WHAT IS “CONSTITUTIONAL EFFICACY”?: CONCEPTUAL OBSTACLES FOR RESEARCH ON THE EFFECTS OF CONSTITUTIONS

ABSTRACT

When and why are codified constitutions efficacious? Answering these key and apparently straightforward questions turns out to be extremely challenging. The road to responding to them is paved with conceptual, theoretical, and empirical difficulties. In this article, I make a modest, but nevertheless hopefully useful, claim: that overlooking certain conceptual difficulties is detrimental to the advancement of the theoretical and empirical agenda on constitutional efficacy. In other words, I posit that empirical and theoretical research linked to these questions can benefit from a clear conceptualization of constitutional (or more broadly formal) efficacy that is consistent with their research objectives. It is not uncommon for social and political science research in this area to overlook the question “how should constitutional efficacy be conceptualized?” A close analysis of academic sources makes it clear that even specialized literature on questions related to constitutional (or more broadly formal) efficacy have assumed conceptualizations that are theoretically problematic given their research objectives, potentially leading to theoretical inconsistencies or inaccurate empirical conclusions. To exemplify this point, I analyze the conceptualization of constitutional efficacy used in two influential political science texts: Barry Weingasts “The Political Foundations of Democracy and the Rule of Law” and Gretchen Helmke and Steven Levitskys Informal Institutions and Democracy. I argue that the conceptualizations of constitutional (or more broadly formal) efficacy used in their theoretical proposals are not adequately suited to their own research objectives, and that this conceptual misfit affects the theoretical consistency and empirical applicability of their conclusions.

RESUMEN

¿Cuándo y por qué las constituciones codificadas son eficaces? Responder estas preguntas cruciales y aparentemente directas han resultado un reto mayúsculo. El camino a su resolución está plagado de dificultades conceptuales, teóricas y metodológicas. En este artículo defiendo una tesis modesta pero, espero, útil: Ignorar ciertas dificultades conceptuales es perjudicial para el progreso de la agenda teórica y empírica sobre la eficacia constitucional. En otros términos, afirmo que la investigación teórica y empírica vinculada a estas preguntas puede beneficiarse de una conceptualización clara de eficacia constitucional (o de manera más general de eficacia formal) que sea consistente con los objetivos de su investigación. Un análisis a detalle de las fuentes académicas muestran que incluso la literatura especializada sobre cuestiones vinculadas a la eficacia constitucional han presupuesto conceptualizaciones que son teóricamente problemáticas con sus objetivos de investigación, y que ello los puede conducir a problemas de orden teórico y empírico. Para ejemplificar este punto analizo la conceptualización de eficacia constitucional utilizada en dos influyentes estudios de ciencia política: “The Political Foundations of Democracy and the Rule of Law” de Weingast e Informal Institutions and Democracy de Helmke y Levitsky. Argumento que las conceptualizaciones de eficacia constitucional (o de manera más general de eficacia formal) empleadas en sus estudios no son adecuadas para los objetivos de su investigación, lo cual genera problemas para la consistencia y aplicabilidad de sus propuestas.

Keywords

Constitutional Efficacy ; Concept Building ; Informal Institutions ; Self-enforcing Constitutions ; Weingast ; Helmke and Levitsky

Palabras clave

Eficacia Constitucional ; Conceptos ; Instituciones Informales ; Constituciones como Equilibrio ; Weingast ; Helmke ; Levitsky

I. INTRODUCTION

With the emergence of codified national constitutions, modern constitutionalism made a critical bet: “that societies of men are really capable… of establishing good government from reflection and choice [, that they are not] forever destined to depend for their political constitutions on accident and force.”¹ Prima-facie this bet has reached a consensus, prolific constitution-making processes all around the globe have marked the last two decades of the 20th Century, and the first years of the 21^st .² Latin America has been a frontrunner in this global trend with a stunning production of half of the world constitutions counting from independence until 2008.³ The stakes of this bet are substantive: constitution-making processes imply considerable social and political risks and costs.⁴ Moreover, from the late 18^th Century to our day there have been important voices willing to bet against the efficacy of codified constitutions as mechanisms of social and political change⁵ and the empirical evidence on constitutional efficacy does not support overly optimistic views.⁶ Therefore, understanding when and why codified constitutions are efficacious is not only a matter of academic interest, but also of great social and political concern.

Unfortunately, answering these key and apparently straightforward questions turns out to be extremely challenging. The road to responding to them is paved with conceptual, theoretical, and empirical difficulties.⁷ In this article, I make a modest, but nevertheless hopefully useful, claim: that overlooking certain conceptual difficulties is detrimental to the advancement of the theoretical and empirical agenda on constitutional efficacy. In other words, I posit that empirical and theoretical research linked to these questions can benefit from a clear conceptualization of constitutional (or more broadly formal or de jure ) efficacy that is consistent with their research objectives.

It is not uncommon for social and political science research in this area to overlook the question “how should constitutional efficacy be conceptualized?” On the one hand, this question may seem odd. “Constitution” and all its associated terms are recurrent elements of our public speech, and “constitutional efficacy” is not the exception. Even if considered a legitimate question, it may be thought of as an issue relevant only to the philosophical research on Law that has little relevance for social or political science research.⁸ Nevertheless, as I argue, a close analysis of academic sources makes it clear that even specialized literature on questions related to constitutional (or more broadly legal) efficacy have assumed conceptualizations that are theoretically problematic given their own research objectives, potentially leading to theoretical inconsistencies or inaccurate empirical conclusions. To exemplify this point, I analyze the conceptualization of constitutional efficacy used in two influential political science texts: Barry Weingasts “The Political Foundations of Democracy and the Rule of Law”⁹ and Gretchen Helmke and Steven Levitskys Informal Institutions and Democracy . ¹⁰ I argue that the conceptualizations of constitutional (or more broadly formal) efficacy used in their theoretical proposals are not adequately suited to their own research objectives, and that this conceptual misfit affects the theoretical consistency and empirical applicability of their conclusions. Specifically, I argue that the theoretical proposals of both texts imply a conceptualization of efficacy that I label as norm-behavior congruence , and that this conceptualization is not adequate for their aims.

The remainder of the article is divided into four sections. The first section discusses the conceptualization of constitutional efficacy as norm-behavior congruence and argues that it is not adequate for research on whether, when, and why constitutions (or more broadly formal institutions) have a causal effect on public officials behavior. In the second section, I analyze Helmke and Levitskys theoretical proposal to account for the different kinds of relations between formal and informal institutions, and how these relations affect formal efficacy. In the third section, I analyze Weingasts theoretical proposal to account for the mechanism that makes constitutions work. In the fourth section I briefly conclude.

II. EFFICACY AS NORM-BEHAVIOR CONGRUENCE

First of all, it is important to make clear that this article is only concerned with the efficacy of constitutional norms that prescribe behavior to public officials. Therefore, we will focus on an important, but limited, subset of the different types of provisions contained in contemporary codified constitutions. The efficacy of these norms is important to political science and political theory research since they are considered fundamental institutions for the realization of constitutionalism (i.e. limited but effective government).¹¹

The centrality of norm-behavior congruence for constitutional efficacy is very intuitive and is present in everyday discourse. Consider the following news report published on March 22^nd 2007 that the BBC Monitoring Kiev Unit entitled “Ukrainian mayor says top presidential official controls home region:”

The mayor of Uzhhorod, Serhiy Ratushnyak, made a resonant statement at a news conference in Kiev today. According to him, the laws and constitution do not work in the Transcarpathian Region. The region is actually controlled by the family of the head of the presidential secretariat, Viktor Baloha.

According to the mayor, the Ukrainian constitution does not work in the Transcarpathian region because the real rules of the game are different from those established in the constitution: the legal authority is impotent and the actual rules are those imposed by the powerful Baloha family. If the constitution is ineffective because political reality differs from the constitutional norms then, under the implied notion of constitutional efficacy, for a constitution to work what is legally prescribed by the constitutional text (de jure ) must correspond with the behavior that actually (de facto ) occurs.

Norm-behavior congruence can be considered either a necessary condition for a constitution to work or a necessary and sufficient condition for constitutional efficacy. ¹² Only the latter implies that observing correspondence between a politys constitutional norms and their prescribed behavior is sufficient to claim that its constitution works. This is the claim of the conceptualization of constitutional efficacy that I label “norm-behavior congruence.”

In this section I defend two theses:

Research concerned with the effects of codified constitutions on political reality must consider agreement between the constitutional norms and the prescribed behavior as a necessary condition for constitutional efficacy and

Given this research objective, norm-behavior congruence should not be considered a sufficient condition for a codified constitution to work. In particular, I argue that norm-behavior congruence as a necessary and sufficient criterion for efficacy is not satisfactory because it is too broad (i.e., under it constitutions that play no role in their polities are considered efficient).

1. Norm-behavior Congruence: a Necessary Condition

The first thesis, that norm-behavior congruence is a necessary element of a plausible conceptualization of constitutional efficacy, is hardly a controversial statement. The constitutional articles that are the center of this enquiry are those concerned with the behavior of public officials. Thus, if norm-behavior congruence is not a necessary condition of constitutional efficacy then it is possible for those articles to be fully ignored and yet to work. In other words, the negation of the first thesis implies that it is possible for norms regulating behavior to do so effectively and for such behavior to be inconsistent with them. This is a contradiction since the very meaning of “regulation” implies agreement between the prescribed behavior and the norm(s) that regulate it. Hence, by reductio ad absurdum , norm-behavior congruence is a necessary condition of any plausible conceptualization of constitutional efficacy.

Of course this does not imply that the only effects (intended or not) of codified constitutions are prescribed behaviors. For instance, it can be the case that a codified constitution is causally linked to political riots or economic growth, but the relation between constitutional norms and those effects would not constitute constitutional efficacy. In other words, constitutional efficacy must minimally involve correspondence between the norm and the prescribed behavior.

2. Norm-behavior Congruence not Sufficient for Constitutional Efficacy

Now, let me focus on the idea that agreement between the constitutional norms and the political behavior is not only necessary but also not sufficient for constitutional efficacy. In what follows, I argue that political science research should not consider that norm-behavior congruence is sufficient to assert constitutional efficacy because under it codified constitutions that have no motivational role on the behavior of public officials are considered effective. In other words, my aim is to show that for this research it makes sense to open the possibility of constitutional inefficacy even if we observe that the relevant public officials behave in accordance with what the provision in question requires.

My first argument has the same form as the classic argument presented by Schumpeter against the claim that “government approved by the people” is a satisfactory definition of “democracy.”¹³ The norm-behavior congruence criterion is too broad in exactly the same way “government approved by the people” is too broad to define democracy. Schumpeter claims that “government approved by the people” is not a satisfactory definition of democracy because “by accepting this solution we should lose the phenomenon we wish to identify: democracies would be merged in a much wider class of political arrangement which contains individuals of clearly non-democratic complexion.”¹⁴ In the same way, I argue that norm-behavior congruence is not a satisfactory criterion for constitutional efficacy since if we accept it, constitutions that work would be merged in a wider political class: that of written constitutions whose content is consistent with the political equilibrium of its polities that contains constitutions that do not matter. In other words, a satisfactory conceptualization of constitutional efficacy for empirical legal studies should not have as part of its extension constitutions that have no effect on their regulatory target.

This last argument bears the question, how can we have norm-behavior congruence without constitutional efficacy? Congruence is a state of agreement. The norm-behavior criterion of constitutional efficacy is satisfied when constitutional norms and the prescribed behavior of public officials agree independently of what is behind such an agreement. Such an agreement can be attained 1) because constitutional norms has some effect on the behavior of public officials, 2) because the behavior (or intended behavior) of public officials has some effect on the constitutional text or 3) because of another non-related cause. Notice that while in these three cases there is congruence between constitutional norms and behavior prescribed by them, only in the first case it makes sense to claim that the constitution is efficacious. In other words, only in the first case the constitutional norms motivate individuals to behave in a certain way. In what follows I discuss in detail scenarios where the criterion text-law is satisfied but the constitution has no effect on the behavior of constitutional role-holders.

3. Ex-post ad-hoc Enactment, Ex-ante ad-hoc Enactment and Parallel Norms

Ex post ad hoc Enactment

An ex post ad hoc enactment occurs when the constitutional text is made to fit an already occurring behavior. The 1980 Chilean constitution is a particularly illuminating case in this respect. This constitution has two parts: the permanent articles that provided the basic framework for a transition to civil rule that did not come into effect until 1989, and the transitory part that dealt with the institutional framework that ruled Chile until the transition. What is important for our current purposes is that, to an important extent, these articles enacted an already established institutional framework, an institutional framework that had ruled Chile from the early years of the dictatorship that had associated behaviors well established by then (the military coup took place in 1973).

By 1980 “the Junta already had agreed to its own rules… The transitory articles enacted did not significantly depart from this prior organization.”¹⁵ Thus, the fact that the constitution entered into effect in March 11, 1981 was not very noticeable: “the organization of power during the transitory period remained largely identical to the period which the regime allegedly was stepping away from.”¹⁶ Take for instance Pinochets executive role and the legislative faculties of the Junta. These roles and faculties emerged in 1973-4 to respond to specific political challenges and from power struggles within the military Junta. As Barros’ account clearly shows the constitution did not constitute the particular equilibrium linked to these institutional roles; actually the equilibrium preceded the constitution.

In sum, the transitory articles of the Chilean constitution of 1980 are a good case of ad hoc ex post enactment, and thus a case where norm-behavior congruence does not provide a sensible foundation to conceptualize constitutional efficacy. Even if there was a high degree of norm-behavior congruence relative to the Juntas legislative powers, it would be misleading to say that the constitutional provisions dealing with that power were efficacious since, arguably the behaviors associated to prescribed by those norms were originated and maintained by means independent of the constitution, that is the de facto power of the Junta members.

4. Ex-ante ad-hoc Enactment

Someone could argue, following Thomas Paines famous dictum,¹⁷ that given that the constitutional norms in question precede the relevant behavior, it would be sufficient to observe norm-behavior congruence to infer that those norms are efficacious. Thus, prima facie , we could say that the norm-behavior congruence criterion could still survive by additionally requiring that the enactment of constitutional provisions precede the prescribed behavior. Let me call this addition the precedence condition , and the criterion that incorporates it the modified norm-behavior criterion.

I believe the precedence condition misses the mark. Even incorporating the precedence requirement, the modified text-reality criterion is not sufficient to ascertain constitutional efficacy. In particular, I argue that such a criterion is still too broad since it leads us to consider efficacious cases that are not such. As I have shown, correspondence between norm and behavior can be reached through different routes. In the previous section I showed that text can be made to fit behavior. In what follows, I show that the text can also be made to fit intended behavior and that this type of fit undermines the modified norm-behavior congruence criterion for constitutional efficacy.

The constitutional norm can be made to fit an individual intended behavior when the intention to behave in a certain way shapes the enactment of the constitutional provision that is supposed to regulate the intended behavior. I call this an ex ante ad hoc enactment , since the enactment of the provision in question precedes the behavior but the content of the provision is expressly made to fit the intention to behave in that way.

Consider the following example:

Imagine a President of a country who is about to finish his term with overwhelming public support, and who heads a party with the political capacity to amend the current codified constitution (e.g. a party with a supermajority in congress). Suppose that the constitution of that country has a provision (CP₁ ) that mandates a term limit that is about to expire and prohibits presidential reelection. Now, suppose that the president intends to seek re-election, and that he knows that given his public support he could ignore the constitutional term limit without any real opposition. Suppose further that nevertheless, the President has a legalistic preference that leads him to instruct the members of his party to amend CP₁ in an ad hoc fashion. CP₁ is amended and a new constitutional provision, CP₂ , enabling indefinite reelection is enacted. If the president stays in office until he finishes his term and then seeks reelection, there would be congruence between the relevant constitutional text (CP₂ ) and the presidents behavior. However, it would be misleading to say that CP₂ was in any way causally linked to such behavior. The amendment was done only because of the Presidents legalistic preference, but if it had not been enacted the Presidents (and other relevant actors’) behavior would have been the same. Therefore, in this case too claiming that norm-behavior congruence is sufficient for a constitution to be effective implies that it is possible for a constitutional norm to work even if the behavior it prescribes has no causal relation to it.

The previous is a counterexample to the modified text-reality criterion showing that even if the precedence condition is met, norm-behavior congruence is too weak to ascertain constitutional efficacy. However, it may be argued that this counterexample does not pose a real problem to the modified criterion since in the real world ex ante ad hoc enactment does not occur. To refute this point consider the following example:

When the administration has the control of the organs required to amend the constitution it has the capacity to surpass the rigidity of codified constitutions without opposition. In such a context, ex ante ad hoc enactment is facilitated. This was the case during what Dominicans call “The Era of Trujillo,” the time during which Rafael L. Trujillo ruled the Dominican Republic (1930-1961). During those times, ex ante ad hoc enactment was not an uncommon practice. ¹⁸ Trujillo was president from 1930 to 1938 and from 1942 to 1952, but he remained “the Supreme Leader of the Dominican Party” and in fact he and his family controlled Dominican politics until his assassination in 1961. Trujillos rule was a bloody and authoritarian period in Dominican history; it was also a time marked by personality cult. However, he had a notable respect for the legal forms and constitutional technicalities that lead him on several occasions to amend the constitution for it to fit his intended actions.¹⁹ Thus, under this legalistic dictator the modified norm-behavior congruence criterion was satisfied, but one could hardly claim that the constitution governed Trujillos behavior. In this case norm-behavior congruence was achieved through the adjustment of law to intended behavior.

5. Parallel Norms

I have argued that the central problem with the norm-behavior congruence criterion is that it is satisfied whenever the constitutional text and the political reality agree independently on what is behind such an agreement. As already discussed, this agreement can be reached without any guarantee of constitutional efficacy when the constitution is made to fit behavior or intended behavior. There is a last logical possibility where the norm-behavior congruence is satisfied but constitutional efficacy is not assured: when there are what I call parallel norms .

A codified constitution is a system of norms. It is a system because its constitutional provisions are interrelated, creating a more or less consistent whole. And that system is of norms because its provisions establish constitutional roles (e.g. that of Supreme Court Justice or President) and regulate the behavior of individuals occupying those roles.

But, codified constitutions are not the only normative systems of political life. Historically, in fact, they are latecomers: they have been present in the political scene only since the late eighteenth century. Moreover, even in countries with codified constitutions, the Constitution is only one among many political normative systems that can potentially regulate interactions of individuals in constitutional roles. Constitutional conventions (non-written norms regulating relations between political parties or governmental branches)²⁰ and intra-political parties’ formal and informal norms are only two of the many normative systems in place in the political scene. Each of these normative systems establishes institutional roles and regulations linked to them. Furthermore, politics is not an isolated sphere, and normative systems are present in all areas of social life. In this way, a complex net of normative systems constitutes social and political life.²¹

Now, any given individual has a number of different roles. For instance, an individual with a constitutional role like that of “the President,” can also be member of a party, a corporations stoke holder, a friend of many, and a parent of two. And therefore, a given interaction between two individuals holding constitutional roles can be regulated by a number of different, potentially conflicting, normative systems.²² For instance, an interaction between two individuals holding the constitutional roles of “Vice-president” and “member of Congress” correspondingly could be regulated by a constitutional provision linked to those roles, by an informal corporative norm if they both are board members of a corporation, by an interpersonal norm if they happen to be friends, among many others.

Here I am interested in what I call parallel norms. This is its definition: Two norms are parallel if an individual holds two roles linked to two independent normative systems, each role belongs to one of these systems and can be satisfied by the same behavior. Note that in this case there is no behavioral conflict derived from the norms associated to two different roles. In what follows, I present an example in which parallel norms present a systematic problem to the empirical assessment of constitutional efficacy.

The PRI (Partido Revolucionario Institucional ) was the hegemonic party in Mexico from 1929 to 1989. During the PRI Era, this political party had control over the administration, the federal Congress, the states’ governments and the judiciary. The President was the head of a very well disciplined political system: he was the head of the government and the head of the PRI. He had the political capacity to violate some provisions of the 1917 Constitution without political opposition. For instance, the Constitution mandated life tenure for Supreme Court judges. However, every six years the incoming President used to appoint as much as 72% of the Court (Ruiz Cortinez, 1952-58) and no less than 36% (Lopez Mateos, 1958-64). “The president could thus somehow create vacancies to be filled by justices he appointed or, put in other terms, he could either dismiss justices or induce early retirements.” ²³ Furthermore, the PRIs supermajoritarian control also gave him the legal capacity to alter the Constitution. Every incoming President amended the Constitution to make it fit his political agenda: as much as 66 constitutional provisions were altered in the presidential term of Miguel de la Madrid Hurtado, 1982-1988.²⁴

Nevertheless, surprisingly during this president-centered era (1929-1989), Article 83 of the constitution that establishes a six-year presidential term without re-election was neither altered nor violated. In 1927, Article 83 had been amended to enable non-consecutive re-election allowing former president Álvaro Obregón to run for a second term, but in 1928 (after the assassination of president elect Obregón) the article was again amended back to its original form, and it was never again touched.

Why did presidents with extraordinary power accept to hand over political power and to retire from public life once their term was over? Arguably, the means by which Article 83 was enforced, at least during the first terms of the PRI era, were independent of constitutional prescriptions. ²⁵ During this period, Article 83 was enforced through the norms of the PRI that also enabled, and in some instances promoted, the violation of some other constitutional provisions and the ad hoc amendment of others. In other words, there was a highly efficient normative system alternative and parallel to the constitution: that of the hegemonic political party, the PRI. If this normative system could totally account for the behavior of presidents facing the end of their term, then Article 83 was ineffective.

The prescriptions of the hegemonic party system sometimes contradicted the constitutional norms, as happened with the party norm that enabled the President to dismiss Supreme Court justices or induce their early retirement. At other times, the norms of the PRI were parallel to the constitutional ones, as was the case with the prohibition of re-election. In this case, norm-behavior congruence would not be sufficient to affirm constitutional efficacy since the Presidents behavior could be fully motivated by the partys norm, the constitutional norm could then have no motivational effect, and it could not work while the norm-behavior congruence would still hold.

In conclusion, if we are interested on the effects of codified constitutions (or more broadly formal institutions) on public officials’ behavior as is most, if not all, political science research in this thematic area, then we need a conceptualization of constitutional efficacy considers norm-behavior congruence as a necessary but not sufficient condition. We need a conceptualization of constitutional efficacy where the norm not only corresponds to the behavior it prescribes, but were it has a causal relation to such a behavior. However, as I will now show, influential works on the effects of constitutions do not take into consideration such conceptual discussions and in fact adopt a criterion of norm-behavior congruence as sufficient for constitutional efficacy.

III. ON THE RELATIONS BETWEEN FORMAL AND INFORMAL INSTITUTIONS

As we discussed in the previous section, constitutions, and more generally formal institutions, are not isolated, they interact in various ways with informal institutions of all sorts, from the reciprocity rules that characterize clientelistic networks, to social norms such as foot binding. Since these informal institutions systematically motivate individual behavior as formal institutions aim to do, if we want to understand what can affect the motivational capacity of formal institutions, we need to understand the different types of relations among formal and informal institutions. In other words, informal institutions can have different kinds of relations with formal institutions, and those relations have important implications for formal efficacy in general, and therefore for constitutional efficacy in particular since codified constitutions are paradigmatic examples of formal institutions. In this section I analyze the influential typology of formal-informal institutions relations by Helmke and Levitsky,²⁶ and I argue that it implies the norm-behavior congruence conceptualization of constitutional efficacy that is not adequate for their research aim that involves understanding institutions as causes of behavior.

Helmke and Levitsky are not the only authors that have dealt with the different relations formal and informal institutions can have, but they are, without doubt, as clear and systematic as any other author. Their commitment to analytic clarity enables the discussion and the criticisms, hopefully constructive, I present here.

To begin let me provide the basic definitions of formal and informal institutions that Helmke and Levitsky give. “We define informal institutions as socially shared rules, usually unwritten that are created, communicated and enforced outside officially sanctioned channels… [and]…formal institutions are rules and procedures that are created, communicated, and enforced though channels that are widely accepted as official.” ²⁷ Here I take these definitions as given. As Helmke and Levitsky note it is important to be clear that:

• Not all informal institutions are linked with cultural or traditional practices.
• It is not the case that the formal-informal distinction coincides with the state-societal distinction (i.e. there are informal state institutions).
• It is not the case that informal rules are not externally enforced while formal rules are.
• Ineffective formal institutions do not always imply the presence of informal institutions.
• And
• Informal institutions should not be mistaken for other informal behavior not rooted on shared expectations or rule bound.²⁸

The typology Helmke and Levitsky present, is based on two dimensions:

First, the degree of convergence between the outcomes of formal and informal institutions:

“The distinction here is whether following the informal rules produces a result substantively similar to or different from that expected form a strict and exclusive adherence to the formal rule…Where following the informal rule leads to a substantively different outcome, formal and informal institutions may be said to diverge. Where the two outcomes are not substantively different, formal and informal institutions converge”.²⁹

What do the authors mean by “outcomes” or “result substantively similar or different” is not very clear. Based on the examples they provide we can conclude that they mean very broad outcomes such as political competiveness, cohesion or stability. As I argue later the lack of specificity of this dimension is problematic.

The second dimension of Helmke and Levitskys typology is the effectiveness of the relevant formal institution. They tell us: “[b]y effectiveness we mean the extent to which rules and procedures that exist on paper are enforced or complied with in practice”³⁰ It is important to note that this explicit definition of formal efficacy is not that of norm-behavior congruence since enforcement and compliance imply more than mere correspondence. That this is the explicit definition of formal efficacy makes sense given their research interests but as will become clear later, their typology does imply a conceptualization of efficacy as norm-behavior congruence, and this conceptual misfit creates theoretical problems for their proposal reducing its empirical usefulness.

As already mentioned the first dimension of the typology, (“whether following the informal rules produces a result substantively similar” to the produced by the formal one), appears to refer to the effects those institutions have vis-à-vis a substantive broad political outcome. For instance, whether it enhances political stability I believe this criterion is problematic for practical and theoretical reasons.

First, I want make two points of a practical nature. Given the first dimension of this typology, establishing what type of relation a formal and an informal institution have could be very taxing in practical terms because determining the effects of institutions is often not an easy task. In fact, an important section of the most sophisticated political science research aims to specify the effects of particular institutional arrangements, and doing so is not trivial most of the times. Furthermore, the outcomes of institutions often vary considerably depending on political or social conditions. The same institution may enhance political stability under some conditions while contribute to instability under others. The prohibition of executive re-election, is a good example of this, it arguably contributed to the political instability of Mexico in the period that immediately followed the Revolution (1917-1934), while it was arguably helpful to that effect under era of hegemonic party (1934-1997). Therefore, pining down this dimension with respect to specific formal and informal institutions in order to establish their relation will often be practically difficult. Second, institutional arrangements often have multiple effects and these are often not unidirectional vis-à-vis a substantive broad outcome as rule of law, or political stability. In these cases, it would be impossible to establish the type of relation institutions has.

Now, my main concern is with the second dimension of the taxonomy, since it is based on a definition of formal efficacy that, when applied to constitutional articles, implies the norm behavior correspondence conceptualization of efficacy which is misleading. To see why this is the case, let us give account in greater detail of the different types of relations informal and formal institutions can have according to this typology.

Complementary informal institutions “shape behavior in ways that neither violate the overarching formal rules nor produce substantively different outcomes.”³¹ According to Helmke and Levitsky the following is one type of complementary informal institutions:

“[A type of complementary informal institution] …serves as the underlying foundations for formal institutions. These informal norms create incentives to comply with formal rules that might otherwise exist merely as pieces of parchment. Thus compliance with formal rules is rooted not in the formal rules per se but rather in shred expectations created by underlying (and often preexisting) informal norms”. ³²

Two parallel norms, as described and discussed in this paper, could perfectly fit this description of an informal complementary institution of this sort vis-à-vis a formal one. As stated before, two norms are parallel if an individual holds two roles linked to two independent normative systems, each of these norms belongs to one of these systems, and both norms can be satisfied by the same behavior. So, in the case of the complementary rule above described, the informal and the formal rules would be part of different normative systems (e.g. the Constitution and the informal hegemonic party political norms), and both could be satisfied by the same behavior, actually the informal complementary one would fully motivate the behavior prescribed by the formal one. The theoretical critique to this typology is now clear: under this circumstances it would be mistaken to claim the formal institution is efficacious if we assume a notion of efficacy that is not reduced to norm-behavior congruence, instead we need to incorporate a causal... incorporate a causal link between norm and behavior.

In this connection, if an informal norm creates “incentives to comply with formal rules that might otherwise exist merely as pieces of parchment presence” it is a misattribution to claim that the formal institution is efficacious (in the stronger sense implied in terms such as “enforced” or “complied”) since the informal norm is fully responsible of producing the prescribed behavior. Helmke and Leviskys theoretical framework is problematic since under it these would be complementary norms and, as the Table 1 shows, the formal parallel norm is claimed to be efficacious in these cases. The function of norms is to motivate specific behaviors. Hence, if they totally fail to do so, it is problematic to claim that they are efficacious in the strong sense, even if the behavior in question happens to be produced by another norm. In sum, claiming that in these cases the formal rule works even if it has no role what so ever in the producing the behavior, makes explicit a implicit assumption of this typology: under it observing the behavior is sufficient for to consider efficacious a formal rule that prescribes such behavior. In other words, it assumes norm-behavior congruence conceptualization of efficacy.

TABLE 1. Helmke and Levitsky typology.

Outcomes/Effectiveness	Effective Formal Institutions	Ineffective Formal Institutions
Convergent	Complementary	Substitutive
Divergent	Accomodating	Competing

Now, notice the implication of the previous argument: if what I have argued is correct in these cases the formal institution should be considered inefficacious and hence it is problematic to claim that the informal parallel norm is complementary to the formal one, what it actually does is to substitute the formal rule, since it plays the role the formal one ought to play (i.e. leading to the behavior it prescribes).

What if both the formal and the informal parallel norms are efficacious? Would they then be complementary? According to the Oxford Dictionary “complementary” means: “[c]ombining in such a way as to enhance or emphasize the qualities of each other or another” ³³ Hence, saying that two things complement each other implies that, they have an effect on their qualities : it implies certain type of interaction. If a formal and an informal rule prescribe the same behavior and both work, the behavior is over-determined: any of the two norms would be sufficient to motivate it. But in these circumstances the efficacy of one has no impact on the efficacy of the other. They are totally independent with respect to their efficacy. For this reason, I believe it would not be advisable to claim that these norms complementary . I think that the best way to characterize their relation is by saying that they are parallel vis-à-vis their efficacy.

Now to complete the discussion on the Helmke and Levitskys typology let me briefly characterize the other types of informal institutions vis-à-vis their relations with formal institutions. An informal institution is accommodating vis-à-vis a formal institution, if the later is effective and they have divergent outcomes. These informal institutions do not directly violate their formal counterparts: “they contradict the spirit, but not the letter of the formal rules.” ³⁴ The substantive outcomes of these rules are incompatible. Competing informal institutions combine ineffective formal institutions and divergent outcomes. These informal institutions “trump their formal counterparts, generating outcomes that diverge markedly form what is expected form the formal rules.”³⁵

The category of competing informal institutions creates another theoretical problem. It assumes that what makes inefficacious the formal norm is the informal norm. But it is possible for a formal institution to be inefficacious and to have divergent outcomes with an informal institution, and for the informal institution to have nothing to do with the formal inefficacy in question. These cases have no place under Helmke and Leviskys typology. Finally, substitutive informal institutions “…combine ineffective formal institutions and compatible outcomes.”³⁶

I have so far presented an account of Helmke and Leviskys typology, I have argued that it implies the conceptualization of formal efficacy (and for implication of constitutional efficacy) as norm-behavior convergence, that this conceptualization is not adequate for their research objectives, and that it differs with the conceptualization they explicitly offer. I have further pointed to some theoretical problems their typology has as a result of their conceptualization of constitutional efficacy.

Finally, it is important to note that using this typology to account for the relations between formal and informal institutions and their implications for formal efficacy would lead us, in certain cases, to problematic empirical conclusions. For instance, take the example of Article 83 of the Mexican Constitution in the PRI era discussed in the previous section. Under Helmke and Levitskys typology during the presidential succession processes that followed the consolidation of the hegemonic party (at least until the election of Ruiz Cortines in 1952) Article 83 and the PRI non-reelection informal norms would be considered “complementary.” They would be complementary because arguably, the PRI norms “…server[ed] as the underlying foundations for formal institutions. These informal norms create[d] incentives to comply with formal rules that might otherwise exist merely as pieces of parchment… compliance with formal rules [was not] rooted not in the formal rules perse but rather in shred expectations created by underlying informal norms.” ³⁷ But, if this was in fact the case, in spite of its complete lack of motivational effect on presidential no-reelection, this taxonomy would lead us to conclude that Article 83 was efficacious at that time. If what I have argued is correct the misfit of the conceptualization of efficacy and Helmke and Levitskys research objectives is problematic both for theoretical and for empirical reasons.

IV. EXOGENOUS FOUNDATIONS OF CONSTITUTIONAL EFFICACY?

Constitutional efficacy as an equilibrium is arguably the most common criterion in the literature of political science. While several authors describe constitutions as equilibria, it is clear that the claim is not that all codified constitutions constitute equilibria, but that all codified constitutions that work are equilibria. Thus, for instance, I believe that Russell Hardins account of constitutions as coordination devices is not that all codified constitutions can be considered as such, but that all constitutions that work do coordinate. ³⁸

A constitution depicts an equilibrium if and only if actors behave in accordance with the constitutional text and they individually have nothing to gain by changing his or her own strategy unilaterally. A constitution that depicts an equilibrium is often characterized as a self-enforcing constitution. In other words, by using the game theoretic notion of equilibrium political scientists have aimed to give an account of the mechanism through with constitutions become efficacious. Now, an account of how codified constitutions become efficacious necessarily implies a conceptualization of constitutional efficacy (in a more general way, you can not account how x becomes an xa without implying an idea of what it means to become a ). The conceptualization constitutional efficacy implied in the account of constitutional efficacy as equilibrium is that of norm-behavior congruence.

As the reader most probably can see by now, the problem with constitutional efficacy as equilibrium is that it tells us nothing of what maintains such an equilibrium. In particular, it can perfectly well be the case that what maintains the correspondence between the constitutional norm and the actors behavior bears no relation to the constitution itself. If this were the case, a constitutional norm that has no effect on the relevant behavior would be considered effective. Hence, given that political science research in this area has as one of its objectives to understand the effects of institutions on behavior, the account of constitutional efficacy as equilibrium is too broad: for constitutional efficacy to make sense endogenous motivations must play a role in the maintenance of the equilibrium. Therefore, if the aim is to account for how codified constitutions can motivate behavior the equilibrium account of constitutional efficacy should make limit themselves to a subset of equilibria, those where the equilibrium is not exclusively maintained by exogenous controls.

To clarify this point consider an account of constitutional efficacy as equilibrium that only incorporates exogenous controls: the one presented in Barry Weingasts very influential article “The Political Foundations of Democracy and the Rule of Law”. Weingasts central question is: “How are democracys limits enforced?” His aim is to give “a unified approach to the political foundations of limited government, democracy and the rule of law- phenomena requiring that political officials respect limits on their own behavior”.³⁹ Political officials respect the limits of their behavior if and only if those limits are self-enforcing. His approach is modeled by a game of the stability of limited government that focuses on the relation between a single political official, called the sovereign, and the citizenry.

To stay in power the sovereign requires sufficient support from the citizens, and each individual supports the sovereign as long as he does not transgress what the citizen believes are her rights.⁴⁰ Different citizens have different “preferences and values” and therefore, different conceptions of what her rights are.⁴¹ So accordingly constitutions are devices that coordinate the citizens on what constitutes a violation of rights so that they can collectively react to transgressions by withdrawing their support from the sovereign. If the constitution is effective, that is if citizens are coordinated on its content, the sovereign will avoid any behavior that violates the constitution because by doing so he risks losing power. Notice that in this model the converse relation also holds: if the sovereign acts in accordance with the constitution, the constitution is efficacious. Therefore, clearly this particular model falls under the conceptualization of constitutional efficacy as norm-behavior congruence: norm-behavior congruence is necessary and sufficient for constitutional efficacy.

Furthermore, in the model the controls are exogenous to the constitution. Weingast claims that whether or not a constitution coordinates individuals on its content is a function of the social consensus of the rights of citizens and the limits of the state.

“In terms of the model, limits become self-enforcing when citizens hold these limits in high enough esteem that they are willing to defend them by withdrawing support from the sovereign when he attempts to violate these limits. To survive a constitution must have more than philosophical or logical appeal; citizens must be willing to defend it”.⁴²

“Because citizens have different views about ideal limits, a unique set of ideal limits is unlikely. Coordination requires that citizens compromise their ideal limit…When the difference between each citizens ideal and the compromise is small relative to the cost of transgression, the compromise makes the citizens better off”.⁴³

Thus according to this account whether or not a constitution works depends on the presence of a common set of citizen attitudes that are totally exogenous to the constitution and its incentives. What maintains the equilibrium of efficacy has therefore nothing to do with the codified constitution and its design. To clarify this point further Weingasts account of why Latin American constitutions “have not worked” while the American has is particularly helpful:

“[Latin American constitutions “have not worked” because] Latin American states are not characterized by a common set of citizen attitudes about the appropriate role of government… [While] citizen reaction implies that US constitutional restrictions on officials are self-enforcing …Latin America states exhibit a complementary set of phenomena: citizens unwilling to defend the constitution, unstable democracy and episodic support for coups”.⁴⁴

In sum, there is a theoretical and conceptual tension between the researchs aim, accounting for how constitutional norms can systematically cause the behavior that characterizes limited governments and the rule of law (how they become efficacious), and the account of constitutional efficacy as an equilibrium maintained only by exogenous controls (citizens’ attitudes) that it offers.

To close let me point out that this lack of adequacy may be the result of conflating two different understandings of what it means to say “constitutions are coordination devices.” As Hardin argues:

“In claiming that a particular constitution is a device for coordination we could be making two quite different claims: that the choice of the content of the constitution was itself a matter of coordination or that the constitution works by successfully coordinating actions under it”.⁴⁵

Claiming that the content of a particular constitution coordinated the most important sectors of a society may be given as an account of a successful constitution-making process and as an explanation of why the content of a particular constitution is such. So following Hardins account, we can claim that the American-constitution making processes coordinated the most important economic interests and, we may add, following Weingast, also the most important attitudes about the appropriate role of government (i.e. that those interests and attitudes were coordinated on the content of the constitution).

Now when we claim that a constitution that works is a coordination device, we are claiming that actions are successfully coordinated under it; i.e. that the behavior that is its regulative target is attained thanks to the incentives the constitution gives to the relevant individuals. That public actors act according to the constitution as a result of their pursuit of individual benefits under constitutional laws.

The need of separating these two senses in which a constitution is a coordination device follows from the recognition that an account of modern constitutional government requires a two-stage theory.⁴⁶ Success in coordinating on a particular constitutional content does not guarantee that the individuals who populate the institutions created by the constitution will coordinate under it. In other words, a successful constitution-making process is not sufficient for constitutional efficacy. This is the case because constitutions create and distribute power in ways that are not predictable ex ante . Moreover, constitutions are complex systems that often have unintended effects. Therefore, what enables coordination on a particular constitutional content and what enables coordination under that constitution require separate accounts. In particular, while the former necessarily deals only with interests and other motivations exogenous to the constitution, the latter requires the incorporation of motivations endogenous to it.

If I am right, the source of Weingasts problematic account of constitutional efficacy lies in his conflation of the two stages required by a satisfactory theory of modern constitutional government. He conflates the determinants of coordination on constitutional content with the determinants of coordination under a modern constitutional order. Weingasts account is then an instance in which I claim we can have norm-behavior congruence without efficacy in the strong sense linked to research on how codified constitutions can motivate politicians to behave in ways consistent with the ideals of constitutionalism and the rule of law. It also exemplifies why I claim constitutional efficacy as equilibrium implies a conceptualization of constitutional efficacy that is too broad for this research agenda, and thus why we would need to make sure to incorporate controls endogenous to the constitution if we what to purse this research aims.

V. CONCLUSION

It is not uncommon for social science research to overlook the importance of the conceptualizations they use in their proposals and how adequate they are given their research objectives. In this paper, I have defended a modest, but nevertheless hopefully useful, claim: that overlooking certain conceptual difficulties is detrimental to the advancement of the theoretical and empirical agenda on constitutional efficacy. I have argued that It is not uncommon for social and political science research in this area to overlook the question “how should constitutional efficacy be conceptualized?” A close analysis of academic sources makes it clear that even specialized literature on questions related to constitutional (or more broadly formal) efficacy have assumed conceptualizations that are theoretically problematic given their research objectives, potentially leading to theoretical inconsistencies or inaccurate empirical conclusions. To exemplify this point, I analyzed the conceptualization of constitutional efficacy used by Barry Weingasts “The Political Foundations of Democracy and the Rule of Law” and Gretchen Helmke and Steven Levitskys Informal Institutions and Democracy . I showed that conceptualizations of constitutional (or more broadly formal) efficacy used in their theoretical proposals are not adequately suited to their own research objectives, and that this conceptual misfit affects the theoretical consistency and empirical applicability of their conclusions.

I believe that the lack of adequacy between the conceptualization of constitutional efficacy used these texts and their research objectives has the following cause: on the one hand we what to understand how and when institutions in general, and codified constitutions in particular, motivate public officials to behave in ways that embody the political ideals of constitutionalism and the rule of law. These research objectives do require a conceptualization of constitutional (or formal) efficacy that is demanding, where norm-behavior congruence is a necessary but not sufficient condition. This adequate conceptualization requires theoretical accounts and empirical methodologies that are less simple and elegant than those in which norm-behavior congruence is both necessary and sufficient. Unfortunately, as I have shown in this paper, though simpler and more elegant this conceptualization is not well suited for such research puropses. In this area, it seems to be true the dictum that says that simplicity is inversely proportional to relevance.⁴⁷

Notes

1. The Federalist No 1 (Hamilton).

2. Half of the worlds constitutions were written or rewritten between 1978 and 2003: Vivien Hart, Democratic Constitution Making, Special Report 107 (2003).

3. See: Jose Luis Cordeiro , Constitutions Around the World: A View From Latin America”, Institute of Developing Economies, Discussion Paper # 164, (2008) available at'http://www.ide.go.jp/English/Publish/Download/Dp/164.html .

4. David Landau , Constitution-Making Gone Wrong , 62 Alabama L. Rev. 923,938 (2013).

5. See for instance: Ferdinand, Lassalle . On the Essence of Constitutions in 3(1) Fourth International . 25,31 (1942).

6. See for instance: Cliford Carrubaet al . When Parchment Barriers Matter: De jure judicial independence and the concentration of power (unpublished manuscript available at: http://polisci.emory.edu/faculty/jkstato/resources/WorkingPapers/Parchment.pdf ).

7. SeeDavid Law , Constitutions , in The Oxford Handbook Of Empirical Legal Research (Peter Cane & Herbert M. Kritzer eds., 2010).

8. Paradoxically, as Pablo Navarro argues, many philosophers of Law have not been interested on the conceptual analysis of legal efficacy on the grounds that that the analysis of legal efficacy is an issue that concerns Sociology of Law and not Jurisprudence see: Pablo E. Navarro, La Eficacia del Derecho 20 (Centro de Estudios Constitucionales, 1990).

9. Barry Weingast, The Political Foundations of Democracy and the Rule of Law, vol. 91, no. 2, The Political Science Review, 245 (1997).

10. Gretchen Helmke and Steven Levitsky, Introduction in Informal Institutions and Democracy (John Hopkins 2006).

11. Stephen Holmes, Constituionalism in The Encyclopedia of Democracy , (Congressional Quarterly ed. 1995). In what follows “constitutional efficacy” and related terms will refer only to the efficacy of these norms.

12. Note that “legal norm” and “behavior” can be conceptualized in very different ways. How to conceptualize them is a mayor concern of jurisprudence and of philosophy of action correspondently, nevertheless for the purposes of this paper is not necessary to take a position of those debates. For a classical analysis see: Joseph Raz, the Concept of the Legal System (Clarendon Press, 1971).

13. Joseph A. Schumpeter, Capitalism, Socialism and Democracy (Rontledge, 1976).

14. Id. at 247.

15. Robert Barros, Constitutionalism and Dictatorship (Cambridge University Press, 2002).

16. Id . at 179.

17. SeeThomas Paine, Common Sense 59 (1751).

18. Jacobo Espinal, Constitutionalism and Democracy in the Dominican Republic (University of Virginia Press, 1997).

19. Idem .

20. J. Jaconelli, The nature of constitutional convention, vol. 19, no. 1, Legal Studies (1999); J. Jaconelli, Do constitutional convetions bind?, vol. 64, no. 1, The Cambridge Law Journal (2005).

21. John R. Searle Making The Social World (Oxford University Press, 2010).

22. Robert Merton, Social Theory and Social Structure (Simon and Schuter, 1968).

23. Beatriz Magaloni , Authoritarianism, Democracy and the Supreme Court: Horizontal Exchange and the Rule of law in Mexico , in Democratic Accountability in Latin America 228-289 (Scott Mainwaring & Christopher Welna eds., 2003)

24. Francisco Valdés Ugalde, La Regla Ausente. Democracia y Conflicto Constitucional [Gedisa-IIS-UNAM] (2010).

25. Magaloni , supra note 23.

26. Helmke & Levitsky, supra note 10.

27. Id . at 5.

28. Id . at 5-8.

29. Id . at 13.

30. Id . at 13.

31. Id. at 13.

32. Helmke and Levitsky, 2006, 14.

33. Oxford Dictionary available at http://www.oxforddictionaries.com/definition/english/complementary?q=complementary (last visited april, 8^th , 2016).

34. Helmke & Levitsky,supra note 10 at 15.

35. Id. at 15.

36. Id . at 16.

37. Id . at 16.

38. In this connection, some criticisms to this theory would be somewhat off the mark. For instance, showing that for most Latin American constitutions the probability of replacement increases as time goes by would not falsify Hardins theory for the Latin-American region (see Gabriel Negretto, Shifting Constitutional designs in Latin America: A Two-Level Explanation, 89 Texas Law Review 1777-1805 (2010)), since we would expect re-coordination costs to decrease the probability of replacement only for constitutions that in fact coordinate. In other words, assuming the interpretation I propose, if it is the case that most Latin American constitutions have a very low degree of efficacy, the empirical implications of the theory would not be in conflict with such empirical findings. Hardin acknowledges that: “Many actual constitutions do have the character of contracts at their core. They cover the agreed resolution of a bargaining process in which interests are compromised. Unfortunately, constitutions that include contracts at their core are typically unstable… If a constitution is to be stable, it must be self-enforcing:” Russell Hardin, Liberalism, Constitutionalism, and Democracy 98 (Oxford University Press, 1998).

39. Weingast, supra note 9.

40. Id. at 246.

41. Id. at 246 y 247.

42. Id. at 251.

43. Id. at 252.

44. Id. at 54.

45. Hardin , supra note 37, at 103.

46. See id. at 83.

47. Navarro , supra note 8. Chap. 1.