Going Local: Parliamentary Questions as a Means of Territorial Representation in the Italian Parliament

Abstract

Parliamentary questions are often employed by Members of Parliament to demonstrate their responsiveness to the needs of the constituency. This article takes advantage from this opportunity to study the determinants shaping the constituency focus of Italian deputies. Previous research has shown that the degree to which Members of Parliament devote attention to their geographical district depends on several factors related to electoral incentives, demands coming from citizens and taste-based preferences. By studying the behaviour of Italian deputies elected in three legislative terms with a system giving few incentives to perform constituency service, this article offers new evidence that non-electoral reasons are important to shape the role played by Members of Parliament.

Keywords

parliament representation constituency service Italy parliamentary questions

The purpose of this article is to examine territorial representation in the Italian Parliament. This is done by looking at the activities undertaken by elected representatives on behalf of their constituencies. In parliamentary systems of government that are typical in Europe, the relationship between parliamentarians and their voters in the district is considered of secondary importance; representation is primarily shaped by the interactions between voters and parties. The indisputable prominence of collective representation, however, should not obscure the fact that representation also has an important dyadic component. When looking at the attitudes and behaviour of parliamentarians, empirical scholars note that some Members of Parliament (MPs) define their role as constituency members (Brouard et al., 2013; Searing, 1994) and devote some of their activity both within parliament (Borghetto et al., 2020; Däubler, 2020; Fernandes et al., 2020; Marangoni and Tronconi, 2011; Martin, 2011; Russo, 2011; Zittel et al., 2019) and outside (Brouard et al., 2013) to speak on behalf of their districts. This apparent paradox – the idea that in European countries, dyadic representation is not theoretically important and the empirical finding that a substantial proportion of MPs devote some time and effort to represent their districts – can be resolved by remembering the lesson learnt from Eulau and Karps (1977): representation has to do with being responsive to both policy and non-policy matters. Territorial representation often takes the form of service or allocation responsiveness (Eulau and Karps, 1977). When parliamentarians speak publicly to advocate for public funding for their constituencies, when they denounce the wrongdoing of a local administration or when they try to solve specific problems for a constituent, they are performing the function of linkage between their geographical constituencies and the government.

The goal of this article is to contribute to the field of study that aims to understand why parliamentarians choose to devote time and effort to representing their particular territorial constituency. There are basically two complementary approaches to analyse this behaviour. The rational-institutional approach, which stresses the importance of self-interest and institutions, looks at the incentives to cultivate a personal vote given by different electoral rules and candidate selection systems (Carey and Shugart, 1995). In contrast, scholars adopting a sociological approach employ concepts such as appropriate behaviour, socialisation (Rush and Giddings, 2011), inner motivations (Searing, 1994) and personal experiences (Burden, 2007) to account for this parliamentarian practice. Both theoretical frameworks, however, share the importance of parliamentarians’ self-perceptions and personal characteristics. This approach can convincingly explain why parliamentarians may considerably differ in their role orientation and behaviour, even when elected in similar contexts. While institutions constrain the range of available strategies, which tends to funnel the range of options towards uniformity of behaviour, personal characteristics represent a source of variability. A comprehensive framework should combine system- and individual-level variables, structure and actors. There are, however, only a few studies that have attempted to disentangle the role of incentives and personal characteristics in shaping territorial representation. Furthermore, this type of study is difficult to conduct in a single country (Louwerse and Otjes, 2016).

The Italian political system offers a great opportunity for studying how incentives and personal characteristics contribute to territorial representation. The electoral system adopted in 2005 provides virtually no incentive to cultivate a personal vote, and yet there is considerable variation in the degree of decentralisation of the candidate selection systems adopted by various parties, a feature that is commonly believed to influence the necessity to build a personal reputation (but see Shomer, 2009). Moreover, political representation is remarkably more personalised in Southern than in Northern Italy, another factor creating different incentives for parliamentarians. Finally, in the Italian Chamber of Deputies, parliamentarians coexist with different profiles in terms of personal and political experiences, which may result in different attitudes towards the territorial constituency. This suggests that MPs’ personal characteristics have a major role in shaping the attention dedicated to the regions in which they are elected. There is some evidence, however, that being elected in regions where political representation is personalised can also have a positive effect on territorial representation.

This article contributes to the literature on the determinants of territorial representation by analysing the content of the Parliamentary Questions (PQs) for Written Answer asked by Italian deputies during three legislative terms, from 2006 to 2018. The analysis of the content of parliamentary questions, a tool which is not subject to strict party discipline, has proved to be a promising strategy to assess the specialisation of different parliamentarians and to measure their focus of representation (Martin, 2011; Martin and Rozenberg, 2012; Russo, 2011). From a methodological point of view, it contributes to the growing body of research that analyses the content of parliamentary documents through computerised content analysis (Van Atteveldt et al., 2019).

The article is structured as follows. The next section summarises the relevant literature and draws some hypotheses. The most important features of the Italian electoral law employed in the 2006, 2008 and 2013 elections are then summarised. It continues with a presentation of the data, an analysis of the research findings and concluding remarks.

Related Literature and Hypotheses

The recognition that both rules and preferences affect the behaviour of legislators is hardly a new one in parliamentary studies, especially since the advent of the new institutionalisms (Martin et al., 2014). The path-breaking work of Searing (1991, 1994) on the British House of Commons intended to integrate economic and sociological theories in an attempt to explain the behaviour of parliamentarians. According to Searing’s theory, personal preferences (motivations) matter more for backbenchers than for frontbenchers, whose behaviour is completely constrained by formal and informal rules. In general, legislative scholars acknowledge that there is a tension between exogenous incentives, to which all MPs wishing to remain in politics are subject, and endogenous preferences, which are unrelated to their career (Fenno, 1978). Examples of non-career-related preferences include acquiring influence, promoting good public policy or behaving according to the expectations of the peers. With few exceptions (Louwerse and Otjes, 2016), most of the empirical research on the topic has focused either on the role of incentives or on the importance of preferences.

The rational neo-institutional turn in political science has pushed legislative scholars to consider how electoral incentives that come from different institutional settings constrain the behaviour of parliamentarians. This paradigm, influenced by the economic tradition, considers parliamentarians as rational actors interested in pursuing certain goals. Although there are different ways to conceptualise these goals, re-election is generally viewed as the primary objective shared by most parliamentarians (Mayhew, 1974). For this reason, the role of electoral systems has been widely acknowledged by scholars of legislative studies. For instance, Carey and Shugart (1995) classified electoral systems according to the incentives they provided to candidates to compete on party or individual reputation. With regard to this study, the Italian electoral system has not changed in the three legislative terms under consideration (see the next section for a description), although this does not mean that all parliamentarians operate under the same set of incentives. As has been noted (Strøm, 1997), to achieve re-election, it is necessary to be first reselected; it is this necessity that explains why candidate selection procedures can have an important role in shaping MP behaviour. This is especially true under closed-list proportional electoral systems, in which the specific placement of a candidate on the list is crucial for re-election (Strøm, 2012).

The most widely adopted framework to study candidate selection has been elaborated by Rahat and Hazan (Hazan and Rahat, 2010; Rahat and Hazan, 2001). Their approach looks more at the formal rules than at what actually happens behind the scenes (Lisi, 2020), which is difficult for external observers to analyse. Among several dimensions included in their analysis, the degree of territorial decentralisation of the candidate selection mechanism is theoretically linked to the focus of representation adopted by parliamentarians and cannot be ignored in a study on geographical representation.¹ Extant studies report mixed empirical evidence. Shomer (2009), studying the Israeli case, did not find empirical evidence linking intraparty selection methods with legislative vote-seeking behaviour. A more recent study on Portugal (Fernandes et al., 2020) found that parliamentarians selected locally often represent their districts, despite being elected with a non-preferential electoral system.

Hypothesis 1 (H1): Decentralised candidate selection systems increase the attention devoted to the territorial constituency.

Incentives are created not only by formal rules but also by the context in which incumbents have to run. Recent studies have focused on how district-level variables influence the incentives of parliamentarians and parties. The logic behind this kind of argument is that the impact of incumbents’ activities on voters’ electoral choices depends on the type of voters that prevails in a district. In a recent study on the German case, Zittel and colleagues (2019) found that the level of constituency-oriented questions asked by German MPs is higher in districts where the voters are less partisan, and the vote, therefore, more volatile. In the case of Italy, Russo (2011) found that MPs in Southern Italy are more likely to ask constituency-oriented questions than their counterparts in the northern part of the country, a result that can be explained by the more personalised nature of political representation in the south. Both results can be achieved by the fact that incumbents decide how to allocate their time and resources based on the likely impact of their activities on voters; loyal party voters are not likely to change their voting preferences based on the activities of an individual parliamentarian. However, the mechanism can also be mediated by parties. Vote-seeking parties may want to select candidates who have proved to cultivate a personal vote in areas where voters are more attached to individual politicians, that is, where the predominant style of representation is more personalised. In Germany, many parliamentarians interviewed by Klingemann and Wessels (2003) believed that their party reselected them for their capacity to win votes in the constituency. In Italy, Marangoni and Russo (2018) found that parties were more likely to reselect productive parliamentarians in Southern Italy, where voters are less attached to parties and more to individual politicians, but not elsewhere. Traditionally, Italy has been considered a country in which parliamentarians acted as brokers between citizens and an inefficient bureaucracy (Golden, 2003). That level of personalisation varies considerably within the Italian territory, however, being greater in Southern Italy and decreasing in other parts of the country. Southern Italy, dominated by a tradition of political patronage and plagued by endemic unemployment, has always offered more incentives to cultivate a personal vote. Evidence of this difference can be found in the diffusion of preference voting for individual candidates; until the abolition of the open-list proportional system adopted in the post-war period, preference voting was primarily a southern phenomenon, as it still is when it comes to regional elections (Marangoni and Russo, 2018).

Hypothesis 2 (H2): Being elected in regions where the style of representation is more personalised increases the degree of attention paid to the district.

An additional argument regarding incentives and career-related goals concerns the political standing of parliamentarians and the need to consolidate their position in the district (Papp, 2020). In the first stages of their career, representatives must devote a significant amount of attention to their constituency to widen their base of supporters; they are in the expansionist phase, according to the terminology introduced by Fenno (1978). Furthermore, because freshmen are often kept at the periphery of more policy-oriented activities, focusing on the constituency has a low opportunity cost (Norton and Wood, 1990). These arguments have been developed keeping in mind MPs elected in single-member districts. There is some evidence, however, that under proportional and mixed systems, newcomers are also more willing to focus on the district (André et al., 2014).

Hypothesis 3 (H3): Candidates with little political experience will focus more on the constituency.

A slightly different version of the same argument takes into account that representatives may have different types of ambitions (Schlesinger, 1966) that are usually linked to their current position. This idea has been developed by Strøm (1997), who, using a rational neo-institutional approach, argued that MPs have four goals that are interrelated and usually hierarchically linked. These are reselection, re-election, party office and legislative office. The attainment of one of these goals normally requires that a previous objective has already been secured. Under closed-list proportional systems, the first two goals are almost indistinguishable because the party has full control over the re-election prospects of each candidate. The standing of a parliamentarian within the party is, however, of the utmost importance. Politicians occupying a relevant party office, or a prominent legislative position, do not have to worry much about reselection and re-election, but will, however, aspire to important positions within the legislature or in the executive. As a consequence, they will pay less attention to their district.

Hypothesis 4 (H4): Parliamentarians with important party offices or legislative positions will devote less attention to their territorial constituency than their colleagues.

Notwithstanding their heterogeneity, the four previous hypotheses are all based on the role of incentives. A complementary way to look at legislative behaviour emphasises that, although most politicians are highly motivated to remain in power, they also have powerful inner motivations that shape their behaviour. Scholars oriented to empirical and normative approaches recognise that political representation occurs not only when representatives behave to please voters but also when they look within and act on the basis of their own beliefs and principles, which typically derive from their background and experiences (Burden, 2007; Mansbridge, 2003). Indeed, several streams of studies have analysed parliamentary behaviour, shifting the focus from exogenous incentives to endogenous motivations and preferences, most of which use the overarching concept of representative or parliamentary role (Blomgren and Rozenberg, 2012).

For the purpose of this study, it is worth noting that scholars working within the above-mentioned tradition emphasise the link between the personal characteristics of MPs, their attitudes and behaviour. The basic logic behind this view is that parliamentarians have inner motivations that are independent from career incentives but very important for influencing their behaviour. Most studies on descriptive representation, which argue that parliamentarians tend to champion the interests of those who share their social characteristics, work with this assumption. As far as gender politics is concerned, since the 1990s, several studies have investigated the role of female legislators in shaping the legislative agenda in favour of women (Taylor-Robinson, 2014), concluding that women are more likely than men to champion women’s interests. Several empirical studies investigated the impact of ethnic minority representation on local and national representative assemblies, suggesting that it fosters substantive policy representation and responsiveness (Saalfeld and Bischof, 2013). More generally, in a study on Nordic parliaments, Narud and Valen (2000) showed that MPs’ political attitudes are indeed related to their social background and that voters are better represented by MPs who belong to their social group. Studying the activities of American Members of Congress, Burden (2007) found that representatives’ behaviour is shaped by their personal experiences, especially when it comes to activities that are not constrained by party discipline.

With regard to territorial representation in Parliament, several studies have found that the degree of localness of incumbents matters. However, much depends on the context and on how localness is operationalised. Working on the German case, Zittel and colleagues (2019) made a distinction between biographical localness (being born and/or residing in the constituency) and political localness (having held offices at the local level). In the end, both proved to be positively associated with MPs’ mentions of the constituency in parliament. In the case of Portugal, Fernandes and colleagues (2018) found that MPs with previous experience in local government asked more district-related questions. According to Borghetto and colleagues (2020), having held a seat in a local assembly made Portuguese MPs more likely to respond to constituency pressures on crime but not on unemployment. Previous studies on the Italian case showed that MPs with local ties were more likely to focus on the constituency in their parliamentary activities. Marangoni and Tronconi (2011) built a four-point scale of localness combining data about candidates’ profiles and their previous local government experience, proving that local MPs introduce constituency-oriented bills more often than their colleagues. Interestingly, the association holds in legislatures elected with different electoral systems. Russo, studying only the first legislature elected with a closed-list proportional system (2006–2008), found similar evidence on parliamentary questions.

In summary, being raised in the constituency and being involved in local politics can have profound consequences on how politicians see their role. Representing the territory in which one was born and still resides creates an identification with the community that is difficult to develop otherwise. In addition, having success in local politics requires great knowledge of the social and economic realities that operate within the district.

Hypothesis 5 (H5): MPs with local ties give more attention to territorial representation.

The Constituency in the Italian Context

Italy between 2006 and 2018 is an interesting case to evaluate the combined effect of career-related incentives and personal characteristics on the behaviour of legislators. The 2005 electoral law, which was adopted in three consecutive general elections (in 2006, 2008 and 2013),² abolished the single-member districts that had constituted the backbone of the previous electoral law adopted in 1993, and replaced them with 27 districts to elect 618 deputies (the remaining 12 deputies were reserved for Italians living abroad). In each district, parties presented blocked lists and citizens had only to indicate the one for which to vote. The average size of the districts was slightly above 27.³ The new law made it possible for candidates to stand in multiple districts, choosing the district in which to be elected only after the election had taken place. According to Chiaramonte

the fundamental element of the new electoral reform was the abolition of the vote for individual candidates (the ‘coalition vote’), replaced with a vote given solely to a party list (which automatically goes to the coalition of which the chosen list forms part) (Chiaramonte, 2015: 15).

In summary, the electoral fate of each candidate was in the hands of the party electorate; based on past results and opinion polls, parties identified safe seats and allocated them to candidates who had to be elected. The electoral system gave no incentive to cultivate a personal vote. Under these conditions, paying attention to the constituency was not a strict necessity for all Italian MPs, but rather a strategy induced by the parties or a choice pursued on the basis of one’s preferences.

Despite the incentives given by this electoral law, a recent survey research on the role of perceptions of Italian parliamentarians has shown that the constituency (either party voters in the constituency or all citizens in the constituency) is the primary focus of representation for a sizable minority of MPs⁴ (Di Virgilio et al., 2015). Between 2006 and 2010, more than one-third of the parliamentary questions for written reply tabled by Italian deputies contained explicit references to their territory (Russo and Verzichelli, 2012: 360). Past research on the Italian case found great variability among parliamentarians’ efforts devoted to initiate parliamentary activities focused on their constituencies (Marangoni and Tronconi, 2011; Russo, 2011, 2013). This confirms that, in a non-preferential electoral context, the choice of dedicating time and effort to represent the constituency is not obvious, but rather depends on several other factors, including the selection rules adopted by parties, the expectations of the voters and personal considerations. All these make Italy an interesting case to understand how incentives provided by parties and personal preferences shape territorial representation.

Data Sources and Operationalisations

The Dependent Variable

In the Italian Chamber of Deputies, there are five different types of questions: written questions (interrogazione a risposta scritta), oral questions (interrogazione a risposta orale), oral questions for question time (interrogazione a risposta immediata), interpellations (interpellanza) and urgent interpellations (interpellanza urgente). Past research has shown that written questions are relatively free from party discipline and frequently employed by parliamentarians to address the problems of their district (Russo, 2011, 2013). For this reason, the following analysis focuses only on this category of question.

Written questions can be presented by any member of the House and should receive an answer within 20 days. Questions are always associated with the name of a first signatory, but very often also have a number of co-sponsors. The government communicates the answer to the President of the Chamber, who ensures that it appears in the official proceedings of the Chamber. For this research, each written question asked in the 15th (2006–2008), 16th (2008–2013) and 17th (2013–2018) Chamber of Deputies has been downloaded from the official website of the Parliament. I built a dataset containing the following information for all questions: name of the questioner (only the first signatory, disregarding the co-sponsors), text and minister who was responsible for answering.

Table 1 shows that Parliamentary Questions for Written Answer are a common tool for Italian deputies, who routinely present thousands of such questions, even if the government manages to only answer a fraction of them. The vast majority of deputies pose questions. Likewise, asking questions related to one’s own district is very common. For these reasons, studying Parliamentary Questions for Written Answer is a useful tool to measure the territorial representation in the Italian Chamber of Deputies.

Table 1.

Parliamentary Questions (PQs) for Written Answer in the Italian Chamber of Deputies (2006–2018).

	15th legislature (2006–2008)	16th legislature (2008–2013)	17th legislature (2013–2018)
Number of PQs	6400	19,499	19,022
Of which: concluded	2233 (34.9%)	7019 (36.0%)	5223 (27.5%)
Of which: constituency oriented	2877 (45.0%)	7037 (36.1%)	9470 (49.8%)
Number of MPs asking at least one question	463	550	571
Number of MPs asking at least one constituency-oriented question	365	456	490
Average number of constituency-oriented PQs	6.39	13.20	17.18

Source: Author’s elaboration of data retrieved from the official data portal of the Italian Chamber of Deputies (camera.dati.it).

To detect the geographical scope of a question, I adopted a dictionary approach and used the Quanteda package for R (Benoit et al., 2018). The geographical dictionary consisted of the names of all regions, provinces and municipalities (comuni). If the text of a question named at least one locality belonging to a given region, I coded it as related to that region. Integrating this dataset with the Circap archive on MPs, which reports the region of election for each MP, I coded the questions to be constituency oriented if their content was related to the region in which the presenter was elected. This approach might lead to an overestimation of the constituency focus for MPs elected in the six largest regions where there are more than one district.⁵ I included a dummy variable in the model to take into account this potential problem. Having the number of constituency-related questions asked by each MP in each legislature, I needed to develop a synthetic measure of constituency orientation.

The sheer number of constituency-oriented questions is not a good indicator because it depends on the activism of each MP. Some parliamentarians table hundreds of questions while others table only a few, but this is mainly a function of how much time and effort they devote to questioning. Another rather intuitive possibility is to compute the proportion of questions related to the constituency (Marangoni and Tronconi, 2011), applying the formula

Proportional index of territorial representation = \frac{P Q_{c}}{P Q_{t o t}}

where PQ_c is the number of constituency-oriented questions and PQ_tot is the total number of questions. This index is the most simple and intuitive but suffers from one main problem: it either overestimates or underestimates the constituency orientation of the many MPs who present a limited number of questions, which would artificially overcrowd the two extremes of the scale. This problem can be easily illustrated with an example. Think of two MPs asking only constituency-oriented questions. MP A tables only one question, while MP B tables 50. According to the Proportional Index of Territorial Orientation, they would each be assigned the maximum score of one. However, during the legislative period, MP B has proved to devote considerable attention to the constituency, probably during an extended period, while MP A has tabled just one question in 5 years. Considering both of them to be completely focused on the constituency is at best difficult to justify.

The limitations implied in the above-mentioned methods highlight the need to find a compromise between relying only on absolute numbers and discarding them in favour of proportions. An alternative index, which has been developed for data generated by count processes and has been successfully applied as a measure of emphasis in the literature on party manifestos (Lowe et al., 2011), is the empirical logit

Empirical logit index of territorial representation = \log \frac{P Q_{c} + 0.5}{P Q_{t o t} - P Q_{c} + 0.5}

This index has value 0 when a deputy presents the same number of constituency and non-constituency-oriented questions (when PQ_c = PQ_tot − PQ_c). Different from the proportional index, the empirical logit index is not constrained on a 0–1 scale, but rather takes positive values when the number of constituency-oriented questions is greater than the number of non-constituency-oriented questions and vice versa. However, the empirical extension of the scale is limited by the logarithm; each additional constituency-oriented question has a marginally decreasing effect on the scale. This is a desirable feature because, while taking into account activism, it acknowledges that reiterating similar questions over time does not add much to the level of constituency orientation. Coming back to the previously mentioned example, MP A would now be assigned a score of 0.48, while MP A would get a score of 2.00.

When computed on all MPs, the two indexes are highly correlated (R = 0.95), and the correlation increases for MPs asking more than one question in the whole legislative term (R = 0.98). The empirical logarithm index is used as a measure of constituency orientation for its superior theoretical value. The descriptive statistics of the empirical log index are reported in Table 2 and its distribution is depicted in Figure 1.

Table 2.

Descriptive Statistics of the Empirical Log Index of Territorial Representation.

	Mean	Median	Standard deviation	Minimum	Maximum
15th legislature (2006–2008)	−0.09	0.00	1.29	−3.55	3.30
16th legislature (2008–2013)	−0.15	−0.04	1.18	−3.98	2.94
17th legislature (2013–2018)	0.03	0.00	1.18	−4.15	3.13

Figure 1.

Distribution of the Empirical Logit Index of Constituency Orientation in the 15th, 16th and 17th Chamber of Deputies.

Independent and Control Variables

To test H1, it is necessary to find a measure of decentralisation of the candidate selection procedures adopted by Italian parties in the three elections of 2006, 2008 and 2013. There is no abundance of comparative studies focused on the selection methods of Italian parties, not only because applying Rahat and Hazan’s framework to the subtle aspects of empirical reality is not easy (Lisi, 2020) but also because the Italian party system in the period under study suffered from a certain degree of instability. In this study, I rely on two complementary sources to rank the degree of decentralisation of candidate selection methods: the formal analysis of the party statutes conducted by Calossi and Pizzimenti (2015) and the empirical study of the 2013 election conducted by Giannetti (2016) on the basis of the Comparative Candidates Survey. The evolution of the organisational arrangements adopted by parties to select candidates to national elections demonstrates that differences among parties were significant and remained almost unchanged in the period under consideration (2006–2013). Calossi and Pizzimenti (2015: 177) assessed the level of involvement of the regional branch of the party in the selection of national candidates, assigning a score varying from 1 (maximum centralisation) to 5 (maximum decentralisation) to each major Italian party. Their study does not cover the 2013 election, for which I rely on the description of the candidate selection methods adopted by Italian parties conducted by Giannetti (2016). Combining their work, Table 3 reports a qualitative judgement of the degree of decentralisation in the three elections under study.⁶ I assigned a value of 1 (low decentralisation) when national bodies had a prominent role, a value of 2 (medium decentralisation) when there was an equilibrium between the role of local and national bodies and a value of 3 (high decentralisation) when candidates were selected primarily by regional-level bodies or in party primaries.

Table 3.

Degree of Decentralisation of the Candidate Selection Procedures.

Electoral list	2006	2008	2013
Alleanza Nazionale (AN)	Medium	–	–
Forza Italia (2006)/Popolo delle Libertà (2008 and 2013)	Low	Low	Low
Rifondazione Comunista (PRC-2006), Sinistra Ecologia e Libertà (SEL-2013)	Low	–	High
Ulivo (2006), Partito Democratico (PD – 2008 and 2013)	Low	Medium	High
Unione di Centro Democratico (UdC)	(Low)	(Low)	Low
Lega Nord (LN)	High	High	High
Movimento 5 Stelle	–	–	High
Scelta Civica	–	–	Low

Source: Author’s elaboration based on data from Calossi and Pizzimenti (2015) and Giannetti (2016).

To test the impact of the level of personalisation on the district, it is necessary to find an objective measure of the propensity of voters to resort to ‘personal votes’. Although voters cannot cast a preference vote in national elections, in most regions, the electoral systems adopted for the election of regional councillors give the possibility of voting for one candidate.⁷ The propensity to cast preference votes is highly variable in the Italian context, generally growing when moving from northern to southern regions. The level of personalisation in each region is measured with the percentage of preference votes cast at regional elections. Considering that voters’ attitudes towards personal votes are very slow to change, I measured them looking only at the last regional election held before 2013. The summary statistics for the variable preferences and for the other continuous independent variables are presented in Table 4.

Table 4.

Summary Statistics of the Continuous Independent Variables.

Variable	Mean	Median	Standard deviation	Minimum	Maximum
Preferences	51.17	43.70	25.28	23.3	90.8
Tenure	1.14	1	3.44	0	6
Age	49.54	50.00	10.21	26	78

With regard to H3, the obvious measure of political experience is Tenure, defined as the number of legislative terms in which an MP had served before the current term. Moving to H4, the standing of MPs is measured using a dummy variable (MajorOffice) having value 1 for parliamentarians who are part of the party national executive and/or have an important legislative office (Speaker, Deputy Speaker, President and Vice President of Permanent Committees, Questor and other similar positions).

To test H5, I built a three-point scale of localness, which takes value 1 (low localness) for MPs who have not being elected in the region in which they were born and live, value 2 (medium localness) for those who have been elected in the region in which they were born and live and value 3 for those who in addition have held offices in local or regional politics. The Circap archive does not contain information on the region in which an MP had political experience; therefore, I assumed that it took place in the same region in which MPs were born, are currently living and in which they have been elected. Table 5 shows the localness level for those MPs included in the empirical analysis.

Table 5.

Localness of Italian Parliamentarians Included in the Analysis.

Localness	15th legislature (2006–2008)	16th legislature (2008–2013)	17th legislature (2013–2018)
Low	31.45	27.41	20.43
Medium	13.12	15.44	33.57
High	55.43	57.14	46.00
Total	100.00	100.00	100.00
N	442	518	563

Finally, control variables for age, opposition and MPs elected in regions with multiple districts have also been included.

Analysis and Results

The dependent variable is continuous, and can be modelled with the classical linear model. A single equation has been estimated for each legislature to evaluate the stability of the results. Given the panel structure of the dataset, I have also estimated random effects and a fixed-effects model. The Hausmann specification test indicated that a fixed-effects model is preferred, so only that one is reported in Table 4. Before commenting on the results, it is worth remembering that fixed-effects models control for the effect of time-invariant (observed or unobserved) variables that affect individual observations. The effects of these variables, however, cannot be estimated. This type of model is extremely useful to understand what happens to individual observations when independent variables change because it utilises only the within-unit variation over time (Mummolo and Peterson, 2018). Several of the independent variables in the dataset analysed in this study vary more between-unit than within-unit over time. For this reason, the panel model needs to be analysed with care.

H1 on candidate selection methods was not consistently supported by the data (see Table 6). In the models for the 15th and 16th legislatures, it seemed that MPs elected in parties in which candidates were selected by a joint decision of national and regional bodies (medium decentralisation) were more inclined to focus on the constituency. As shown in Table 3, the parties that fall in this category are Alleanza Nazionale (AN – National Alliance) in 2006 and Partito Democratico (PD – Democratico Party) in 2008. However, the effect for parties with a high level of decentralisation was seldom statistically significant and never in the expected direction. In addition, the variables in the panel model measuring candidate selection decentralisation were not significant. All in all, there was no support for H1.

The support for H2 was mixed. The coefficient for the variable preferences, measuring the level of personalisation of the style of representation in different Italian regions, was positive and statistically significant in the 15th and 17th legislatures, but not in the 16th or in the panel model. Only a few MPs had changed regions of elections; hence, the results of the panel model were not surprising. However, even with regard to the legislature where its effect was maximum (the 15th), the variable preference had a substantively limited effect; in passing from a region where only 30% of the voters cast a preference vote (for instance, Tuscany or Emilia-Romagna) to a region where as many as 80% of the voters did so (several southern regions), the index of territorial representation increased by only about a half point.

H3 and H4 are concerned with the relationship between the seniority of parliamentarians and their orientation to the constituency. H3 was not supported by the data; the effect of tenure was always indistinguishable from zero. In contrast, the coefficient for the variable distinguishing MPs with important party or legislative offices (MajorOffice) found support in the 15th and 17th legislature, but not in the 16th or the panel model. Together with the results of H2, these findings might suggest that the16th legislature is somehow different from the others. However, the difference should not be exaggerated: the coefficients usually point in the same direction, and the effect of the measures tapping the level of personalisation and MPs’ seniority are substantively small even in the 15th and 17th legislatures.

By contrast, the hypotheses linking the localness of candidates to territorial representation were strongly supported, both in models for single legislatures and the panel model. The coefficients were positive and statistically significant in all cases. Having been born and living in the constituency, even without having held offices in local politics, were associated with an increase in the index of territorial representation of 1.1 points (panel model). Having held an office in local politics further increased the effort dedicated to the district. With regard to the single legislature models, this indicates that MPs with local ties were clearly more dedicated to the constituency than their colleagues without such ties. The results of the panel model were of particular interest because they referred to the cases in which an MP changed the level of localness in two different elections. This may happen, for example, when a candidate who had been parachuted into a district in a previous election is eventually elected in the region in which she was born (or vice versa). Changing the region of election is not very common in Italian politics, but it may happen, especially in the case of prominent political figures. Among the MPs included in the analysis, 56 (of 328 re-elected) in the 16th legislature and 72 (of 220 re-elected) in the 17th legislature had been re-elected in a different district.

The effect of the control variables, age and opposition, which were sometimes hypothesised to be correlated with territorial orientation, was never significantly different from zero. Also, the variable introduced to control for the potential overestimation of the constituency orientation of those MPs elected in multidistrict regions had no significant effect on the single legislature model and it had a negative value in the panel model.⁸

To check the robustness of the findings to alternative specifications, I also estimated a negative binomial model for each legislature using the sheer number of constituency-oriented questions as the dependent variable and including the total number of questions asked as an exposure variable. This last approach is state of the art for the analysis of the constituency focus of MPs’ parliamentary behaviour, although it does have the disadvantage of producing results that are less intuitive than those of linear regression. Nevertheless, the results obtained using negative binomial regressions (which are reported in the Appendix 1) were comparable to those found using linear regressions (Table 6).

Table 6.

Regression Analyses Estimating the Index of Territorial Representation.

	15th legislature (2006–2008)	16th legislature (2008–2013)	17th legislature (2013–2018)	Panel model (fixed-effects)
Candidate selection decentralisation (reference category: low level)
Medium level	0.372**	0.204*		0.176
	(0.177)	(0.121)		(0.151)
High level	−0.303**	0.167	−0.085	−0.162
	(0.153)	(0.132)	(0.142)	(0.141)
Preferences	0.007***	0.003	0.005***	−0.004
	(0.002)	(0.002)	(0.002)	(0.005)
MajorOffice (dummy)	−0.213	−0.054	−0.294**	−0.101
	(0.161)	(0.146)	(0.126)	(0.086)
Tenure	−0.057	−0.015	−0.057	−0.085
	(0.052)	(0.041)	(0.061)	(0.136)
Localness (reference category: low level)
Medium level	0.596***	1.005***	0.589***	1.187***
	(0.208)	(0.176)	(0.182)	(0.348)
High level	1.027***	1.154***	1.012***	1.400***
	0.596***	1.005***	0.589***	(0.239)
Opposition (dummy)	−0.222	(omitted)	0.213	−0.185
	(0.137)		(0.140)	(0.116)
Age	−0.007	0.004	−0.008	0.007
	(0.007)	(0.006)	(0.006)	(0.040)
Multi_district_Region	0.020	0.063	0.006	−0.617***
	(0.123)	(0.107)	(0.102)	(0.224)
Constant	−0.531	−1.451***	−0.394	−0.590
	(0.393)	(0.341)	(0.347)	(1.820)
Observations	397	459	507	1363
R²	0.207	0.190	0.139	0.186
Number single observations				1054

Robust standard errors are in parentheses.

***

p < 0.01, **p < 0.05, *p < 0.1.

Conclusion

Individual parliamentarians have an important role when it comes to representing the districts in which they are elected. This article shows that in the Italian Chamber of Deputies, parliamentary questions are often utilised by MPs to draw the attention of the government to the problems of the constituency. The inclination of parliamentarians to be responsive to the needs of their territory is generally explained by the incentives provided by electoral systems. Differences among countries can be explained, at least partially, by looking at electoral systems. However, within-country variation is generally considerable. To make sense of it, scholars look at both the candidate selection procedures adopted by parties and at the personal characteristics of MPs. By studying three legislative terms in which all parliamentarians were elected through a closed-list proportional system, this article assessed the extent to which several variables influencing the incentives faced by MPs (candidate selection, the level of personalisation in the district, seniority) and the factors shaping their inner motivations were able to explain the emphasis they gave to territorial representation.

The result is straightforward; in the Italian context, the personal background and political experiences of MPs exert the most significant influence. The analyses presented in this article demonstrate that the more a parliamentarian is personally and politically rooted in her constituency, the more she will devote effort and time to raise local matters through parliamentary questions. Substantively, MPs born and raised in the district in which they have been elected are much more likely to care for their geographical constituency. In addition, having held institutional positions in local politics also has positive effects on territorial representation. Scholarly analyses of the process leading to the composition of electoral lists often report that local parties try to resist the imposition of parachuted candidates (Di Virgilio, 2007, 2010); if they do so to defend the importance of territorial representation, their concerns are empirically justified.

This finding fits well with the idea that personal motivations play a significant role in accounting for MPs who work on behalf of their districts, especially when institutions do not impose strong incentives for MP behaviour. In the Italian case, at least between 2006 and 2018, electoral laws did not compel parliamentarians to cultivate a personal vote. Reselection was either not a concern for incumbents or, more plausibly, was not achieved by paying attention to the constituency. Our results offer some mixed evidence that this may have depended on the characteristics of the district. In such a context, personal motivations play a great role in explaining the extent to which MPs focus on their districts.

Due to questions of data availability, this article leaves some unanswered questions that should be explored by future research. First, some objective biographical variables were used as proxy measures for capturing subjective inclinations. This strategy has some advantages when compared to attitudinal measures drawn from survey data; biographical data are available for all MPs, and there is no risk of obtaining socially desirable answers. However, the validity of these measures needs to be validated to make sure that their effect is actually due to their capacity to influence individual attitudes. Second, the ideal strategy to measure the impact of candidate selection mechanisms would be to use individual data about how each individual MP was selected, as was done by Shomer (2009). Unfortunately, attitudinal data available for Italian parliamentarians cannot be matched with data on their activities, because – to the best of my knowledge – all the available elite surveys containing information on the subjective perception regarding the candidate selection process are anonymised.

Footnotes

Appendix 1

Table A1.

Negative Binomial Regressions Estimating the number of Constituency Oriented PQs.

Variables	15th legislature	16th legislature	17th legislature
Candidate selection decentralisation (reference category: low level)
Medium level	0.204**	0.121*
	(0.084)	(0.063)
High level	−0.097	0.026	−0.030
	(0.086)	(0.074)	(0.071)
Preferences	0.003***	0.002	0.002**
	(0.001)	(0.001)	(0.001)
MajorOffice (dummy)	−0.211**	0.038	−0.148**
	(0.085)	(0.079)	(0.063)
Tenure	−0.033	0.006	−0.049
	(0.029)	(0.024)	(0.031)
Localness (reference category: low level)
Medium level	0.403***	0.492***	0.254**
	(0.132)	(0.122)	(0.103)
High level	0.616***	0.603***	0.416***
	(0.110)	(0.112)	(0.102)
Opposition	−0.066	(omitted)	0.036
	(0.071)		(0.061)
Age	−0.005	−0.000	−0.004
	(0.004)	(0.003)	(0.003)
1.Multiple_districts	−0.019	0.044	0.045
	(0.064)	(0.060)	(0.046)
Constant	−0.990***	−1.342***	−0.819***
	(0.231)	(0.209)	(0.181)
Log–pseudo likelihood	−805.1	−1159.9	−1383.6
Observations	397	459	507

Robust standard errors are in parentheses.

***

p < 0.01, **p < 0.05, *p < 0.1.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

ORCID iD

Federico Russo

Notes

Author Biography

Federico Russo is associate professor of political science at the University of Salento. His research focuses on legislative behaviour, political representation and European integration.

References

André

Gallagher

Sandri

(2014) Legislators’ Constituency Orientation. In: Deschouwer

Depauw

(eds) Representing the People. Oxford: Oxford University Press, pp.166–187.

Benoit

Watanabe

Wang

, et al. (2018) quanteda: An R Package for the Quantitative Analysis of Textual Data. Journal of Open Source Software 3 (30): 774.

Blomgren

Rozenberg

(2012) Parliamentary Roles in Modern Legislatures, 1 edn. London ; New York: Routledge.

Borghetto

Santana Pereira

-J

Freire

(2020) Parliamentary Questions as an Instrument for Geographic Representation: The Hard Case of Portugal. Swiss Political Science Review 26 (1): 10–30.

Brouard

Costa

Kerrouche

, et al. (2013) Why Do French MPs Focus More on Constituency Work Than on Parliamentary Work? The Journal of Legislative Studies 19 (2): 141–159.

Burden

(2007) Personal Roots of Representation. Princeton, NJ: Princeton University Press.

Calossi

Pizzimenti

(2015) Party Organizational Change: Formal Distribution of Power between National and Regional Levels in Italian Political Parties (1991-2012). Partecipazione e Conflitto 8 (1):167–189.

Carey

Shugart

(1995) Incentives to Cultivate a Personal Vote: A Rank Ordering of Electoral Formulas. Electoral Studies 14 (4): 417–439.

Chiaramonte

(2015) The Unfinished Story of Electoral Reforms in Italy. Contemporary Italian Politics 7 (1): 10–26.

10.

Däubler

(2020) National Policy for Local Reasons: How MPs Represent Party and Geographical Constituency through Initiatives on Social Security. Acta Politica 55 (3): 472–491.

11.

Di Virgilio

(2007) Nuovo sistema elettorale e strategie di competizion: quanto è cambiata l’offerta politica? In: D’Alimonte

Chiaramonte

(eds) Proporzionale Ma Non Solo. Le Elezioni Politiche Del 2006. Bologna: Il Mulino, pp.191–242.

12.

Di Virgilio

(2010) Cambiare strategia a regole invariate. La rivoluzione dell’offerta. In: D’Alimonte

Chiaramonte

(eds) Proporzionale Se vi Pare. Bologna: Il Mulino, pp.33–74.

13.

Di Virgilio

Giannetti

Pedrazzani

, et al. (2015) Candidates in 2013 Italian General Election: Evidence from the Italian Candidate Survey. Italian Political Science Review/Rivista Italiana di Scienza Politica 45 (3): 317–342.

14.

Eulau

Karps

(1977) The Puzzle of Representation: Specifying Components of Responsiveness. Legislative Studies Quarterly 2 (3): 233.

15.

Fenno

(1978) Home Style: House Members in Their Districts. Boston: Little Brown.

16.

Fernandes

Leston-Bandeira

Schwemmer

(2018) Election Proximity and Representation Focus in Party-Constrained Environments. Party Politics 24 (6): 674–685.

17.

Fernandes

Won

Martins

(2020) Speechmaking and the Selectorate: Persuasion in Nonpreferential Electoral Systems. Comparative Political Studies 53 (5): 667–699.

18.

Giannetti

(2016) Stili di rappresentanza politica. In: Di Virgilio

Segatti

(eds) Metodi di selezione dei candidati e democrazia infra-partitica. Bologna: Il Mulino, pp.295–321.

19.

Golden

(2003) Electoral Connections: The Effects of the Personal Vote on Political Patronage, Bureaucracy and Legislation in Postwar Italy. British Journal of Political Science 33 (02): 189–212.

20.

Hazan

Rahat

(2010) Democracy within Parties: Candidate Selection Methods and Their Political Consequences. Oxford: Oxford University Press.

21.

Klingemann

H-D

Wessels

(2003) The Political Consequences of Germany’s Mixed-Member System: Personalization at the Grass Roots? In: Shugart

Wattenberg

(eds) Mixed-Member Electoral Systems. Oxford: Oxford University Press, pp.279–296.

22.

Lisi

(2020) Selecting Representatives: Methods and Practices. In: Cotta

Russo

(eds) Research Handbook on Political Representation. Cheltenham: Edward Elgar Publishing Limited, pp.253–265.

23.

Louwerse

Otjes

(2016) Personalised Parliamentary Behaviour without Electoral Incentives: The Case of the Netherlands. West European Politics 39 (4): 778–799.

24.

Lowe

Benoit

Mikhaylov

, et al. (2011) Scaling Policy Preferences from Coded Political Texts: Scaling Policy Preferences from Coded Political Texts. Legislative Studies Quarterly 36 (1): 123–155.

25.

Mansbridge

(2003) Rethinking Representation. The American Political Science Review 97 (4): 515–528.

26.

Marangoni

Russo

(2018) Not All Roads Lead to Rome: The Conditional Effect of Legislative Activity on Reselection Prospects in Italy. Parliamentary Affairs 71 (4): 888–907.

27.

Marangoni

Tronconi

(2011) When Territory Matters: Parliamentary Profiles and Legislative Behaviour in Italy (1987–2008). The Journal of Legislative Studies 17 (4): 415–434.

28.

Martin

(2011) Using Parliamentary Questions to Measure Constituency Focus: An Application to the Irish Case. Political Studies 59 (2): 472–488.

29.

Martin

Rozenberg

(2012) The Roles and Function of Parliamentary Questions. London: Routledge.

30.

Martin

Saalfeld

Strøm

, et al. (2014) Introduction. In: Martin

Saalfeld

Strøm

(eds) The Oxford Handbook of Legislative Studies. Oxford: Oxford University Press, pp.1–28.

31.

Mayhew

(1974) Congress: The Electoral Connection. Yale: Yale University Press.

32.

Mummolo

Peterson

(2018) Improving the Interpretation of Fixed Effects Regression Results. Political Science Research and Methods 6 (4): 829–835.

33.

Narud

Valen

(2000) Does Background Matter? In: Esaiasson

Heidar

(eds) Beyond Westminster and Congress. The Nordic Experience. Columbus, OH: Ohio State University Press, pp.83–106.

34.

Norton

Wood

(1990) Constituency Service by Members of Parliament: Does It Contribute to a Personal Vote? Parliamentary Affairs 43 (2): 196–208.

35.

Papp

(2020) Constituency Service. In: Cotta

Russo

(eds) Research Handbook on Political Representation. Cheltenham: Edward Elgar Publishing Limited, pp.266–275.

36.

Pedrazzani

Pinto

(2018) Italian candidates under the Rosato mixed electoral system. In search of personal votes in the plurality tier? Journal of Modern Italian Studies, 23 (4): 400–417.

37.

Rahat

Hazan

(2001) Candidate Selection Methods An Analytical Framework. Party Politics 7 (3): 297–322.

38.

Rush

Giddings

(2011) Parliamentary Socialisation: Learning the Ropes or Determining Behaviour? Understanding Governance. Houndmills; New York: Palgrave Macmillan.

39.

Russo

(2011) The Constituency as a Focus of Representation: Studying the Italian Case through the Analysis of Parliamentary Questions. The Journal of Legislative Studies 17 (3): 290–301.

40.

Russo

(2013) Gli onorevoli. Cosa fanno e come ci rappresentano i nostri parlamentari. Bologna: Il Mulino.

41.

Russo

Verzichelli

(2012) Parliament and Citizens in Italy: An Unfilled Gap. The Journal of Legislative Studies 18 (3–4): 351–367.

42.

Saalfeld

Bischof

(2013) Minority-Ethnic MPs and the Substantive Representation of Minority Interests in the House of Commons, 2005-2011. Parliamentary Affairs 66 (2): 305–328.

43.

Schlesinger

(1966) Ambition and Politics; Political Careers in the United States. Chicago, IL: Rand McNally.

44.

Searing

(1991) Roles, Rules, and Rationality in the New Institutionalism. The American Political Science Review 85 (4): 1239.

45.

Searing

(1994) Westminster’s World: Understanding Political Roles. Cambridge, MA: Harvard University Press.

46.

Shomer

(2009) Candidate Selection Procedures, Seniority, and Vote-Seeking Behavior. Comparative Political Studies 42 (7): 945–970.

47.

Strøm

(1997) Rules, Reasons and Routines: Legislative Roles in Parliamentary Democracies. The Journal of Legislative Studies 3 (1): 155–174.

48.

Strøm

(2012) Roles as Strategies: Towards a Logic of Legislative Behavior. In: Blomgren

Rozenberg

(eds) Parliamentary Roles in Modern Legislatures. Abingdon: Routledge, pp.85–100.

49.

Taylor-Robinson

(2014) Gender and Legislatures. In: Martin

Saalfeld

Strøm

(eds) The Oxford Handbook of Legislative Studies. Oxford: Oxford University Press, pp.250–266.

50.

Van Atteveldt

Welbers

van der Velden

(2019) Studying Political Decision Making With Automatic Text Analysis. In: Oxford Research Encyclopedia of Politics. Oxford: Oxford University Press.

51.

Zittel

Nyhuis

Baumann

(2019) Geographic Representation in Party-Dominated Legislatures: A Quantitative Text Analysis of Parliamentary Questions in the German Bundestag. Legislative Studies Quarterly 44 (4): 681–711.