This essay discusses the use of culture in comparative constitutional amendment theory. In particular it assesses the use of culture in normative-culturalist as opposed to empirical approaches of constitutional amendment. It discusses the notion of ‘amendment culture’ put forth by Ginsburg and Melton, delineating how this represents a shift compared to their earlier work (Elkins, Ginsburg and Melton). Moreover, it probes their idea of a measurable amendment culture in light of the recent work by L. Epstein and A.D. Martin on empirical methodology on one hand, and as regards our own approach to constitutional change through amendment models on the other. This essay examines these approaches to explore their possible contribution to the design of formal amendment rules.
What matters in constitutional change?
When we talk about constitutional amendment probably what matters most is our motivation for doing so. Writing constitutional theory is different from making choices among constitutional design options. In making constitutional design choices the crucial dilemma is whether we should focus on the impact of constitutional culture and history on constitutional amendment or rather stick to empirical, metrics-based approaches. Could those two paths be reconcilable and complementary?
Regardless of the questionable status of certain results, empirical studies produce interesting data and often impose the use of novel lines of reasoning in constitutional thinking. An example of this is Elkins, Ginsburg and Melton’s work on constitutional endurance.Footnote 1 Their approach to the prerequisites and consequences of constitutional endurance introduces important new questions on the relationship of constitutional identity and the passage of time, and has become a point of reference on the topic of constitutional change. Nonetheless, it has also been severely criticised by Epstein and Martin due to problems in their use of empirical methodology.Footnote 2 In their recent work Ginsburg and Melton address such methodological problems by introducing the concept of amendment culture.Footnote 3 We shall attempt to assess what this concept contributes to their previous work and explore whether it creates a bridge between quantitative empirical analysis and conceptual analysis in the comparative study of constitutional amendment.
The study of constitutional amendment is an area of constitutional law frequently subject to quantitative empirical methods, those typically used when investigating the ‘small c’ constitution.Footnote 4 It could, however, be argued that the study of constitutional amendment includes both the formal constitution and its informal counterpart, that is, the capital C as well as the small c constitution. The reason for this lies in the elusive correlation between formal and informal change, which in itself has posed a conundrum to scholars seeking to establish a causal relationship between the two. It is far from clear whether it is true that the more difficult it is to formally amend a constitution, the more frequently informal change will take place. Empiricists studying constitutional amendment are not in agreement as to the methods to be used and consequently on the validity of their results. Such disagreements are often triggered because the ways in which constitutional culture influences amendment are ignored.
By contrast, comparatists have begun to come to terms with the situation whereby constitutional culture largely defines the way institutions work and have accepted that they can achieve better results by embracing cultural differences rather than ignoring them. The universalist-culturalist divide between the comparative approaches illustrates how opposite theories have become reconciled over time: despite incompatibilities a certain degree of borrowing does take place. Yet culture still matters: understanding constitutional identity has the ability to inform rather than fetter comparative methodology.Footnote 5
Quantitative empirical research intensifies the difficulties inherent to comparison. Comparison entails taking into consideration factors that cannot always be quantified, and the possibility that important variables are omitted is high. The recent creation of databases collating information on constitutions encourages the use of data merely ‘because it is there’, while all too frequently ignoring comparability issues. When doctrinal constitutional analysis makes use of comparative methodology, comparability issues are bound to emerge and be addressed, whereas those issues can easily be overlooked or consciously ignored when applying quantitative empirical approaches. However, empirical research performed at a cross-national level can only produce useful results if it adopts and further elaborates on comparative methodology.
This adoption and elaboration raises many questions. Is constitutional culture measurable and topic-specific? Are there micro-cultures and subcultures within a given constitutional culture, such as an amendment culture, a constitutional review culture, an electoral culture, a constitutional interpretation culture, etc., that are clearly distinguishable? According to Jack Balkin, a constitutional culture is dynamic, features many subcultures, and is capable of reaching a new state of equilibrium, and of developing a ‘new constitutional common sense’.Footnote 6 Can culture be seen as a variable which helps explain constitutional amendment? Does the use of culture as a variable imply or presuppose that culture can be measured? Would accepting the non-measurability of constitutional culture necessarily negate the value of quantitative empirical research on constitutional amendment?Footnote 7Can comparative and quantitative empirical research work in tandem to produce results?
In consideration of these questions, we shall begin by juxtaposing the culturalist approach and the quantitative empirical approach to the study of comparative constitutional amendment. We shall then focus on the degrees of hindrance to constitutional change, and discuss the sources of those difficulties. The differences between viewing constitutional amendment through the prisms of metrics and of culture are pointed out, this in light of a recent attempt to operationalise culture. Finally, the essay analyses opposing views as to what is, or can be brought, under the control of designers. We conclude that it is possible for constitutional designers to perform an ex ante impact assessment by taking into consideration a list of parameters that influence constitutional change, including both measurable and non-measurable determinants.
Hindrances and barriers to constitutional amendment
Varying approaches are used to address constitutional change in the field of constitutional scholarship, some based on conceptualisation,Footnote 8 others on metrics.Footnote 9 Much of the literature employs the expression ‘constitutions change’, rather than stating outright that ‘constitutions are changed’, since, once enacted, constitutions convey the impression that they have acquired a life of their own.
In our 2013 work on amendment models we aimed to build models of constitutional change based on the detection of constitutional amendment patterns.Footnote 10 The level of abstraction necessary to build amendment models does not preclude degrees of differentiation between countries that share similar characteristics. Differentiation is bound to exist due to the multiplicity of parameters taken into account. Modelling does not ignore or play down these differences. On the contrary, by examining amendment models, differences are explored and brought to light. That something can be understood through aggregation and differentiation does not, however, necessarily mean that it is measurable or can provide a firm basis for predictions. Comprehensible does not mean measurable. The question raised is whether reliable conclusions may be drawn by ignoring the non-measurable, or by attempting to force measurability on non-measurable concepts.
By contrast, quantitative empirical approaches aim for an understanding of constitutional change through the use of metrics. Recently, Tom Ginsburg and John Melton, the foremost advocates of a quantitative empirical analysis of constitutional change, have revised their amendment metrics methodology. This is of particular importance, because criticisms of empiricism, made by sworn culturalists, may employ powerful rhetoric, yet lack deep knowledge of that enterprise, which is a culture unto itself. Candidly and methodically, Ginsburg and Melton attempt to ‘explain why several proposed solutions’, including their ‘own in earlier work’, have ‘flaws’. Ginsburg and Melton accept that the various measurements of ‘amendment difficulty’ offered in literature ‘are poorly correlated, suggesting potential validity problems’. In their analysis, ‘[t]he powerful force of institutional incentives can always be overcome by political or social factors that drive behavior. […] [A]mendment difficulty is subject to such extra-institutional forces’.Footnote 11
What is thought-provoking in Ginsburg and Melton’s analysis is that, at first glance, they appear to adopt a line of reasoning which explains constitutional change by understanding the context in which it takes place. They thus seem to come closer to our own approach to constitutional change, which is founded upon a matrix of interacting parameters, in the belief that ‘the way formal change works in practice depends on a country’s legal culture, party system, judicial review system, constitutional ethos, etc’.Footnote 12 This approach has allowed us to categorise constitutional change into distinct amendment models: the elastic model; the evolutionary model; the pragmatic model; the distrust model; and the direct democratic model.Footnote 13 By contrast, Ginsburg and Melton attempt to turn all the dynamic interacting factors into one single indicator, which they have named ‘amendment culture’. They thus try to reconcile quantitative empiricism with the non-measurability and multifaceted nature of constitutional culture.
An important point of view that Ginsburg and Melton seem to have in common with our context-based, culturalist analysis of constitutional change is that they acknowledge that ‘barriers to amendment are not merely institutional’.Footnote 14 In their book The Endurance of National Constitutions, Elkins, Ginsburg and Melton, while distinguishing design and environmental factors, were optimistic about the ability of designers to affect constitutional durability.Footnote 15 Ginsburg and Melton now question how much the amendment procedure matters. Do they now also accept that ‘the amending formula may not be the only determining factor with regard to the level of difficulty [of] formal constitutional change.’?Footnote 16 This is a crucial distinction. US scholars came to terms long ago with the idea that formal amendment rules were not, to say the least, the only determining factor with regard to constitutional change. David Strauss’ aphorism on ‘the irrelevance of constitutional amendments’Footnote 17 in the US vividly encapsulated the importance of informal change in the US. Ginsburg and Melton acknowledge just how context-sensitive amendment formulas are. Does coming to terms with the idea that ‘ultimately, the normative content and operation of amending formulas turn out to be highly versatile, unpredictable, and interacting with parameters that the constitutional legislator had failed to afford their due consideration’Footnote 18 have an impact on how we view the possibility of rational constitutional design?
Although Elkins, Ginsburg and Melton placed more weight on constitutional design in 2009, in 2014 Ginsburg and Melton clearly revisit this approach by questioning whether the amending formula matters after all. Yet, even while discerning environmental factors, they appear reluctant to recognise the non-quantitative empirical parameters of what is considered to be the environment within which constitutions operate. Historical legacy is linked to an enduring constitution, but this is viewed through the prism of metrics. They find that the elapse of each decade between the previous and the current constitution reduces the likelihood of replacement by 12 per cent. Here the amendment rate is taken to reflect history, however it is unclear whether such a measurement suffices per se to reflect historical and constitutional tradition.
The landscape is obscured even more by the fact that qualitative differences exist between amendments: some may be of great importance while others are trivial, some extensive and some limited. What is more, the importance of each constitutional amendment can only be ascertained in relationship to the political context that produced it. It is also important to consider that what is being explored is the life-expectancy of constitutions, regardless of other characteristics of this life, including its quality. As Hutchinson and Colon-Rios observe, a long life is neither the sole, nor the best, measure of a good life, asserting that the answer to such questions can only be normative, and not empirical.Footnote 19 Just as having a good life connotes something quite different from having a long life, amendment difficulty connotes different things, dependent on varying characteristics.
To address the complexity of constitutional change we disassembled ‘rigidity’, suggesting that multiple ‘rigidities’ exist. Institutional rigidities are institutionally embedded ways of increasing the level of difficulty in enacting constitutional change.Footnote 20 The amending formula is the most obvious among such rigidities, while the system of government and the judicial review system, with regard primarily to constitutional review, also fall into this category. Factual rigidities are sources of impediments to constitutional change situated in the practices, attitudes and behaviour patterns of different actors, which emerge through the application of institutional requirements, or address areas that fall outside the scope of institutional regulation, or stem directly from the political, legal or social culture and constitutional ethos.
To stress the importance of culture as a constraint on constitutional amendment we looked to the UK as an example of culture-driven rigidity. Hindrances to amendment exist even in the absence of formal amendment rules. We argued that ‘culture and practice-driven rigidity appears in the UK, [which demonstrates] that political self-restraint may fetter change even in the absence of an entrenched written document’.Footnote 21 In the UK, rigidity stems from the constitutional culture, which obliges actors to show self-restraint in handling constitutional matters. Ginsburg and Melton also use the UK as an example to demonstrate that ‘barriers to amendment are not merely institutional’ and that ‘political barriers to changing rules are the source of stability, and these political barriers function so well that additional institutional protections are unnecessary.’Footnote 22
The problem with culture: pitfalls and misunderstandings
Understanding the link between amendment formulas and their context
Ginsburg and Melton assert that ‘cultural explanations have been out of fashion in the social sciences for some time’,Footnote 23 which can be explained by the fact that ‘culture is difficult to measure’. Nevertheless, by articulating this indicator of amendment culture, they remedy the flaws of their earlier work and allegedly add a new parameter to constitutional scholarship.
Yet the claim that cultural explanations have fallen out of favour is rather unfair to the bulk of constitutional law literature, which strives to approach constitutional change by endeavouring to grasp the zeitgeist, and to understand the impact of constitutional culture on the way constitutions operate and change. As has been stressed by Michel Rosenfeld, who elaborated on constitutional models and models of constitution-making from the point of view of the identity of the constitutional subject: ‘Just as each constitution is unique in the way it produces constitutional identity, the making of a constitution is a singular historical event.Footnote 24 […] So is the structuring of each constitution and its relationship with its socio-political environment’.Footnote 25 It therefore makes sense that attempts to identify how constitutions change tend to focus on this relationship. Notably, Elkins, Ginsburg and Melton, in their influential work on constitutional endurance, also structured their account of constitutional change by emphasising ‘the role of politics in constitutional formation and maintenance’.Footnote 26
Nonetheless, it is true that ever since Lutz employed empirical methodology in order to analyse comparative constitutional amendment, this particular area of constitutional law research has increasingly been subjected to metrics.Footnote 27 Still, constitutional law remains an essentially normative venture, focused on interpretation and abstract conceptualisation, while paying due to cultural matters. Concepts that are difficult to measure do not present much of a problem to the average constitutional scholar. Rather the opposite is true, as constitutional scholars are used to the incommensurablity of values, liberties, principles, etc. Once it has been accepted that formal amendment rules must be read in the context in which they operate in order to grasp the way they interact with other institutional and factual hindrances to constitutional change, the problem becomes how to produce a coherent image.
Ginsburg and Melton have put forth the notion that amendment culture, which they use as a single indicator, better explains patterns of amendment, than do the pre-existing variables and indices. They discern the idea of ‘an amendment culture at the level of a constitutional system’.Footnote 28 They describe amendment culture as:
the set of shared attitudes about the desirability of amendment, independent of the substantive issue under consideration and the degree of pressure for change. In other words, we posit a baseline level of resistance to formal constitutional change in any particular system; as this baseline level increases, the viscosity of the constitutional amendment process decreases even under identical institutional arrangements. Footnote 29
The realisation that ‘something deeper’ is missing from an understanding of constitutional amendment if one focuses solely on institutions, led Ginsburg and Melton to seek new determinants. However, they appear to be in search of something as straightforward and tangible as institutions and formal rules, fitting in nicely with their metrical framework for measuring amendment difficulty.
In our work on amendment models we have stressed that the taxonomy could not be based on the hurdles set by the amending formula (since formulas might be identical, yet countries be placed within different amendment models), nor on quantitative data analysis, but should rather be based on correlations between amending processes, the political system, the constitutional ethos, and the legal culture.Footnote 30 It is clear that formal amendment rules do not sufficiently account for how constitutions change. Similarly, Ginsburg and Melton accept that ‘two countries with the same amendment procedure, but with different political configurations’Footnote 31 are likely to have different outcomes with regard to constitutional change. That may happen, as they explain, because one may have ‘a dominant political party that wants to change the constitution and is able to do so regularly, while the other features an array of small parties, most of which oppose constitutional amendment’.Footnote 32
However, the impression that Ginsburg and Melton have finally come to terms with the notion that studying amending formulas without reference to their context cannot produce results, at least not reliable ones, is misleading. They cite the example of the Greek amending formula to discuss the impact of lengthy procedures on constitutional amendment practices, but without reference being made to its working environment, and here they get it wrong. They point to one of the procedural prerequisites of Article 110 of the Greek Constitution,Footnote 33 which dictates that constitutional revision take place in two parliamentary phases, between which general elections are held. In the Greek experience, however, the intervening elections, while prolonging the process, hardly influence ‘public opinion towards the amendment,’ which, as they suggest, ‘might change over time’Footnote 34. This view is based on the possibility that voters may change their opinion on a proposed amendment over the course of time, or be influenced by successful opposition to the amendment.
Greek reality is quite different. The prerequisite of popular elections was designed to enhance the role of the people in constitutional revision, but had totally different and unintended consequences. Normally, constitutional amendment issues do not play a role in Greek general elections. The Greeks vote for the political party whose general agenda conforms most to their political convictions, or simply for their favourite candidate. Within a polarised political system, distrust among political actors is the major obstacle to achieving flexibility through consent, while polarisation does not leave room for voters to reflect on constitutional amendment issues. The distrust culture explains the divide between the rationale of the Greek amending formula, and its unintended consequences. The requirement of intervening general elections indeed lengthens the procedure. However, to understand what truly affects the procedure one must view this hurdle within the context of its execution, and in conjunction with other procedural limits, such as a five-year mandatory time lapse between revisions, and an enhanced majorities requirement.Footnote 35
Intended and unintended consequences of amending formulas
Ginsburg and Melton’s acknowledgement that ‘the formal amendment rule may, in the end, not matter at all, or at least may not matter in predictable ways across countries’Footnote 36 confirms the importance of the divide between intended and unintended consequences. The previously-cited example demonstrates how amending formulas can produce unintended results. In our view, the normative content and operation of formal amendment rules can be ‘highly versatile, unpredictable, and interacting with parameters that the constitutional legislator had failed to afford their due consideration’.Footnote 37 Can this unpredictability be addressed with the ‘amendment culture’ indicator?
Ginsburg and Melton provide a second definition of amendment culture in the same essay, referring ‘not only [to] institutional factors, or the baseline pressures caused by political and social change, but also [to the] different weights ascribed to the constitution itself’,Footnote 38 reflected in the constitutional amendment rate in a particular country. Focus on culture raises the question as to whether treating the constitution as a sacred text influences constitutional change. Ginsburg and Melton note the distinction between perceiving the constitution as a sacred document, or as a document of little normative significance. In analysing the effect of the passage of time on constitutions, we focused on the issue of ‘constitutional sacredness’, and discussed how eternity clauses confer sacredness on a constitution. We argued that such clauses ‘correspond to the image of a sanctified constitution, which resists the demystification that stems from disconnecting the amendment process from the memory of the constituent moment’.Footnote 39 Ginsburg and Melton observe that ‘in some countries, the constitution is treated as a sacred text, never to be touched except in matters of major importance, while in other countries, the constitution is of little normative significance’. Countries in the latter category ‘would observe different values on entrenchment’.Footnote 40
Ginsburg and Melton’s conceptualisation of sacredness is rather vague. What is more, prima facie, it recognises no middle ground between a sanctified constitution and a constitution that has little normativity. It suggests that constitutions must be perceived as mystical, sacred documents in order to be accorded normative significance. Normativity, however, can stem from the ability of the constitution to perform the functions it is designed to perform, none of which are dependent on sacredness. The opposite of a sacred text is, in their view, a text of little normative significance. That is, constitutional normativity stems from sacredness. It is possible, of course, to read this as referring to a scale where normativity decreases as the document becomes less sacred. Nonetheless, the link between sacredness and normativity is somewhat dubious. The perception of sacredness is one-dimensional, with ‘sacred’ defined as ‘never to be touched except for matters of major importance’.Footnote 41 This is, however, but one way among many to describe sacredness, one alternative being that eternity clauses render some, but not all, things untouchable, leaving the ‘mundane’ provisions subject to regular maintenance.
Sacredness can also be linked to constitutional culture, to the way political actors and citizens view the constitution, and so on. Alternatively, ‘untouched’ could also imply precluded from democratic re-negotiation, which could cause a loss of normativity. Adaptability might also account for normative efficiency, while the democratic pedigree of constitutional norms, including constantly letting the people have their say by way of constitutional revision, poses as a serious alternative to sacredness.Footnote 42 Not only can a down-to-earth, adaptable constitution have great normative force, it can also gain the people’s affection and inspire constitutional faith.
It is also unclear whether the notion that cultural elements impact amendment practice allows for working with samples that contain both democratic and undemocratic constitutional orders. Would not the adoption of the parameter of culture, even a simplified and one-dimensional operationalised notion of culture, render this grouping impossible? Does constitutional change in authoritarian regimes obey the same logic as constitutional change in democracies? Is there such a thing as an amendment culture detectable through the application of identical criteria for both democracies and authoritarian regimes? Does the amendment rate suffice to pin down an ‘amendment culture’ that consists of similar sets of embedded attitudes indifferent to political regimes? Is it true that, if ‘amendment culture’ is equated with ‘the amendment rate of the last constitution enacted’, it is even possible to trace it using the same toolkit when applied to fundamentally different regimes? Is such an approach really compatible with an emphasis on constitutional culture?
Culture seems to be the focal point, yet it is measured in terms of amendment rates. This problematic operationalisation may explain why it is somewhat disconcerting when the theory of constitutional moments is mentioned. Ginsburg and Melton refer to ‘constitutional theorists [who] wrestle with the mechanisms of constitutional change and the relative importance of formal constraints’.Footnote 43 This is a reference to Bruce Ackerman’s We the people,Footnote 44 where Ackerman introduces his theory of constitutional moments, and traces the way history and culture affect constitutional change. Precisely because of this feature, Ackerman’s theory is subject to constant analysis, and has been a catalyst for important discussions on how to define constitutional moments. The dialogue on constitutional moments demonstrates both the assets and the shortcomings of Ackerman’s methodology. His work poses questions, triggers debate, and yet does not aspire to convince on the basis of quantitative data-based objectivity, as would otherwise be the case with measuring.
Why is it important to grasp Ginsburg and Melton’s approach to constitutional moments, given that they make such cursory mention of it? The answer is that it reveals how they actually view cultural accounts of transformative constitutional change. In order to be able to explain constitutional change, Ackerman emphasises the study of history, and his theory ‘adjusts accordingly’.Footnote 45 Constitutional moments as a theory of constitutional change relies on understandings of mechanisms that operate under the influence exerted by various political actors on constitutional culture. As such, it is hardly compatible with the attempt to transform culture into a quantitative indicator.
Another question that needs to be addressed concerns the institutional factors designers can effectively control. The amending formula itself is not the only institutional factor in play. Other institutional rigidities exist that lie within the grasp of the constitutional designer, and even of the lawmaker. What we suggest is that altering the amending formula can only be successful if due consideration is paid to all the factors that determine the degree of difficulty of constitutional change, that is to say, to all institutional and factual rigidities. In our view, ‘institutional and factual rigidities are communicating vessels, and the exchange between them is unavoidable. Consequently, changes at the institutional level influence factual rigidities and vice versa’.Footnote 46 Conversely, altering other parameters that influence constitutional change, for example the electoral law or the judicial review system, can potentially affect (the use of) the amendment formula. ‘The interaction between institutional and factual rigidities may be exploited to address detected dysfunctionalities by intentionally inducing change. At the institutional level, modifications are effected formally and intentionally, and can be used with the additional purpose of transforming political culture-induced rigidities’.Footnote 47
Ginsburg and Melton refer to ‘institutional factors and cultural factors’ that account for a ‘higher or lower amendment rate’, and contend that ‘theories focusing on amendment procedures and those that focus on amendment culture are not mutually exclusive’, because ‘it is possible (and maybe even likely) that both procedure and culture affect flexibility’.Footnote 48 However, what they do not analyse is how understanding the interaction between culture and procedure can be of use to constitutional designers. In addition, Ginsburg and Melton state in their analysis that de jure constitutional review may not substitute for constitutional amendments – a finding which they believe should be further explored. What they should take into consideration is whether a dialogic relationship between constitutional lawmaker and judge exists, forming a dialogic culture in constitutional change. This kind of exchange may entail formal amendments as a response to judicial decisions inducing informal constitutional change.
Can such considerations aid constitutional designers? Helping constitution drafting is a worthy aspiration, but it remains to be seen whether this is an attainable goal. Ginsburg and Melton seem rather sceptical, as they conclude their paper by stating that constitutional designers are influenced by their understanding of ‘cultural barriers to amendment’ and that ‘the formal amendment rule may, in the end, not matter at all, or at least may not matter in predictable ways across countries’.Footnote 49 However, we assert that it is feasible to perform an impact assessment with regard to the design of amendment formulas based on the analysis of multiple, interacting factors. Addressing the complexity of the parameters that impact constitutional change, and scrutinising the mechanisms that channel such change, can facilitate the design of formal amendment rules. ‘The possible consequences of altering the amending formula can be identified, while changes to other institutional rigidities that influence constitutional change, such as the system of judicial review or the electoral system, can also be assessed as to their potential side effects with regard to constitutional change’.Footnote 50 Each time formal amendment rules are modified, the balance between factual and institutional barriers to constitutional change is altered. Constitutional designers must focus on possible outcomes and trace the consequences of each separate intervention. We offer no generalisations, no instruction manual, but suggest that building awareness of the factors that influence constitutional change may prove to be particularly useful for the constitutional designer.
As regards the tracking of distinct amendment cultures, the importance of its precise definition and operationalisation surfaces. Ginsburg and Melton recognise the difficulty inherent in the concretisation of the concept of amendment culture. Their stated ideal would be to have data available on the actual attitude of individuals towards constitutional change. Since no such data exists they use a proxy, and operationalise amendment culture using ‘the rate at which a country’s previous constitution was amended’, assigning a value of zero to countries’ first constitutions.Footnote 51 This attempt to make (amendment) culture measurable reduces it to one measurable element, and actually negates the potential importance of culture. However, amendment culture may prove to be an extremely useful concept if elaborated upon with the intent of accommodating the multiple and often non-measurable parameters that influence constitutional change.
Constitutional amendment between metrics and culture
Empirical research has taken root in comparative constitutional law and has been applied to the study of the rules and practices of constitutional amendment. However, the question as to whether metrics are suitable for understanding comparative constitutional amendment remains unanswered. To answer that question, a series comprising the characteristics of constitutional amendment, as well as series of characteristics of quantitative empirical research, should be taken into account. It must be kept in mind that the question to be addressed is not whether quantitative empirical approaches, in general, have something to offer legal scholarship, but whether the determinants of constitutional change can be decoded through metrics. When questions become more complex, issues such as ‘avoiding conflicting evidence’ emerge.Footnote 52 So, to discuss the value of quantitative empirical approaches to constitutional amendment we must first establish what exactly is being examined through empirical methodology, and what the purpose of this examination is.
Constitutional amendment depends on multifarious interacting parameters. Some can be quantified, whereas others cannot. How many times a constitution has been amended can be tallied, but constitutional culture cannot. Attempting to oversimplify the cultural and historical features of each polity that have an impact on the route of constitutional change does not remedy the omitted variable bias problem. It addresses the danger inherent to empirical studies, that is, the drawing of conclusions merely on the basis of what can be counted, by counting what cannot be counted. To paraphrase a well-known maxim,Footnote 53 what matters in constitutional amendment cannot be counted and what can be counted does not matter. In understanding constitutional amendment it is what Hirschl calls ‘crucial yet fuzzy factors’Footnote 54 that matter most. Too many ‘ifs’ render the endeavour of studying comparative constitutional amendment through metrics highly risky: the approach is valid if variables are not omitted, if crucial parameters are not overlooked, if the metrics are properly executed, and if they are not arbitrary in their classifications. The prerequisites are manifold.
Ginsburg and Melton write well. Their grasp of conceptual constitutional frameworks is impressive, and the ‘methods sections’ do not spoil the rest of the prose. And they do not just sweep the problematic elements of their account of constitutional change under the carpet. Still, the transcultural and transnational data sets they use, plus the operationalisation of culture through metrics, render their approach misleading.
Amendment culture is not a ‘tricky concept to measure’ as Ginsburg and Melton state.Footnote 55 It is, rather, an unmeasurable concept.Footnote 56 What seems to have gone wrong in operationalising it on the basis of how many times a constitution has been revised is the ‘move from the abstract to the concrete’,Footnote 57 that is, the actual measurement. This seems to be a recurring problem with quantitative empirical approaches to constitutional change. As stated by Lee Epstein and Andrew Martin, in Ginsburg and Melton’s previous work on constitutional endurance the catch lies in translating the concept of specificity ‘into some precise indicator of specificity’.Footnote 58 The concretisation of the abstract concept of specificity proves to be quite problematic. The same applies to the effort to concretise amendment culture. Is there any need to measure amendment culture? Why should one make this effort? As posited by Pierre Legrand, incommensurability is not unintelligibility.Footnote 59
Constitutional cultures reflect common understandings and attitudes towards the constitution.Footnote 60 Constitutions emerge and are enacted within a specific legal culture and also operate, are enforced, interpreted, and developed within and through that legal culture.Footnote 61 The disposition of citizensFootnote 62 is part of that culture, but not the only component. According to Ginsburg and Melton, the ideal way to assess amendment culture would be to amass data on citizens’ attitudes regarding constitutional change. It is not clear, however, how this attitude can be separated from the general attitude of the citizen towards the constitution, and whether it would suffice to delineate only this citizen-based amendment culture, or whether the attitudes of other parties involved in constitutional change, for example, politicians and judges, would be also relevant.
Supposing a distinct amendment culture even exists, separable from the constitutional culture, should we view it as a monolith, or can it be broken down into its component parts?Footnote 63 First of all, insofar as it is indeed distinguishable from constitutional identity and culture, its creation would have to be placed within a time frame. While the enactment of a constitution takes place within the context of a constitutional culture, it is not clear whether this implies that an amendment culture exists as well, or whether it takes some experience with amendment practice before an amendment culture materialises. Or is it perhaps a specific amendment culture that generates the formal amendment rules?
The previous questions are challenging, and appear to defy measurement. Any attempt to answer them without reference to history, constitutional identity and the use of abstract concepts would appear to be, and perhaps is, doomed. The parameters that impact constitutional change are multifarious, thus ‘all things equal’ might not allow the numbers to have any precision. The problem of bidirectional causationFootnote 64 poses a serious danger, as misunderstandings may arise when dealing with constitutional amendment issues on the basis of numerical evidence. The relationship between the number of amendments and the so-called amendment culture, as well as with the environment in which they take place, is without doubt bidirectional.
The goal of scholarly endeavour is decisive. The intent when engaging in empirical research is critical to the determination of how grave the perils of bidirectional causation are. The distinction to be made between persuasion and inference proves illuminating. Lee Epstein and Garry King suggest that lawyers specialise in persuasion, whereas social and natural scientists aim to make inferences.Footnote 65 Empirical research aims, in that sense, to gain knowledge about the world. Beyond a discussion about how accurate the distinction is – to what extent persuasion and inference overlap etc. – lies the issue of detecting what the study of constitutional amendment entails. In other words, persuasion, as part of an adversarial legal system, is not the exclusive topic of all constitutional law scholarship.
As Epstein and King observe, empirical articles do attempt to persuade by making use of empirical evidence.Footnote 66 This makes sense, given the ‘performative value of numbers’.Footnote 67 Numbers can be powerful tools of persuasion, as they are symbols of precision, accuracy and objectivity. They can create a common denominator where there is none.Footnote 68 These qualities are dream tools for constitutional scholars who want to convince us of something. What if, however, the intent is to offer practical suggestions for the drafting of constitutions?
Ginsburg and Melton appear pessimistic about the possibilities and limitations of constitutional design. However, the realisation that cultural factors influence the way formal amendment rules work does not diminish the value of empirical evidence for drafting amending formulas. It suggests that empirical methodology cannot offer an isolated, contained context for constitution-writing, the challenge being to correlate it with other comparative constitutional law methodologies.
The design of amending formulas
The intent of research into constitutional change matters. If the purpose is to help constitution drafters to design a workable amendment formula, the focus must be placed on what is, or can be, brought under their control.
Consider the alternatives: (a) in the event that amending formulas cannot really have an effect on the outcome, there is no need to put much effort into their design. Any attempt to seek ways to improve formal amendment rules would appear to be pointless. Constitutional designers may just allow for a random combination of material or procedural limits, and let the factors that matter most take over; (b) in the event that the amending formula does have a decisive effect on the outcome, but other factors do so as well, or even more so, the question at hand is whether there is a correlation between the formula and those other factors. The next step is to distinguish between factors that can be brought under control and those that cannot, to set realistic goals for the endeavour, and to avoid mistakes caused by a failure to take vital parameters into consideration.
One option would, therefore, be to abandon all efforts at generating better formal amendment rules, and to control the paths of constitutional change through legal design and intervening with the other institutional rigidities. Alternatively, ways to attune the amendment rules to their environment could be sought. This endeavour demands a deeper understanding of the correlation between such rules and their context. Empirical approaches to constitutional change can offer insight with regard to constitutional design.Footnote 69 However, they can also lead to constitutional nihilism by suggesting that important constitutional provisions such as amending formulas have no impact on constitutional change. This is misleading and is only logical if one construes the relevance of amendment culture to be measurable, for instance by limiting it to a measurement of amendment rates. We, however, maintain that, although the cultural factors that affect formal and informal constitutional change are not measurable, and that the correlation between them is incommensurable, they are perceptible.
Moreover, patterns do exist in the realm of constitutional change and are detectable. The descriptive models we have used to classify constitutional change are based on correlations between an amending formula, the political system, the constitutional ethos, the judicial system, the legal culture, etc., and which can facilitate constitutional design.Footnote 70 The key lies in the observation of the way mechanisms of constitutional change operate when producing change. For example, in the amendment model which we labelled ‘pragmatic’, formal change is facilitated by the amending formula or a consensual legal culture that allows for constitutional adaptability.Footnote 71 Detecting how constitutional adaptability works in a consensus-dominated culture, and how it relates to the amending formula, may for instance provide insight into how one can render constitutions resilient, resistant to external pressures.Footnote 72 Culture is crucial here, and entails much more than amendment rates. We do not imply that formal rules are not important, but only that a similar rule may operate differently in an environment where a consensus to effectuate change exists, as opposed to an environment in which such consensus is lacking. Intentional efforts to control the mode of constitutional change by drafting or changing amendment rules must therefore be culture-conscious.
Culture-consciousness is even more crucial to an understanding of how constitutional change mechanisms can interfere with formal amendment rules within the context of the distrust model of constitutional amendment, as we labelled it. In this model, political conflicts, distrust, polarisation, and veto strategies create an environment of distrust, within which the legal stringency of amendment rules may be heightened. Especially when, in such an environment, amendment rules requiring augmented consensus are drafted, multiple unintended consequences may ensue. The very process of placing a country within a specific model, and of examining where it conforms or deviates from that model, will impose on constitutional designers a disciplined approach to the specific parameters they are attempting to tame. Awareness of the various interacting mechanisms of constitutional change, and of the way they operate within a specific context, enables the constitutional designer to predict the overall effect of interventions made to the amending formula.
In other words, constitutional design can benefit from the use of a flexible matrix for constitution-writing in order to detect whether an amendment formula is attuned to its environment. Such a matrix must not fail to take into account that the unpredictable nature of the future in general, and unforeseen events in particular, may have an impact on the political and legal environment. Nevertheless, a rational ex ante impact assessment of the effects of amendment rules is attainable, based upon a juxtaposition of the rationale behind amendment rules, and of the context within which they will operate, or have operated. That the amendment difficulty does not depend solely on the amendment formula itself does not mean that constitutional designers cannot influence it at all. It means that in order to achieve the desired level of flexibility, they have to take into consideration the entire ensemble of hindrances to constitutional change.
Amending formulas keep amendment rates in check, but serve other purposes as well,Footnote 73 and are in fact an essential characteristic of the whole document. For instance, unamendability is important regardless of the possibility that it might not in reality preclude amendment, since the removal of an eternity clause necessitates deliberations regarding the change of what the clause aims to protect.Footnote 74 In that sense, the limited degree to which drafters can impact amendment rates does not lessen the importance of carefully examining design options. The amending formula influences much more than the flexibility of the constitution itself, and successful design depends on an awareness of the possible impact it may have on all the many functions served by formal amendment rules.
An ex ante impact assessment of behaviour under certain amendment rules entails making a readily-available check list of the functions formal amendment rules have, and of the parameters that influence constitutional change, that is, the embedded practices and norms, or otherwise the institutional and factual rigidities (including the political and constitutional culture). Non-measurable elements may rightfully be included in such lists, which are, after all, drawn up to aid in making choices that, although subject to rationalisation, are also political and ethical.
Incommensurability and even intangibility do not preclude the inclusion of parameters in the list, thus avoiding potential omitted variable bias problems. Nor is there any need to dress the old debate on incommensurability, science, and empiricism in new (constitutional) clothes. To understand the idiosyncrasies of constitutional change, we should embrace perplexity and attempt to solve the sliding puzzle by studying its mechanics.
