The ‘population ecology perspective’ in the literature

The population ecology perspective specifically investigates the inequality of political voice of interest organizations starting from data of the whole interest population. Differently from the ‘pressure system’ of Schattschneider (Reference Schattschneider1960), population studies focus on all types of organizations active in seeking to influence public policy and take into account the actions taken by organizations in order to enter the political fray (Schlozman, Reference Schlozman2010: 430).

From this perspective it comes to light that organizations seeking political influence are not only associations of individuals, but also corporations, firms, and institutions. Furthermore, the composition of interest communities is affected by decisions – taken by organized interests – on whether to mobilize and lobby policymakers; yet those same decisions may vary over the course of years. As a result, exploring the composition of interest systems in a systematic way and over time becomes a principal analytical task (Binderkrantz et al., Reference Binderkrantz, Christiansen and Pedersen2015).

This perspective has become analytically and empirically relevant in the last two decades. Scholars now know that populations matter a great deal in terms of defining the severity of the collective action problems organizations face. They also know that interest system density influences the mortality risks of interest groups, as well as the kind and number of issues they lobby on. Finally, they are well aware that the level of bias of a system is a complex result of ongoing factors rather than the simple product of wealth as power. Political scientists currently focus on not only explaining where the groups come from, but also on the kinds of interests and concerns that have vigorous representation and those that do not (Schlozman, Reference Schlozman2010: 426). As a consequence, empirical studies focussing on interest systems and populations have become both more numerous, as well as more theoretically refined (Lowery and Gray, Reference Lowery and Gray1995; Lowery et al., Reference Lowery, Poppelaars and Berkhout2008; Berkhout and Lowery, Reference Berkhout and Lowery2010; Messer et al., Reference Messer, Berkhout and Lowery2011; Halpin and Jordan, Reference Halpin and Jordan2012).

Yet, the population ecology perspective and related research designs become feasible only where data sets are accessible and group categories can be investigated (Poppelaars, Reference Poppelaars2009). This is why in Europe it has only been since the last decade that this perspective has gained ground. Indeed, as Berkhout (Reference Berkhout2014: 1) underlines in a review article on this literature, scholars have only very recently started data collection on interest group systems and populations. Yet, the lack of public registers and data sets is still the main challenge for interest group scholars in many European countries. Apart from the Danish case, which has been the subject of several studies (Christiansen, Reference Christiansen2012; Beyers and Braun, Reference Beyers and Braun2014; Binderkrantz et al., Reference Binderkrantz, Christiansen and Pedersen2015), scholars have carried out very few investigations into national systems (see e.g. Halpin et al., Reference Halpin, Baxter and MacLeod2012; Jordan and Greenan, Reference Jordan and Greenan2012). On the other hand, precisely due to the availability of official sources on interest groups that are active at the supranational level, many scholars have recently developed interesting and fine-grained analyses of the EU interest system (Berkhout and Lowery, Reference Berkhout and Lowery2010; Wonka et al., Reference Wonka, Baumgartner, Mahoney and Berkhout2010; Messer et al., Reference Messer, Berkhout and Lowery2011).

Regardless of either their national or supranational focus, all these new studies take into account the interest system as the unit of analysis; they thus differ from ‘traditional’ interest group research, which generally focussed on organizational and membership features of individual groups or group categories. Any interest system – at the national, local, EU level, or issue level – consists of the entirety of all organized interests that are steadily activated and seek to influence public policy.

A shared preliminary belief of all these studies is that ‘assessing the size and shape of the organised interest system is crucial to addressing a range of fundamental questions about the plurality of interests voiced in policymaking arenas that potentially shape policy outputs’ (Halpin and Jordan, Reference Halpin and Jordan2012: 9). Therefore, empirical works look at system density and diversity, not only with respect to interests populating a given political system, but also with regard to its sub-units, such as, for example, policy domains and economic sectors (Halpin and Jordan, Reference Halpin and Jordan2012; Nownes, Reference Nownes2012; Berkhout, Reference Berkhout2014).

An essential dimension of any given interest system is thus represented by its (potential) bias in representation, which is generally analysed in terms of inequality or overrepresentation of privileged groups such as business interests (Danielian and Page, Reference Danielian and Page1994; Schlozman, Reference Schlozman2010). With respect to this, scholars have identified clear patterns over time: citizen groups (Berry, Reference Berry1999; Jordan et al., Reference Jordan, Baumgartner, McCarthy, Bevan and Greenan2012), as well as health and institutional interests (Schlozman, Reference Schlozman2010) are now better represented than they were in the past. Yet, an equal/unbiased representation for all group types – a fundamental principle to democracy – remains a sort of chimera in contemporary interest systems.

However, studies that follow the abovementioned perspective also differ in some relevant analytical and theoretical respects. The first relates to the definition of what a population is: indeed, that same definition varies on the basis of theoretical questions to be answered (Lowery, Reference Lowery2012: 50). A second difference relates to the types of interest group included; with respect to this, scholars can use either a behavioural or a structural definition of interest group (Baroni et al., Reference Baroni, Carroll, Chalmers, Muñoz Marquez and Rasmussen2014): of course, this choice directly affects both the selection of cases and the classification of interest organizations (Berkhout, Reference Berkhout2014). Third, within this bulk of literature, different theories have been used to explain the size and shape of interest communities: ‘bottom up’ vs. ‘top down’ models or supply vs. demand explanations privilege social and organizational features in the former view, with policy- and political-related factors privileged in the latter. Even within the population ecology perspective developed by Gray and Lowery (Reference Gray and Lowery1997, Reference Gray and Lowery1998) we can encounter different modes/approaches in examining group environments.

In conclusion, the research questions that population studies pose are not new; yet they are differently posed. Within this perspective, size (density) and shape (diversity), volatility and turnover, interest group stability, become the main analytical dimensions and basic features of any interest system. These dimensions are correctly considered as being crucial: indeed, counting interest groups and how to count them does not only follow descriptive purposes. On the contrary, demographic features are directly connected to crucial political science concepts such as interest representation and democratic quality more broadly.

The application of the ‘ESA model’ to the Italian case: a ‘plausibility probe’

As we have repeatedly said, the main aim of this work is to present and discuss the size and shape of the Italian interest system over the course of the last three decades: it thus represents a case study. Generally, the term ‘case study’ is used loosely in the literature, as an impressive but redundant synonym for ‘study’. Yet, the difference between a case study and a study helps to clarify the nature of comparative politics itself. A case is an instance of a more general category. To conduct a case study is therefore to investigate something that has significance beyond its boundaries. By contrast, a study is undertaken for its own sake, without any pretence to wider relevance. Therefore, between case studies and theory testing, there is a clear (but often neglected) link. With respect to this, more than 40 years ago Eckstein (Reference Eckstein1975) suggested running a sort of trial test of a given theory on a particular case: in his words, this test should represent a ‘plausibility probe’ for that same theory to be empirically validated (Eckstein, Reference Eckstein1975: 108).

Within population ecology studies, one of the main theoretical models (maybe the most famous one) is represented by the so-called ‘ESA model’, which was originally proposed by Lowery and Gray (Reference Lowery and Gray1995) more than 20 years ago. More specifically, this framework is premised on the expectation that environmental constraints ultimately determine the contours of interest populations by bringing selective pressures to bear so that not all interest groups survive (Messer et al., Reference Messer, Berkhout and Lowery2011: 166). In Lowery and Gray’s view, three environmental factors positively influence the carrying capacity of interest systems for numbers of interest groups: the number of potential constituents of an interest (the ‘area’ or ‘supply’ term of their model); the level of legislative activity in the interest group’s field of interest (the first of the two ‘energy’ terms of their model); the policy uncertainty of the given field of interest (the second ‘energy’ term of their model). In more detail, this theoretical model assumes that: (i) the broader the ‘area’, the higher the number of interest groups that mobilize; (ii) the more the government is active in a given interest field, the more interest groups ‘occupying’ that same interest field can readily mobilize their constituents; (iii) policy uncertainty or the likelihood that policy will change provides ‘energy’ that interest groups can use in order to mobilize.

While the ESA model was initially conceived by American scholars and tested within the American state interest systems, it is not an ‘American’ theory per se, and its theoretical concepts are quite general (Messer et al., Reference Messer, Berkhout and Lowery2011: 169). Indeed, in recent years many scholars recurred to the ESA model in order to explain variation, not only in density and diversity, but also in mobilization and stability, as well as in the ‘degree of pluralism’ characterizing both different interest systems (Lowery, Reference Lowery2012), as well as their sub-units, such as, for example, policy domains and economic sectors (Halpin and Jordan, Reference Halpin and Jordan2012; Nownes, Reference Nownes2012; Berkhout et al., Reference Berkhout, Carroll, Braun, Chalmers, Destrooper, Lowery, Otjes and Rasmussen2015).

Following Eckstein (Reference Eckstein1975), we therefore believe that the Italian interest system might represent a good case to conduct a ‘plausibility probe’ for the ESA model. In other terms, we are quite confident that Lowery and Gray’s theoretical approach – albeit with some specifications – could help very much in understanding long-term features of the Italian case. Indeed, many sources of ‘energy’ (in authors’ terms) became active in recent decades in Italy. As for the level of legislative activity (the first of the ‘energy’ terms of the Lowery and Gray’s model), in the so-called ‘Second Republic’ (since 1996) governments have started to pass ‘significant’ legislation (Pritoni, Reference Pritoni2015a). With respect to this, many authors have convincingly demonstrated that laws have become bigger and more heterogeneous in their contentFootnote ⁸ (De Micheli and Verzichelli, Reference De Micheli and Verzichelli2004; Borghetto and Visconti, Reference Borghetto and Visconti2014). Moreover, both the necessity to pass wide-ranging reforms in complex policy areas, as well as the need to repeatedly strike deals among parties in office, led to perpetual modifications of legislative agreements, in order to mirror a continuously changing balance of power among those same governing parties. This type of legislative productivity should therefore push more and more interest groups into activation in order to monitor – and when possible to influence – the law-making process and the policymaking as a whole.

Where policy uncertainty (the second of the ‘energy’ terms of the ESA model) is concerned, alternation in government – which was virtually impossible during the First Republic (Sartori, Reference Sartori1976) – has boosted the likelihood that right-wing executives change what left-wing executives previously decided, and vice versa (Cotta and Verzichelli, Reference Cotta and Verzichelli2007). Therefore, this new bipolar logic of political competition, as well as the necessity for governments to gain legitimacy also through policy responses to their respective constituencies, have produced a growing policy uncertainty.

In our view, the abovementioned tendencies should have led to relevant transformations in both interest system density and diversity: on the one hand, the number of politically active interest groups is likely to grow over time; on the other (and in turn), the more an interest system becomes crowded (i.e. the more it increases its density), the more we should expect that only well-off interest groups (namely, business groups and large unions) survive. In other words, the application of the ESA model to the Italian case results in the following theoretical expectations:

1. i. a growing density of the Italian interest system over the course of time;

2. ii. a changing diversity of the system too, with more and more business groups and large unions (which are usually considered to be groups with more resources), on one side, and fewer and fewer identity groups and public interest groups (which are usually considered to be groups with fewer resources), on the other.

The more empirical findings resemble those expectations, the more the ‘plausibility probe’ will be considered successfully passed.

Research design

As we claimed in the introduction, we focus on the Italian interest system as a whole, which consists of all interest groups revealing some political activity in Italy over time. But what kind of interest groups? In the literature, scholars tend to choose one definition of interest group over another (Baroni et al., Reference Baroni, Carroll, Chalmers, Muñoz Marquez and Rasmussen2014). A key distinction can be made between a ‘behavioural definition’ (Baumgartner et al., Reference Baumgartner, Berry, Hojnacki, Kimball and Leech2009) and an ‘organizational definition’ (Jordan and Greenan, Reference Jordan and Greenan2012; Binderkrantz et al., Reference Binderkrantz, Christiansen and Pedersen2015). In the first case, groups are defined based on their observable, policy-related activities; in the second case, the ‘interest group’ term is reserved only for membership associations. The most recent studies applying the ESA model recur to the so-called ‘behavioural definition’ (Messer et al., Reference Messer, Berkhout and Lowery2011; Berkhout et al., Reference Berkhout, Carroll, Braun, Chalmers, Destrooper, Lowery, Otjes and Rasmussen2015): therefore, with the aim of developing future ‘cautious comparison’ between the Italian case and other interest systems, we have decided to make use of this definition, too.

In doing so, we started from ‘Guida Monaci’, which is to date the only available data set collecting basic information on Italian companies, associations, public administrations, non-territorial bodies, and non-profit organizations in Italy. This periodical publication relies upon voluntary registration and general information provided by organizations themselves. Broadly speaking, ‘Guida Monaci’ mimics encyclopaedias and directories which are available in other countries, such as the Encyclopedia of Associations and the UK Directory of British Associations (Jordan and Greenan, Reference Jordan and Greenan2012), or the Washington Representatives directory (Berry, Reference Berry1999; Schlozman, Reference Schlozman2010).

The main advantages of this source lie in both its inclusiveness – thousands of different organizations are listed – as well as in its periodicity: indeed, ‘Guida Monaci sul Sistema Italia’ is available for the years 1971, 1979, 1986, 2001, 2005, and for 2013 there is an online edition. Our first sample of interest groups thus derives from the editions of 1979 and 2013: more precisely, we originally collected all interest groups (i.e. all interest associations listed in one of the following sub-categories: (i) Associations; (ii) Committees; (iii) Confederations; (iv) Federations; (v) Foundations; (vi) Leagues; (vii) Unions; (viii) Professional Orders) that were either listed in the 1979 edition, or in the 2013 edition, or both. Yet – as scholars usually suggest (Jordan and Greenan, Reference Jordan and Greenan2012; Nownes, Reference Nownes2012: 102–103) – each empirical analysis dealing with national interest systems has to focus on interest groups that are politically active at the national level.Footnote ⁹ Therefore, we first excluded all interest groups that were/are sub-nationally focussed. Second, we ignored multiple entries (many groups were listed in more than one sub-category). Third, we chose to include not only associations and groups with individual membership, but also big companies. However, we established a threshold for the inclusion of corporations: only big firms have the opportunity to lobby at the national level, whereas small firms usually focus on their territorial niche; therefore, among the thousands of firms that are registered in ‘Guida Monaci’, we selected only the ones listed on the Italian Stock Exchange.Footnote ¹⁰ Fourth, in order to distinguish between interest groups that are politically active and interest groups that are not, we decided to conduct an empirical analysis of interest group appearances in the media. Put differently, our choice has been to recur to media appearances as a checking procedure for actual interest group mobilization and policy activity over a given time span.

We are well aware that this research design is both challenging and risky: above all, not all interest groups are equally ‘newsworthy’ (Binderkrantz, Reference Binderkrantz2012); furthermore, some interest groups might seek not to appear in the media. Yet, we think these are not insurmountable issues. First, by developing a diachronic analysis which focusses on evolutionary trends, we can sterilize the ‘newsworthiness bias’: labour unions – for example – were probably more newsworthy than business groups in the 1980s as well as nowadays. Secondly, many empirical analyses have demonstrated that interest groups, rather than employing either direct lobbying strategies or indirect lobbying strategies (i.e. appeal to the media), usually recur to a mix of both (Thrall, 2006; Binderkrantz et al., Reference Binderkrantz, Christiansen and Pedersen2015): especially where an extended time span is concerned, it is virtually impossible that an interest group systematically tries not to appear in the media for decades. Third, this procedure is anything but exceptional: indeed, other scholars have so far followed a similar research design (Berry, Reference Berry1999; Nownes, Reference Nownes2012; Binderkrantz et al., Reference Binderkrantz, Christiansen and Pedersen2015; Binderkrantz et al., Reference Binderkrantz, Bonafont and Halpin2016).

More precisely, we chose to focus on national newspapers. This is because their archives are easily accessible and provide an extremely reliable source for our investigation. With respect to this, we selected two national dailies – ‘la Repubblica’ and ‘il Sole 24 Ore’ – on the basis of three criteria. First, we need to take into account different editorial lines: in this regard, we opted for two newspapers with quite clear opposite political leanings, one in favour of business groups (‘il Sole 24 Ore’), and one in favour of labour unions and public interest groups (‘la Repubblica’). The second criterion deals with the national prestige of the newspaper,Footnote ¹¹ while the third concerns the reliability and completeness of its historical archive.

As for the search terms, we recurred to both the exact name (e.g. Confederazione Generale Italiana del Lavoro) as well as the usual acronym (e.g. CGIL). Where alternative spellings of the same interest group were possible (e.g. Snals-Confsal, which is also known as simply Snals), we searched for each of them. Finally, we collected data separately and solved potential contradictions by looking carefully at the web site of the interest group under scrutiny; in this way, we believe that data reliability is very high.

Consequently to this complex procedure, our unit of analysis – the Italian interest system – thus consists of 594 interest groups,Footnote ¹² sampled into six categories (business groups; identity groups; groups of institutions and authorities; labour unions; other sectional groups; public interest groups), which in turn can be further differentiated in 30 sub-categories (Table 1).Footnote ¹³

As for the time span, we compare two different periods of 5 years: 1984–88 and 2010–14. This choice is easily justifiable, both theoretically as well as pragmatically. From a theoretical point of view, we have already mentioned how many political, as well as institutional features have changed in Italy during the last 30 years: it is therefore extremely interesting to verify whether similar transformations occurred in the Italian interest system as well. Pragmatically, the data we obtained from ‘Guida Monaci’ are available and reliable from late 1970s onwards, while the archives of the most important national newspapers start precisely in 1984: thus, it is impossible to broaden the analysis further back in time. However, we are convinced that our time span is sufficiently far-reaching as to highlight substantial diachronic trends.

More in detail, the 1984–88 Italian interest system thus consists of all interest groups (of our sample) which were cited at least once from 1 January 1984 to 31 December 1988, whereas the 2010–14 Italian interest system consists of all interest groups (of our sample) which were cited at least once from 1 January 2010 to 31 December 2014. We opted for this solution – rather than simply comparing two particular points in time (i.e. 1984 and 2014) – because in this latter case empirical findings could suffer from some random fluctuation: for example, a given interest group which received media coverage in both 1983 and 1985, while having no citations in 1984, would have been put out of the picture; the same in the case of an interest group receiving media coverage in both 2013 and 2015, but not in 2014. In other words, comparing 2 years would have unduly restricted the number of interest groups which we consider as ‘politically active’: being cited in only 2 years is logically more difficult than being cited just once over the course of 10 years. Therefore, the simple comparison between 1984 and 2014 would have made little sense.

Empirical findings

In the literature, scholars have mainly analysed Italian interest groups because of their subordinated role to parties: in particular, empirical studies have focussed on the main organizations (of farmers, workers, and business) with respect to the period of Italian democratic consolidation (Morlino, Reference Morlino1991). ‘Clientelism’Footnote ¹⁴ between groups and bureaucracies (LaPalombara, Reference LaPalombara1964) and ‘collateralism’Footnote ¹⁵ between groups and parties, as well as the role of policy gatekeepers exercised by the latter, represented for decades the dominant analytical framework within which interest group politics was explained.

Starting from the early 1990s, some scholars revived the topic of interest groups by focussing on trade unions and – above all – entrepreneurial associations (Lanzalaco, Reference Lanzalaco1990; Mattina, Reference Mattina1991); at the same time, empirical analyses on legislative production and public policies provided wider but sectional knowledge about dynamics characterizing interest group access to policymakers and interest group policy success (Lange and Regini, Reference Lange and Regini1986). With the remarkable exception of Lanzalaco (Reference Lanzalaco1993), who defined the Italian interest system as characterized by a distorted version of pluralism – that is, ‘oligopolistic pluralism’Footnote ¹⁶ – there is no empirical research on the unequal representation, differentiated mobilization capacity, lobbying, and influence of Italian interest groups. In other terms, the Italian interest system has never been considered a research subject per se to date.

Therefore, this work represents the first attempt to describe and analyse the size and shape of the Italian interest system over the last 30 years. It can also provide the starting point for a broader research project, which has Italian interest groups as its units of analysis. Even though this could be considered deserving, the simple presentation of the Italian case is not sufficient to avoid considering aspects common to other countries as if they were Italian peculiarities; therefore, all the empirical findings we are going to present will be compared to those of similar analyses developed by scholars who focussed on other cases. That said, please see Table 2, which highlights how the Italian interest system changed from 1984–88 to 2010–14.

By examining Table 2, several interesting considerations can be made. First of all, the Italian interest system has experienced a real explosion in the last 30 years: in 1984–88, only 316 interest groups (of the 894 associational groups and big companies listed in ‘Guida Monaci 1979’) appeared at least once in at least one out of the two surveyed dailies, whereas in 2010–14 this number is almost double that, and equal to 564 (+78.5%) (of the 1759 associational groups and big companies listed in ‘Guida Monaci 2013’). This expansion is both greater than the one highlighted by Berkhout and Lowery (Reference Berkhout and Lowery2010) with respect to the EU interest population, and the one stressed by Schlozman (Reference Schlozman2010) with respect to the US interest system.Footnote ¹⁷ Moreover, similar studies on the British (Jordan and Greenan, Reference Jordan and Greenan2012), Danish (Christiansen, Reference Christiansen2012), and Scottish (Halpin et al., Reference Halpin, Baxter and MacLeod2012) interest system evidence significantly lower growth rates.Footnote ¹⁸

In our view, this empirical finding represents a first – preliminary but remarkable – confirmation for our ‘probability probe’: as expected, more legislative productivity and more policy uncertainty have led to a huge increase in the Italian interest system density. More and more interest groups mobilized in order to monitor – and when possible to influence – the law-making process and policymaking as a whole, in a context in which competing governments succeeding one another have produced many more wide-ranging reforms touching many policy areas.Footnote ¹⁹

However, we are also convinced that the original ESA model could benefit from a further refinement that takes into account a fundamental Italian peculiarity; in our view, this peculiarity has multiplied the effect of original sources of energy, such as legislative activity and policy uncertainty. More precisely, we do believe that policy uncertainty, in particular, has been increased by the political turmoil of the early 1990s, as well as by the never-ending transition that derived from it. Differently from the past, when only interest groups that were ‘collateral’ to parties could have a role in the policymaking (Lanzalaco, Reference Lanzalaco1993), now a growing number of interest organizations can realistically expect to exert influence in the policy process. This more autonomous role of interest groups can be convincingly linked to the fact that political parties have lost their previous role of policy gatekeepers (Koff and Koff, Reference Koff and Koff2000; Lanza and Lavdas, Reference Lanza and Lavdas2000; Fabbrini, Reference Fabbrini2003), in so powering up the release of interest groups (Mattina, Reference Mattina2011). In more detail, over the course of the last 20 years ‘traditional’ parties disappeared, whereas new and different ones appeared. While ‘old’ parties were characterized by clear ideological stances, ‘new’ ones display less neat positions in terms of political and policy preferences. Furthermore, both parties and the party system are more and more fragmented (Fasano and Pasini, Reference Fasano and Pasini2014: 114). For all these reasons, political parties in the Second Republic have been defined as ‘partiti liquidi’ (fluid parties) (Cotta and Verzichelli, Reference Cotta and Verzichelli2016). This is why the policy-making process appears to be now more chaotic and more open to different actors than it was in the First Republic (Capano et al., Reference Capano, Lizzi and Pritoni2014) and, in turn, is probably why our empirical analysis has shown such an impressive growth in the number of politically active interest groups: the ‘unregulated’ nature of the policy process acted as a multiplier for traditional sources of energy.

What about interest system diversity? With regard to this crucial aspect, many considerations can be proposed: first, as many other interest systems (Berkhout and Lowery, Reference Berkhout and Lowery2010; Schlozman, Reference Schlozman2010; Halpin and Jordan, Reference Halpin and Jordan2012; Binderkrantz et al., Reference Binderkrantz, Christiansen and Pedersen2015), the Italian system appears to be biased in favour of economic groups: both in 1984–88 as well as in 2010–14, in fact, more than two-thirds of interest groups that entered our survey are either business groups or other sectional groups. This empirical finding largely confirms what has been repeatedly argued in the literature (Danielian and Page, Reference Danielian and Page1994; Schlozman, Reference Schlozman2010; Berkhout, Reference Berkhout2014). Second, the enlargement of the Italian interest system has not been particularly uneven across various categories of interest groups. On the contrary, the overall picture shows a good deal of stability.Footnote ²⁰ At first glance, this empirical finding appears to be the legacy of a past interest system that – in the words of Lanzalaco (Reference Lanzalaco1993) – was characterized by ‘oligopolistic pluralism’ and, in turn, by long-lasting centrality of some groups and some policy sectors. Simultaneously, this apparent stability contrasts with our theoretical expectations, which postulated a growing presence of well-off groups (i.e. business groups and large labour unions) to the detriment of groups that are relatively less well-equipped (i.e. identity groups and public interest groups).

Yet, even in a context of clear stability, some interesting diachronic trends have to be stressed. But do those trends confirm or contradict our hypotheses? Answering this question is not an easy task: empirical results are rather contradictory. On the one hand, business groups have indeed experienced a larger growth, going from representing 38.6 to 40.2% of the interest system, thus confirming what we expected. On the other hand, a similar increase is the prerogative of public interest groups – increasing from 7.3% in 1984–88 to 8.9% in 2010–14 – which is precisely one of the group categories that, on the contrary, was supposed to present declining numbers. Finally, even large unions and identity groups do not follow the expected path: indeed, both group categories appear impressively stable (−0.1%), but the former should have increased its percentage with respect to the whole interest system, whereas the latter was expected to have a lower percentage.

These patterns do not seem to give sufficient support to our theoretical expectations, but are completely in line with what has been stressed by other scholars with regard to other interest systems: business groups continue to represent a great part of interest communities (Schlozman, Reference Schlozman2010; Christiansen, Reference Christiansen2012; Jordan and Greenan, Reference Jordan and Greenan2012), while the growth of public interest groups is anything but an Italian peculiarity (Kohler-Koch, Reference Kohler-Koch2007; Berkhout and Lowery, Reference Berkhout and Lowery2010).

Yet, we should not be too precipitous in claiming that density patterns do not follow the ESA model’s theoretical hypotheses in the Italian case. In fact, Lowery and Gray (Reference Lowery and Gray1995) argued that there is an indirect correlation between interest system density and underprivileged interest group survival ‘when interest systems become crowded’. However, as Wonka et al. (Reference Wonka, Baumgartner, Mahoney and Berkhout2010: 468) noted, the Italian interest system is less crowded than not only the EU system, but also in comparison to the German, British, and French cases. In other terms, competition for resources among interest groups still appears to be relatively low in Italy and there is sufficient room of manoeuvre for ‘new’ (and relatively ‘poor’) interest groups that want to mobilize.

Several further considerations can be made if we look at sub-categories of interest groups. We would like to stress three aspects in particular. First, where business groups are concerned, the overall growth is the result of two opposite trends: on the one hand, the number of corporations has increased very much in the last 30 years; on the other, business associations (of entrepreneurs, artisans, and merchants) have grown at a slower pace. With regard to other sectional groups, organizations in the energy sector showed an extraordinary growth rate, whereas in particular groups in the food and beverage sector, as well as in the manufacturing sector, diminished their relative weight over the course of time. Finally, the increase of public interests depends almost exclusively on the extraordinary growth rate characterizing interest groups populating the ‘civil rights and solidarity’ sub-category.

Without any doubt, at this stage of our research any potential explanation of these trends cannot be statistically tested and, in turn, might be considered as substantially impressionistic. However, many of the abovementioned long-period trends seem to give some support to our theoretical expectations: for example, as for interest groups in the energy sector, their remarkable increase has probably been favoured by huge legislative activity. In recent decades, different pieces of legislation have indeed restructured the policy field by incentivizing renewable energy (Farinelli, Reference Farinelli2004). As a result, more and more interest groups emerged in order to politicize both sides of the ‘renewable issue’: traditional groups with the aim of limiting investments in renewable energy, as opposed to newer groups with the aim of taking advantage from those same new investments.

As claimed above, a declining trend characterizes interest groups populating both the manufacturing sector as well as the food and beverage sector. In Lowery and Gray’s terms, these findings are perfectly in line with a decline of ‘supply’ in the former case (the Italian manufacturing sector still going through a long-lasting economic crisis), and with the fact that both sectors were already crowded 30 years ago. As predicted by the model, in other terms, the number of interest groups decreases (or, in this case, grows at a slower pace when compared to other policy sectors) when there is less supply and/or when the system is too crowded.Footnote ²¹

Finally, the main reason for the ‘explosion’ of civil rights associations might also be linked to the recent policy agenda in Italy and, in turn, to Lowery and Gray’s analytical framework. Although the Italian parliament has legislated on issues such as same-sex marriage and common law couples only very recently (i.e. in March 2016), these issues have indeed entered the political agenda many times over the last 15 years (Paternotte and Kollman, Reference Paternotte and Kollman2013).

Overall, empirical findings point out that, even though the Italian interest system has greatly increased in size over the course of the last three decades, it has substantially maintained the structural shape of an ‘oligopolistic pluralism’. Interest groups are now more numerous than they were 30 years ago; however, there is no real competition among them yet. Moreover, the abovementioned growth has benefited business groups and, above all, big corporations, in so maintaining what has been defined in the literature as an ‘enduring imbalance’ (Mattina, Reference Mattina2011).Footnote ²²

Therefore, what is really interesting in our results is that – following a different (we would say an ‘original’) path – we still came to the same conclusions that many policy studies have proposed so far: in fact, scholars who focussed on the many transformations that have occurred in Italy since the early 1990s had already noted the growing fragmentation (balkanization?) of interest representation in many policy sectors (Carrieri, Reference Carrieri2009; Fabbrini, Reference Fabbrini2009: 39; Capano et al., Reference Capano, Lizzi and Pritoni2014; Pritoni, Reference Pritoni2015b). We can thus confirm from a broader perspective what these sectional analyses originally evidenced: this growing fragmentation (in population ecology terms, the increasing density of the system) characterizes the whole Italian interest system. Furthermore, we are also convinced that this ‘balkanization’ of policymaking is going to be a long-lasting feature of the Italian interest system, mainly because of the absence of either new political parties with a gate-keeping role, or of institutionalized procedures of consultation and inclusion of interests in policymaking: neither the former, nor the latter, seem to be particularly likely in the near future in Italy.

In any case, with this article we have only conducted a ‘probability probe’ (Eckstein, Reference Eckstein1975) for the ESA model: its theoretical hypotheses need much more research for them to be fully validated. More precisely, a statistical test of the ESA model among policy sectors could be very interesting in the Italian case and will be undoubtedly object of future research. Yet, our preliminary results appear to be both reasonable and promising.

Concluding remarks and future research

Even though, in recent years, a substantive amount of empirical research on interest groups has been produced (Hojnacki et al., Reference Hojnacki, Kimball, Baumgartner, Berry and Leech2012; Bunea and Baumgartner, Reference Bunea and Baumgartner2014), Italian scholars have not paid sufficient attention to interest group politics and the Italian case is still absent in any comparative research. More and more international scholars – following a new ‘population ecology perspective’ – have focussed on the size (density) and shape (diversity) of interest systems (Berkhout, Reference Berkhout2014 for a review). Yet, there has been neither systematic analysis of, nor empirical research into, the Italian interest system so far. We are convinced that the population interest approach is important for a number of reasons, so the aim of this article has been precisely to focus on the demographic features of the Italian interest system over the last 30 years.

With respect to this, the main added value of this study therefore lies in presenting the first database capturing the density and diversity of the Italian interest system over a period characterized by intense political and institutional change. We are rather proud of our research effort, for at least three reasons: first, no one has done anything like this to date. In other words, our study could represent a sort of ‘landmark’ for all scholars who will focus on the Italian interest system from now on. Second, our database represents the final outcome of a very demanding and methodologically rigorous procedure: differently from the United States, the EU, and many other European countries, in fact, lobbying registers do not exist in Italy. Therefore, in order to provide a reliable description of all interest groups that were/are politically active in Italy, we have been forced to find an original path, which has been mainly based on the interrelation between two different sources: ‘Guida Monaci’ and newspapers archives. Third, this study addresses a literature that is more and more relevant within interest group research: that of interest systems and interest populations.

As for our empirical findings, they may be summarized as follows: first, from 1984–88 to 2010–14, the number of interest groups that were politically active (i.e. which received media coverage by at least one out of the two most relevant Italian dailies) increased from 316 to 564 (+78.5%). This impressive growth has probably been due not only to the factors originally highlighted by Lowery and Gray through their well-known ‘ESA Model’ – above all, legislative activity and policy uncertainty – but also to the unique combination of those factors with some Italian peculiarities: in particular, we would like to cite the long-lasting policy legacies of an interest system previously characterized by ‘oligopolistic pluralism’ (Lanzalaco, Reference Lanzalaco1993), the fact that political parties irremediably lost their previous role of policy gatekeepers (Koff and Koff, Reference Koff and Koff2000; Lanza and Lavdas, Reference Lanza and Lavdas2000), and the consequences of the never-ending Italian politico-institutional ‘transition’ (Cotta and Verzichelli, Reference Cotta and Verzichelli2007). These ‘intervening factors’ in particular entailed a much more chaotic policy-making process. Policy opportunities have been growing impressively over the last decades, thus Italian interest groups are now more unrestrained than they were in the past in their competing and mobilizing in order to get what they want.

Second, the expansion of the Italian interest system has not been particularly uneven across various categories of interest groups: thus, the diversity characterizing the Italian case has not changed so much over time. Business interests and other sectional groups still dominate the picture. However, similar to other interest systems (Berry, Reference Berry1999; Berkhout and Lowery, Reference Berkhout and Lowery2010; Schlozman, Reference Schlozman2010), public interests and, above all, civil rights associations have gained ground.

Yet, this study represents a preliminary and mainly descriptive analysis. In order to accurately analyse and explain the demography of the Italian interest system, further research into a number of issues is required. First of all, our data set – as representing the first attempt of listing politically active interest groups in Italy to date – would benefit from a further validation: in particular, we are going to examine interest groups participating in parliamentary and governmental hearings, as well as interest organizations which are engaged in the bureaucratic arena. In other words, the aim is to cross-check different data collected and, in doing so, to aggregate groups which were previously absent and to remove duplicates, following the example of Halpin et al. (Reference Halpin, Baxter and MacLeod2012). Furthermore, in order to take a further step towards ‘cautious comparison’ (Lowery et al., Reference Lowery, Poppelaars and Berkhout2008), it could be interesting to apply Lowery and Gray’s ESA model to the analysis of policy sectors in Italy. With regard to this, more data are thus required, differentiated per policy guild: data on potential constituents, on legislative activity, on policy uncertainty. In conclusion, still much has to be done in order to fully explain demographic features of the Italian interest system. Yet, this article may well represent the first step forward in this direction.