubbins, Rodriguez & Weingast, supra note 6, at 1463. Second, the rule of law requires a specific set of institutions and governance structures. 12 Id. Third, the rule of law is fundamentally antithetical to autocratic governance. 13 Id. Fourth, the rule of law is more or less easily measured across states. 14 Id. at 1463–64. McCubbins, Rodriguez, and Weingast hors question each of these assumptions. The table below suggests why they seem so plausible.

The table displays the states with the ten highest and ten lowest rule-of-law scores in 2007, 15 The year 2007 was chosen because it was the last year for which I have available data for all the relevant indicators. as measured by the World Bank’s well-known scale. 16 Daniel Kaufmann et al., Governance Matters VIII: Aggregate and Individual Governance Indicators 1996–2008 (World Bank Dev. Research Group, Policy Research Working Paper No. 4978, 2009). The columns classify the states’ political regimes according to information provided by the Polity IV project. 17 Monty G. Marshall & Keith Jaggers, Political Regime Characteristics and Transitions, 1800–2008, Polity IV Project, http://www.systemicpeace.org/polity/polity4.htm (last visited March 2, 2010) (“The Polity IV Project . . . [codes] . . . the authority characteristics of states in the world system for purposes of comparative, quantitative analysis . . . The Polity conceptual scheme is unique in that it examines concomitant qualities of democratic and autocratic authority in governing institutions, rather than discreet and mutually exclusive forms of governance.”). Democratic states are listed on the left; autocratic states are listed on the right. States that are believed to have independent judiciaries, as indicated by the Henisz and Tate and Keith scores, receive stars. 18 The Henisz measure derives from the Polity IV project’s indicator of executive constraints and the measure of law and order according to Political Risk Services (PRS). Witold J. Henisz, The Institutional Environment for Economic Growth, 12 Econ. & Pol. 1 (2000) (providing details of the measures). Henisz codes a judiciary as independent if the executive constraints measure is high enough to reflect the presence of an independent judiciary and the PRS measure of law and order is sufficiently high as well. Id. at 27–28. Otherwise a judiciary is coded as dependent. Id. Insofar as the Polity regime score is a function of the executive constraints subcomponent, the Henisz score is clearly endogenous to the measure of democracy. For this reason, I included another measure of judicial independence provided by Tate and Keith. C. Neal Tate & Linda Camp Keith, Conceptualizing and Operationalizing Judicial Independence Globally (Sept. 1, 2007) (unpublished manuscript) (on file with Emory Law Journal) (paper delivered at the annual meeting of the American Political Science Association, Aug. 30–Sep. 1, 2007, Chicago, Illinois). The Tate and Keith measure is derived from U.S. State Department reports on human right. Id. at 15–17. It is a three-category ordinal scale (0=dependent; 1=partially dependent; 2= fully independent). Id. at 17. Three features of the table are immediately striking. First and most obviously, the states at the top and bottom of the rule-of-law scale seem appropriately placed, at least relative to each other. It is hard to imagine an argument defending the validity of an indicator in which Denmark and Switzerland fall below Chad or Myanmar in 2007, under any concept of the rule of law. Second, the states in the top ten are all democracies, whereas all but two states in the bottom ten are autocracies. Even the low rule-of-law democracies are borderline cases on the democratic dimension, receiving a five on the Polity IV scale, which some scholars might treat as an indicator of autocracy. 19 The Polity IV team suggests treating states with a score at or above six as “democratic” regimes. Because the scale is ordinal and includes a large number of categories, it is worth questioning any choice to define regime type with respect to a specific threshold, as this strategy has us do. Likewise, not a single state in the top ten has a judiciary that is notoriously dependent on its government, 20 Although I will not push the point here, another implication of institutional research on judicial independence is that no court is fully autonomous of the political context in which it is situated. See, e.g., John Ferejohn, Independent Judges, Dependent Judiciary: Explaining Judicial Independence, 72 S. Cal. L. Rev. 353 (1999). The claim is relatively non-controversial if we limit ourselves to the developing world, but there is also strong evidence of extra-legal influences on judicial decision making in domestic courts as well regarded as the German Federal Constitutional Court and supra-national bodies like the European Court of Justice. See Georg Vanberg, The Politics of Constitutional Review in Germany (2005); Clifford J. Carrubba et al., Judicial Behavior Under Political Constraints: Evidence from the European Court of Justice, 102 Am. Pol. Sci. Rev. 435 (2008). In light of this work, the cross-national measures of judicial independence referenced in the table are best thought of as providing rough and relative information on autonomy. See table infra note 22 whereas every state in the bottom ten possesses a judiciary about which such concerns have been raised. 21 Tate & Keith, supra note 18.Table: The Rule of Law, Democracy, and Judicial Independence (2007) 22 This table displays the states with the ten highest scores on the World Governance Indicator rule-of-law scale by regime type as measured by the Polity IV scale, where any state above a score of five is treated as a democracy, and all states below five are treated as autocracies. See Marshall & Jaggers, supra note 17. Iceland, Luxembourg, Iraq, and Afghanistan are excluded from the Table for lack of information on regime type or judicial independence

 

 

Regime Type

 

Democracy

Autocracy

 

 

 

 

Rule of Law Top Ten

 

Denmark*

Norway*

Switzerland*

Sweden*

New Zealand*

Finland*

Austria*

Canada*

Australia*

Ireland*

 

 

 

 

 

Rule of Law Bottom Ten

 

 

 

Venezuela

Congo D.R.

 

 

 

Haiti

Myanmar

Guinea

Sudan

Central African Republic

Ivory Coast#

Zimbabwe

Somalia

* An asterisk (*) indicates states whose judiciaries are classified as independent by both Henisz as well as Tate and Keith. 23 See supra discussion at note 18. No state other than the Ivory Coast was coded as independent by either Henisz or Tate and Keith. Id.

# The judiciary is coded as independent by the Henisz measure, but not by Tate and Keith. 24 See supra discussion at note 18.

This is the kind of empirical information that has been used to support core assumptions about the rule of law in the advocacy community. Indeed, we probably did not need the quantitative measures to tell us that democracy, judicial independence, and the rule of law hang together to some extent around the world. Any thin historical sense of the institutional histories of Europe and Africa would have suggested what is plainly evident in the Table. Nevertheless, the table underscores why the MRW task is far from trivial—they are challenging a conventional wisdom grounded in a number of indisputable historical facts. At first blush at least, the data give no reason to question the notion that the rule of law is tightly linked to democracy, that an independent judiciary—among other institutional features—is essential to its promotion, or much less that we can measure the core concept at issue. 25 For a discussion of common measures of the rule of law, see Svend-Erik Skaaning, Measuring the Rule of Law, 63 Pol. Res. Q. 449 (2010).

Despite what the table suggests, the correlation between the Polity IV democracy score and the World Bank rule-of-law indicator for all available years turns out to be only .45 (.40 in 2007), 26 Justin Esarey & Neza Sakari, Political Stability, Ethnic Fractionalization, and Political Institutions 8 (Jan. 24, 2010) (unpublished manuscript) (on file with the Emory Law Journal). This is a level of association that seems consistent with a positive link between the rule of law and democracy, but far from an essential one. Further, Esarey and Sarkari also find that the relationship between the rule of law and democracy is likely curvilinear. 27 Id. As they write, the “rule of law is strongest in the most autocratic and the most democratic states.” 28 Id. This result suggests either that the indicator is not tapping into the rule of law concept reformers have in mind, which calls into question the fourth assumption, or that autocracy and the rule of law are not incompatible, which calls into question the third assumption. And, of course, nothing in this simple exercise can speak to the process by which the rule of law is constructed, much less whether there are reasons to believe that particular institutional arrangements are required. So, despite initial appearances, it is unclear whether we should accept without criticism the core underlying assumptions identified by MRW.

II. The Claims

McCubbins, Rodriguez, and Weingast advance three principle claims about rule-of-law-reformism. First, they argue that any useful concept of the rule of law must take into account substantive values, incorporate a normative theory of law, and clarify how the dimensions of the concept are to be valued. 29See McCubbins, Rodriguez & Weingast, supra note 6, at 1458. Second, they suggest that the rule of law must be measured appropriately in light of this concept. 30 Id. at 1458–59. Third, they contend that reform advice must be subject to careful scientific inquiry. 31 Id. at 1459. This will require a good theory of institutional reform as well as empirical evidence, which is consistent with the predictions from that theory. Any theory of that sort, they argue, will likely require that reformers confront a series of tradeoffs of institutional design. On their face, none of these claims are particularly controversial. They are merely asking that rule-of-law advocacy be connected to scientific values. In practice, however, the implications of such a move turn out to be consequential for reformism.

A. The Rule of Law Requires Substance and a Theory of Law

McCubbins, Rodriguez, and Weingast argue primarily that any useful concept of the rule of law must be connected to minimal substantive values and married to a normative theory of law. 32 Id. at 1471. Beyond its core procedural features, which might be derived from Fuller’s list (e.g., clarity, transparency, generality, stability) and which allow individuals to be guided by the law, the rule of law dictates that “society should have decent confidence that their expressed desires should be accommodated and respected by their representatives.” 33 Id. If we accept this conceptual assessment, the implication for the rule-of-law project is immediate. In order to sustain the assumption that the rule of law is a universal value, which is applicable equally across the globe and over time, we now have to come to an agreement on what constitutes an appropriate link between societal values and the policies pursued by leaders.

There is a strong and a weak version of this point’s implication. Under the strong version, where we might insist that the connection between societal values and policy outputs should be chosen via some sort of democratic process, MRW’s point considerably frustrates the effort to measure the rule of law. Fundamental results derived from the theory of collective choice raise the question whether any normatively appealing rules for aggregating individual preferences can ensure that social choices are stable and free from manipulation. 34 See generally Kenneth J. Arrow, Social Choice and Individual Values (2d ed. 1963); Richard D. McKelvey, Intransivities in Multidimensional Voting Models and Some Implications for Agenda Control, 12 J. Econ. Theory 472 (1976). Within this tradition, it is unclear how we could come to an agreement on what constitutes an “appropriate connection” between underlying individual values and policy outputs, or whether any society could meet our ideal. Articulating this standard is going to be especially problematic since we are asking for a connection between societal values and policies across multiple rule-of-law dimensions. If we cannot come to an agreement about what the right connection would look like, we will have to jettison our efforts to measure the rule of law. And if we do that, we threaten removing the rule of law from the realm of scientific inquiry. If this is what MRW have in mind, it is difficult to see how to reconstruct our measurement strategy.

A weaker implication of the point is possible, though. Perhaps MRW simply wish to say that the rule of law requires some minimal commitment to outcomes that are substantively fair. If we can agree on, say, a fundamental, universal notion of fairness, then we can sustain the assumption that the rule of law is a universal value. Yet, if we are willing to admit that there can be at least some variance across cultures and communities with respect to what is fair, then it is again hard to sustain the notion that the rule of law can be a universal value.

The claim that a normative theory of law must be married to the rule-of-law concept further complicates the effort to sustain the universality assumption. If we connect a theory of law to the rule of law, we immediately confront tensions between rule-of-law values themselves. The most obvious example involves how we should evaluate the possibility of replacing a legal standard with a rule. Whereas the rule might advance the rule-of-law value of “clarity,” the doctrinal disruption associated with replacing a vague standard with a clear rule undermines the rule-of-law value of “stability.” How we evaluate this tradeoff seems inextricably linked to our understanding of what law should be.

Although I am persuaded by MRW’s first two conceptual claims, they are far from unassailable. McCubbins, Rodriguez, and Weingast may very well be right about the substantive nature of the rule of law and the corresponding need to link it to a normative theory of law, but these claims target the substance of the concept rather than its clarity, which is of greater concern to the remainder of their argument. A perfectly reasonable response is that the rule of law is a procedural concept, regardless of MRW’s interest in making it substantive. The real bite in their conceptual argument lies with the final claim they proffer: If we want to measure the rule of law appropriately, we are going to have to clarify how we value the different dimensions of the concept. 35 McCubbins, Rodriguez & Weingast, supra note 6, at Part II: Measuring the Rule of Law. It is with this point in mind that I turn to their measurement claim.

B. The Rule of Law Must Be Measured

It is hard to imagine a less controversial claim than that the rule of law must be measured. To appreciate its value, we must wed it to MRW’s final conceptual point—that rule-of-law concepts do not explicitly articulate the relationships between sub-dimensions. 36 Id. Raz’s concept provides an immediate example. 37 See Joseph Raz, The Authority of Law: Essays on Law and Morality 213 (2d ed. 2009). He contends that the rule of law “has two aspects: (1) that people should be ruled by the law and obey it, and (2) that the law should be such that people will be able to be guided by it.” 38 Id. The focus of his analysis is on the second condition, on what is required for people to be guided by law. But the first condition is no less important. Thus, the question arises: How do we value the extent to which individuals follow the law in practice relative to the extent to which it is possible for them to do so? We can ask the same of Fuller’s list of appropriate characteristics of the law. 39 Lon L. Fuller, The Morality of Law (rev. ed. 1969). How should we value generality relative to clarity? If there is a conflict between a general statement of law and a clear statement, how do we evaluate the tradeoff? Insofar as the world is awash with examples of vague but general laws 40 For a discussion about vagueness in the law, see Timothy Endicott, Law Is Necessarily Vague, 7 Legal Theory 379 (2001). On the incentives for legislators to construct vague law, see John Huber & Charles Shipan, Deliberate Discretion: The Institutional Foundations of Bureaucratic Autonomy (2002). and insofar as individuals and public officials alike commonly violate what are undoubtedly clear rules, these questions are far from abstract or inconsequential.

The essential idea is that moving from the typical list-making activity, which is associated with defining the rule of law, to an operational measure of the concept requires a statement about the hierarchy of elements on the list. If the sub-dimensions are in tension with each other, the task becomes harder still. Even if the statement MRW is looking for is that each dimension of the rule of law (e.g., generality, stability, social order) is as important as any other, the statement must be explicit. It is important to stress why.

Imagine that our measurement goal is to develop a one-dimensional indicator of the rule of law, like that produced by the World Bank and described in the table. If our concept is really multidimensional, then our measurement task is going to require that we aggregate information across several dimensions. And absent a statement about the relative importance of the dimensions, it is impossible to validate whatever measure we derive. When one aggregation choice is as good as any other, it becomes impossible to know whether a particular aggregation choice is appropriate. We need not even restrict ourselves to the assumption that the rule of law is a linear or even a smooth function of the underlying dimensions. We might imagine that there are many combinations of the rule-of-law traits, which result in a context that we would like to describe as reflecting the rule of law. 41 For example, we might imagine that a community where the law is typically vague but general and largely obeyed is just as much a “rule-of-law community” as one where the law is typically clear, but highly specific to individuals and occasionally disregarded. It is not so much that this must be true, but rather that it is possible under some concept of the rule of law. But what we cannot do is merely say “we know the rule of law when we see it.”

In light of the aggregation issue, it is worth returning to the World Bank’s measure, which aggregates information from individual and expert surveys, credit rating agencies, non-governmental organizations, and selected states. 42 Daniel Kaufmann et al., Governance Matters VIII: Aggregate and Individual Governance Indicators for 1996–2008, at 12–13 (World Bank Policy Research, Working Paper No. 4978, 2009) (on file with Emory Law Journal). The central measurement problem the team tackles involves how to weight the information provided by each source about each state for each year when there is a lack of agreement across the sources. This is a tricky aggregation problem for sure. The estimates they derive are the product of a statistical model, which assumes that the information from the sources all provide information on the latent trait (i.e., the rule of law), and which allows the team to sort out stronger signals about that trait from noisier ones. Critically, although this approach is appropriate given the problem the team identifies, it is quite a different aggregation problem from that which I have been discussing. The world governance indicators approach makes no explicit conceptual choice about the relationship among underlying dimensions. 43 See Governance Matters 2009: Worldwide Governance Indicators, 1996–2008, World Bank Group, http://info.worldbank.org/governance/wgi/sc_country.asp (last visited June 12, 2010). And the sources provide information on multiple elements of the rule of law—from confidence in the police to land rights and beliefs in judicial impartiality. 44 See Kaufmann et al., supra note 42, at 39–71. This feature of the measure sheds some light on the curvilinear relationship between democracy and the rule of law, which leads back to the findings of Esarey and Sarkari. 45 Esarey & Sakari, supra note 26, at 25–27. If we value social order or crime equally with the enforceability of contracts or the impartiality of the judiciary, we should expect to observe authoritarian states with reasonably high rule-of-law scores, particularly ones that repress dissent yet control crime (e.g., Singapore or perhaps Chile under Pinochet). Whether that is what we mean by the rule of law is an open question. And the openness of the question is a problem.

C. Reformism Must Be Better Connected to Research

McCubbins, Rodriguez, and Weingast write, “[R]ule-of-law reform must contemplate the relationship between means and ends. A satisfactory understanding of this relationship requires positive theory and empirical support, not merely normative leaps of faith or ipse dixit.” 46 McCubbins, Rodriguez & Weingast, supra note 6, at 1459. The authors remind us that key lines of research on elements of the rule of law highlight the role of political context and require designers to confront difficult tradeoffs. The authors develop this argument most forcefully in the context of the debate over the unitary executive and the separation of powers. 47 Id. at 1486–89. I will stress these points in the context of work on judicial independence. I divide this section in two parts. In the first, I consider the extent to which we believe that judicial independence is designable. In the second, I assume that is, and ask how we should go about it.

1. Political Context and the Emergence of Independent Courts

The first challenge to our ability to construct the rule of law via judicial independence is that judicial independence depends on features outside a state’s legal system. A body of scholarship has proposed, and provided evidence to support, the notion that independence depends on features of political context, most notably, government fragmentation. 48 See Rebecca Bill Chavez, Rule of Law in Nascent Democracies: Judicial Politics in Argentina (2004); Julio Ríos-Figueroa, Fragmentation of Power and the Emergence of an Effective Judiciary in Mexico, 1994–2002, 49 Latin Am. Pol. & Soc’y 31 (2007). When the set of officials with control over judicial resources or whose participation is required for the faithful implementation of judicial decisions are limited by explicit or implicit vetoes from other officials, judges are freed to resolve decisions consistently with their sincere evaluations of the record. 49 Ríos-Figueroa, supra note 48, at 31–57 (observing that the Supreme Court of Mexico Court is far more likely to strike down national laws under periods of divided government than periods of unified government). Unfortunately, though fragmentation can be designed, or at least encouraged, via carefully selected electoral institutions, it is unclear whether we should advocate such a change because fragmentation is also associated with inefficient social spending, 50 See generally Nouriel Roubini & Jeffrey Sachs, Government Spending and Budget Deficits in the Industrial Countries, 8 Econ. Pol’y 99 (1989). and perhaps of greater concern, regime instability. 51 See generally Matt Golder, Presidential Coattails and Legislative Fragmentation, 50 Am. J. Pol. Sci. 34 (2006); Scott Mainwaring, Presidentialism, Multipartism, and Democracy, 26 Comp. Pol. Stud. 198 (1993). Thus, increasing fragmentation might encourage judicial independence, but for related reasons, pork and the propensity for political disorder would also increase.

Other sources of judicial independence are not designable, even indirectly. For example, scholars have suggested that a deep public commitment to the institutional integrity of the judiciary is a necessary condition for courts to constrain government behavior. 52 See James L. Gibson et al., On the Legitimacy of National High Courts, 92 Am. Pol. Sci. Rev. 343 (1998); Matthew C. Stephenson, Court of Public Opinion: Government Accountability and Judicial Independence, 20 J.L. Econ. & Org. 379 (2004). Still others have suggested that if public belief in judicial legitimacy is necessary to constrain government, the conflicts courts resolve will have to be transparent. Where the public is unaware of government activities, public pressure can be no constraint. 53 See generally Vanberg, supra note 20. On this account, the promotion of judicial independence will involve a clear strategy for building judicial legitimacy and perhaps ensuring clear and accurate media coverage of the law. More problematic is that in some political contexts, transparency and judicial legitimacy may be in tension with each other. If the legitimacy theory account is correct, then transparency triggers the public pressure needed for independent judging. And for this reason, judges have strong incentives to invite and shape media coverage. 54 See generally Richard Davis, Decisions and Images: The Supreme Court and the Press (1994). But if judges face significant political pressures in particular cases, which overwhelm whatever source of power they might derive from legitimacy and undermine independence, then transparency will only expose the public to non-independent judicial behavior. That kind of information can undermine legitimacy beliefs. 55 See generally Jeffrey K. Staton, Judicial Power and Strategic Communication in Mexico (2010).

Of perhaps greater importance for our interests in promoting rule-of-law reform are the growing accounts of judge-led efforts to construct judicial independence, which posit that courts can gain autonomy over time by using careful strategies of prudence. 56 Tom Ginsburg, Judicial Review in New Democracies 73 (2003); Clifford J. Carrubba et al., Judicial Behavior Under Political Constraints: Evidence from the European Court of Justice, 71 J. Pol. 1 (2009). Consider Ginsburg, who has evaluated court-building projects in contexts as distinct as the United States, South Korea, and Taiwan. 57 Ginsburg, supra note 56. According to Ginsburg’s account, managing political pressures in a court’s infancy—by not asking the government to implement unacceptable policies when it would simply ignore such orders or attack the bench—can help courts simultaneously avoid political clashes that raise questions about independence and subtly induce a norm of compliance. 58 Id. at 90–96. By pursuing this strategy over time, judges ultimately expand the boundaries of their authority and evolve into powerful, independent actors. The empirical evidence in support of this mechanism is historical in nature, but it has proven difficult to test the model in a systematic way either on field data or in the laboratory. But if the idea is right, the implication for reform is significant. We may advance the rule of law via judicial behavior that clearly violates core rule-of-law values. The strategy of prudence requires the incremental exercise of judicial authority. If Ginsburg is right, our reform interests require that we ask questions of the following sort: How long must a court engage in the Ginsburg strategy before it can begin resolving cases sincerely? Will this kind of strategy strain efforts to build judicial legitimacy? And if it does, what amount of legitimacy is worth building a norm of compliance? Is that even possible?

2. Judicial Rules and Judicial Behavior

What if we assume that though contextual influences like fragmentation or judicial legitimacy matter at the margin, we can nevertheless construct judicial independence via institutional design? This assumption would require us to confront our uncertain knowledge about the link between rules and behavior.

Conventional wisdom suggests that if we can induce independence via institutional design, it will be through the rules that insulate judges from external political interference. These rules include fixed, relatively long tenure, budgetary autonomy and multilateral appointment procedures, among others. 59 See generally Gerald N. Rosenberg, The Hollow Hope: Can Courts Bring About Social Change (1991). Unfortunately, the empirical evidence is far from consistent with this pathway. Consider the figure below, which shows zero-order correlations between Ríos-Figueroa’s detailed measure of constitutional provisions for judicial insulation in Latin American states from 1945 to 2005, and seven measures of de facto judicial independence. 60 See Andrea Pozas-Loyo & Julio Ríos Figueroa, Enacting Constitutionalism: The Origins of Independent Judicial Institutions in Latin America, 42 Comp. Pol. 293 (2010). A diamond indicates the result for democracies and a circle indicates the result for all states in the region. The associations are obviously quite small—a few are even negative. On the whole, there would appear to be no direct relationship between formal constitutional rules and judicial behavior, at least among this sample of Latin American states. Yet, these are just zero-order correlations; perhaps a well-specified, theoretically informed model would suggest different results. Alarmingly, however, scholars lack consensus about the influence of formal institutions on judicial behavior. Although there are some studies demonstrating a positive relationship between de jure and de facto judicial independence, as estimated by expert surveys, 61 See Bernd Hayo & Stefan Voigt, Explaining De Facto Judicial Independence, 27 Int’l Rev. L. & Econ. 269, 282–88 (2007). studies that have focused explicitly on judicial behavior—as captured by actual decisions in real controversies—and on particular courts rather than general impressions about the system commonly fail to uncover a connection. 62 See Erik S. Herron & Kirk A. Randazzo, The Relationship Between Independence and Judicial Review in Post-Communist Courts, 65 J. Pol. 422 (2003).

Figure: Judicial Rules and Judicial Behavior in Latin America 63 This figure displays zero-order correlations between Ríos-Figueroa’s measure of formal constitutional provisions designed to create judicial independence and behavioral measures of judicial independence for eighteen Latin American states between 1945 and 2005. A circle indicates the result for all states; a diamond represents the result for democracies.

One unsurprising explanation for these results is that there is little relationship between formal rules and judicial behavior. Informal norms are far more important. 64 Daniel M. Brinks, The Judicial Response to Police Killings in Latin America: Inequality and the Rule of Law (2008). Another possible explanation is that the relationship between formal rules and behavior is highly dependent on political conditions like divided government. 65 Andrea Pozas-Loyo & Julio Ríos-Figueroa, When and Why ‘Law’ and ‘Reality’ Coincide? De Jure and De Facto Judicial Independence in Chile and Mexico, in Judicial Independence and Accountability: A Debate 168, 178–197 (P. Sabiha Khanum ed., 2008). In contrast, Gretchen Helmke and I have argued that a single formal rule can incentivize two very different kinds of behaviors. Consider judicial tenure. 66 See Gretchen Helmke & Jeffrey K. Staton, The Puzzle of Judicial Politics in Latin America: A Theory of Litigation, Judicial Decisions and Inter-Branch Conflict, in Courts in Latin America (Gretchen Helmke & Julio Ríos-Figueroa eds., forthcoming Dec. 2010). Ostensibly, the logic behind fixing and increasing judicial tenure is that job security lowers the incentive for a judge to curry favor with politicians who enjoy control over her post. Yet, insofar as the post itself is valuable, life tenure raises its value by increasing the stream of perks and privileges associated with the seat. As long as there are ways to remove judges—even judges with life tenure—lengthening the term of employment can also increase incentives for prudence. 67 Removal can be through constitutional or extra-constitutional means, of course. Yet as we look around the world, it is clear that both approaches have been used to remove particular judicial obstacles to political coalitions. See Gretchen Helmke, Courts Under Constraints: Judges, Generals, and Presidents in Argentina (2005). For this reason, and especially in a context of political conflict where impeachment or imprisonment are meaningful possibilities, increasing tenure simultaneously advances judicial independence by increasing job security and undermines judicial independence by increasing the value of the seat.

This argument has clear implications for research design. If rules induce competing behavioral incentives, then behavioral studies that make use of field data will not be able to sort out the competition easily. A study using cross-national measures of constitutional rules governing judicial insulation to estimate variance in judicial behavior will find it challenging to pull apart judicial incentives associated with the same rule. A study using cross-national expert opinions of the many courts in a system will only compound the problem. To properly sort out the effects of judicial institutions, we are likely going to have to move to experimental research designs.

D. Implications

I conclude by stressing the key implications of the MRW argument. First and foremost, if we are unwilling or unable to state clearly how it is that we value the sub-dimensions of the rule of law, we cannot develop a valid indicator. Thus, until we formulate a coherent concept, even the most basic effort to describe the rule of law globally will result in an unclear picture.

Second, the multidimensionality of the concept calls into question whether it should be conceptualized independently of context. People might reasonably disagree over whether their community values legal clarity more than equity, or generality more than systematic compliance. It even seems reasonable to ask whether one might value a commitment to social order at the expense of arbitrary governance. Of course, current measures do not allow us to observe precisely how states have made these tradeoffs in practice, and it is unclear whether policy advocates are pursuing the matter on the ground. The commitment to a universal notion of the rule of law would seem to preclude such a strategy, but there are good reasons to question this universality.

Third, and quite critically, it is unclear whether advocacy programs are incorporating lessons from institutional research on those aspects of the state that are allegedly essential to the rule of law. I have focused on judicial independence. For the most part, our understanding of how specific institutional rules influence judicial behavior is guided by conventional wisdom, not careful theory. And the empirical record testing the conventional wisdom is extremely mixed. As we have begun to develop more sophisticated accounts of these rules, we have learned that they create competing incentives for judicial behavior. These accounts suggest the need for research designs that are different than those currently in vogue, which rely on time-series, cross-sectional regression techniques applied to field data. Although institutions believed to promote judicial independence are just one element of the suite of rules associated with the rule of law, they are pretty consequential ones. The implication is that we should question how resolutely we can claim that we know what rules are required to promote the rule of law and why.

Fourth, it is possible that the marginal effect of institutional design on the construction of the rule of law is terrifically small relative to other strategies. Yet, it may be that indirect approaches to building the rule of law (say by incentivizing more fragmented government) risks causing other socio-political failures. For this reason, reformers must have some way of evaluating corresponding tradeoffs, and we should accept that different communities might evaluate their tradeoffs differently. Finally, a stream of research on the construction of the rule of law suggests that judges may have to violate core rule-of-law tenets in order to build it. It is unclear precisely how this should be done, and advocating such a strategy in the context of the broader rule-of-law narrative is bound to be uncomfortable. It will certainly prove challenging.

In summary, the rule-of-law concept involves measurement challenges that have yet to be solved. Our theories about its construction suggest that we may not be able to design it directly, and if we can, we may have to balance that against other values and political goals. These are pretty serious implications. We find ourselves in a position to question much of the template around which rule-of-law scholarship is framed.

What should we do about this? Aside from a general call to address these problems and a request for the reform community to more tightly link their recommendations to the scientific record, MRW are largely vague. In my view, a reasonable preliminary step would involve carving up the dimensions of the rule of law. Although the focus on the aggregate concept and its aggregate measures is alluring, its multidimensional nature asks a lot of an empirical researcher. Perhaps it asks too much. Instead of studying the rule of law in toto, we might study its sub-dimensions more carefully.

Of course, this is already being done. It is the relative quality of extant work on corruption or property rights or police professionalism that seems to induce a sense that we can combine lessons from each of these fields and easily develop a coherent science of the whole. However, we have yet to take many of the necessary preliminary steps on the road to a coherent theory of the rule of law. Most obviously, before we can ask how to sequence a series of major institutional changes, we really must understand institutional effects better than we do. Doing this right may require that we de-emphasize our efforts to develop international scales of expert opinions or to collect cross-national maps of constitutional rules, and instead focus on the development of careful theoretical claims to be tested in the laboratory. If we want good answers to our questions, it seems crucial that we slow down and rescale our work.

Footnotes

Assistant Professor, Department of Political Science, Emory University.

1 See, e.g., Jeremy Waldron, Is the Rule of Law an Essentially Contested Concept (in Florida)?, 21 L. & Phil. 137 (2002) (providing a useful summary of the multiple definitions of the rule of law).

2 See Daron Acemoglu et al., The Colonial Origins of Comparative Development: An Empirical Investigation, 91 Am. Econ. Rev. 1369 (2001) (using an instrumental variables approach to present evidence that the quality of a state’s formal legal institutions is an important cause of economic development).

3 See, e.g., Douglass C. North et al., Order, Disorder, and Economic Change: Latin America vs. North America, in Governing for Prosperity 17 (Bruce Bueno de Mesquita & Hilton L. Root eds., 2000) (suggesting that legal institutions can lower the stakes of holding power, and in that way, reduce incentives to capture or maintain control of the state).

4 See, e.g., Charles R. Epp, The Rights Revolution: Lawyers, Activists, and Supreme Courts in Comparative Perspective (1998) (observing that an independent judiciary is a necessary, if not sufficient, condition for a rights revolution).

5 Thomas Carothers, The Rule of Law Revival, in Promoting the Rule of Law Abroad: In Search of Knowledge 3, 4, 10–11 (Thomas Carothers ed., 2006).

6 Mathew D. McCubbins, Daniel B. Rodriguez & Barry R. Weingast, The Rule of Law Unplugged, 59 Emory L.J. 1455 (2010).

7 See id. at 1458–59.

8 Id. at 1459.

9 See id. at 1456–58.

10 The authors largely put aside another fundamental concern—that we have yet to identify the right causal effect of the rule of law on the elements of human welfare that we believe it advances (for example, development and human rights). It may be that the effect is small or non-existent, and we have yet to develop an identification strategy that is fully satisfying. The authors do cite Stephen Haggard, et al., who summarizes this debate nicely. Id. at 1457 n.12 (citing Stephan Haggard et al., The Rule of Law and Economic Development, 11 Ann. Rev. Pol. Sci. 205 (2008)).

11 McCubbins, Rodriguez & Weingast, supra note 6, at 1463.

12 Id.

13 Id.

14 Id. at 1463–64.

15 The year 2007 was chosen because it was the last year for which I have available data for all the relevant indicators.

16 Daniel Kaufmann et al., Governance Matters VIII: Aggregate and Individual Governance Indicators 1996–2008 (World Bank Dev. Research Group, Policy Research Working Paper No. 4978, 2009).

17 Monty G. Marshall & Keith Jaggers, Political Regime Characteristics and Transitions, 1800–2008, Polity IV Project, http://www.systemicpeace.org/polity/polity4.htm (last visited March 2, 2010) (“The Polity IV Project . . . [codes] . . . the authority characteristics of states in the world system for purposes of comparative, quantitative analysis . . . The Polity conceptual scheme is unique in that it examines concomitant qualities of democratic and autocratic authority in governing institutions, rather than discreet and mutually exclusive forms of governance.”).

18 The Henisz measure derives from the Polity IV project’s indicator of executive constraints and the measure of law and order according to Political Risk Services (PRS). Witold J. Henisz, The Institutional Environment for Economic Growth, 12 Econ. & Pol. 1 (2000) (providing details of the measures). Henisz codes a judiciary as independent if the executive constraints measure is high enough to reflect the presence of an independent judiciary and the PRS measure of law and order is sufficiently high as well. Id. at 27–28. Otherwise a judiciary is coded as dependent. Id. Insofar as the Polity regime score is a function of the executive constraints subcomponent, the Henisz score is clearly endogenous to the measure of democracy. For this reason, I included another measure of judicial independence provided by Tate and Keith. C. Neal Tate & Linda Camp Keith, Conceptualizing and Operationalizing Judicial Independence Globally (Sept. 1, 2007) (unpublished manuscript) (on file with Emory Law Journal) (paper delivered at the annual meeting of the American Political Science Association, Aug. 30–Sep. 1, 2007, Chicago, Illinois). The Tate and Keith measure is derived from U.S. State Department reports on human right. Id. at 15–17. It is a three-category ordinal scale (0=dependent; 1=partially dependent; 2= fully independent). Id. at 17.

19 The Polity IV team suggests treating states with a score at or above six as “democratic” regimes. Because the scale is ordinal and includes a large number of categories, it is worth questioning any choice to define regime type with respect to a specific threshold, as this strategy has us do.

20 Although I will not push the point here, another implication of institutional research on judicial independence is that no court is fully autonomous of the political context in which it is situated. See, e.g., John Ferejohn, Independent Judges, Dependent Judiciary: Explaining Judicial Independence, 72 S. Cal. L. Rev. 353 (1999). The claim is relatively non-controversial if we limit ourselves to the developing world, but there is also strong evidence of extra-legal influences on judicial decision making in domestic courts as well regarded as the German Federal Constitutional Court and supra-national bodies like the European Court of Justice. See Georg Vanberg, The Politics of Constitutional Review in Germany (2005); Clifford J. Carrubba et al., Judicial Behavior Under Political Constraints: Evidence from the European Court of Justice, 102 Am. Pol. Sci. Rev. 435 (2008). In light of this work, the cross-national measures of judicial independence referenced in the table are best thought of as providing rough and relative information on autonomy. See table infra note 22

21 Tate & Keith, supra note 18.

22 This table displays the states with the ten highest scores on the World Governance Indicator rule-of-law scale by regime type as measured by the Polity IV scale, where any state above a score of five is treated as a democracy, and all states below five are treated as autocracies. See Marshall & Jaggers, supra note 17. Iceland, Luxembourg, Iraq, and Afghanistan are excluded from the Table for lack of information on regime type or judicial independence

23 See supra discussion at note 18. No state other than the Ivory Coast was coded as independent by either Henisz or Tate and Keith. Id.

24 See supra discussion at note 18.

25 For a discussion of common measures of the rule of law, see Svend-Erik Skaaning, Measuring the Rule of Law, 63 Pol. Res. Q. 449 (2010).

26 Justin Esarey & Neza Sakari, Political Stability, Ethnic Fractionalization, and Political Institutions 8 (Jan. 24, 2010) (unpublished manuscript) (on file with the Emory Law Journal).

27 Id.

28 Id.

29See McCubbins, Rodriguez & Weingast, supra note 6, at 1458.

30 Id. at 1458–59.

31 Id. at 1459.

32 Id. at 1471.

33 Id.

34 See generally Kenneth J. Arrow, Social Choice and Individual Values (2d ed. 1963); Richard D. McKelvey, Intransivities in Multidimensional Voting Models and Some Implications for Agenda Control, 12 J. Econ. Theory 472 (1976).

35 McCubbins, Rodriguez & Weingast, supra note 6, at Part II: Measuring the Rule of Law.

36 Id.

37 See Joseph Raz, The Authority of Law: Essays on Law and Morality 213 (2d ed. 2009).

38 Id.

39 Lon L. Fuller, The Morality of Law (rev. ed. 1969).

40 For a discussion about vagueness in the law, see Timothy Endicott, Law Is Necessarily Vague, 7 Legal Theory 379 (2001). On the incentives for legislators to construct vague law, see John Huber & Charles Shipan, Deliberate Discretion: The Institutional Foundations of Bureaucratic Autonomy (2002).

41 For example, we might imagine that a community where the law is typically vague but general and largely obeyed is just as much a “rule-of-law community” as one where the law is typically clear, but highly specific to individuals and occasionally disregarded. It is not so much that this must be true, but rather that it is possible under some concept of the rule of law.

42 Daniel Kaufmann et al., Governance Matters VIII: Aggregate and Individual Governance Indicators for 1996–2008, at 12–13 (World Bank Policy Research, Working Paper No. 4978, 2009) (on file with Emory Law Journal).

43 See Governance Matters 2009: Worldwide Governance Indicators, 1996–2008, World Bank Group, http://info.worldbank.org/governance/wgi/sc_country.asp (last visited June 12, 2010).

44 See Kaufmann et al., supra note 42, at 39–71.

45 Esarey & Sakari, supra note 26, at 25–27.

46 McCubbins, Rodriguez & Weingast, supra note 6, at 1459.

47 Id. at 1486–89.

48 See Rebecca Bill Chavez, Rule of Law in Nascent Democracies: Judicial Politics in Argentina (2004); Julio Ríos-Figueroa, Fragmentation of Power and the Emergence of an Effective Judiciary in Mexico, 1994–2002, 49 Latin Am. Pol. & Soc’y 31 (2007).

49 Ríos-Figueroa, supra note 48, at 31–57 (observing that the Supreme Court of Mexico Court is far more likely to strike down national laws under periods of divided government than periods of unified government).

50 See generally Nouriel Roubini & Jeffrey Sachs, Government Spending and Budget Deficits in the Industrial Countries, 8 Econ. Pol’y 99 (1989).

51 See generally Matt Golder, Presidential Coattails and Legislative Fragmentation, 50 Am. J. Pol. Sci. 34 (2006); Scott Mainwaring, Presidentialism, Multipartism, and Democracy, 26 Comp. Pol. Stud. 198 (1993).

52 See James L. Gibson et al., On the Legitimacy of National High Courts, 92 Am. Pol. Sci. Rev. 343 (1998); Matthew C. Stephenson, Court of Public Opinion: Government Accountability and Judicial Independence, 20 J.L. Econ. & Org. 379 (2004).

53 See generally Vanberg, supra note 20.

54 See generally Richard Davis, Decisions and Images: The Supreme Court and the Press (1994).

55 See generally Jeffrey K. Staton, Judicial Power and Strategic Communication in Mexico (2010).

56 Tom Ginsburg, Judicial Review in New Democracies 73 (2003); Clifford J. Carrubba et al., Judicial Behavior Under Political Constraints: Evidence from the European Court of Justice, 71 J. Pol. 1 (2009).

57 Ginsburg, supra note 56.

58 Id. at 90–96.

59 See generally Gerald N. Rosenberg, The Hollow Hope: Can Courts Bring About Social Change (1991).

60 See Andrea Pozas-Loyo & Julio Ríos Figueroa, Enacting Constitutionalism: The Origins of Independent Judicial Institutions in Latin America, 42 Comp. Pol. 293 (2010).

61 See Bernd Hayo & Stefan Voigt, Explaining De Facto Judicial Independence, 27 Int’l Rev. L. & Econ. 269, 282–88 (2007).

62 See Erik S. Herron & Kirk A. Randazzo, The Relationship Between Independence and Judicial Review in Post-Communist Courts, 65 J. Pol. 422 (2003).

63 This figure displays zero-order correlations between Ríos-Figueroa’s measure of formal constitutional provisions designed to create judicial independence and behavioral measures of judicial independence for eighteen Latin American states between 1945 and 2005. A circle indicates the result for all states; a diamond represents the result for democracies.

64 Daniel M. Brinks, The Judicial Response to Police Killings in Latin America: Inequality and the Rule of Law (2008).

65 Andrea Pozas-Loyo & Julio Ríos-Figueroa, When and Why ‘Law’ and ‘Reality’ Coincide? De Jure and De Facto Judicial Independence in Chile and Mexico, in Judicial Independence and Accountability: A Debate 168, 178–197 (P. Sabiha Khanum ed., 2008).

66 See Gretchen Helmke & Jeffrey K. Staton, The Puzzle of Judicial Politics in Latin America: A Theory of Litigation, Judicial Decisions and Inter-Branch Conflict, in Courts in Latin America (Gretchen Helmke & Julio Ríos-Figueroa eds., forthcoming Dec. 2010).

67 Removal can be through constitutional or extra-constitutional means, of course. Yet as we look around the world, it is clear that both approaches have been used to remove particular judicial obstacles to political coalitions. See Gretchen Helmke, Courts Under Constraints: Judges, Generals, and Presidents in Argentina (2005).