Prosecuting high-level political corruption. Mapping judicial investigations in three Italian cabinets (2008-2013).

Cristina Dallara (Irsig-CNR/ Università di Bologna)1

Very first draft, please do not cite or circulate

Introduction

While a growing body of research examines the causes and consequences of corruption, using sociological, economic, political and anthropological approaches (Mungiu-Pippidi, 2006; Holmes 2006; Vachudova, 2009; De Ridder, 2009; Kostadinova, 2012) less has been written about the institutions and the judicial procedures applied to tackle high-level political corruption. For this reason, we propose to shift the attention from the causes and effects of corruption to the state’s action to curb it, through criminal prosecution. Until now, no systematic studies have been specifically developed on high-level corruption prosecution and very few and reliable information are available on the amount investigations involving high-level politicians, on the timing of such investigations and on the number of indictments for political corruption crimes. We believe that shedding light specifically on the prosecution side could offer relevant information for studying corruption as a whole. As an example, a pilot study focusing on the prosecution of cabinet ministers in Eastern and Southern Europe (Popova and Post 2013) has generated some hypotheses about those factors that could be associated with a higher probability of investigation for corruption charges. With this paper, we propose a pilot study mapping judicial investigations in Italy, using primarily media sources, for corruption related charges against cabinet ministers, deputy ministers and undersecretaries form 2008 to 2013. Our goal is to map out the entire corruption prosecution process, from the appearance of the first allegations to the opening of a formal investigation, to the filing of an indictment by the public prosecution, and finally to the court proceedings during a trial. The aim of the paper is to assess in the Italian case some hypotheses developed in the literature on political corruption, concerning factors that could be associated with a higher probability of indictment on corruption charges, such as: type of ministerial portfolios, affiliation to a strong/weak party and timing of investigative action (before/during/after tenure). To this end, we will proceed as follow: the first section offers a brief introduction about corruption in Italy and its relevance of this topic for the whole relation between justice and politics (§1); then, some reflections on how to deal with data and sources for measuring corruption will be presented (§2); finally, the last section presents and focuses on the mapping exercise we have conducted on the Italian

1 This paper has been written with the collaboration of Luigi Rullo (internship student of the University of Bologna). He specifically accomplished the analysis of the judicial investigations in relation to deputy-ministers and under-secretaries of the IV Berlusconi government and the Monti executive. He also has co-written the paragraph 1. 1 case, following the study of Popova and Post (2013) and analyzing, using media sources, all the corruption-related investigations involving cabinet-ministers, deputy-ministers and undersecretaries or of three executives: the center-left II Prodi government (2006-2008), the center-right IV Berlusconi government (2008- 2011) and the technocratic government led by Monti (2011-2013). The information drawn from this mapping exercise, such as timing of the investigations, portfolio of the minister or vice-minister investigated and party of the minister or vice-minister investigate, will allow us to test some relevant hypothesis developed by the study of Popova and Post (2013).

1. Political corruption in Italy and the M-factor2

Italy enjoys the reputation of having one of the biggest and hugest problems of corruption among Western Democracies. Nevertheless, the corruption issue did not have a strong impact on politics until the Tangentopoli scandal in the early nineties. In fact, the media coverage of Tangentopoli's scandal and the investigative action called Clean Hands' “pushed the issues of corruption and public ethics to prominence” (Hine 2015, 608). However, as Heywood (1997, 418) points out “the fact that corruption should have generated such public concern cannot be divorced from the wider political context within which it was uncovered”. In fact, the crisis of Italian parties is strictly connected with the weakening of their social and ideological bases caused by the end of the cold-war order (Vannucci 2009). The latter assured political actors with ideological reference points providing the Italian political system an high degree of stability, with the monopoly of Christian Democrats in government. Yet, the fall of the Berlin wall and the dissolution of the Soviet Union undermined the ideological bases of citizens' support that characterized the Italian party sistem in the post-war world (Di Virgilio 2006). In such a scenario, the Tangentopoli scandal acted as detonators of the crisis, bringing out an unprecedented corruptive system (della Porta 2001). It is quite telling that in the summer of 1993, 205 out of 630 deputies and 81 out of 326 senators were under judicial investigation, and “at the end of 1994 the number of individuals involved had already reached more than 7000, including 338 ex- deputies, 100 ex-senators, 331 regional, 122 provincial and 1525 municipal administrators, and 1373 public functionaries” (della Porta and Vannucci 1997, 516), with over 4000 preventive custody orders. Yet, the crisis of the party system paved the way to Berlusconi’s decision “to enter into politics” in 1994 and to the establishment of the so-called Second Republic. The media certainly played an essential role in this transition (Campus 2010). In fact the mediatization of politics (Mazzoleni and Schulz 1999; Esser et al. 2014) transformed the media into amplifiers and brokers of ideas. Indeed, it displayed a centralization, professionalisation, and personalization of the political messages, with a greater emphasis on the leaders. In such a scenario, Berlusconi encouraged

2 This section has been written with the collaboration of Luigi Rullo. 2 “the establishment of a trust relationship between the leader and the voters that bypass the traditional intermediary role of parties” (Campus 2010, 227). Substantially, the personalization of politics led to a more and more direct and immediate relationship between the institutional apices and the citizens, enabling the path towards the predominance of “personal leaders” (Musella 2015). Another result of the transition from the First to the Second Republic has been an increasing judicialization of politics (Guarnieri 2001, Della Porta 2001, Tate and Vallinder 1995), complemented by an already lasting expansion of the judicial function's (Guarnieri and Pederzoli 2002; Newell 2011). The judicialization of politics is often confused with, or used as a synonym for, the concept of judicial activism. The latter is a term coined to refer to the behavioural patterns followed by American judges appointed to the U.S. Supreme Court in the area of constitutional interpretation. It initially referred to the tendency of these judges to go beyond the literal meaning of constitutional provisions and expand the meaning of the rights that they offer. In spite of this generalization, it is worth mentioning that the concept of judicial activism refers more to the single actor side of the phenomenon, whereas that of the judicialization of politics refers principally to its systemic side (Dallara 2015). The empowerment of judicial institutions is explained as the result or a consequence of the expansion of judicial power “at the expense of the politicians and/or administrators” (Tate and Vallinder, 1995, p.13). Judges are increasingly making decisions that have significant political implications, and they are assuming a progressively more central role in political controversies. This growing demand for justice entrusts judges for resolving problems that other institutions are unable or unwilling to deal with effectively (Guarnieri and Pederzoli 2002). According to Guarnieri and Pederzoli (1997), there are a number of factors that have contributed to a significant increase in the role of the judiciary in Italy. These factors are rooted in the constitutional arrangements that establish the institutional power of the judiciary. The features of the 1948 Constitution that most contributed to strengthening the judiciary in Italy are the principle of compulsory criminal prosecution, 3 powerful guarantees of external and internal institutional independence, and the establishment of the High Judicial Council (CSM – the Consiglio Superiore della Magistratura), 4 and the rules on its composition. Della Porta (2001) added that the growing legitimation of the magistracy was favoured by the action she had fought against mafia and terrorism, and by a “weakening complicity of some judges with those political forces which had partly hindered the activities of the magistracy” (della Porta 2001, 5). This trend was further supported by the increasing role of the media, during the 1990s, in shaping public opinion in the area of justice and the relations between judicial and political actors. Judges and prosecutors saw the media as a new channel to gain legitimacy and visibility during a period (the 1990s) in which the power of the judicial hierarchy decreased rapidly (Catino 2001; Pizzorno 1998).

3 Article 112 of the Italian Constitution obliges prosecutors to investigate all cases that come to their attention where there is a suspicion that a criminal act has been committed. This means that prosecutors are granted discretionary power to initiate investigations on their own initiative (Guarnieri 1997; Newell 2005). 4 The High Judicial Council was envisaged in the 1948 Constitution and established in 1959. 3 The Tangentopoli investigations proved to be the critical juncture for a massive mediatization of justice to begin. In fact, the “transformation of judicial and political events into a scandal is by and large the result of media activity, which filters and communicates, but also simplifies, personalises and sensationalises information at high speed” (Pujas and Rhodes 1999, 56). Yet, these processes have led to elaborate a narrow connection between media and magistracy (Chang et al. 2010; Asquer 2015; Guarnieri 2008), which finds his principal gear in the episodes of corruption involving high-level politicians (Calise 2016). In fact, considering the gravity and strong resonance of these events, the mass media begun to recognise the judges as defenders of morality triggering a mechanism of mutual support (Giglioli 1996; Amoretti 2000). The result of this was the creation of a media-judiciary circle that contributed to a further increase in the capacity of judges to build public opinion in the field of justice (Mazzoleni 2004; Giglioli 1997). While the media's intrusion in the relation between judiciary and politics could have positive effects for governmental accountability (Chang et al. 2010; Camaj 2013), guaranteeing external controls on political corruption, Calise (2016) has stressed the progressive emergence of an “M-factor” 5 , highlighting the affinities between judicial power and mass media and the dangers that this symbiotic relationship could provoke. The increasing relevance of the M-factor reached its peak with the entry into the scene of Berlusconi as the main actor of the Italian political life (Calise 2010). Successful entrepreneur as media, sport and finance tycoon and political leader with great communicative skills and charisma, Berlusconi quickly captured the attention of M-factor. The latter have played a leading role during the “Ventennio Berlusconiano” as the duel between toghe rosse and il Cavaliere (Amoretti 2000; Dallara 2015) was permanently transposed both within the judicial offices and on the media. Indeed, “every judicial act relating to Berlusconi, his family or his interests assumed genuine political significance” (Dallara 2015, 72). On the whole, the Berlusconi's long-running personal crusade against the judiciary has distracted the policy makers from the needed reforms for a better functioning of the Italian judicial system and for an effective fighting against corruption. In such a context, the issue of corruption, and moreover the need of an anti-corruption systemic policy, lasted and remained in the backstage for several years.

1.2 Anti-corruption policies and specialized institutions

In summarizing the development of the anti-corruption policies in Italy, Mattarella (2013), remembers that “the need for a specific and comprehensive anti-corruption policy was not always obvious in Italy”, although it has been in other jurisdictions, that, before us, approved a similar legislation reacting to facts of widespread corruption. In France, for example, a specific anti-corruption law was adopted as a result of scandals and judicial investigations occurred in 1993. The record of the anti-corruption policy in Italy is characterized, as in many other policy sectors, by

5 M-factor stand for media and magistracy factor. 4 many proposals with a permanent gap in approving and implementing concrete measures. In Italy, in 1996, as reaction to Mani Pulite, two "expert commissions” were established with the aim of proposing a new comprehensive anti-corruption framework. One Commission within the Presidenza del Consiglio (staff-structure for the Premier office) and another one with the Chamber of Deputies. Both the commissions elaborated many proposals and touched the key points for the Italian anti-corruption frameworks. These remained almost ignored for fifteen years. Only several changes in the Criminal code definitions were settled. The first comprehensive Anti-Corruption Law (190/2012), entered into force only on 28 November 2012, under the technocratic government led by . A first draft of this bill was elaborated by the Berlusconi ministers in 2010 but lasted two years passing from Chamber of Deputies and Senate without any adoption. This bill was very much echoed by Berlusconi and his government as the first real step towards an anti-corruption law. In truth, there were few and unclear information about the content and the procedures for its drafting (Merloni and Vandelli 2010). With the Monti government, the new Minister of Justice (Severino) immediately resumed the draft-bill with the aim to modify it, leading to its adoption. Although the Severino text was really different from the original ones, it was drafted throughout a long process of consultation of all the political parties trying to adjust all the provisions according to all parties’ visions. Even the center-right wing parties were involved in this process. In fact, in the end, the law was adopted in October 2012. A part from all the critics it is possible to move to this law (Vannucci 2015), it represents the first comprehensive anti-corruption text ever approved in Italy, collecting methodically international organizations’ suggestions and recommendations from academics, scholars and expert commissions. It also introduces relevant provisions related to the rules preventing politicians convicted of corruption from running for election and for public offices charges. Anyone who has been sentenced to more than two years of punishment must lose his or her position in political office. New rules regarding those convicted of public administration crimes, even if with non-definitive sentence, are banned from assignments in the field of financial resources to become members of tender commissions. Those sentenced for serious offences such as corruption and mafia will be permanently banned from public procurement contracts. Some scholars (Vannucci 2012; Mattarella 2013) argue that one of the main concerns of the law is that it deals mainly with administrative corruption and not with the political one. As an example, the law stresses a lot the need of administrative transparency, while is silent on the transparency in funding political parties; promotes the use of codes of conduct for civil servants, but it still does not predict for politicians; it takes care of conflicts of interests for private actors, not those of parliamentarians, who continue to be almost the only category of civil servants for which, in our system, the conflict of interest is legally irrelevant. This approach, according to Mattarella (2013), clashes with the experience of recent years, in which the majority of judicial investigations involved political actors, such as ministers and their staff, regional councilors and bankers of political parties. Another interesting story is the one related to the Anti-corruption specialized agency, a targeted

5 national institution, today present in the majority of the European countries, having as main task that of preventing corruption and supporting the judicial institutions in the prosecution action. The first move on that way was reached under the II Berlusconi government when the Law n. 15/2003 established a ‘High Commissioner for the Prevention of Corruption’. This provision was adopted as a specific requirement6 of the Criminal Law Convention on Corruption (signed in 1999, but nor ratified until 2012). The Commissioner was appointed only in 2004 (a judge of the Cassation Court) and sonly he claimed against lack of funding and scarce “institutional relevance” of the body. He resigned just after. Two other Commissioners were appointed until 2008. One of them declared to the media that ”the Commissioner was too dependent and controlled by the Government and it is conceived as an institution without powers to act concretely.”. In 2008 Berlusconi decided to close this office, replacing it with another “new” body called Anti-corruption and Transparency Service (Servizio Anticorruzione e Trasparenza, SAET)7. Then, in 2012, another “new” and “specialized office” was created inside the Ministry for the Public Administration (CiVIT – the independent Commission for Evaluation, Transparency and Integrity of Public Administrations) acting as the National Anti-Corruption Authority until 2012. Only in 2013, the National Anti-Corruption Authority (ANAC) was finally created as an independent and autonomous agency for the assessment and transparency in public administrations. The institution is headed by the judge Raffaele Cantone, former deputy prosecutor in the District Anti-Mafia Directorate of Naples. The institution was gradually empowered in the last three years and started to function as an efficient hub to collect claims about corruption suspects and to support the judicial authorities in the prosecution activity.

2. Data, sources and measurement of corruption in Italy.

The action at the basis of the corruption phenomena is not clearly visible and traceable for its own nature. To be perceptible and measured that act must be brought to light and made visible [Cazzola 2007]. Thus, it is difficult to measure it if not being manifested. Corruption is, in fact, a sort of “silent crime” and, as Vannucci (2009) efficiently explains, “like other victimless crimes nobody has an interest in reporting the phenomenon, it is difficult to measure and official statistics do not represent a reliable source of information on its extent”. The same type of data and information could generate different and divergent interpretations. A major number of denounced cases could stand for a major judicial capacity in prosecuting corruption crimes. However, it could also mean that corruption is really widespread. Similarly, a few cases could account for a declining trend in corruption crimes or, on the contrary, for an increased ability of corrupting actors in hindering their traffics. In spite of that, the problem of measuring corruption is crucial and

6 Article 20 – Specialised authorities. 7 The office in 2011 went through a big scandal as an Italian journal published some pages from the “confidential report” done by the USA Ambassador (R.Spogli) after visiting the office. Before this visit even OECD asked for more information and clarifications on the new body. 6 essential, both for the scientific analysis of the phenomenon and for setting preventive and fighting strategies (Vannucci 2012). Up to now, the data more frequently used to measure corruption derive from the crossing of different sources of information such as, surveys, statistical analysis, legal documents and media reports, from which we can gain some evidences on the qualitative and quantitative characteristics of the phenomenon (Vannucci 2010). However none of these sources allows obtaining a precise measurement of the phenomenon. In Italy, two types of data are mainly used to measure corruption and to report it: 1. International surveys and indexes; 2. Judicial statistics and data. Starting from the first one, we could only remind here that these types of methodological exercise often measure individual perceptions about the existence of corruptive practices, or about direct experiences and do not represent an effective measurement of corruptive actions. These tools are the ones that are most frequently cited by the media as they present data in a comparative way and provide standardized information, easily understandable to the general public. However, these surveys are not organized on regular annual basis. As such, they offer only a spot description capturing only some aspects of the phenomenon. Moreover, they often represent a small sample of the entire population or only certain categories of professionals (often enterpreneurs) enquired, throughout interviews or questionnaires, about their percepiton and/or experiences. Among the most famous and cited are the corruption perception index of Transparency International (Corruption Perception Index, CPI) and some specific Eurobarometer surveys. The data for Italy suggests a dramatic deterioration over the last years. In fact, the CPI shows that Italy is fallen from the 40th position in 2005 to the 61st position in 2015. Furthermore, the World Bank's World Governance Indicator (WGI) reveals a considerable decrease of the control of corruption over the recent years. These data about perception were confirmed also by the 2013 Eurobarometer survey on Corruption stating that corruption is still widespread within EU Member States: in Italy, 97% of respondents believe that it is a common phenomenon, rooted in the national culture, while 88% considers that corruption is one of the quickest ways to access the public service. In addition, according to the Global Competitiveness Index 2013-2014, Italy is the country where corruption has the most negative effect on the economy, beeing indicated as the major obstacle when it comes to doing business in country.

The other main source of information about corruption in Italy is represented by the judicial statistics counting for the number of investigations, proceedings, and trials for corruption crimes, over the time. Although these types of data could be potentially crucial to understand the phenomenon, they provide only a partial representation, as they are essentially based on the amount of prosecution action taken by judicial authorities. In other words, these data represent only those corruptive facts for which someone has decided to blaming and denouncing. Moreover, data on corruption crimes present many concerns in term of reliability and such a problem could mislead the tendency analysis.

7 First of all, in Italy, judicial statistics and aggregated database are not updated and the way data are processed is not the same for each judicial office. The Ministry of Justice is always laggard in collecting data on proceedings and, according to judges and prosecutors8, the method applied to collect data is not well organized. The level of reliability could be really different for each office and “each chief public prosecutor could organize and collect data differently and slightly modify whether or not to insert within the statistics some (may be not so relevant) proceedings”9. Huge differences could also derive from the way cases and proceedings are registered. Each office could adopt slightly different methods of classifying and registering the claims, producing, involuntarily, a jeopardized landscape of data and statistics. According to judges and prosecutors, data about corruption related crimes are among the more difficult to find out as, being complex crimes to disentangle, much depends on the way each office decide to classify the claims and to register them, when the public prosecutor starts an investigation. In addition, it is worth to notice that judicial statistics are updated very slowly and it is difficult to extract from the official statistics of the Ministry of Justice fresh data. To date, in 2016, the whole available and aggregated data on the number of investigations and/or indictments for corruption related crimes at national level are limited to 2011. As regards to the amount indictments, there are many factors to consider in interpreting it. In Italy, a consistent part of the public opinion10 and several judges agree that some of the so called Ad personam laws, approved during the years of Berlusconi governments, had a huge impact in decreasing the amount of final indictments for corruption charges. Among those laws, the Ex-Cirielli Law foreseeing the limitation period for many so-called “minor offences” reduced at 7½ years instead of 15 impacting on indictments for corruption, tax evasion, bankruptcy, fraud, and, generally speaking the bulk of offences relating to crimes against the public administration. This presumably means that a significant number of crimes would have not been prosecuted due to the new law, but unfortunately it is very difficult to provide reliable data on this matter. The available data are often conflicting, in that they show either a slight increase in the number of proceedings being concluded because the limitation period has expired or even a quite significant decrease.11 According to Davigo (2007), the different surveys or interviews given by specialized bodies in the measurement of the phenomenon confirmed the mass perception of corruption, but only a small part

8 In order to reinforce this argumentation I will cite some extracts drawn from several hours of training I have personally held at the Italian National School of the Magistracy (Scandicci, Florence). The training sessions I have held, devoted to judges applying for court-president or chief-prosecutors, were specifically focused on “how to collect and process data about the office” and consisted in some hours of lecturing and some other hours of focus-groups (see http://www.scuolamagistratura.it/formazione-dirigenti/formazione-dirigenti-anno-2016.html). The extracts cited here are drawn from the minutes of the focus-groups. 9 See Footnote 8. 10 See, as an example, https://www.transparency.it/il-paese-degli-impuniti/ (the country of the unpunished). 11 Some data on the number of crimes cancelled due to the new statute of limitations are available on the DG-STAT website of the Statistical Office of the Ministry of Justice, which was created in 2001. However, these data are incomplete, because they are only available for certain judicial districts in Italy. See http://webstat.giustizia.it/AreaPubblica/default.asp. On this specific aspect see also Dallara (2015). 8 of this phenomenon comes to the attention of the judicial authorities and a intrinsic lack of evidence makes harder to reaching a final judgment. Moreover, corruption is difficult to be measured and proved in a country having its flaw of submerged economy and having the primacy for the amount of contradictory and ambiguous laws. Much has been made after the scandal of Tangentopoli both in term of research and study of the phenomenon, to understand the mechanisms and the actors involved, and to trace the amount of repressive measures (quantity and quality of the sanctions), but it is still not enough for a deepened analysis. A systematic study on the investigations, from the opening of the procedure (when the judicial authorities decide to start investigating a case involving one or more subject – avviso di garanzia) until the end (if filed in court – rinvio a giudizio and the subsequent final decision - sentenza) could offer a new perspective to analyze this phenomenon. Although it is true that the exercise we are proposing here only provides a limited picture, focused on investigations related to cabinet ministers, vice- ministers and undersecretaries, as covered by the media, it offers some interesting insights on the amount of judicial activity in relation to corruptions crimes, involving high-level politicians. Besides, it reinforce the statement about the difficulty to prosecute political corruption reaching a final judgment in court, especially with a conviction verdict. Moreover, it could suggest new reflections about judicial activism of the Italian magistracy and its perceived politicization, adding new hints to some existing studies (Ceron and Mainenti 2015).

3. The study of Popova and Post (2013) and its replication in Italy

The research of Popova and Post (2013) represents an important pilot study to analyze the prosecution of cabinet ministers. They provide an empirical analysis of the phenomenon, by examining seven Eastern European democracies (Bulgaria, Croatia, the Czech Republic, Macedonia, Poland, Romania, and the Slovak Republic) from 1999 to 2013. More specifically, they focus on the extent of corrupt cabinet ministers' and they observe the factors linked to a higher probability of indictment. To this aim, the authors formulate three hypotheses to understand what leads to corruption indictments. In more detail:

1) The first hypothesis is that “engaging in corruption increases the probability of indictment”. Even if, criminal justice systems may produce some distortions because sometimes-guilty individuals avoid sanction and innocents get punished. They argue that some ministers have greater opportunities to commit corruption-related offenses because their portfolios involve more discretionary spending than average, or contracts in less competitive sectors. Thus, they posit that corruption activity could be proxied by the minister’s portfolio assignment. 2) The second hypothesis is focused on cabinet ministers' party membership. They hypothesize that “belonging to a stronger party lowers a minister’s chances of an indictment because

9 stronger political parties can protect their members from indictment”. 3) The third hypothesis is that “ministers in countries with independent anti-corruption prosecution agencies will face a higher probability of being indicted, ceteris paribus”.

To test the three hypotheses, they have built the “Eastern Europe Corruption Prosecution Database” which contains information on 863 cabinet ministers in the selected countries. The database includes names, portfolio, tenure (starting and ending), party membership, gender, and the corruption prosecution experiences, which describes the type of investigation, the timing, if it was filed in court and finally its conclusion (distinguishing between dismissal, acquittal or conviction). The descriptive analysis highlights that 59 cabinet ministers out of 863 have faced corruption-related criminal charges over a period dating from 1999 to 2013. Observing the indictments rates by country, they found that Bulgaria and Romania have roughly the 10% of cabinet ministers criminally indicted; Macedonia, Croatia, Poland and the Czech Republic the 5-7%, while in Slovakia just two ministers have been prosecuted. Furthermore, Popova and Post (2013) underline that most indictments take place while ministers were no longer in office, noticing an average lag between minister's term in office and indictment of 21 months. For this reason, they include a control variable that reflects the amount of time spent in the office. Then, in order to test the first hypothesis they assume that corruption activity is linked with the minister’s portfolio assignment. They identify the corruption-prone portfolios using Transparency International’s Bribe Payers Index ranks sectors. In addition, they consider the corruption's perception in the country during the minister’s tenure in office, using the World Bank’s World Governance Indicator Control of Corruption. On this basis, they point out that the ministries of agriculture, transportation/communications, and economy/industry would be furnished by more corruption opportunities. In fact, these portfolios are characterized by more discretionary spending, and they could conclude contracts in less competitive sectors. Testing the second hypothesis, they operationalize two proxies of party strength—junior/senior status in the governing coalition and being a communist successor party. The rationale for the first variable is that senior governing coalition parties would have greater access to state resources than their smaller partners. Furthermore, with regard to the third hypothesis, they have examined the presence of anti-corruption agencies in each country. The analysis confirms the central role played by anti-corruption agencies to tackle political corruption as indictment rates' increase where they operate.

3.1 The replication of the study in the Italian case

The work presented here is based on an empirical exercise, put in place in 2015-2016, conducted by

10 the author, together with two students of the University of Bologna12. The time period analyzed covers from 2006 to 2013. In these seven years, Italy has experienced a peculiar political phase, since three governments, with three “different orientations”, were successively in office. All of them ended in advance for the resignation of their leaders. The first was the the center-left Prodi II government, from 17 May 2006 to 8 May 2008; this was followed by the center-right Berlusconi IV government, in office from 8 May 2008 to 16 November 2011, and finally, from 16 November 2011 to 28 April 2013, the technocratic government headed by Monti. The second Prodi government, remained in office for 1 year, 11 months and 21 days; it was composed by 26 ministers, 10 deputy ministers and 66 undersecretaries, for a total of 103 individuals13, including the Prime Minister, Romano Prodi. The governing coalition was named the Union (Unione), composed by many and heterogeneous parties. Among the major parties the Democrats of the left (DS), Democracy is freedom – the Daisy (DL), Communist Refoundation Party (PRC), the Greens (FdV), (IdV). The Berlusconi IV governing coalition was composed by (PdL), , Movement for Autonomy (PT-MpA) and, since March 2011, the Populars of Italy Tomorrow (PID); other small groups and movements joined the coalition that was characterized by several changes in its composition. Despite that, the two major parties of the coalition remained PdL and Lega Nord. This cabinet remained in office for 3 years, 6 months and 8 days and it was initially composed of 21 ministers, 37 undersecretaries and no deputy minister, for a total amount of 58 members, with Silvio Berlusconi as head of government. At the time of the resignation of the President in 2011, the Government was composed by 22 ministers, three deputy-ministers and 37 undersecretaries (63 components). Considering these changes in the government composition, we have worked on a total amount of 83 political actors. Finally, the technocratic government of Monti, which lasted 1 year, five months and 12 days, was composed by 20 ministers, 32 undersecretaries, 4 of which acted also as deputy minister, for a total amount of 52 political actors. Mario Monti, acted as Prime Minister and as Minister of Economy and Finance until 11 July 2012, when the office was assigned to . The lists of the individuals considered for our mapping exercise were constructed with the names of each Minister and Undersecretary from the website of the Italian Government, specifying the portfolio they held and/or the Ministry in which they worked, how long the tenure lasted, the party affiliation and gender. In order to test the hypotheses previously described in the Italian case we applied the same method followed by Popova and Post (2013). We used the Google search engine employing some

12 Luigi Rullo, collaborating also in writing this paper, and Veronica Gamper working for her dissertation and doing the analysis for the Prodi government (2006-2008). 13 The Prodi government was characterized by the early dismissal of two political actors: Marco Verzaschi, Undersecretary of Defense that resigned on December 7, 2007. He was be investigated for corruption and bribery three days after his appointment; and Paolo Giaretta, Under-secretary for the economic development, who opted for a seat in Parliament. For this reason, the considered total amount of individuals that served this cabined is 105. 11 fixed keywords, adopted together with the name of each member of the three governments presented above. The sequence used was the following one: "name of the Minister / Deputy Minister / or Undersecretary + investigation + corruption", for example: "Mastella + investigated + corruption." This was the input for the first level research, then, after having found some initial information, a snow-ball strategy has been applied using more specific keywords. Often, after the case has been identified it was used directly the Minister (or Undersecretary) name + the kind of crime or the specific offense, for example: "Mastella + investigated + abuse of office." The same logic was applied to get more specific information about the steps of the judicial investigation using the name of the investigation "Mastella +investigation + name of the inquiry". In selecting the sources of information we adopt a “hierarchical method”. Firstly, newspaper articles, then, websites and finally blogs and. We focused in particular on the four widely diffused national newspapers, such as "La Repubblica", "La Stampa", "Corriere della Sera" and "Il Sole 24 Ore". Additional information was drawn from opinion magazines like "Il Fatto Quotidiano", "Il Giornale", or "The Huffington Post Italy". Then to verify some aspects mentioned in the national press, we also used blogs or web-pages specifically focused on corruption of the Italian political leaders14, although they cannot be considered as totally unbiased and reliable sources of information. As for undersecretaries, in addition to the aforementioned sources, also web-sites and blogs of local newspapers were consulted, as frequently the news were reported in the territorial area to which undersecretaries belong to. Table 1 shows the structure of the Database as it was filled.

Table 1. Structure of the database

Name of Portfolio Tenure Party- Gender Corruption- When? If yes, Ended How? the (starting membership related (1)Before Filed in (1)dismissal Minister and investigations (2)during court? (2)acquittal ending) (3)after (3)conviction tenure (4) pending (5) statute of limitation 1. Franco External Full PdL M YES During Procura di DISMISSAL- Frattini Affairs (Abuse of (2011) Roma CLEARED office) (then Trib. dei Min) 2. Roberto Interior Full Lega M YES After Procura di PENDING Maroni Minister Nord (Abuse of (2014) Milano, (Home office) Inchiesta affairs) Expo 3. Ignazio La Defence Full PdL M YES During Corte dei CLEARED Russa (Embezzelment) (2011) Conti 4. Nitto Justice Until PdL M NO ----- Francesco 27/07/2011 Palma 5. Angelino Justice From PdL M YES (Abuse of During Procura di DISMISSAL Alfano 28/07/2011 office) (2008) Roma

Source: extract from our database.

14 As an example, “L’incredibile parlamento italiano (the unbelievable )”, Politicicorrotti.it (CorruptPoliticians) or “Il Blog di Beppe Grillo (the Blog of Beppe Grillo, Five star movement)”. 12 Overall, our mapping was conducted on 240 subjects who served as Ministers, Deputy ministers and Undersecretaries, from 2006 to 2013 (see Table 2). 60 people out of 240 have been investigated (Table 3) for one or more of the following offences: corruption, embezzlement, fraud, organized crime (associazione a delinquere), loss of revenue, defamation, vote trading, facilitation, market manipulation, fraudulent misrepresentation, slander and threats, bid rigging, illegal financing, induction by compulsion, fraudulent misrepresentation, false accounting, improper influence, fraudulent bankruptcy and false reporting15

Table 2. Amount of individuals considered for mapping in each government (2006-2013)

Total Ministers and Undersecretaries considered Government for the mapping exercize

105 Prodi II (2008-2008)

83 Berlusconi IV (2008-2011)

Monti (2011-2013) 52

Total 240

Source: our elaboration from the database.

Table 3. Total of individuals under investigations for corruption related charges.

Under investigations? Prodi II Berlusconi IV Monti Totale (1 y, 11 m) (3y, 6m) (1y, 5 m) No 85 50 45 180

Yes 20 33 7 60

% Yes 19,1% 39,7% 13,4% 25%

Total cabinet members 105 83 52 240

Source: our elaboration from the database.

Another interesting data, to be recalled later on, in order to draw some conclusions and further reflections, is the one concerning the total number of investigations for each cabinet, bearing in mind that there are several individuals that were subject to more than one investigation each one (see Table 4).

15 The translation of the specific terms related to the offences could be not completely correct due to the peculiarity of the juridical meanings. Here, as follow, the list of the Italian terms: corruzione, peculato, truffa, danno erariale, diffamazione, voto di scambio, favoreggiamento, aggiotaggio, falso ideologico, calunnia e minacce, turbativa d’asta, finanziamento illecito, induzione per costrizione, dichiarazione fraudolenta, falso in bilancio, influenze illecite, ricettazione e appropriazione indebita, bancarotta fraudolenta e false comunicazioni. 13 Table 4. Total amount of investigations in each cabinet

Prodi II Berlusconi IV Monti (1 y, 11 m) (3y, 6m) (1y, 5 m) Ministers 18 25* 3

Undersecretaries 14 18 4

Total 32 43 7

Source: our elaboration form the database.

It is worth to stress that, data concerning the total amount of investigations related to ministers of the Berlusconi IV cabinet, do not consider the investigations against Berlusconi that are 11*. In the In the Table 4bis we tried to “normalize” the total amount of the investigations for the months each government spent in office.

Table 4bis, number of investigations normalized for the months in office

Prodi II Berlusconi IV Monti (23 months) (42 months) (17 months) Number of investigations normalized for the months in office 1,4 1,3* 0,4

Source: our elaboration from the database.

As for the Berlusconi IV we considered also the 11 investigations related to Berlusconi himself and thus counting 54 investigations (43 + 11). As concerns members subject to more than one investigation for corruption related charges these are: 7 in the Prodi II cabinet (5 Ministers and 1 Undersecretaries); 9 (6 Ministers and 2 Undersecretaries) in the Berlusconi IV and none in the Monti Cabinet.

3.2 Testing the three hypotheses

The information collected through the media analysis presented above has been used to test the three hypotheses advanced by Popova and Post (2013) in relation to the Eastern Euorope countries.

HP 1: The corruption-prone portfolios are more frequently subject to investigations

Concerning the first HP, Popova and Post (2013) have assumed that the ministers, who hold office in a ministry with a portfolio in itself more susceptible to corruption, are those having major probability to be subject to investigations; this is because certain portfolios and departments are much more exposed to substantial flows of money or because there is a greater chance to take crucial decisions on

14 substantive and structural policies. The portfolios identified in the study of the two authors are those suggested by the Transparency International Bribe Payers Index16: 1. Ministry of agriculture; 2. Ministry of transport/ communications; 3. Ministry of economy and/or Industry; 4. Ministry of defense. The information gathered in our database allow us to test this hypothesis for all the three cabinets. Table 5 shows comparatively the portfolios with major number of investigations in each one of the cabinets.

Table 5, portfolios with major number of investigations in each one of the cabinets.

Government Portfolios with more investigations Portfolios with more Total (ministers) investigations (undersecretaries) Justice: 6 investig. (*) Economic development: 6 Justice: 7 Environment: 2 investig. Defence:2 Economic development: 6 Prodi II Infrastructures: 2 investig. Finance and economy: 1 Infrastructures: 3 External affairs: 2 investig. Justice:1 Cultural Heritage: 3 Cultural Heritage: 2 investig. Cultural Heritage:1 Environment: 2 Social Solidary: 1 investig. Infrastructures:1 Defense: 2 Education: 1 investig. Transports:1 External affairs: 2 Regional affairs:1 investig. Youth:1 Finance and economy: 1 Healthcare: 1 investigation Social Solidary: 1 Education: 1 Regional affairs: 1 Healthcare: 1 Transports: 1 Youth: 1

Economic development: 5 Finance: 4 Economic development: 9 Agriculture: 4 Economic development: 4 Agriculture: 7 Berlusconi IV Infrastructures: 2 Agricultural policies: 3 Finance: 6 Regional affair: 2 Justice: 1 Infrastructures: 3 Finance: 2 Infrastructures: 1 Bureaucratic simplification: 2 Environment: 1 Education: 1 Environment: 2 External affair: 1 Reforms: 1 Justice: 2 Interior: 1 Bureaucratic simplification:1 Regional affair: 2 Defense: 1 Environment: 1 Interior: 1 Justice: 1 Work: 1 Defense: 1 Health: 1 Health: 1 Semplificazione:1 Federalism: 1 Federalism: 1 Tourism: 1 Tourism: 1 Decentralization: 1 Decentralization:1 Work: 1 Reforms:1 Education: 1 External affair: 1 Interior:1 Undersecretary to the Undersecretary to the Environment:2 Presidency:1 Presidency:1 Monti Health:1 Health:1 Justice:1 Justice:1 Culture:1 Culture:1 Interior: 1 Environment: 1 *: 5 of them were in charge to the Minister Mastella. Source: our elaboration from the database.

16 The Transparency International’s Bribe Payers Index ranks sectors of the economy according to the frequency with which survey respondents identify them as dominated by bribe-demanding bureaucrats and politicians and bribe-offering business actors. See http://www.transparency.org/bpi2011. 15 As for the Prodi II executive data show that of 32 total inquiries, 6 were related to activities involving the Ministry for economic development, 3 the Ministry for Infrastructures, while 2 investigations pertain to the defense Ministry. The validity of the first hypothesis is also confirmed by the data collected for the Berlusconi IV and Monti cabinets. These data show that the portfolios with the highest number of investigations are the Ministries of economic development, both for the Berlusconi IV government and the Monti one; followed by Infrastructures and Finance. The Ministry of Agriculture presents high scores only for the Berlusconi government.

HP 2: Stronger parties are able to insulate their members from investigations

In order to verify the hypothesis 2, we need to adapt the methodology followed by Popova and Post (2013) to the characteristics of the Italian political system. The concept of "strong party", as operationalized by the two cited authors, cannot be adequately transposed into the Italian parties system. In fact, in order to assess the party strength they proposed two proxies: junior/senior status within the governing coalition and being/not being communist successor party. The rationale for the first variable is that senior governing coalition parties would have greater access to state resources than smaller and newly created parties. Although, it is a really interesting reflection, its usefulness is limited to the post-communist political context.

For the Italian case, we propose to test this hypothesis verifying if the individuals belonging to the major parties in governmental coalitions are those less charged of investigations. The aim is to check whether or not the number of investigations opened is lower against the individuals belonging to the major parties of the governing coalitions.

The results for this HP testing are not highly significant for the Italian case as the majority parties of the governing coalitions are those displaying more inquiries, as total number of individuals under inquiry, even if also the other “smaller” parties are interested by an high number of investigations (Table 6).

As for the coalition of the Prodi II government, called “the Union”, it was formed by the Ulivo (composed by the Left Democrats - DS, by Democracy is Freedom - DL, plus some independent members), plus a number of other minors parties, including Udeur, Italy of Values, Communist Re- foundation and others. The “major parties” could be identified in those forming the Ulivo. On the total amount of individuals (20) investigated in the second Prodi government, 14 belong to the Ulivo group. More precisely, 5 belong to the DS party, 8 to DL and 1 was an Independent for the Ulivo.

If we try to consider this data in relation to the whole number of the cabinet members for each party, adopting a sort of normalization (Table 6bis), the results display that 19% of the individuals belonging

16 to the major parties (14 out of 74 members) were under investigations; while the percentage is almost the same for the “smaller” parties of the collation: 20% of the individuals were under investigations (6 out of 30 members).

Table 6, Party membership of Ministers and Undersecretaries under investigation

Government N° of ministers under investigation for each N° of undersecretaries under investigation for each political party political party Prodi II DL: 3 (out of 7 DL ministers) DL: 5 (out of 21 vice-ministers and undersecret.) DS: 1 (out of 9 DS Ministers) DS: 4 (out of 28 vice-ministers and undersecret.) Indip. ULIVO: 1 (out of 3 ULIVO Ministers) Indip. ULIVO: 0 (out of 4 undersecret.)

UDEUR:1/1 UDEUR:1/1 PrC:1/1 RnP:1/4 IdV:1/1 PrC:0/7 FdV:1/1 IdV:0/2 Others: 0/3 FdV:0/2 Others: 0/7 Berlusconi IV PDL: 13 (out of 17 Ministers) PDL: 11 (out of 53 vice-ministers and undersecret. considered) Lega Nord:3 (out of 3 Ministers) Lega Nord: 3 (out of 5 vice-ministers and undersecret.) PT:1 (out of 1 Ministers) PT:2 (out of 2 vice-ministers and undersecret) Monti Independents Independents

Source: our elaboration from the database.

Table 6bis, N° of individual (ministers and undersecretaries) under investigation in the major parties vs “smaller parties” out of the total of the members

Government N° of individuals under investigations N° of individuals under investigations belonging to belonging to the major parties the “smaller” parties Prodi II ULIVO (DL +DS +Indip. ULIVO): 14 out of 74 UDEUR+RnP+PrC+IdV+FdV+Others: 6 out of 30 members (19%) members (20%)

Berlusconi IV PDL: 24 out of 56 members (43%) Lega Nord + PT + others: 9 out of 27 (34%)

Monti Independents Independents

Source: our elaboration from the database.

The situation is slightly different for the Berlusconi IV cabinet. Here the overall number of investigated members is 33, including ministers and undersecretaries, 24 belong to the PDL (People of Freedom), the founding party of the ruling coalition between 2008 and 2011, followed by the Northern League, the second-largest party in the government composition, with a total of 6 members investigated, and 3 members investigated in the small PT party (see Table 6). If we consider these data in relation to the total amount of the PdL member present in the cabinet (both as Ministers and Undersecretaries) the results show that 24 out of 56 PdL members were under investigations (43%) while, in the “smaller” parties, the individuals under investigations were 9 out of 27 members (34%). Thus, for this cabinet

17 the HP2 propoved by Popova and Post (2013) is not validated. Here the major party (PdL) was the one more subject to the judicial action. The HP2 cannot be checked in relation to the Monti cabinet in reason of its technocratic nature. The members of that government were not professional politicians, and only 6 of 52, were members of a political party, while one was a military. In fact, they are properly classified as independent.

The second hypothesis, therefore, is not expressively validated in the Italian case. Besides, the data collected testify a strong bipartisan judicial attention on the members of the major political parties, both in the center-left and the center-right governing coalitions (see Table 6). What it is clear from the data is that, both in the center-left and in the center-right executive, the major parties of the governing coalition were massively interested by judicial investigations for corruption related crimes.

HP 3: the presence of an independent anti-corruption agency would increase the probability of indictment

The hypothesis 3 is not applicable to the Italian context. The Italian anti-corruption authority was officially in place only from 2014 (Decree Law 90/2014, converted into Law No. 114/2014) as ANAC (National Anti-corruption Authority). It is functionally independent from the executive, but it is only entrusted with supervision and prevention of corruption within the public administration, in checking companies and tenders (following public contracts), in spreading transparency and legality criteria within those sectors sensitive to corruption phenomena (Canton and Merloni 2015). Moreover, this hypothesis could be checked only in comparing more country cases.

Timing and final outcome of the investigations: the contentious topic

Some interesting reflections could emerge if one looks at the timing of the investigations in relation to the cabinet’ tenure and, even more, at their final outcome, both in term of how many (on the total of the investigations opened) have been then filed in court and how they ended (meaning if they ended with conviction or with other conclusion). As for the timing, Table 7 shows that the majority of the investigations started during the tenure in office of the cabinets. However, the number of the investigations started after the end of the tenure is quite high.

Table 7, Timing of the investigations in each cabinet, Ministers (Min) and Undersecretaries (Und)

Ministers Prodi II Berlusconi IV Monti

Min Und Min Und Min Und Before 2 7 2 3 1 1

18 During 12 3 12 6 2 2 After 8 4 11 9 -- 1 Total 18 14 25 18 3 4

Source: our elaboration from the database

One of the most controversial point concerns the number of investigations that, after the period of judicial inquires, are then filed in court. This is particularly relevant as, only for the investigation in which concrete elements of suspect are found, the judge for the preliminary hearing decides to file them in court starting with the real trial phase. Table 8 shows that the number of investigations filed in court is quite low compared to the number of investigations opened.

Table 8, Total amount of investigations filed in court (rinvio a giudizio)

Government Total Ministers' Undersecretaries Total Investig filed in Investig investigations filed in investigations filed in court court opend court 32 8 9 17 Prodi II 43 16 12 28 Berlusconi IV 7 2 2 4 Monti

Source: our elaboration

Data about the conclusion of the investigations are the most relevant to be considered, but at the same time the less reliable to be assessed. This is somewhat due to the problem mentioned in Section 2 of this paper (official statistics not updated and not homogeneous), but also due to the Italian media system and its attitude in reporting judicial cases about corruption.

With reference to the media sources it was quite easy to find information about the opening step of the investigations, meanwhile it was difficult, due to the lack of news, to trace the final steps of the inquires. For the Prodi II cabinet, due to the time elapsed, we were able to find a good number of final sentences and thus we are sure about the conclusion of such investigations. For the Berlusconi IV cabinet we encountered more difficulties due to the high number of cases still pending.

As already mentioned in the previous sections, data on the amount of cases ended due to the statute of limitation are really difficult to find out as this information is not frequently spotted in media, except for the cases that catch permanently the public opinion attention, such as Berlusconi’s trials. Moreover, no official information could be found on that type of conclusion in the official judicial decision databases. Further analysis and additional sources would be needed to verify this important aspect. Thus, data presented in Table 9 have to be read with caution, especially those referring to the

19 fifth (Pending cases) and sixth (Statute of limitation) columns, as some cases that we codifed as still pending could then be ended due to the statute of limitation expiration.

Table 9, Final outcome of the investigations in each cabinet

Government Dismissal Acquittal Convinction Pending Statute of Tot (archiviata) (Assolto) (Condanna) (Pendente) limitation (Prescrizione) 15 4 4 7 2 32 Prodi II % 47% 12% 12% 22% 7% 100

Berlusconi IV 15 3 13 12 - 43 % 35% 7% 30% 28% 100

Monti 3 1 - 3 - 7 % 43% 14% 43% 100

Italy 2006-2013 33 8 17 22 2 82

% 40% 10% 21% 27% 2%

Source: our elaboration form the database

On the contrary, data about the amount of dismissed cases, about acquittals and convictions are quite reliable as we were able to find precise news or the final decision in the official databases.

Conclusion (to be further developed)

What the data collected throughout the mapping exercise mean for Italy? What in term of relation between courts and democracy, in term of judicial activism and its effect on the fighting against corruption? A confirmation of an high level of “judicial activism”, meaning judicial activity in prosecuting political corruption, if compared with the other countries considered by Popova and Post (2013) in Table 10.

Table 10. Indictment rates by country according to Popova and Post (2013)

Czech Slovak Indicted Bulgaria Croatia Rep Macedonia Poland Romania Rep Total No 94 87 119 115 159 149 81 804 Yes 12 6 7 9 8 15 2 59 Indictment % 11.3% 6.5% 5.1% 7.3% 4.8% 9.1% 2.4% 6.8% Total 106 93 126 124 167 164 83 863

Source: Popova and Post (2013)

20 Data about Italy show a peculiar attention of the magistracy on the high-level members of the cabinets. More than a politicized justice – for two decade the target of the Berlusconi rhetoric (Dallara 2015) – it seems a bipartisan prosecution of the governing coalition members, independently from the political orientation. A magistracy that perceives itself as charged by a civic mission towards the society?

In this respect, these first results are somehow contrasting with the study of Ceron and Mainenti (2015) that have analysed the behaviour of the magistrates depending on their political inclination, studying the requests for authorization to serve against deputies of the Chamber from 1983 until 2013. A total of 1,256 requests (limited to crimes as corruption or against the Public administration) towards 1,399 deputies was analysed, showing that the more the judge’s political position diverges from the party’s ideological positions, the more the inclination to investigate against that party grows. Conversely, the results of this mapping exercise show an high number of investigations, independently to the political orientation of the cabinet.

Then, another main point to the notice is that, in spite of a high number of investigations we found out a few amount of convictions. In fact, a high number of inquires ended with a dismissal conclusion.

What are the effects of such a situation in term of political and electoral accountability? (Sberna and Vannucci 2013). An effective and “impartial” prosecution against corruption practices carried out by high-level elected politicians could foster electoral accountability. More specifically, prosecution of corruptive practices might impact electoral accountability if effective prosecution of corrupt practices turn finally into convictions and legal sanctions. If not, the risk is the opposite. Sberna and Vannucci (2013) define this situation as the politicization of anti-corruption; Political actors would start to profiling themselves as victims, increasing the intensity of the debate and opposing against or fostering the judiciary. Party leaders will tend to systematically discharge suspected politicians to minimize the feared electoral drawback. A loopholes effect of the judicial system that makes the prosecution of corruption crimes ineffective most of the times. Some judges and some scholars find in the statute of limitation the main cause of this situation. In truth, this is not clearly detectable. No official data are available for corruption-related cases. The lack of conviction negatively impacts the legitimacy of the judicial activism and consequently weakens the fighting against corruption.

M-factor (Calise 2016) !

Media: obsessive attention of the media in the first phases of the investigations. Just after the opening of the inquiry iper-detailed news about all the political actors involved even if not directly touched by the investigations. Frequent cases of ministers and undersecretaries name only mentioned in telephone tapping that for the media are considered the protagonist of the inquires. Meanwhile, only few and uncorrected information about the following steps of the investigation path. Our exercise

21 show as it is difficult to trace the entire process using only media sources. The ending phase of the investigations is reported more frequently if there is a conviction, rarely if the inquiry has been dismissed.

References: (to be completed)

22