Institutional Possibilities for Democratic Constitutional Change

This section begins to build a framework of evaluative criteria for assessing mecha­nisms of constitutional change out of the consideration of a few case studies of vari­ous institutional experiments.

9.3.1 Direct Democratic Constitutional Change

Let me start by considering an imaginary institutional arrangement: namely, some form of anti-constitutional direct democracy. The idea here involves direct democ­racy—such as regular periodic assemblies of the entire enfranchised populace— where that assembly has plenary power over all of the law governing the populace. Hence in this scenario, the legislative power is entirely in the hands of the assembled demos. And that legislative power is indistinguishable from a constituent power, since exactly the same procedures apply to passing all forms of law, statutory and regulatory as well as constitutional.

Prima facie such an arrangement satisfies many of the conditions I indicated earlier as central to deliberative democratic constitutionalism. Fundamentally, it is a quite straightforward way of structuring the idea of co-authorship of laws. Citizens are here directly involved in giving themselves the laws, enabling them to under­stand themselves as simultaneous subjects and authors of the law. Furthermore, suitably designed decision procedures for the assembly should be able to track closely other key conditions. Political equality can be easily secured when all citi­zens have roughly equal opportunities to influence the lawmaking process. Such political equality is extended to fundamental matters since no law is off limits. And the plenary authority over the entire legal corpus at every assembly means that revisability is likewise ensured. As described so far, these arrangements do not necessarily involve structured deliberation; the assembly could use simple majority rules on secret ballots for both initiating and enacting proposals. More naturally however, we would expect practices of debate and deliberation to arise and it should not be difficult to structure them by procedures sensitive to the other conditions indicated. In particular, with deliberative mechanisms for the exchange of informa­tion, opinions and reasons, political equality is enriched beyond a simple equal vote, encompassing real opportunities for equal voice and qualitative input into the law­making process. And because the entire enfranchised citizenry is involved, we should expect the process of opinion formation and decision making to cast as wide an epistemic net as possible: the assembly of all makes it possible for all kinds of different information and opinions from the broadest swath of citizens to be can­vassed and included.

Nevertheless, there is one crucial missing condition: namely, some form of con­stitutionalism, some form of formal or institutional separation of the exercise of ordinary legislative and constitutional legislative powers. This could of course be relatively easily remedied by the adoption of a formal distinction between ordinary legislative and constitutional legislative activity, a distinction reinforced by making the process of constitutional change more difficult and subject to higher standards of opinion formation. Such moves would then constitutionalize the procedures of direct democracy, thereby enabling citizens to understand the outcomes of those procedures as legitimate and binding even when citizens disagree with the sub­stance of those decisions—as many inevitably will under conditions of persistent pluralistic disagreement on matters of substance.

ity—according to the trade-offs needed between alacrity and the seriousness of the issue at stake (Rousseau 1997: Book IV, chapter 2). Even so, however, his arrangement is still anti-constitutional in this sense: no law is put out of reach of a current assembly. In fact, every assembly opens first with this question: “whether it please the Sovereign to retain the present form of Government?” (Rousseau 1997: 120, Book III, chapter 18).

9.3.2 Legislative Constitutional Change

and Operationalizability

Its no accident that so far I have been referring only to a merely imaginary arrange­ment. Despite whatever normative attractions some form of direct democracy might have, the fact is that all existing national systems for democratic lawmaking employ elected representative bodies to carry out legislative functions. And the reasons for this are not at all obscure. The costs of operationalizing direct democracy are simply so high as to make it unfeasible for populous, complex and extended nation-states.

From a reconstructive perspective it might seem perverse to consider an unreal­ized arrangement: what does a fantasy have to do with the normative content immi­nent in historically actualized institutions? I would argue however, that unanimous rejection of the most direct institutional realization of democratic self-rule tells us about a key normative criterion: operationalizability. Whatever other values they promise to realize, institutions that cannot be actualized are deficient. It is thus reconstructively clear why all national democratic legislative systems employ indi­rect modes of democracy.

9.3.3 Agency Problems and Structural Independence

With the move from direct to representative systems, however, new normative con­cerns arise. Most pressingly, there is the general problem of tying officials’ actions to the interests, opinions and reasonings of the demos—the central problem of agent-principle relations. It is well beyond the reach of this paper to say anything in general about the problems of agency encountered by representative democracies. But it does seem to me that with agents for law-making, it becomes ever more important to insist on mechanisms for separating the function of ordinary and con­stitutional legislation. This can be seen most simply by recalling that constitutional procedures are those which structure not only ordinary lawmaking procedures, but also regulate the elections of representatives and structure the workings of govern­ment. As Ely among others have made abundantly clear, then, representative democracy is subject to a particular form of procedural distrust: distrust that legisla­tors will manipulate constitutional procedures to freeze the ordinary mechanisms of democratic change and to insulate themselves or status quo arrangements from challenge (Ely 1980).

Agency problems then recommend a real form of structural independence between ordinary legislative and constitutional legislative processes. There ought to be a clear institutional demarcation of the difference between the ordinary exercise of government business through established procedures, and, the people’s constitu­ent power of changing the procedures—the fundamental institutions and basic rights protections—that are the procedural warrant for the legitimacy of the out­comes of ordinary political processes. Hence whatever mechanisms are available for constitutional change, they ought not to be easily manipulated by current repre­sentative majorities in order to lock in future constitutional procedures that system­atically favor them or systematically foreclose ongoing possibilities for democratic change.

The need for structural independence is perhaps most easily seen in the recent constitutional history of Venezuela, where the elected officials of one political party (the PSUV forcefully led by the charismatic Hugo Chavez) were able in subsequent rounds of constitutional change to effectively close off avenues of political change and defang opposition candidates and parties. The sequence of changes was inaugu­rated by the new 1999 Venezuelan constitution, which laid the grounds for collaps­ing any independence of ordinary and constitutional legislative functions. The 1999 constitution significantly weakened the legislature in relation to the executive, it significantly centralized and strengthened the executive in the direction of strong presidentialism, and importantly, it specified a quite easy threshold for constitu­tional amendment.

But even that bare recitation of the formal amendment procedures makes it look harder than it is, given the particularly robust and overlapping forms of the centralization of power in the presidency. Consider for instance a major constitu­tional amendment achieved 10 years after the new constitution: Amendment 1. In that case not only was the majority in the National Assembly effectively voting in lock step with the wishes of president Chavez, but the entire apparatus of the state was brought to bear in a one-sided propaganda campaign to convince voters to ratify the amendment. And the content of 2009’s Amendment no. 1? The abolition of term limits for the president, for national and regional legislators, and regional and municipal governors—effectively closing paths of political change and ensuring the long-term single-party dominance of the PSUV.^[269]

It is perhaps not overly dramatic to say that because of a lack of structural inde­pendence between ordinary and constitutional mechanisms, Venezuelan constitu­tionalism has enabled apparently democratic mechanisms to be used strategically in order to foreclose the ongoing possibility of open and competitive democracy for the future.

9.3.4 Entrenchments and Democratic Co-authorship

In the light of such dangers, one might think that constitutional obduracy is a pre­ferred way to ensure the structural independence of constitutional change mecha­nisms from current regime office holders. Making constitutional provisions very hard to change—even making some impossible to change in the form of hard entrenchments—would seem to protect against future agency problems where office holders attempt to change political procedures in order to capture the political system and remain in power. Constitutional amendment procedures might then be set to require a very high bar to enactment—for instance, as in the United States or Australian constitutions—or even set aside certain portions of the constitution as formally not subject to amendment—as in the hard entrenchments of senatorial rep­resentation in the U.S. or of certain fundamental individual rights in the German basic law. Comparative scholarship has established, however, that there are signifi­cant perils associated with overly obdurate constitutions. For instance, there is a significant correlation between constitutional flexibility and constitutional longev­ity (Elkins et al. 2009).^{[270] [271] Overly rigid constitutions are, to be blunt, more likely to suffer an early death.}

More recent constitutions have apparently avoided hard entrenchments.¹¹ For instance, the exemplary South African constitution of 1996 does make one part more difficult to change than all the other parts: Sect. 2.1 of Chap. 2 concerning the foundational principles of the republic (democracy, human dignity, constitutional supremacy) is harder to change than all other parts of the constitution, subject to 75 % rather than 66 % of the legislature (as well as the normal 6 of 9 regional prov­inces for amendments affecting regional powers). But even then, these foundational principles are not impossible to change, only harder than other constitutional prin­ciples. From a normative perspective, it seems clear why hard entrenchments are to be avoided: they violate the criterion of democratic co-authorship. In effect, hard entrenchments establish that the people are to be subject to some laws that they themselves cannot alter, or at least cannot alter without a revolutionary replacement of the constitution in its entirety. Thus even the foundational principles of the South African constitutional settlement are in-principle open to democratic renegotiation into the future, even as they are set aside as especially fundamental to the republic— as one would predict from the need for constitutional structuration itself. Democratic co-authorship ought not stop at ordinary legislation, or even at some subset of con­stitutional law, but must extend to all fundamental matters of law, otherwise sub­jects can only understand themselves as passive subjects of the lawmaking of others. Apart from hard entrenchments, very difficult procedures for constitutional amend­ment can also effectively foreclose possibilities for democratic co-authorship of constitutional law, even if they remain in-principle possibilities.

9.3.5 Judicial Interpretation and Political Equality

In light of both pressures for constitutional adaptation to changing conditions and the negative correlation between constitutional obduracy and longevity, it should be no surprise that constitutions with formally rigid change procedures have in fact adopted a number of mechanisms for constitutional change apart from formal amendment procedures. Most prominent here is, of course, constitutional change carried out by judiciaries, usually through the exercise of powers for the judicial review of legislation, regulation, and administrative action. For instance, in his comparative study of 36 democratic nation-states between the end of World War II and the mid 1990s, Lijphart found a statistically significant positive correlation between increasing constitutional rigidity and the likelihood of strong judicial review, that is, assertive forms of judicial policy making with respect to constitu­tional issues (Lijphart 1999). More recent literature on the judicialization of poli­tics—including constitutional politics—shows that there is a real shift in constitutional legislation away from more democratically accountable actors and towards more politically insulated judiciaries (Ginsburg 2003; Hirschl 2006; Shapiro and Sweet 2002; Stone Sweet 2000; Tate and Vallinder 1995).

It is well beyond the scope of this paper to fully evaluate issues of judicial review; instead I will make just three points concerning the employment of the judiciary as constitutional legislators. The first point is that the criteria of both oper­ationalizability and structural independence speak in favor of constitutional change through judicial interpretation of constitutional law. On the one hand, judiciaries must already specify legal provisions of whatever form in the routine application of those provisions to concrete cases—constitutional provisions no less than any other. It is then an easy mechanism to operationalize. On the other hand, judiciaries are regularly insulated in a number of ways from the vicissitudes of politics and from pressures facing electorally accountable political actors in order to ensure fairness to individual litigants. Judiciaries involved in constitutional change through inter­pretation are therefore already structurally independent of the ordinary process of legislation carried out by electorally accountable politicians. This structural inde­pendence is, of course, the basis for proceduralist justifications for placing the power of constitutional review in the hands of the judiciary: they are to be, in effect, the unelected guardians of the very procedures of democracy, that is, of the consti­tutional rules which proceduralists take to warrant the legitimacy of democratic outcomes in the first place (Dahl 1989; Ely 1980; Habermas 1996; Zurn 2007).

The second point is that, nevertheless, it will be very difficult, if not impossible, to cabin courts with powers of constitutional review to the pure function of consti­tutional protection. Because of several different reasons—the abstract and under­theorized character of constitutional norms, judicial responses to informal political changes in a constitutional system and to general social changes, doctrinal develop­ment and legal path dependence—court-based constitutional protection will inevi­tably transmute into positive constitutional elaboration.^[272] The clear line between judicial protection of a legal provision and judicial elaboration of the content of law will be constantly undermined: protection will inevitably bleed into elaboration for both ordinary and constitutional law. In the course of enforcing the (constitutional) rules of the political game, then, judiciaries with powers of constitutional review will inevitably become much more than referees: they will become constitutional legislators.

The holders of constituent power however, thirdly, are emphatically supposed to be the entirety of the citizenry in democratic theory (of whatever form). If only a small subset of citizens are the decisive constitutional legislators, and if those legis­lators are institutionally positioned exactly so that they are not subject to attempts to influence them by the demos, then constitutional change through the judiciary emphatically violates a baseline criterion of political equality. Even if that constitu­tional elaboration is carried out conscientiously and benevolently, it is still a pater­nalist institutionalization of the power for constitutional change. This worry about judicial paternalism with respect to fundamental constitutional procedures is, I think, the real basis of the democratic complaint against judicial review, and not the extremely misleading idea that judicial review is suspect because it is counter­majoritarian. For there are any number of counter-majoritarian political procedures which are fully consistent with political equality. For instance, counter-majoritarian voting rules requiring either full consensual unanimity or various levels of super­majorities nevertheless afford each voter an opportunity equal with all other voters to influence the outcome of a decision. The democratic problem with constitutional change through judicial interpretation is that every citizen is distinctly not afforded an equal opportunity to influence the law-making occurring—the problem is then one of political equality, not majoritarianism.^[273] When judges are empowered as constitutional legislators—perhaps out of the necessity for some agents of change in overly obdurate and rigid constitutional systems—enfranchised citizens are effec­tively shut out of that constitutional law-making process and citizens thereby become mere subjects of laws authored and paternalistically imposed by others.^[274]

9.3.6 Veto Players and Inclusive Sensitivity

To my knowledge, no democratic constitution formally places the power of consti­tutional amendment in the hands of courts. Rather, the overwhelming majority pro­vide for amendment through either elected legislatures and executives, or various forms of popular initiative from citizens themselves, or various forms of special constituent assembly of democratically accountable representatives—or frequently from some combination of the three.^[275] This is not a mere coincidence: the constitu­ent power is always formally recognized as resting directly or indirectly in the hands of the citizenry, at least in democratic systems. These institutional arrangements lend support to the reconstructive hypothesis that they embody the ideals of demo­cratic co-authorship and political equality.

Furthermore almost all democratic constitutions make constitutional legislation more difficult to pass than ordinary law—lending supporting to the hypothesis con­cerning structural independence. And even the notable exceptions where there is no formal difference between making constitutional and ordinary law—e.g., the United Kingdom and New Zealand—evince robust informal traditions, norms, and custom­ary practices that distinguish between the two, rendering constitutional changes more difficult.^[276] There are however several different characteristic mechanisms for increasing the difficulty of enacting constitutional change. For instance, there can simply be higher supermajority thresholds in the legislature for amendment, typi­cally three-fifths or two-thirds, and less frequently three-quarters. Bicameral sys­tems usually require such supermajorities in both houses. Second or third readings of amendment proposals might be required; intervening elections between those readings can further increase difficulty. All of these amendment mechanisms alone, however, in essence employ the same legislative system—and usually the same legislative players—as used for ordinary lawmaking.

By contrast, empirical research has highlighted a different set of mechanisms as important, giving roles to various actors who are differently situated than normal legislators. For instance, amendment proposals might need to be ratified by regional sub-units of the nation, usually the legislatures of federal states, and usually requir­ing a slight supermajority of such states. Quite characteristic of newer constitutions, especially in Latin America, amendment proposals must be ratified in popular ref­erenda, usually by majorities or slight supermajorities of ballots cast by ordinary citizens. Finally, many newer constitutions—for example, those of Bolivia (1999), Bulgaria (1991), Colombia (1991), Ecuador (2008), and Venezuela (1999)—require or permit a form of special constituent assembly for constitutional change propos­als. While the former arrangements simply make it harder for normal legislative officials to pass amendments, the latter arrangements introduce ‘veto players’ into the mix. Empirical research suggests that only the introduction of veto players into amendment schemes actually significantly increases the difficulty of amendment. Rasch and Congleton have shown for OECD countries (and others have confirmed in EU countries (Closa 2012)) that just making it harder for legislatures to ratify amendments (e.g., from three-fifths to two-thirds to three quarters) doesn’t much change the amendment rate (Rasch and Congleton 2006). What really affects the amendability of constitutions seems to be the presence or absence of veto players in the process.^[277] Because currently empowered political parties can frequently muster supermajorities in the legislature in subsequent elections, blocks to constitutional amendment ratification such as moderate legislative supermajorities over a period of time and after subsequent readings are not very different than blocks to enacting ordinary legislation. Hence, “in the absence of powerful external veto players, it seems that political parties’ agreements may sail through even the most stringent constitutional reform procedure” (Closa 2012, 309).

Normatively speaking, the difference in amendment mechanisms with veto play­ers is, I want to suggest, significant. In particular, such a difference speaks to the inclusive sensitivity of the mechanism: the presence of veto players ensures that amendments are acceptable to a broad diversity of constituencies with distinct inter­ests, ideological positions, opinions, values and perspectives.^[278] The arrangements for changing the fundamental procedures of politics and lawmaking ought to struc­turally incorporate sureties that the full diversity of affected persons and interests will be accounted for. Hence the difficulty-increasing procedures for amendments are not just about increasing difficulty—even as this is important for maintaining structural independence. Many of those procedures are better understood as broad­ening the usual pool of information available for—and the sphere of influencers of—constitutional legislation beyond the current party regime and beyond the usual way in which representation is structured across the national legislature and the executive branch. Ratification in the federal sub-units, for instance, should enable a different set of political representatives to have their specific concerns taken into account. And ratification by popular referendum—beyond the way it serves the cri­teria of democratic co-authorship and political equality—promises some greater sensitivity to the opinions of all those affected by the proposal beyond the normal channels available for citizen influence on elite politicians and political parties.

To be sure, this greater inclusive sensitivity should not be oversold: after all, if the normal legislature is largely responsible for proposing and writing the amend­ment in the first place, then the role of veto players is largely confined to a simple ex post thumbs-up or thumbs-down, rather than direct ab initio substantive input into the qualitative content of the initiative. But if the political public sphere is working well and the legislature is at least partly attuned to the likely opinions of veto players, then we can hope at least for some degree of increasing inclusive sen­sitivity through the use of ratification veto players, even where the original amend­ment drafting process is driven exclusively by the legislature.

9.3.7 Constituent Assemblies and Reasons-Responsiveness

This last concern about the degree to which a broad spectrum of the citizenry have real effective input into the substantive content of constitutional amendments—as opposed to a simple power of after-the-fact veto or endorsement of that which has already been authored by others—speaks to a central difference between the way political equality is conceived between aggregative and deliberative conceptions of democracy. In particular, while aggregative conceptions emphasize the equal vot­ing power of each in a process of aggregating over the population’s simple endorse­ments or rejections, deliberative conceptions put more emphasis on the equal access all have to the processes of reason collection and evaluation that lead up to and ensue in the design of a particular proposal. Political equality is not then merely a matter of equal impact registered in an equally weighted vote—even as that is quite important to political equality—but must also involve the equal effective part each can play in the processes of deliberation that ensue in policy creation. Voting is then seen as an egalitarian mechanism for temporarily bringing to a halt ongoing pro­cesses of collective reasoning when a decision is needed under constraints of time, knowledge, and reasonable pluralism.

Returning to constitutional amendment procedures, the question then is whether we can envision procedures that not only are broadly sensitive to the voting impact of a wide diversity of citizens—as are constitutional ratification mechanisms sub­ject to veto players—but also sensitive to a wide diversity of politically relevant reasons from a broad spectrum of citizens. Is there a way of making amendment procedures specifically reasons-responsive? Clearly one central way in which democracies can be reasons-responsive is by connecting the actual workings and outputs of representative legislatures to robust processes of public opinion forma­tion in free, open and diverse political public spheres (Habermas 1996: especially chapters 7 and 8). However, if we are concerned about two agency problems regu­larly faced by legislatures—as I think we should be from everyday experience— then we might worry about whether legislatures alone are sufficiently responsive to a wide diversity of relevant reasons, especially when they are taking the lead role in authoring the substantive content of constitutional proposals. First, given that elec­toral politics as we know it is largely shaped through political parties and party competition, it turns out that legislatures are frequently captured by currently domi­nant political parties. In these cases, a dominant party will be able to effectively ignore relevant reasons from other parties that are contrary to their preferred policy outcomes. Second, even if representatives do account for the reasons of other like political elites, they may still be wholly insensitive to the reasons of broad swaths of ordinary citizens who are not able to make effective use of the communications media of the public sphere. Most obviously this comparative communicative dis­ability falls along socioeconomic lines, but it also quite frequently falls along indig­enous, national, ethnic, religious and/or racial lines. Hence legislative deliberative processes may suffer from both dominant party capture and elite opinion selectiv­ity. Both problems become normatively more serious the more fundamental the matters are for legislative decision, in particular, when they concern matters of basic constitutional law that is to structure ordinary politics.

It seems to me that constituent assemblies—independently elected bodies with a specific mandate to write proposals for constitutional reform either in the form of amendments or a new constitution—promise to improve reasons- responsiveness over constitutional drafting processes that are legislatively driven. Three features in particular would seem to promote reasons-responsiveness. Because constituent assemblies are specifically designed to consider only issues of constitutional change, their deliberative processes are likely to be better focused on constitutionally rele­vant reasons. Second, because the elected members are not the same as elected legislators and because they do not stand for re-election to the assembly, their delib­erations are likely to be less systematically distorted by the incentives of ordinary electoral and party politics. Third, because the assembly is almost always elected through procedures that ensure a wide representation of different segments of the populace, they are likely to be more sensitive to a broader diversity of reasons, interests and opinions than is a legislature controlled by political party elites. For instance Colombia’s 1991 constitution has provisions enabling the convocation of a constituent assembly if both one third of the electorate and both houses of the legislature vote in favor of convening one.^[279] Members of the assembly are to be directly elected by citizens through a ballot separate from ordinary legislative elec­tions. While the assembly meets, the legislature’s powers are suspended. Reform proposals from the assembly are then ratified when agreed to by both a legislative majority and a popular referendum. Such arrangements promise the three delibera­tive advantages indicated above of an exclusive constitutional focus, of insulation from ordinary electoral politics, and of broader representation of the diversity of available reasons.

That at least is the theory, even if it is not always born out in practice—actual cases are decidedly mixed from the normative point of view of democratic constitu­tionalism.^[280] Brazil’s successful transition from military dictatorship to stable consti­tutional democracy was formally achieved through the adoption in 1988 of a constitution written over 2 years through a national constituent assembly. While the members of the constituent assembly were in fact simply the current members of the legislature meeting in special sessions as an assembly, the procedures adopted in the drafting phase not only required input from a diverse representation of social move­ments, political interests and ideological positions, they also ensured a great deal of public input through comments, hearings and largely open proceedings.

The successfully democratic Brazilian experience contrasts, however, with Venezuela’s 1999 constituent assembly process. While coming into power in 1998, Venezuelan president Chavez promised a referendum to call for a constituent assembly to replace the then-in-force 1961 constitution, even though the latter had no provisions for such an assembly. With very strong support in the referendum (92 % and 86 % on the two questions), an assembly was convened under electoral laws that strongly favored members of the president’s party—the party gained 120 of the assembly’s 131 seats. The assembly itself wrote the new constitution very quickly, in 2 months. The assembly’s debates were well publicized in the drafting phase and, once drafted, the constitutional proposals were subject to inclusive debate with many different opinions sectors of society represented (Landau 2012, 941). The proposal was ratified by a significant majority of voters (over 70 %) in a national referendum. Nevertheless, the new constitution created a government with political power strongly centralized under the authority of a charismatic president, a centralization that has increased as that constitutional settlement has developed— with dramatic results for the loss of structural independence, as discussed above in Sect. 9.3.3. While the Venezuelan case presents a fairly good picture of the way constituent assemblies can heighten broad and inclusive democratic sensitivity, it certainly did not avoid the problem of dominant party capture: indeed, the process made it worse by constitutionalizing capture.

Another even more cautionary tale is provided by Bolivia where an irregular and complicated process between 2006 and 2009 led to the formation of a constituent assembly and the eventual ratification of a new constitution. Simplified, the story begins in 2006 after newly elected president Morales took office in 2005. Employing provisions for constitutional replacement in the 1967 constitution, the legislature approved, by the required two thirds majority, the convocation of a constituent assembly for the total reform of the constitution. After convening in 2007, the assembly was subject to a great deal of disagreement, power struggles and contro­versy, ensuing in sometimes violent protest. Ominously, after the diverse parties in the assembly failed to come to a agreement, the assembly moved locations twice. After the first move, opposition members refused to participate and, after the second move, opposition members were forcibly prevented from entering the assembly. Nevertheless, by the end of 2007 the remainder of the assembly delivered a draft to the legislature. More political troubles engulfed the process during 2008 until finally a compromise was reached by elites, and in 2009 a popular referendum was finally held that ratified the new constitution with a 61 % majority of the voters. Even with­out all the necessary detail, it is hard to consider the recent, troublesomely violent and irregular Bolivian process of constitutional change particularly reasons- responsive (not to mention concerns about political equality and inclusive sensitivity).

Perhaps these cautionary tales should not surprise, since constituent assemblies are usually not called in times of political calm and citizen satisfaction with govern­ment; they tend rather to be products of crises of governance of one form or another (Negretto 2012). But Iceland’s recent experience with a constituent assembly—one born out of the deeply impactful 2008 financial crisis—shows that, when suitably designed and taking advantage of the latest forms of communications technology, such assemblies can evince real improvements in both inclusive sensitivity and reasons-responsiveness. Told briefly, the story is that a collective of grassroots movements organized a kind of proto-constituent assembly called the National Assembly in 2009, three fourths of whose membership was drawn from randomly generated citizens and one fourth from political institutions and associations. The purpose was to brainstorm the key ideals for the future of Iceland through well- designed deliberative small-group discussions combined with larger plenary ses­sions. In 2010, the legislature established a formal constituent assembly comprised of 25 individuals elected in national elections—the ‘Constitutional Council’—in order to revise the 1944 constitution.^[281] The legislature also organized a one-day ‘Constitutional Gathering’ as a participatory event for ordinary citizens before the elections to the assembly. The constituent assembly itself drew heavily on citizen input into its deliberations, particularly through the use of internet communications media. A draft constitution ensued from a full consensus of the assembly and was presented to the legislature in 2011.^[282] The draft was endorsed in a non-binding advi­sory referendum in 2012 (with a 67 % popular majority), but to this date, the pro­posed constitution is in limbo, as it has not been ratified by the legislature.

This was very much a process of proposed constitutional change that began ‘from below’ and it maintained throughout a remarkable openness to and constitu­tive connections with broad and diverse populations, interests and opinions through­out the populace. “The originality and unprecedented nature of the whole process lies clearly in the explicit emphasis on citizen-driven constitutional reform, a form of ‘crowd-sourcing’ in the form of a civic brain-storming session, and the explicit exclusion of members of political parties to participate in either the National Gathering or to stand for elections for the Constitutional Council. The citizen-driven constitutional revision process is unique in any established democratic society” (Bergsson and Blokker 2014, 161). In short, it seems to me, that the Icelandic con­stitutional revision process institutionally approximated quite closely the ideal of reasons-responsiveness in manifold ways.^[283] It also achieves this responsiveness pre­cisely by institutionally approximating the other ideals I have highlighted of inclu­sive sensitivity, democratic co-authorship, political equality and structural independence—and its operationalizability is shown by the fact that it has worked.

In considering a few constituent assembly processes, we have then evidence of both successes (Brazil and Iceland) as well as failures (Venezuela and Bolivia). However, when viewed with a bit more discernment and in the light of striking new empirical evidence, the divergence of the cases might be explainable in a way that precisely supports the stress I have been placing on ab initio democratic input into the substantive content of constitutional change proposals, as opposed to mere ex post democratic ratification. An important recently published paper establishes the crucial causal importance to a polity’s future prospects for democratization of the presence or absence of substantive and widely inclusive democratic input during the drafting stage of constitution-making processes (Eisenstadt et al. 2015). Comparing the outcomes of 138 constitutions in 118 different countries over the last 40 years, the study focused on two questions. First, does a high level of demo­cratic participation in general in the constitution-making process make any difference to the prospects for democracy in that country after promulgation? Second, does citizen involvement in the earliest drafting stages of constitution-mak­ing lead to differences in prospects for democracy, in contrast to citizen involve­ment during later stages of debate on and ratification of elite drafted proposals? Their evidence is quite striking: the answer to both question is yes, and democratic involvement at “the earliest stage, drafting, has a greater impact on democratization than the debate stage or the modalities of ratification” (Eisenstadt et al. 2015, 599). These results seem to confirm a basic hypothesis of deliberative conceptions of democracy, namely that when it comes to constitutional change “direct [ab initio] participation through public debate is more important than [ ex post] voting for deep­ening democracy” (Eisenstadt et al. 2015, 593). While I certainly do not want to claim that this single causal factor helps explain all of the relevant differences between constituent assembly successes in Brazil and Iceland and failures in Bolivia and Venezuela, the evidence from my four case studies is largely congruent with the broader trends found through comparative constitutional analysis. The failed pro­cesses incorporated inclusive democratic participation only while debating and rati­fying proposals that had already been drafted by small groups of elites dominated by a single party; the successful processes, by contrast, incorporated wide, inclusive and diverse participation at the drafting stages as well. It seems a quite plausible hypothesis, then, that institutions of constitutional change that incorporate demo­cratic input into substantive constitutional content as it is being drafted in fact embody the crucial deliberative ideal of reasons-responsiveness, and embody it in a way that effectively contributes to a polity’s on-going democratization.

9.4