Keywords

11.1 Introduction

Constitutions, which regulate many of the fundamental structures of government and enunciate certain societies’ more revered values and principles, are central to political life. It is the expectation that all political actors, in democratic societies at least, operate within constitutional structures, which represent “high-laws”, superior to and, in the event of conflict, having priority over other rules or laws. Constitutions are, generally, “laws for making law” (Kelsen 1945: 124; Congleton 2003: 11). Determining the allocation, sharing, and limitations of political power renders constitutions critically important for understanding the role of political parties in modern democracies.

Constitutions shape many of the challenges political parties facing today. For example, constitutions typically specify at least the broad parameters of the electoral system—the all-important means for translating votes into legislative seats or the means for selecting other elected officials for office (Farrell 2011). Constitutions thus shape what political parties must do in order to win elected office. Shaping those electoral systems through constitutions contributes to the configuration of the party system (Duverger 1954; Lijphart 1994). By extension therefore, constitutions determine the degree to which political parties must compete or cooperate in governing a country by determining the prevalence of single party or coalition governments and majority/unified versus minority/divided governments (Lijphart 1999). Constitutions also tend to specify whether or not the system of government is presidential, parliamentary, dual executive and unified, or federal in nature, any of which determine the degree to which a party or groups of parties seek to control, or share political influence (Elgie 1998; Müller 2002). Some constitutions even provide for banning or restraining “extreme parties” or “anti-democratic parties”, even in liberal democracies (Pildes 2010).

Surprisingly then, a degree of obscurity surrounds the origins of constitutional design. Exceptions certainly exist, not least the understanding of the intentions and preferences of some of the Founding Fathers who drafted the United States Constitution—a topic closely studied by historians, scholars of American political development, and constitutional lawyers (Kelly 1983). Yet as Elster (1995) observed, constitution-making is a stagecraft not well studied or well understood.

If the constitutions’ origins are obscure, so are the dynamics of constitutional evolution and constitutional change. Constitutions are living documents. Change can occur in many ways, for example, by judicial interpretation and activism or by formal changes in the wording of the written texts. Some political systems forego amendments or updating in favour of complete constitutional overhaul: The French Republic has a history of both moderate change through amendment and more complete change, such as the adoption of the 5th Constitution in 1958 following the relatively short-lived constitution of the 4th Republic (1946–1958). More recently, in the wake of its banking and financial crisis, Iceland established a constitutional convention which proposed adopting a new constitution (Hardarson and Kristinsson 2011). Recent regime change in the Middle East and Africa has resulted in new constitutional orders (Rubin 2004; Carey and Reynolds 2011). In short, constitutional change is a reoccurring feature of many well-established democracies and increasingly prevalent with each new wave of democratization.

The aim of this chapter is to explore the politics of constitutional change and, in particular, the role of political parties, who tend to be active agents in this process. An understanding of political institutions and policy outcomes requires an understanding of the influence of preferences and interests of partisans for shaping and reshaping constitutions.

Political scientists have been slow to appreciate and understand the nature of constitutional change, despite a renewed interest in the origins of political institutions as part of the neo-institutional revolution (Peters 1998). As explored subsequently, legal scholarship tended to be the source of most understanding of constitutional change. This has had important consequences for the evolution of this field of inquiry. Even political scientists studying the rules governing constitutional change and the rates thereof have tended to ignore the role of political parties and party systems in configuring the agendas for change.

The next section describes the mechanisms by which constitutions evolve and change. While primary attention focuses on amendments, political parties potentially play key roles in constitutional change through other avenues, such as influencing judicial interpretation through judicial appointments. Section 11.3 reviews the lines of development in understanding why constitutional change occurs, with an emphasis on formal changes to the constitutional texts. Perhaps, most notable is the observation that political parties have been largely sidelined in accounts of constitutional amendments, which may partly explain the lack of success in explaining observed patterns of amendment. Section 11.4 suggests the need to “bring the party in” and proposes that incorporating the preferences of parties and the shape of the party system can advance understanding of constitutional change. The introduction of a number of brief qualitative cases provides evidence of the degree to which constitutional change can be a party-driven phenomenon. Moving from the specific to the general, it is suggested that the veto-player approach provides opportunities to better understand the role of political parties in constitutional change. The chapter concludes with suggestions for further research and the need to move beyond treating constitutions and constitutionally mandated political institutions as exogenous variables when trying to understand the impact of political parties on politics and public policy.

11.2 How Constitutions Change

Virtually, every political system allows for modification of its constitution, at least occasionally. Economic, technological, and demographic changes within which the political system operates may render present structures and obsolete rules. Political preferences are not constant over time, and associated changes in values and attitudes of the population may generate a need to update any principles and ideals enshrined in the constitution. For example, as the awareness of human rights has grown over the last decades, many amendments to constitutions include an increasing range of individual rights and freedoms. Also, major realignments in the political arena may expectedly generate demand for institutional reforms. Citizens may seek to modify the system arising from awareness of unintended, unexpected, and unwanted consequences of current constitutional texts.

Four main types of change in the constitutional arrangement of a country exist, as shown in Fig. 11.1 (see also Voigt 1999: 70; Giovannoni 2003). The foundation of the simple matrix relies on two dimensions. One dimension focuses on the formality (altering the text or not) of constitutional change, with the other dimension focused on its legality (legal change in a strict sense or not). As indicated, this gives us four possible combinations.

Fig. 11.1
figure 1

Amendment rates (yearly) for select countries. Sources Long-term series based on Lutz (1994, 1995) are corrected for Denmark and have been updated for Norway (1814–2001), Sweden (Instrument of Government only, 1975–2000), and Germany (1949–1994). Short-term series 1993–2002 are taken from Lorenz (2005), Table A3. (No data for Iceland and Luxembourg). Notes Correlations between series: Pearson’s r = 0.072 (sig 0.776) and Spearman’s rho = 0.580 (sig 0.012)

Full size image

The first possibility is revision or replacement of the constitutional document by means of a formal amendment procedure specified in the constitution itself (Bergman et al. 2003, pp. 120–127). Although amending processes are often strikingly complex, usually a relatively small set of devices are common among constitutions around the world (see Maddex 1996).Footnote 1 Appendix I outlines the formal amendment process for a number of selected countries. Several scholars suggested ways to summarize the complexity of rules governing constitutional amendments and developed lists of hurdles for constitutional amendments. For example, Elster (2000: 101) applied the categories: absolute entrenchment, adoption by a supermajority in parliament, requirement for a higher quorum than for ordinary legislation, delays, state ratification (in federal systems), and ratification by referendum. Hylland (1994: 197) identified four main techniques: delays, confirmation by a second decision, adoption by qualified majorities, and participation from actors other than the national assembly. Lane (1996: 114) listed six mechanisms: no change permitted, referendum, delay, confirmation by a second decision, confirmation by qualified majorities, and confirmation by sub-national government. Lutz (1994: 363) differentiated between four general amendment strategies: legislative supremacy, intervening election (double vote), legislative complexity (referendum threat), and required referendum or the equivalent.

The various instruments provide constitutions with different degrees of rigidity. In other words, the inflexibility of constitutions depends on the difficulty of overcoming formal amendment provisions. The rigidity of amendment processes, in turn, reflects a previous commitment by political forces to entrench certain political structures and values. Rigidity assists in providing commitments with credibility. This technique institutes a higher legal system that will stand above and limit ordinary legislation (Ferejohn 1997). On the other hand, if amending the constitution is too difficult, change by other means becomes more likely (as discussed below).

The second possibility is (gradual) revision of the constitutional framework by means of judicial interpretation. Most constitutions require interpretation because the language of constitutions is often vague and non-specific. Moreover, constitutions may contain internal inconsistencies, with seemingly contradictory sentences or articles. Typically, a country’s legal system has the responsibility for being the ultimate arbiter of constitutions’ interpretations in cases of conflict. Usually, the Supreme Court or in some countries a special Constitutional Court stands at the apex of the legal system, with the power to render final judgements for the meaning of the constitution (Epstein et al. 2001).

Constitutional jurisprudence and the politics of constitutional change via judicial adjudication are perhaps most closely associated with the United States. The 1803 landmark Marbury versus Madison decision of the US Supreme Court established the principle of judicial review (Murphy 2000). However, not all agree with constitutional change via judicial interpretation: Literalism is a judicial and political philosophy, which suggests that decisions of constitutionality ought to be based solely on the written text of the constitution (Kannar 1990). Wording should not gain credence by conjecturing the drafters’ intentions or revision caused by changing society, technology, or political developments. In contrast, originalists demand that judges faithfully interpret the discoverable intentions of those who drafted the constitution (Whittington 1999). Such conservative and minimalist-oriented perspectives argue that the judicial system should discount interpretation of a constitution in the context of modern society. Constitutional change through interpretation has the reputation of being an unwanted opportunity for judges to dictate public policy and act more as lawmakers than judges. In contrast, judicial pragmatists argue that judges should make decisions reflecting the needs of modern society and politics, even if this requires forgoing reliance on the written word or interpretation of intent (Chemerinsky 1997). Advocates of the “living constitution” argue that in the absence of judicial pragmatism, constitutions would become obsolete.

Variation in jurisprudence concerning constitutional change underlies deep conflict in American politics and society regarding the appropriate role of the United States Supreme Court in upholding and interpreting the constitution. For many, the expansion of individual rights under the Warren Court affirmed their fear of a juristocracy—the idea that the Supreme Court interprets the constitution to the extent that the Supreme Court itself becomes a political institution (Hodder-William 1992).Footnote 2 The decision in Roe versus Wade divided the country’s population and ever since remains the subject of questions from senators to perspective Supreme Court justices (Kastellec et al. 2010).

Indeed, party politicization of the selection process for Supreme Court justices in the United States reflects the degree to which voters and politicians accept the Supreme Court as the protector of, or threat to, the constitution (Moraski and Shipan 1999). Recent presidents eagerly nominated Supreme Court justices who align themselves closely with the president’s policies and attitudes towards constitutional law and change. An American president, serving no longer than 8 years, may have continuing influence long afterwards from decisions and judicial philosophies of their Supreme Court nominees (Gibson and Caldeira 2009). Clearly, individual politicians (the president and senators) and political parties in American seek to influence and shape the level of constitutional change by controlling nominations and Senate confirmations (Segal and Cover 1989). Of course, from a principal-agent perspective, politicians may err and appoint justices who then behave at odds with their appointers’ political philosophies (Szmer and Songer 2005).

Although the politics of judicial interpretation is perhaps the greatest in the United States, judiciaries in other countries have also developed the notion of judicial interpretation. Although the Norwegian Constitution does not mention judicial review, the courts introduced it through interpretation during the first half of the nineteenth century (Smith 1993). A truly comparative framework to measure the level of constitutional change via judicial adjudication is still wanting, but area or country-specific studies highlight the prevalence of the practice in places as diverse as Asia (Ginsburg 2003), Chile (Couso 2003), Germany (Kommers 1997), Hungary (Brunner 2000), and Mexico (Domingo 2000).

The third possibility is revision or replacement of the constitutional text by irregular means. The 13th and 14th Amendments of the US Constitution in the 1860s, which emancipated the slaves and bestowed suffrage (Mueller 1999), are illustrations. Since southern states had sufficient votes to block change, the amendments would have failed ratification if the process, established in Article 5 of the constitution, had strictly followed its dictates. A similar example of ignoring the formal amendment procedure is the changes in wording of Article 1 of the Norwegian Constitution in November 1814, reflecting the union with Sweden, and in 1905, marking the dissolution of the union. Likewise, many questioned the legal basis of the 1962 referendum to elect directly the President of France. Some suggest that the procedure used by de Gaulle’s to call the referendum was extra-constitutional (Stone 1992). Popular opinion, particularly if expressed through a plebiscite or referendum, can render decisions valid even if the outcome is contrary to structures governing constitutional change. Currently, no known systematic measure for the level of constitutional change through such extra-legal behaviour seems to exist.

The fourth and final possibility mentioned in Table 11.1 is an intended or unintended revision of the constitutional framework by means of political adaptation by legislative and executive bodies. An important example in Norway and many other European countries is the introduction, or rather evolution, of forming parliamentary government (e.g. Congleton 2001). The example is, however, ambiguous, as the Norwegian case illustrates. The first instance of formation of parliamentary government in Norway occurred as early as in 1884, but this change had no reflection in a revision of any article in the constitution. After a generation or two, lawyers and politicians came to accept parliamentarism as a constitutional custom to which governments must abide. In other words, parliamentarism became a constitutional principle even though the constitution itself had no mention of it whatsoever. In 2007, however, negative parliamentarism—practiced consistently for over a hundred years—gained codification (Article 15). In general, the lack of reference to political parties in many constitutions, despite the centrality of party government, indicates a gap between the formal constitution and the practice of constitutional government. As indicated earlier, several examples exist for constitutional changes that do not follow the formal regulations established in the constitution. However, reasonably, such examples are rare, at least in established democracies, and irregular forms of change in the constitutional text represent highly exceptional circumstances.

Table 11.1 Main types of constitutional change
Full size table

11.3 Measuring and Explaining Constitutional Change

Interest in political aspects of constitutional change and the characteristics and effects of amendment procedures has expanded over time. Simultaneously, empirical measurement of constitutional stability faces significant challenges. A new constitutional order following regime change may be very noticeable, but measuring constitutional change through judicial interpretation and especially (often unobservable) political practice is particularly difficult and has not been comparatively investigated in any significant detail. Consequently, the nature and causes of constitutional change tend to focus on formal amendments.

Even comparative (cross-national) literature, which relies primarily on amendment rates—for instance yearly averages—to indicate the degree of change in constitutional rules over time, is sparse. Figure 11.1 shows such amendment rates for a selection of countries. The data are from Lutz (1994), updated, or corrected for some countries, and Lorenz (2005). The latter is based on the time 1993–2002. Lutz used the entire lifespan of constitutions from their origins until the early 1990s. As a glance at the data in Fig. 11.1 confirms, the two time series are oddly unrelated (Pearson’s r = 0.072).Footnote 3

Lorenz (2005: 351) found Lutz’ data for Germany, France, and Ireland to be inaccurate. Another explanation for the different rates reported in Fig. 11.1 is that different authors apply different operational definitions of “change” in their calculations. Counting instances of amendments and identifying a single instance of constitutional change may seem to have obvious answers, but the reality is the opposite. In the context of the US Constitution, identifying an amendment is relatively easy, since each amendment—so far 27—appears at the end of the constitution; the wording of the original document has no revision. Some of the amendments are rather broad and complex, with several sections (e.g. the 14th Amendment), and only one amendment occurs at a specific ratification date, except for December 1791, the ratification of the first ten amendments, which represent more than one-third of the total number of amendments. Perhaps, counting the first ten amendments as a single change in the constitution is reasonable, or perhaps, counting some of the later amendments as more than one change is also reasonable?

In contrast to the American case, changes take the form of textual revisions in most national constitutions, sometimes as simple as deleting, changing, or adding one word in a single sentence of an article. If we observe simultaneous reformulation of several articles in different parts of a constitution, is it still to be counted as one change? Norway is an example: In February 2007, amendment occurred to several articles. In one committee report based on proposals submitted before the 2005 election, recommendations suggested changes to six articles (two of them new articles); the assembly unanimously accepted them later. Most of the changes concerned the court of impeachment and the legal foundation of its operation.Footnote 4 On the very same day, abolishing the quasi-bicameral organization of the parliament (against through just one vote) involved revisions of wording to seven articles in different parts of the constitution. The actual number of changes is obscure: one, two, or, perhaps, thirteen. The amendments concerned two issues and required two roll-call votes. Had preferences been more diverse, allowing some legislators to support only various subsets of the articles’ amendments, as many as thirteen or more roll-calls, could (and would) have been arranged. In any case, counting constitutional reform issues (“packages” involving several articles) or counting numbers of changes to single articles produces different empirical measures of constitutional change. In the 2007 example, counting the changes as only two might seem reasonable; however, reproducing the Lutz (1994, 2006) amendment rate is not possible this way.

Regardless of the method for counting amendments, distinctions between small and large reforms, or between important or unimportant changes, may be significant. If the interest is the extent to which amendment procedures affect changes to the status quo, the point becomes significant. Symbolic, small, or virtually inconsequential amendments do not represent real changes in the constitutional status quo. In a sense, such reforms constitute distorting “noise” in the data. For example, in 1962, establishment of the office of the Parliamentary Ombudsman occurred in Norway, and the constitution (Article 75) enshrined the office in 1995. The actions produced exactly what existed before, and the reforms substantially changed nothing (although, of course, the Ombudsman from then on gained constitutional protection, and the office could not be abolished by a simple majority). The new Article 15 from 2007 (parliamentary government) was simply a codification of constitutional custom and did not represent any change in the status quo.Footnote 5 Arguably, the 22nd amendment of the US Constitution also was insignificant, in that it simply codified a political norm that only one (Franklin D. Roosevelt) of the previous presidents had disregarded; Roosevelt could have remained the only exception even if the amendment had failed to be adopted.

The literature identifies several measures of constitutional rigidity. The most complex approach appears in Lutz (1994, 2006, pp. 145–182) who established an index of difficulty by specifying the added value of nearly seventy aspects of amending rules. The index ranges from 0.50 (New Zealand) to 5.10 (the United States) in Lutz’ cross-national material (N = 32 countries). Although the approach is highly systematic, some of the results are counterintuitive. Constitutional scholars seem to agree that amending the Danish constitution is very difficult, but the entry for Denmark (2.75) is close to the average (2.50). Amending the Norwegian constitution is certainly easier, but the index value is significantly higher for Norway (3.35). According to the index, amending the Japanese constitution (3.10) is easier than the Norwegian constitution. This is clearly erroneous, since Japan requires a two-thirds majority in each legislative chamber as well as a referendum; Norway requires a two-thirds majority once, but after an intervening election (no majority decision before the election). Austria and Portugal require two-thirds majorities in single legislative decisions, but appear close to majoritarian New Zealand on the index (0.80 versus 0.50 for New Zealand). Clearly, the index is not entirely satisfactory; it is, perhaps, overly complicated, and on their face, some of the resulting scores lack validity.

Lijphart (1999: 219) reduced the great variety of methods for amendment to four basic types: ordinary majorities, between two-thirds and ordinary majorities, two-thirds majorities or equivalent, and supermajorities greater than two-thirds. In effect, that research disregarded aspects of amending provisions other than majority requirements and focused on a one-dimensional approach. Anckar and Karvonen (2002) suggested a slightly more complex measure with nine values, involving either the legislature or the people (in referendum) or both (or even none) in constitutional changes. If involved, the requirement is either an ordinary or a qualified majority. Crossing the dimensions gives nine cells, but the numbering of them (which represent the values of the resulting rigidity variable) is arbitrary and difficult to validate. Lorenz (2005: 346) created a two-dimensional additive index; a slightly modified version of Lijphart’s measure combines scores for the number of “arenas with different voters”. In a set of 39 countries, the index ranges from 1 to 9.5.

The current study cannot attempt to resolve the debate surrounding measurement of constitutional change, but the discussion of the causes of constitutional change requires sensitivity to the fact that employing different measures of constitutional change provide evidence for and against competing explanations.

In a cross-national analysis, Lutz (1994) demonstrated that the degree of flexibility or rigidity of a constitution influences the amendment rate. Leaving aside Lutz’s measures for a moment and referring to Appendix I, New Zealand is prominent as an example of a country with a flexible constitution.Footnote 6 Japan, the United States, Finland, and Greece clearly are more rigid because they require qualified majorities in one form or another, as well as referenda or an intervening election. The amendment procedures in Denmark, France, and Italy may also—less obviously—gain consideration as examples of quite rigid rules. The source of rigidity in Denmark is the referendum requirement and the fact that at least 40 % of the electorate needs to vote in favour of a constitutional amendment for it to pass. This is actually a super majoritarian element. In France, the main procedure involves the president as a veto player; alternatively, the referendum is used. Italy requires double decisions in both chambers (initiation of a referendum occurs if decisions in the national assembly are by majority). Ireland and Sweden have a multiple actor approach within a majoritarian framework, whereas Germany and Portugal allow qualified majorities of the legislatures alone to amend the constitution.

Focusing on US State Constitutions, Lutz (1995) found that the more the procedural difficulty in amending state constitutions, the more the amendments were made. This counter-intuitive finding supports his earlier cross-national work (Lutz 1994) that found the relationship between ease of amendment process and rate of amendment to be negative and curvilinear. For that study, Lutz used information from 36 national constitutions. The dataset included a wide range of countries, from Western Samoa (1962–1984), Kenya (1964–1981), and Argentina (1853–1940) to many well-established Western democracies. Beyond measurement problems, a common criticism of the Lutz research is the lack of control variables employed: To some extent, only length and age of constitutions were controlling variables [in most studies, both the length of the constitution and the age of the constitution correlate positively with rates of amendment, as Dixon (2011) notes].

After disaggregating the Lutz index of difficulty, Ferejohn (1997: 523), in a reanalysis, claimed that “[T]he requirement of special majorities or separate majorities in different legislative sessions or bicamerality—is the key variable to explaining amendment rates”. He continued by saying that “[T]here is no evidence that a ratification requirement, whether involving states or a popular referendum, has any significant impact on amendment rates” (Ferejohn 1997: 523). In other words, special majorities in the legislature may be both necessary and sufficient to achieve a moderate amendment rate. Lorenz (2005) considered the effects of several measures of rigidity on both of the amendment rates. The results appear mixed, especially with regard to the 1993–2002 amendment rates.Footnote 7 Rasch and Congleton (2006) reanalysed the Lutz amendment rates for a small set of 19 OECD countries. Veto authorities (or points) and supermajority requirements are among the variables, but the latter surprisingly has no significant effect on changes in any of the models. Dixon and Holden (2012), revising the issue of amendment rules in state constitutions, found that having a supermajority requirement at the legislative stage reduced the rate of constitutional amendments. In addition, US States’ Constitutions tended to be amended more when states allowed popular initiatives, such as in California.

Clearly, the results from studies seeking to explain the rate of constitutional change through formal amendments are ambiguous and inconclusive. Overall, it has not been demonstrated that the pace of change in constitutions in practice decreased with higher hurdles of constitutional amendment procedures. More reliable cross-nation results do however require that the measurement issues related to the dependent variable (amendment rate) and the main independent variables (various aspects of rigidity) be addressed. Some other challenges require consideration: First is the question of selection bias. Empirical studies so far exclusively focus on successful reforms, that is, the cases for study are selected on the basis of outcomes on the dependent variable (Geddes 2003). Either failed constitutional proposals need inclusion in the analysis, or negative cases in the form of periods or years of stability need consideration (e.g. by using country-years as observations). Second is the problem of controls. At the present stage, then, achieving reliable results in single-country studies (e.g. analyses of time series) with carefully crafted comparative case studies might be easier. In small-N comparative designs, avoiding selection bias and selecting only reasonably similar cases are perhaps possible.

None of the studies so far has included any extra-constitutional explanatory variables, which means those variables not generated by the constitution itself (as its age, length, and, of course, amendment procedure). Additionally, avoiding incorporation of political actors’ interests and preferences in seeking to explain change would produce skewed results. Noting the general absence of political parties from previously proffered theories, the next consideration is attending to the role of political parties in constitutional change.

11.4 Bringing the Party in

As discussed earlier, most of the existing quantitative literature on constitutional change largely ignores political parties. This section discusses the rationale for, and means by which, political parties ought to have consideration as central to the process of constitutional change via formal amendment. Arguably, the interests and preferences of parties and the shape of the party system are necessary elements of any model of constitutional change. To understand the influence of political parties on constitutions, our initial spotlight is on one common constitutional prescription—the electoral system—as a focus for political parties’ agendas for reform and change. The Irish case is used to illustrate the relationship between political parties, electoral systems, and constitutional reform. A discussion of how scholars of electoral system change have incorporated constitutional amendment procedures into their research follows. The section concludes with suggestions for how the veto-player framework provides insights into how parties and the party systems shape demand for, and patterns of, constitutional change.

The case of Irish electoral reform highlights the degree to which political parties can actively promote constitutional change in an attempt to advance a party’s interests. The 1937 constitution was the brainchild of the Fianna Fáil party and in particular its charismatic leader, Eamon de Valera. That constitution included an article directing that the electoral system be proportional representation by means of single transferable vote (STV). As Gallagher (1987: 27) noted, STV came to Ireland earlier in the century at the behest of electoral reformers in England and became well established in Ireland as the preferred electoral system by the 1920s. During the 1937 constitutional debate in the Irish parliament, de Valera indicated that he wanted the constitution to reflect the details of the electoral system, “as the matter was too important to be left to the vagaries of party warfare” (Sinnott 2010: 113). Notably, de Valera implied that in-office politicians recognize the enticement of future politicians to manipulate the electoral system for partisan gain. Enshrining electoral rules in the constitution was a way of ensuring serving governments could not change the electoral rules for future electoral benefit.

Within two decades, de Valera proposed a constitutional amendment to change the electoral system with the aim of enhancing his party’s electoral fortune. In 1959, Fianna Fáil sought to switch from a proportional representation system to a plurality system. Fianna Fáil had maintained power for virtually all of the time, but of the nine governments formed since 1937, the party secured majority status on only four occasions (Sinnott 2010). Fianna Fáil and de Valera wanted an electoral system that would produce for greater electoral rewards for themselves. In the subsequent referendum, voters narrowly rejected the amendment. The narrowness of the result motivated Fianna Fáil to attempt the same constitutional reform within the decade; the results again were a defeat at the referendum stage—this time by a much more significant margin. What this case highlights is the degree to which partisan interest can motivate the desire for constitutional change. The demand for constitutional change among political actors, including political parties, is likely an important factor in determining the rate of attempted constitutional change.

While constitutional scholars generally ignored the preferences of political parties in shaping and reshaping constitutions, scholars of electoral studies have long been aware of the desire of political parties to influence constitutional rules. Boix’s (1999) seminal contribution is a reminder that self-interest is central to political parties’ preferences for constitutional design and, by logical extension, redesign. Parties are strategic actors who attempt to shape political structures to maximize the potential future access to power. The earlier mentioned Irish case clearly illustrates this preference of political actors in this regard. The frequent changes to the French electoral system by incumbent parties attempting to gain a political or power advantage (Elgie 2005) are further evidence of the general desire parties harbour to “fix” constitutions to maintain or expand upon elected office.

Shugart (2008) suggested that demand for constitutional change by political parties may be driven less by a desire to hold future power than by a sense of dissatisfaction with the past performance of established constitutional structures. Shugart asserted that parties who assumed power and who tended to suffer electoral disadvantage (in terms of the disproportionality between seats obtained and votes won) for a long period are most likely to seek constitutional change to remedy what they perceive as unfairness in the system. Britain’s Liberal Democrats seemed to follow this pattern exactly. Shortly after entering into a coalition agreement with the Conservative Party, a referendum on electoral reform occurred in May 2011 with the proposal that the alternative vote replaces the single-member plurality system. Neither the Labour Party nor the Conservatives were strong advocates of reform (although the Labour Party leader campaigned for electoral reform in 2011 and the Labour Party agreed with the Liberal Democrats to electoral reform before the 1997 general election—but not pursued once in power). The Liberal Democrats in government placed their trust in voters to reform Britain’s electoral system—but in a significant defeat for the party, voters decided against such change.

Renwick (2009) observed a crucial phenomenon: Constitutional reform regarding electoral systems is elite-driven, and voters tend to have weak preferences for initiating change. Why we do not see more political parties using their time in government to propel constitutional change, in particular electoral reform that would enhance the incumbent’s advantage, remains unresolved. Pilet and Bol (2011) found evidence that self-interest mixed with an analysis of risk and satisfaction drives political parties’ demands for electoral reform. Political parties know the impact of current constitutional features and may be slow to alter these, even if change would be beneficial to the party in power. As such, political parties may act conservatively in terms of constitutional change. Psychological factors may also be at play—mirroring Shugart’s (2008) “dissatisfaction” explanation, Pilet and Bol (2011) determined that the degree to which political parties feel cheated by the system partly determines the desirability to change the electoral system. Interestingly, when considering the rate of electoral system change, research on constitutional electoral reform fails to consider the likely impact of rules governing constitutional amendments. Indeed, in general, scholars of party politics generally ignored the institutional obstacles to constitutional change.

One exception is Hooghe and Deschouwer (2011) who explore the difficulty of constitutional change in the Belgium system. The consociational nature of the Belgian Constitution makes amending the constitution very difficult. Regionally defined political parties are effective wielders of vetoes (see below). As a result, and despite interest among political parties in power to reform the constitution, very little has changed in Belgium. The significant point is that scholars of political institutions must account for both the preference of political actors and the processes by which constitutional change can occur.

As Colomer (2005) suggested, the party system is as likely, if not more likely, to shape the constitutional structure, as the constitutional structure is likely to shape the party system. Colomer’s “behavioural–institutional equilibrium” theory argued that existing parties attempt to ensure (new) constitutions maintain the status quo in terms of party systems. Although political parties are fundamental to representative government in Europe, scholars with a different regional focus have been quicker to recognize the significance of political parties’ influence on constitutional change. Many Latin American countries have experienced constitutional shifts both away from plurality electoral rules and towards stronger presidential power. Negretto (2009) argued that pre-existing party competition and party organization are crucial variables for explaining the substance of constitutional change. Specifically, higher levels of factionalism in the party system lead to relatively more inclusive electoral rules. Similarly, party decentralization associates with strengthening the president vis-à-vis the legislature. In short, the preferences of existing parties in power shape the type of amendments proposed.

Beyond the preferences of political parties over constitutional change, one way to consider the likely actual impact of political parties on constitutional change is to apply the veto-player approach (Tsebelis 2002) to constitutional amendment procedures. By using this approach, and considering political parties as one of many potential sources of vetoes, identifying connections between the different procedural devices and roles of political parties as actors in constitutional redesign becomes easier. Furthermore, it is so general that any part of amendment procedures can be discussed with respect to its effect on one important variable: the capacity or potential for change in the status quo (i.e. change in the present constitutional norms at any point in time).

According to Tsebelis (2002: 19), veto players are “individual or collective actors whose agreement is necessary for a change in the status quo”. Veto players can be either institutional or partisan (for an overview, see Strøm 2003: 77). In our context, the former type is specified in the amendment clause of the constitution. The parliament, a legislative chamber (in a bicameral parliament), voters in a referendum, a constitutional court, or a president are typical examples of institutional veto players. Parties or other actors inside an institution are (potential) partisan veto players. A disciplined majority party within an assembly that renders decision based on majority rule is an example of a partisan veto player. Disciplined political parties are a common feature of most European parliaments and many legislatures in other parts of the world (Depauw and Martin 2009). Certainly, identification of partisan veto players may be problematic and ambiguous (Ganghof 2003), and the ambiguousness may partly explain the focus on institutional veto players in the literature. Still, identification of individual veto actors inside institutional (collective) veto players is necessary to gain accurate understanding of the consequences of constitutional amendment procedures. Veto players create constraints on decision-making, and therefore, accounting for all relevant actors is a requirement for in any complete model of institutional change.

Political actors try to further their interests, and political parties with veto power should be expected to block proposals that go against their interests. Veto players will not accept changes that make the status quo worse from their perspectives, and no changes occur. A potential for change only exists if all veto players prefer certain outcomes that modify the status quo, that is, if the winset of status quo (the set of alternatives or outcomes that can alter the status quo) is not empty. Figure 11.2 illustrates the set of constitutional amendments that can replace some part of an existing constitution, for instance, one of its articles. The size of the winset is, perhaps, a proxy for stability (Tsebelis 1995: 295). If the winset is empty, the situation is stable; no relevant actor prefers to overturn the status quo. If the winset is small, only incremental changes are possible. The existence of transaction costs and external constraints may also preclude changes in this situation. If the winset is large, a lot of proposals potentially can defeat the status quo. Thus, the larger the winset of the status quo, the more susceptible to change the current constitutional framework becomes.

Fig. 11.2
figure 2

Ideal points of Actors A, B, C, and D. Status quo (SQ) and a constitutional proposal in the winset of status quo WAB(SQ)

Full size image

Repeated decisions are a common technique in constitutional amendment procedures. For example, a parliament needing to render a decision for any constitutional amendment twice creates, in a sense, a parliamentary status at time t 1 and at time t 2, which represent two veto players. Amendments to the Italian constitution require adoption by each of the two parliamentary chambers (Chamber of Deputies and the Senate) twice. The Swedish Constitution can only be amended if the Riksdag approves the changes twice, with one general election having been held in-between the two votes. Another way to describe it is to say that it is two veto points with the same institutional veto player. Any proposal encounters defeat in either the first or the second vote; consent (by a sufficient number of members) on both occasions is the requirement for altering the status quo. Preferences of course may change between the two points in time. This can occur simply due to the passage of time, the tempering of passions, the presence of new information, or an electoral event. In the last instance, the composition of the second parliament may differ from the first with respect to members and party composition. Ideological differences between the parliaments (at t 1 and t 2) may or may not appear as a consequence of the intervening election. This in turn has the potential to affect stability by making adoption of constitutional amendments more difficult. A recent trend in well-established democracies is increased instability at the polls (higher volatility), which creates difficulty for amending those constitutions that require consent of the pre-election and post-election parliaments. Thus, an easily overlooked external factor (shifts in the partisan composition of one or more chambers) may significantly affect the difficulty of amendment processes.

Choice of agenda-setting rules for constitutional proposals is an alternative way to highlight the impact of political parties on constitutional change. The potential power of the agenda setter is illustrated by Romer and Rosenthal (1978), who formulate a setter model with two players—a proposer and a veto player—and two stages of decision-making. In the first stage, a committee—in our case, one which formulates a constitutional amendment—sets the agenda by introducing a proposal to the parliament. Then, in the second stage, the parliament votes on whether to accept the proposal. Political parties are likely to dominate at least one of these stages. If the parliament uses its veto, rejecting the proposal, the status quo prevails. In the model, the parliament as a second-stage actor is not allowed to amend the first-stage proposal. Thus, the decision-making power of the parliament is severely restricted and actually reduced to a take-it-or-leave-it choice. If the ideal points of the proposer and the legislative assembly (median legislator) deviate, the agenda control described above makes it possible for the proposer to move the status quo towards its own ideal point.

The agenda setter in the Romer and Rosenthal (1978) model has both positive power and negative power over the agenda (Cox and McCubbins 2004; cf. also Denzau and Mackay 1983; Shepsle and Weingast 1987; Heller 2001). Positive power over an agenda is the authority to propose changes to the status quo and to ensure that these proposals become part of the legislative schedule for consideration. Negative power over an agenda is the ability to prevent certain proposals from entering the legislative docket (gate-keeping power), the ability to delay considerations of proposals (a weak form of gate-keeping), or the ability to block changes to the status quo (veto power). In the setter model, the first mover has proposal and gate-keeping power; if it decides to close the gates, no proposal emerges. The second mover can neither introduce proposals nor make amendments to proposals. In the vetoing parliament, proposals are considered under a closed rule rather than an open rule, meaning that no amendments to the original proposal are allowed.

The details surrounding agenda setting are important for the outcome of decision-making processes. In particular, it is essential whether or not a veto player can amend a proposal that is already on the agenda. We can again illustrate this by Table 11.1. Suppose that only actor B has proposal rights and that A is a veto player without rights to amend proposals (as in the setter model of Romer and Rosenthal 1978). Then, B will propose an amendment as close as possible to its own ideal point, that is, P in the figure. Seen from actor A’s perspective, P is marginally better than SQ and P will be accepted. What happens if veto player A can amend the proposals that B places on the agenda? If P is proposed, A can revise it so that the decision reflects its own ideal point. But the ideal point of A is worse than the status quo of the proposer B, and no proposal should be forthcoming in the first place as it is not in the proposers interest. In this case where the veto player operates under open rule, the proposer will only make proposals if the ideal point is preferred to the status quo. Thus, even if the winset of the status quo is non-empty, the situation is entirely stable because of the agenda-setting rules.

Finally and notably, the structure of the party system may also impact the ability of political parties to institute constitutional reform, depending on the type of majority required. For example, the purpose of the device of a qualified majority is to protect the (formal) status quo or the existing constitutional provisions. Obviously, achieving adoption of a constitutional amendment is more difficult according to the degree of majority required. Importantly, the impact of requirements for majority will depend on the party system. Contrast, for example, the likely difficulty of constitutional amendments under qualified majority in countries with many parties of relatively equal size, compared to countries with a dominant party (a political system where one large party tends to dominate politics over a significant period of time). As Dixon (2011) suggested, bipartisanship may be a requirement for constitutional change in a multi-party system, whereas, in contrast, qualified majority rules are largely irrelevant in dominant party systems. The suggestion is, then, that political parties gain relevance not only from their preferences for constitutional change, but also from the configuration of the party system which determines the effectiveness of veto players in decision-making processes.

11.5 Conclusion

Constitutions shape the actions and behaviour of political parties. In liberal democracies, the steps necessary to win elected office and share power are of great concern for political parties, their elected officials, their memberships, and their wider base of support. By defining the rules of political encounters, constitutions create incentives and rewards for political parties. Modern political science focuses heavily on explaining political outcomes by examining the consequences of the preferences of actors’ interacted with instructions and rules. Within this research agenda, the understandable practice has been to treat the rules and structures as exogenous. Citing a political institution or rule as having a constitutional basis is a typical way to assure readers that the institution in questions is truly exogenous of party and political influence.

In reality, political parties are not just shaped by constitutions, but also shaped by constitutions. Political interests shape constitutions. More importantly, perhaps, political interests have the potential to reshape constitutions. Constitutional change can occur in a number of ways, from formal amendment procedures (the focus of this chapter), to extra-legal regime change, a change in the practice of politics, or through judicial interpretation. In each one of these methods, political parties are crucial in many democracies. Partisan politicians decide who ascends to a country’s Supreme or Constitutional Court and in so doing influence, potentially for many generations, the level and nature of constitutional evolution or rigidity. Political parties tend to be, collectively or individually (dependent on the shape of the party systems), potential veto players in most procedures for formal constitutional amendment. Even in situations in which citizens’ initiatives set the agenda for constitutional reform, political parties may be key players in the referendum campaign, as in the case of Switzerland. In many ways then, political parties, as self-interested players in political systems, have the potential to impact constitutional continuity or constitutional change. Accepting this proposition may seem obvious, but the consequences and challenges of this are significant for scholars of constitutional law, for constitutional change, and for political scientists interested in the impact of institutions and preferences. The dominant approach of treating constitutionally prescribed institutions as exogenous to particular models and theories of political behaviour and political outcomes requires rethinking. Political parties are not just shaped by constitutions but actively shaped and reshaped by them. Clearly, further investigation into political parties’ interactions with amendment structures and other political variables is necessary. An understanding of how constitutional change reflects changing preferences or political opportunities is a first step towards a fuller appreciation of how modern political parties impact the organization and operation of politics. However, even for such inquiry, the origins of the mechanisms by which constitutions change cannot be divorced, original rules governing change—which have likely been shaped by parties and the party system.