Italian Political Science, VOLUME 17 ISSUE 1 Published in 2022 under Creative Commons BY-NC-ND license by: Italian Political Science. ISSN 2420-8434. All copyrights of article contents are retained by the authors. Volume 17, Issue 1, 16–33. Contact Author: Luca Pinto E-mail address: luca.pinto@unibo.it Party system change at the legislative level: evidence from the 18th Italian legislature Luca Pinto UNIVERSITY OF BOLOGNA Abstract Party systems and party system change have long been one of the most studied topics within the comparative politics literature, yet most work in this field focuses on changes that occur between elections, neglecting the possibility that parties and party systems may reconfigure during the inter-election period. Building on the studies on party system change, this paper aims to analyse how individual changes in party affiliation can aggregate into changes at the level of parliamentary party system in the 18th Italian legislature. To achieve this goal, data on individual movements in the membership of parliamentary parties in the Senate and the Chamber of Deputies are used to track the extent of legislative party system change over time, as indicated by changes in the number of relevant parties, their relative size and strength, and the bargaining power of parliamentary groups in the Italian parliament. Overall, our results show major changes in the main features of the Italian parliamentary party system and the structure of competition in the 18th legislature, returning the image of a fluid, unstable and constantly moving system. 1. Introduction olitical parties and party systems can be considered the pillars of democracy. On the one hand, according to the famous quotation of Schattschneider (1942), par- ties are a necessary condition for the very existence of democracy. Parties aggregate social interests and organise electoral competition; they guarantee the func- tioning of legislative assemblies, organising the agenda of parliaments and translating preferences into policies. Finally, parties form governments, whose survival depends on their support. On the other hand, party systems largely characterise the quality and func- tioning of democracy (Wolinetz 2006; Casal Bértoa 2017), as the number of individual parties and the way they interact and compete for electoral office and control of the gov- ernment define voters’ alternative choices and influence patterns of coalition formation and the survival of cabinets (Sartori 1976; Mair 1997, 2001, 2006). Given the strong relationship between parties, party systems and the functioning of democracy, a substantial body of work within political science has been concerned with understanding what conditions influence the democratic process, both in terms of its le- gitimacy and effectiveness, identifying one of these conditions in party system institutionalisation (Mainwaring and Scully 1995; Morlino 1998; Mainwaring and Tor- cal 2006; Lindberg 2007; Casal Bértoa and Mair 2012; Casal Bértoa and Enyedi 2016; Casal Bértoa 2017; Chiaramonte and Emanuele 2017, 2019; Emanuele and Chiaramonte P PINTO 17 2018). Moving from Sartori’s (1976, 44) classic definition of a party system as ‘the system of interactions resulting from inter-party competition’, Casal Bértoa (2017, 407) defined party system institutionalisation as ‘the process by which the patterns of interaction among political parties become routine, predictable and stable over time’. According to this definition, institutionalisation is higher when parties compete in a structured and standardised way, thus reflecting a high level of ‘systemness’. On the contrary, party sys- tems are under-institutionalised when the patterns of interactions between parties are not predictable, revealing therefore an ‘open’ structure of competition. In other words, it is only when voters become accustomed to a certain set of political alternatives that a system becomes structured, which also implies that votes are channelled through the party rather than through personalistic leadership (Sartori 1994). In the absence of these conditions, there can be no structure in the party system. This line of reasoning also es- tablishes a connection between institutionalisation and party system stability and change, identifying stability when there is a strong degree of institutionalisation and, conversely, change when there is a shift in the ‘prevailing structure of competition’ (Mair 2006, 66). Although there is a broad agreement among scholars in the general definition of party system change, there is less consensus on how to identify it. These differences arise from how party systems are defined in the first place (for a review see Wolinetz 2006; Enyedi and Casal Bértoa 2020). On the one hand, comparative scholars have built clas- sifications and typologies of party systems (Duverger 1954; Blondel 1968; Sartori 1976), developing criteria that have proven useful in categorising party systems ‘into distinct classes or types, such as two-party systems, systems of moderate pluralism, multiparty systems, or whatever’ (Mair 2006, p. 63). According to this approach, changes imply moving from one discrete category to another. On the other hand, party systems can be summarised by a number of distinct empirical indicators that allow party system change to be tracked as a continuous process (Wolinetz 2006). Finally, Mair (1997, 2001, 2006) emphasises the structure of competition for control of the executive as the key criterion for distinguishing party systems and therefore identifying party system change. Despite different approaches to defining party systems and analysing their changes, most schol- ars agree that the main characteristics of party systems derive almost exclusively from electoral competition, thus ignoring that, between elections, party systems can change due to party mergers, splits and individual party switching of Members of Parliament (MPs) (Heller and Mershon 2009).1 To balance the ‘bias in favour of an elections-dominant understanding of what par- ties and party systems are and do’ (Mershon and Shvetsova 2013, I), legislative party switching has recently become the subject of an expanding literature within political sci- ence. Scholarship studying defections has mainly focused on the factors that drive legislators to change their party affiliation and on the interplay between MPs’ ambitions and institutions (Heller and Mershon 2005, 2008; Desposato 2006; McMenamin and Gwiazda 2011; Pinto 2015; Klein 2016, 2018, 2021; Volpi 2019). At the same time, schol- ars have started to investigate the implications of party switching for parties (Ceron and 1 Focusing on the structure of competition for the executive, Mair (2006) implicitly acknowledged the electoral bias of most work on party system change (see Mershon and Shvetsova 2013, 7-8 for a discus- sion). Party system change at the legislative level 18 Volpi 2021, 2022) and more generally for party systems and party system change be- tween elections (Mershon and Shvetsova 2008, 2013, 2014). This paper is intended to fit into the latter line of research, analysing how individual changes in party affiliation can aggregate into changes at the level of the parliamentary party system. This contribution is therefore mainly empirical and aims to track the evolution of the Italian parliamen- tary party system during the 18th legislature (2018-2022). This legislature is the product of the electoral success of two different populist parties - the League and the M5S - whose fortunes are linked to the multiple crises (political, economic and migratory) that have shaken Italy (and Europe) in recent years (Caiani and Padoan 2021). It is therefore a case worth studying through the lens of party switching, which is often perceived as a failure of democratic representation (Heller and Mershon 2009). To achieve this goal, data on individual movements in the membership of parlia- mentary parties in the Senate and the Chamber of Deputies are used to trace the extent of legislative party system change over time, as indicated by changes in the number of relevant parliamentary groups, their relative size and strength, and the general structure of competition in the Italian parliament. In addition, given the close relationship be- tween stability, change and institutionalisation of party systems, this contribution also adapts two indices originally developed to measure the institutionalisation/de-institu- tionalisation of party systems in the electoral arena − the volatility and the degree of party system innovation (Powell and Tucker 2014; Chiaramonte and Emanuele 2017, 2019; Emanuele and Chiaramonte 2018) − to be applied in the legislative context. It should be emphasised that, by doing this, this paper does not intend to equate party sys- tem change and legislative party system change as two identical concepts. Rather, the aim of this work is to provide more insights into the study of party system change by an- alysing one single dimension of this broader phenomenon, namely the evolution of legislative parties between elections as a consequence of party switching.2 Hence, it can- not say anything about patterns of party competition as it is more commonly understood, i.e. outside parliament through electoral competition and parties’ supply-side strategies. Overall, quantitative indicators show significant changes in the main features of the Ital- ian parliamentary party system and the structure of competition in the 18th legislature, returning the image of a fluid, unstable and constantly moving system. The argument and evidence presented in this paper are structured as follows. After this introduction, the next section offers a brief review of the main approaches to the study of party system and party system change. The third section provides evidence on inter-electoral change in legislative party membership in the Italian parliament. In the fourth section, the main indicators to track the degree of parliamentary party system change are presented and their results are discussed, illustrating whether the changes identified by the quantitative indicators are also reflected in changes in the prevailing structure of competition in the lower and upper chambers. Finally, the implications for the analysis of the Italian case and more generally for a better understanding of party system change are discussed in the concluding section. 2 In the paper, legislative party and parliamentary group are used as synonyms, in the knowledge that the party in public office represents only one ‘face’ of the broader concept that is the party, together with the party on the ground and the party in central office (Katz and Mair 1993). PINTO 19 2. Party systems and party system change: a brief review Starting with the seminal work of Sartori (1976, 44), a party system has been conceptu- alised as the product of regular and recurring interactions among its constituent parties deriving from inter-party competition (see also Wolinetz 2006, 52). The type and quality of these interactions depend on various features which account for the variance observed across different party systems. These features typically include: the number of relevant parties contesting elections and winning offices, their size and relative strength both in terms of votes and seats, the number and the content of the policy dimensions that shape the space of political competition and, finally, parties’ policy preferences and the dis- tances that separate them on the most salient dimensions of competition. Other features on which party systems may differ are related to the level of institutionalisation or de- institutionalisation of the party system itself. This also implies considering the stability or instability of the patterns of parties’ electoral support and the entry of new parties into the competition (Casal Bértoa 2017; Chiaramonte and Emanuele 2017, 2019; Emanuele and Chiaramonte 2018). In order to grasp the differences between the various party systems, prominent scholars such as Duverger (1954), Blodel (1968) and Sartori (1976) have combined one or more of the above characteristics to construct typologies. However, while these criteria have proved useful in classifying party systems into specific types, they prove rather in- effective in capturing changes in party systems over time (Mair 2006, 63). On the other hand, a second approach to the study of party systems has completely avoided the issue of classification, instead using empirical indicators to summarise the main characteris- tics of party systems and show how they vary over time and space. In the panorama of the studies on party systems and party system change, Mair (1997, 2001, 2006) stands out for challenging both those who resort to discrete categories to define party systems and those who use quantitative indicators to summarise the properties of such systems. According to Mair, quantitative indicators only matter if they affect the structure of competition for control of the executive, which he considers the core of any party system and the only criterion for identifying party system change. More precisely, Mair defines party competition for the executive in terms of three fac- tors: the degree of alternation in office, the stability of government alternatives over time, and, finally, the extent to which access to government is open to new parties. The combination of these elements makes it possible to classify party systems as closed or open structures of competition for the executive and, consequently, to define party sys- tem change when there is a change in the prevailing structure of competition. Building on these insights, in this paper we address the phenomenon of party sys- tem change by focusing on a specific dimension that has usually been ignored in the comparative literature, namely the evolution of legislative parties as a consequence of party switching. 3. Party switching in the 18th legislature Defined as an umbrella label for any registered change in party affiliation by elected poli- ticians (Heller and Mershon 2009, 10), party switching encompasses different phenomena involving varying degrees of coordination between legislators: party Party system change at the legislative level 20 mergers, splits and individual ‘jumps’ of MPs from one party to another (Ceron 2013; Golder et al. 2022). Building on the Müller and Strøm (1999) classification of political goals, a first strand of literature has shown that Members of Parliament (MPs) may change party affiliation to secure re-election, obtain office benefits and influence policies (Heller and Mershon 2005, 2008; Desposato 2006; McMenamin and Gwiazda 2011). A second strand of literature has focused instead on how the utility of promoting MPs’ elec- toral, office and policy benefits can vary according to electoral institutions (Klein 2016, 2018), the level of institutionalization of the party system (Klein 2021), and parties’ ide- ology (Volpi 2019). Finally, a third strand of literature has concentrated more on the implications of party switching for sending and receiving parties (Ceron and Volpi 2021, 2022) and for party systems, party system change, and the structure of competition be- tween elections (Mershon and Shvetsova 2008, 2013, 2014). More generally, the study of defections can provide insights into party system change and the dynamics of political competition in the inter-electoral period (Laver and Benoit 2003; Laver 2005), also offer- ing a new understanding of political parties as clusters of legislators rather than as unitary actors (Benoit and Giannetti 2009; Ceron 2016). This work positions itself in this latter line of research, focusing not so much on the factors that drive individual party switching, but on the implications of defections for party system change between elections. Although legislators are supposed to remain loyal to their party for the duration of the entire legislature, this is far from being completely true. Comparative data show that party switching is quite a common phenomenon in legislatures around the world (see O’Brien and Shomer 2013; Volpi 2019; Klein 2021, for data on defections measured at the party- level). These data show that both in Western Europe and in the younger democracies of Central and Eastern Europe, the relevance of party switching varies greatly between coun- tries, both in terms of the number of legislators involved and the number of parties affected by some defection. The main point that emerges when comparing the two groups of coun- tries is certainly the greater recurrence of the phenomenon in Central and Eastern Europe in the period following the collapse of the Soviet bloc (Sbabad and Slomczynski 2004; McMenamin and Gwiazda 2011). In the group of Western European countries, however, Italy emerges as a deviant case due to the remarkable number of defections that have been registered in the last few years. Given this uniqueness, the Italian parliament has come to be one of the most studied cases within the empirical literature on party switching, espe- cially since 1993, when the composition and labels of legislative parties became increasingly unstable following the passage from the so-called First to Second Republic (Verzichelli 1996; Heller and Mershon 2005, 2008; Giannetti, Di Virgilio and Pinto 2012; Russo, Tronconi and Verzichelli 2014; Valbruzzi 2014; Pinto 2015, 2021). Our data confirm the trends registered by previous empirical research and update the analysis to the 18th Italian legislature, also extending the investigation to the Senate between 2018 and 2022.3 Following the definition of party switching presented above, we 3 Most of the studies on party switching in Italy are focused on the Chamber of Deputies, probably in the belief that the Senate repeats the patterns observed for the lower chamber. It should be noted, however, that a number of studies have registered a growing incongruence in the distribution of seats between the two assemblies (Pedrazzani 2017; Giannetti, Pedrazzani e Pinto 2020). Moreover, the Chamber of Depu- ties and the Senate differ according to the Rules of Procedures, in particular in relation to the norms regulating party switching and the formation of new parliamentary groups (Pedrazzani and Zucchini 2020). PINTO 21 register as a defection event any recorded change in party affiliation on the part of MPs, as reported in the Open Data of the Chamber of Deputies and Senate.4 Table 1 provides all the relevant information about party switching in the two chambers under examina- tion, including the period of observation, the number of MPs, the number of switching events, the number of unique switchers, the incidence of party switching in relation to the entire population of MPs, and finally the percentage of defectors with respect to the total number of legislators. These last two figures are useful for comparing the switching behaviour in the lower and upper chambers, as the former comprises twice as many MPs as the latter. Table 1. Summary data on party switching in the Italian Chamber of deputies and the Senate (18th legislature) Chamber Senate Start 2018-03-27 2018-03-27 Stop 2022-02-09 2022-09-02 Days 1620 1620 Months 53 53 MPs 660 334 Switches 294 147 Switchers 215 87 Switches/MPs (%) 44.54 44.01 Switchers/MPs (%) 32.57 26.04 Notes: the number of MPs includes all legislators who held a seat during the legislature including those who resigned or died and those who took over. For the Senate, also life senators are included. Both legislatures are still running with the end scheduled after the national elections on 25 September 2022. Source: own elaboration on the Open Data of the Chamber of Deputies and the Senate. Table 1 shows that the absolute number of defections in the Chamber is double that in the Senate (294 vs. 147). However, taking into account that the lower house includes far more legislators than the upper chamber, the incidence of the phenomenon is more or less the same: 44.54 percent for the Chamber of Deputies and 44.01 percent for the Senate. This result is quite surprising as the Senate introduced a reform of the Rules of Procedures in 2017, with new regulations aimed at reducing the formation of new parlia- mentary groups during the legislature, which should also have had an impact on party switching. Although a rigorous test to study the impact of these new rules on party switching is beyond the scope of this paper, data seem to suggest that this reform was ineffective. The main difference between the two assemblies is in the percentage of unique switchers, which is higher in the Chamber. This means that in the Senate, a 4 Raw data on the movements between parliamentary groups can be retrieved from https://dati.cam- era.it/it/ and https://dati.senato.it/sito/home. These data record all the changes occurring in the composition of parliamentary groups in the two chambers starting from the first day of the legislature (27-3-18), without distinguishing between voluntary and involuntary defections. At the time of writing this paper, the legislature has not yet ended (general elections are scheduled for 25-9-22). For this reason, we have arbitrarily retrieved data up to 2-9-22. We cannot exclude further defections before the end of the legislature. https://dati.camera.it/it/ https://dati.camera.it/it/ https://dati.senato.it/sito/home Party system change at the legislative level 22 smaller group of senators − about a quarter − are responsible for the total number of de- fections, changing party affiliation several times during the legislature. In the Chamber, on the other hand, the phenomenon is more widespread, involving almost a third of MPs. Figure 1. Alluvial plots of party switching in the Chamber of Deputies and the Senate (18th legislature) Notes: PD: Democratic Party; FI: Go Italy; League: League; FdI: Brothers of Italy; M5S: Five Star Movement; IV: Italy Alive; Mixed: Mixed Group; LeU: Free and Equal; CI: Cheer up Italy; SVP, UV: South Tyrol Peoples Party, Valdotanian Union; MAIE-CD: Asso- ciative Movement of Italians Abroad-Democratic Centre; C.A.L.: Constitution, Environment, Labour; IpF: Together for the Future. Source: own elaboration on the Open Data of the Chamber of Deputies and the Senate. A clearer picture of the dynamics of party switching in the 18th legislature is pro- vided by Figure 1, which employs alluvial plots to show switching flows during the legislature. This picture can be useful to get a first idea of how the occurrence of the PINTO 23 phenomenon reshaped the parliamentary party system between 2018 and 2022. Col- umns represents the distribution of legislators in the different parliamentary groups observed at a yearly base, starting from the first day of the legislature. Flows show how MPs regroup in parliamentary parties from year to year. Thicker flows identify party splits, while thinner flows identify individual ‘jumps’. Figure 1 identifies three common trends shared by both chambers. Firstly, the first year of the legislature is characterised by a very low level of defections, which instead in- creases from the second year onwards. Second, the size of the mixed group – which by definition is characterised by low levels of unity and party discipline and therefore can constitute a potential problem for the proper functioning of the assemblies – grows throughout the years. Third, there is an increase in fragmentation due to the emergence of new parliamentary groups during the legislature. If we take the parliamentary party system in place at the beginning of the legislature as a reference point, Figure 1 provides initial evidence of its evolution and, consequently, of a change in its structure. In the next section, we will investigate these changes in a more systematic way using a series of in- dicators to summarise them. 4. Legislative party system change in the 18th legislature As mentioned above, one approach to the study of party systems relies on quantitative indicators to summarise the main characteristics of party systems and track changes in party systems over time. A second approach focuses more on changes in the structure of competition for control of the executive as an indicator of change in the party system. In this section, we first present and discuss the main indicators for tracking the degree of change in the parliamentary party system during the 18th parliamentary term. Then, we will test whether the changes revealed by these indicators have an impact on the prevail- ing structure of the competition in the lower and upper chambers. 4.1. Quantitative indicators of party system change As a first step in our analysis of parliamentary party system change, we rely on four indi- cators, most of them adapted from the literature on party systems and party system change in the electoral context: the effective number of parliamentary parties, within- term seat volatility, the size of new parliamentary party groups, and, finally, the size of the mixed group. All the indicators are based on the relative strength of parliamentary parties. Although important, the size of parties is only one of the criteria for studying party systems. Other important aspects are the ideological preferences of parties and the number of salient dimensions of competition (Wolinetz 2006). Unfortunately, the main methodologies for studying parties’ policy preferences – the expert survey and the con- tent analysis of parties’ manifestos (Budge et al. 2001; Benoit and Laver 2006) − do not allow us to track the evolution of parties’ ideal points and salient dimensions during the legislative term.5 For this reason we exclude in our analysis any consideration related to the spatial structure of the space of competition.6 All indicators are measured on a daily 5 Both expert surveys and the analysis of party manifestos calculate party policy positions at fixed points in time, which usually correspond to elections (see Budge et al. 2001; Benoit and Laver 2006). 6 For an account of party system change in Italy from the perspective of the spatial approach to elections and party competition, see Giannetti, Pedrazzani and Pinto (2017, 2022). Party system change at the legislative level 24 basis. This choice makes it possible to trace specific events that can be linked to the rec- orded changes. In addition, daily measurements can form the basis for further aggregations. A first indicator we use in our study concerns the number of competing parties in a system. Counting the number of relevant parties has always been an issue debated in the literature on party systems and party system changes (Blondel 1968; Sartori 1976; Si- aroff 2000). A standard method for counting parties that avoids discretionary decisions is the Laakso and Taagepera (1979) index of effective number of parties. The index is measured by dividing one by the sum of the squares of proportions of votes (effective number of electoral parties, ENEP) or seats (effective number of parliamentary parties, ENPP) won by each party competing in the system. Given our focus on the evolution of the parliamentary party system, we use this latest version of the index, measuring the ENPP per day, starting from the first day of the 18th legislature until the last one of the observation periods. A second indicator used in our analysis to track the changes in the party system over the inter-election period is a measure of volatility. The phenomenon of volatility is con- ceptually related to the stability or instability of voting behaviour patterns and is measured by averaging the vote shifts between parties in two successive elections (Pedersen 1979).7 For the purpose of our study, we use a measure of within-term volatil- ity which averages shifts in party seat shares from one day to the next (Mershon and Shvetsova 2013). Volatility is the product of two different trends: the first one is when voters switch their vote between existing parties (type B volatility or ‘alternation’); the second one occurs when voters shift their vote from existing parties to new ones (type A volatility or ‘regeneration’) (Powell and Tucker 2014; Chiaramonte and Emanuele 2017, 2019). While these two types of volatility have different implications in electoral con- texts, this distinction is less relevant in the legislative context, as both are indicators of instability in the structure of parliamentary party competition. For this reason, in the subsequent pages we focus our attention on total within-term volatility. Finally, to complete our analysis of party system change between elections, we em- ploy two further indicators. The first is a measure of the degree of 'newness' in the parliamentary party system and is calculated as the share of seats held by parliamentary groups that were formed after the first day of the legislature (Emanuele and Chiara- monte 2018). As the emergence of new parties can make inter-party competition increasingly unstable and unpredictable, party system innovation has often been used as an indicator of party system change (Casal Bértoa 2017). The second measure tracks the share of seats held by the mixed group. Given the fundamental importance of party unity in legislative voting, the size of the mixed group, which is by definition heterogeneous and not very cohesive, increases unpredictability in parliamentary decision-making. Both indicators capture relevant aspects of the changing structure of party competition in the legislature. The trends identified by the four measures of party system change in the Chamber of Deputies and the Senate are illustrated by Figures 2 and 3 respectively. 7 Wolinetz (2006, 60) contests the use of volatility as an indicator of party system change or de-institu- tionalisation, as it measures ‘the ability of parties to build loyal followings and collectively structure the electorate’, which ‘are properties of parties, individually and collectively, rather than aspects of the party system’. Despite this, volatility has been widely used to measure the regeneration and de-institutionali- sation of party systems (Mainwaring and Scully 1995; Chiaramonte and Emanuele 2017, 2019). PINTO 25 Figure 2. Indicators of party system change in the Chamber of Deputies (18th legislature) Notes: in Panel A, the Y-axis measures the effective number of parliamentary parties; in Panel B, the percentage of seats shifted between party groups from one day to the next; in Panel C, the percentage of seats held by new parties formed during legislature; in Panel D, the percentage of seats held by the mixed group. Data refer to the Chamber of Deputies. Source: own elaboration on the Open Data of the Chamber of Deputies and the Senate. Figures 2 and 3 show that in both the assemblies, the ENPP almost doubles during the legislative term (Panel A). On the first day of the legislature, the ENPP was 4.39 in the House and 4.48 in the Senate; on the last day of the observation period, these figures rose to 7.03 and 7.23 respectively. This is a relevant increase which, using Siaroff’s clas- sification (2000) of party systems, can be interpreted in substantial terms as a shift from a case of extreme multipartitism with two dominant players – the M5S and the League which together hold more than 50 percent of the seats – to a case of extreme multiparti- tism with a greater balance between parties, as the seats controlled by the M5S and the League fell to about 35 percent. The observed trends in the within-term volatility provide further evidence in favour of party system change in the 18th legislature (Panel B). Both figures show a fairly constant rate of volatility calculated on a daily basis, punctuated by more consistent peaks of over 8 percent in the case of the Chamber and 4 percent for the Senate. The average volatility each year is about 12 and 31 percent for the lower and the upper houses respectively, figures that analysists of volatility in the electoral context would consider high (Chiaramonte and Emanuele 2017). Moving to our third indicator, party system innovation offers information about the regeneration of the parliamentary party system (Panel C). In both the assemblies, the share of seats held by groups formed after the first day of the legislature increases over time, rising in the Chamber from less than 4 percent to more than 15 at the end of the observation period. Similarly, in the Sen- ate there is a shift from less than 2 to more than 12 percent. Finally, comparably to what Party system change at the legislative level 26 we observe for party system innovation, in both assemblies, the size and weight of the mixed group grow over time (Panel D), exceeding 12 percent of the seats: on the last day of the observation period, the mixed group is the second largest group in the Chamber and the fourth in the Senate. Figure 3. Indicators of party system change in the Senate (18th legislature) Notes: in Panel A, the Y-axis measures the effective number of parliamentary parties; in Panel B, the percentage of seats shifted between party groups from one day to the next one; in Panel C, the percentage of seats held by new parties formed during legislature; in Panel D, the percentage of seats held by the mixed group. Data refer to the Senate. Source: own elaboration on the Open Data of the Chamber of Deputies and the Senate. Overall, the four indicators provide evidence in support of party system change dur- ing the 18th legislature. The data show that, after about a year of relative stability in the parliamentary party system, the environment started to become more dynamic, in par- ticular from the second half of 2019, coinciding with the dissolution of the Conte I government and the formation of the second cabinet led by Giuseppe Conte. 4.2. Changes in the structure of competition As a second step in our analysis, we shift our attention to the three governments formed during the 18th legislature, to see whether the changes in the number of parties and the degree of party system innovation detected above had an impact on the competition for the executive. To this purpose, Table 2 reports information on the composition of the governments, together with data on the extent of the support for governments during the confidence vote given by new parties and the mixed group. Table 2. Governments in the 18th legislature Term Members Support PINTO 27 Conte I 2018-06-01 − 2019-08-20 (445 days) M5S-League 1.71 Conte II 2019-09-05 − 2021-01-26 (509 days) M5S-PD-LeU-IV 13.70 Draghi 2021-02-13 – 2022-07-21 (523 days) M5S-PD-LeU-IV- League-FI-IpF 15.63 Notes: membership refers to parties with at least one minister or deputy minister at the time of the end of the government. ‘Support’ reports data on the percentage of deputies belonging to new groups or to the mixed group who voted in favour of the government over the total votes of confidence in favour of the government. In bold the new groups formed during the legislature. Acronyms: M5S: Five Star Movement; League: League; PD: Democratic Party; LeU: Free and Equal; IV: Italy Alive; FI: Go Italy; IpF: Together for the Future. Source: data adapted from ParlGov (Döring, Huber and Manow 2022) and own elaboration on the Open Data of the Chamber of Deputies and the Senate. Table 2 shows that the two inter-electoral governments following the first cabinet led by Conte are based either on a reshuffling of the original coalition, with the PD re- placing the League as the main partner of the M5S in the Conte II government, or on the incorporation of additional parties into the pre-existing alliance, in particular the cen- tre-right League and FI, in the executive headed by Mario Draghi. We observe, therefore, a high degree of alternation and instability in the government formula. Moreover, both the Conte II and the Draghi governments include new parliamentary parties created during the legislature, showing that the competition for the executive is open to brand new parliamentary groups. In particular, the second government headed by Conte incor- porates IV, a splinter of the PD.8 Further information on the evolution of competition for the executive is provided by the last column of Table 2, which shows the percentage of confidence votes given by new groups and the mixed group in support of the govern- ments over the total number of favourable votes. While out of 350 votes in favour of the Conte I government in the Chamber of Deputies, only six came from the mixed group or new parties (1.71 percent), in the Conte II and Draghi governments this number rises considerably (13.70 and 15.63 percent respectively). Considering that the mixed group includes independents and small party labels that do not reach the threshold necessary to form a parliamentary group, these figures indicate that an increasing share of support for the governments comes from components that are ideologically heterogeneous and poorly disciplined by definition. Overall, this brief analysis configures the Italian parlia- mentary party system of the 18th legislature as highly unstructured, with unpredictable patterns of competition. An alternative and more elegant way to summarise how the structure of the compe- tition for the executive evolved during the 18th legislature is provided by Figure 4, which shows the distribution of the Shapley-Shubik (1954) power index among the parliamen- tary parties in the Chamber and the Senate, calculated at the beginning of the legislature and then after each year. The Shapley-Shubik index is an index of bargaining expecta- tions in an office-seeking coalition system as it measures the power of a given party in coalition bargaining, based on the probability that the party can turn a winning coalition 8 According to our coding also LeU can be classified as a new parliamentary group since it was created in April 2018 only in the Chamber of Deputies. However, this group had existed since the beginning of the legislature as part of the mixed group. IpF instead can be considered a new party; however, it is very re- sidual in terms of the period of the Draghi cabinet covered. Party system change at the legislative level 28 into a losing one (Laver and Benoit 2003, 217). Thus, the more likely a party is to play a pivotal role in the government formation process, the more power it will have in coalition bargaining and the higher the power index will be. The index is computed on the basis of the seats controlled by each party and varies between 0 and 1. It can be interpreted as the proportion of all potential coalitions in which a party is pivotal. The sum of the powers of all parties is always equal to 1. Figure 4. Shapley-Shubik power index in the Chamber of Deputies and the Senate (18th legislature) Notes: PD: Democratic Party; FI: Go Italy; League: League; FdI: Brothers of Italy; M5S: Five Star Movement; IV: Italy Alive; Mixed: Mixed Group; LeU: Free and Equal; CI: Cheer up Italy; SVP, UV: South Tyrol Peoples Party, Valdotanian Union; MAIE-CD: Asso- ciative Movement of Italians Abroad-Democratic Centre; C.A.L.: Constitution, Environment, Labour; IpF: Together for the Future. Source: own elaboration on the Open Data of the Chamber of Deputies and the Senate. Figure 4 shows that at the beginning of the legislature, the M5S is the most ‘power- ful’ actor in both the Chamber and the Senate. With a power index of 0.5, the M5S is PINTO 29 expected to be decisive in half of all potential coalitions. The League, the PD and the FI, independently from the seats they control, have an index of 0.17, capturing their minor strategic importance in the competition for the executive. All other parliamentary groups are not relevant at all, having an index of 0. After the first year of the legislature, the balance of power between parties begins to change. Starting from the second year, the strategic power of the M5S gradually diminishes, while that of the other groups – in particular the League – increases over time. At the end of the observation period, the League is the most decisive party in the Chamber, with an index of 0.23. Following nu- merous defections to the mixed group, the latter became the second most powerful actor (0.17). The M5S ranks only third along with the PD (0.16). In contrast, in the Senate, the M5S is still the most powerful party (0.21), but with an index that is less than half of what it was at the start of the legislature. In conclusion, our analysis reveals that over the course of the legislature, the structure of the competition changed drastically, becoming more open and less predictable. 5. Conclusions Party systems and party system change have long been one of the most studied topics within the comparative politics literature, yet most work in this field focuses on changes that occur between elections, overlooking the possibility that parties and party systems may reconfigure during the inter-election period. This paper explores a specific dimen- sion of the broader phenomenon of party system change, namely the evolution of legislative party system in the Italian Parliament between 2018 and 2022. The configu- ration of parliaments is usually considered rather stable: elections allocate seats and define the balance of power between parties until the next election. However, individual party switching can subtly alter this configuration, to the point of changing the bargain- ing power of parliamentary groups and support for government policy through parliamentary voting. According to the analysis presented in this paper, this is exactly what happened in Italy during the 18th legislature. An impressive number of defections in both the Chamber of Deputies and the Senate, despite the fact that the latter had im- plemented anti-defection rules, led to significant changes in the relative strength of parliamentary party groups and in the number of parties, followed by the entry of new political formations in parliamentary arena. These changes contributed to transforming the structure of competition in parliament, which proved to be open and unstable, with patterns of inter-party relations that can hardly be predictable. Our findings have two major implications. The first challenges the value of elections as a mechanism for holding governments accountable and ensuring that citizens’ views and interests are adequately represented in the democratic process. It seems clear from our analysis that elections do not define the structure of party competition once and for all, but, if anything, they are only an intermediate step in a broader process in which the main actors – parties and individual legislators − continually review and adapt their po- sition. The second implication concerns the functioning of parliaments. Institutionalisation is usually associated with strong institutions. Therefore, the insta- bility and unpredictability of the structure of party competition detected for the Italian parliament contribute to undermining its role vis-à-vis the government, further fuelling a crisis of political representation. Party system change at the legislative level 30 More generally, this work shows that inter-electoral politics is important. Future re- search should therefore extend the analysis to previous legislatures in order to verify whether the changes in the parliamentary party system noted for the 18th legislature are an isolated case or are instead the rule. Secondly, although our analysis offers an original perspective on the topic of party system change, it is limited to the evolution of legislative party system in the parliamentary arena. As repeatedly emphasised above, this consti- tutes only one dimension of a broader phenomenon which is commonly understood through electoral competition. This is the reason why future research should better ex- plore the link between party system change in the electoral and parliamentary contexts, in order to investigate whether the two are in any way related. In addition to being a study of only one legislative term, this work is also limited by the fact that it ignores changes in parties’ policy preferences over time, which are an important element of party systems. Consequently, future empirical works should overcome current limitations, developing techniques, such as the content analysis of parliamentary debates, to track the evolution of policy preferences during legislatures. Acknowledgements This work was supported by the Italian Ministry for Education, University and Research, with the grant [2020NK2YHL_002], ‘DEMOPE: DEMOcracy under PressurE ’. References Benoit, K. & Laver, M. (2006). Party Policy in Modern Democracies. London, New York: Routledge. Blondel, J. (1968). Party systems and patterns of government in western democracies. Cana- dian Journal of Political Science, 1:180–203. Budge, I., Klingemann, H.-D., Volkens, A., Bara, J. & Tanenbaum, E. (2001). Mapping Policy Preferences: Estimates for Parties, Electors, and Governments 1945–1998. Oxford: Oxford University Press. Caiani, M. & Padoan, E. (2021). Populism and the (Italian) crisis: The voters and the context. Politics 41(3): 334–350. Casal Bértoa, F. & Enyedi, Z. (2016). Party system closure and openness: Conceptualization, operationalization and validation. Party Politics 22(3): 265–277. Casal Bértoa, F. (2017). Political parties or party systems? Assessing the ‘myth’ of institution- alisation and democracy. West European Politics 40(2): 402–429. Casal Bértoa, F., & Mair, P. (2012). Party System Institutionalization across Time in Post- Communist Europe. In F. Müller-Rommel & H. Keman (eds.), Party Government in the New Europe. London: Routledge, 85–112. Ceron, A. & Volpi, E. (2021). Breakups hurt: Party switching and perceived proximity between politicians and their party. Party Politics 27(4): 656–666. Ceron, A. & Volpi, E. (2022). How do parties react to defections? Electoral strategies after a valence loss. European Journal of Political Research 1–18. PINTO 31 Ceron, A. (2013). The politics of fission: An analysis of faction breakaways among Italian par- ties (1946-2011). British Journal of Political Science 45(1): 121–139. Ceron, A. (2016). Inter-factional conflicts and government formation: Do party leaders sort out ideological heterogeneity? Party Politics 22(6): 797–808. Chiaramonte, A. & Emanuele, V. (2017). Party system volatility, regeneration and de-institu- tionalization in Western Europe (1945-2015). Party Politics 23(4): 376–388. Chiaramonte, A. & Emanuele, V. (2019). Towards turbulent times: Measuring and explaining party system (de-)institutionalization in Western Europe (1945-2015). Italian Political Science Review/Rivista Italiana di Scienza Politica 49(1): 1–23. Desposato, S.W. (2006). Parties for Rent? Ambition, Ideology, and Party Switching in Brazil’s Chamber of Deputies. American Journal of Political Science 50(1): 62–80. Di Virgilio, A., Giannetti, D. & Pinto, L. (2012). Patterns of party switching in the Italian Chamber of Deputies 2008-2011. Rivista Italiana di Scienza Politica 42(1): 29–58. Duverger, M. (1954). Political Parties: Their Organization and Activity in the Modern State. New York: Wiley. Emanuele, V. & Chiaramonte, A. (2018). A Growing Impact of New Parties: Myth or Reality? Party System Innovation in Western Europe After 1945. Party Politics 24(5): 1–36. Enyedi, Z., & Casal Bértoa, F. (2020). Party Systems: Types, Dimensions, and Explanations. Oxford Research Encyclopedia of Politics. Giannetti, D. & Benoit, K. (eds.) (2009). Intra-party Politics and Coalition Government. New York: Routledge. Giannetti, D., Pedrazzani, A. & Pinto, L. (2017). Party System Change in Italy: Politicising the EU and the Rise of Eccentric Parties. South European Society and Politics 22(1): 21–42. Giannetti, D., Pedrazzani, A. & Pinto, L. (2020). Bicameralism and government formation: does bicameral incongruence affect bargaining delays? European Political Science Review 12(4): 469–484. Giannetti, D., Pedrazzani, A. & Pinto, L. (2022). Lo spazio della politica: Partiti e politiche da Berlusconi ai populisti. Bologna, Il Mulino. Golder, S., Ibenskas, R., Salek, P. & Sikk, A. (2022). Understanding the Complexity of Party In- stability in Parliaments. Paper delivered at the Annual Meeting of the American Political Science Association, Montréal. Heller, W.B. & Mershon, C. (2005). Party Switching in the Italian Chamber of Deputies, 1996- 2001. The Journal of Politics 67(2): 536–559. Heller, W.B. & Mershon, C. (2008). Dealing in Discipline: Party Switching and Legislative Voting in the Italian Chamber of Deputies, 1988–2000. American Journal of Political Sci- ence 52(4): 910–924. Heller, W.B. & Mershon, C. (eds.). (2009). Political Parties and Legislative Party Switching. Basingstoke: Palgrave Macmillan. Katz, R.S. & Mair, P. (1994). The Evolution of Party Organizations in Europe: The Three Faces of Party Organization. American Review of Politics 14: 593–617. Klein, E. (2016). Electoral Rules and Party Switching: How Legislators Prioritize Their Goals. Legislative Studies Quarterly 41(May): 715–738. Klein, E. (2018). The personal vote and legislative party switching. Party Politics 24(5): 501– 510. Klein, E. (2021). Explaining legislative party switching in advanced and new democracies. Party Politics 27(2): 329–340. Party system change at the legislative level 32 Laakso, M. & Taagepera, R. (1979). ‘Effective’ Number of Parties. A Measure with Application to West Europe. Comparative Political Studies 12(1): 3–27. Laver, M. & Benoit, K. (2003). The Evolution of Party Systems between Elections. American Journal of Political Science 47(2): 215–233. Laver, M. (2005). Policy and the Dynamics of Political Competition. American Political Sci- ence Review 99(02): 263–281. Lindberg, S. (2007). Institutionalization of Party Systems? Stability and Fluidity among Leg- islative Parties in Africa’s Democracies. Government and Opposition 42(2): 215–241. Mainwaring, S. & Scully, T. (1995). Building Democratic Institutions: Party Systems in Latin America. Stanford: Stanford University Press. Mainwaring, S., & Torcal, M. (2006). Party System Institutionalization and Party System The- ory after the Third Wave of Democratization. In R. Katz & W. Crotty (eds.), Handbook of Party Politics. London: Sage, 204–227. Mair, P. (1997). Party System Change. Approaches and Interpretations. Oxford: Oxford Uni- versity Press. Mair, P. (2001). The Freezing Hypothesis: An Evaluation. In L. Karvonen and S. Kuhnle (eds.), Party Systems and Voter Alignments Revisited. London: Routledge, 24–41. Mair, P. (2006). Party System Change. In R. Katz & W. Crotty (eds.), Handbook of Party Poli- tics. London: Sage, 63–73. McMenamin, I. & Gwiazda, A. (2011). Three roads to institutionalisation: Vote-, office- and policy-seeking explanations of party switching in Poland. European Journal of Political Research 50(6): 838–866. Mershon, C. & Shvetsova, O. (2008). Parliamentary Cycles and Party Switching in Legisla- tures. Comparative Political Studies 41(1): 99–127. Mershon, C. & Shvetsova, O. (2013). Party System Change in Legislatures Worldwide. Moving Outside the Electoral Arena. Cambridge: Cambridge University Press. Mershon, C. & Shvetsova, O. (2014). Change in parliamentary party systems and policy out- comes: Hunting the core. Journal of Theoretical Politics 26(2): 331–351. Morlino, L. (1998). Democracy between Consolidation and Crisis: Parties, Groups, and Citizens in Southern Europe. Oxford: Oxford University Press. Müller, W.C. & Strøm, K. (1999). Policy, office, or votes? How political parties in western Eu- rope make hard decisions. Cambridge: Cambridge University Press. O’Brien, D.Z. & Shomer, Y. (2013). A Cross-National Analysis of Party Switching. Legislative Studies Quarterly 38(1): 111–141. Pedersen, M. (1979). The dynamics of European party systems: changing patterns of electoral volatility. European Journal of Political Research 7(1): 1−26. Pedrazzani, A. & Zucchini, F. (2020). Ineffective changes for hard times. The 2017 reform of the Italian Senate’s Rule of Procedure and its effects. Italian Political Science 15(2): 173– 190. Pedrazzani, A. (2017). Fare le leggi nella Seconda Repubblica: Come cambia il Parlamento. Mi- lano: Egea. Pinto, L. (2015). The Time Path of Legislative Party Switching and the Dynamics of Political Competition: The Italian Case (1996–2011). The Journal of Legislative Studies 21(3): 323–341. Pinto, L. (2021). Libertà di mandato e party switching. Spunti dalla letteratura politologica. Quaderni di scienza politica 28(3): 335–355. PINTO 33 Powell, E.N. & Tucker, J.A. (2014). Revisiting Electoral Volatility in Post-Communist Coun- tries: New Data, New Results and New Approaches. British Journal of Political Science 44(1): 123–147. Russo, F., Tronconi, F. & Verzichelli, L. (2014). Snipers And Switchers. The Difficulties Of Parliamentary Representation In The Italian XVII Legislature. Polis 28(1): 85–106. Sartori, G. (1976). Parties and Party Systems. A Framework for Analysis, Volume I. Cambridge: Cambridge University Press. Sartori, G. (1994). Comparative Constitutional Engineering: An Inquiry into Structures, Incen- tives and Outcomes. London: MacMillan. Sbabad, G. & Slomczynski, K.M. (2004). Inter-party mobility among parliamentary candi- dates in post-communist east central Europe. Party Politics 10(2): 151–176. Schattschneider, E.E. (1942). Party Government. New York: Rinehart & Company. Shapley, L. & Shubik, M. (1954). A Method for Evaluating the Distribution of Power in a Com- mittee System. American Political Science Review 48(3): 787−792. Siaroff, A. (2000). Comparative European Party systems: An Analysis of Parliamentary Elec- tions since 1945. New York: Garland Publishing. Valbruzzi, M. (2014). Is trasformismo a useful category for analysing modern Italian politics? Journal of Modern Italian Studies 19(2): 169–185. Verzichelli, L. (1996). I gruppi parlamentari dopo il 1994: fluidità e riaggregazioni. Italian Political Science Review/Rivista Italiana di Scienza Politica 26(2): 391−413. Volpi, E. (2019). Ideology and Party Switching: A Comparison of 12 West European Countries. Parliamentary Affairs 72(1): 1–20. Wolinetz, S.B. (2006). Party systems and party system types. In R. S. Katz & W. Crotty (eds.), Handbook of Party Politics. London: Sage, 51–62.