Cleaning mafia cash: An empirical analysis of the money laundering behaviour of 2800 Italian criminals

Abstract

Despite the wide reach of anti-money laundering legislation worldwide and increasing media attention, fostered by journalistic leaks such as Panama Papers, empirical knowledge on how criminals launder their illicit proceeds is still scarce. The few available empirical studies show that money laundering (ML) schemes are often less sophisticated than they are depicted in the political and media debate. To contribute to the empirical knowledge of ML behaviour, and test this hypothesis, the present study analyses the ML activities related to 2818 Italian offenders included in the ML section of the LexisNexis’ WorldCompliance database. Through a quantitative content analysis of textual information related to each offender's profile, it highlights the characteristics of the ML offenders, the methods (or ‘typologies’ in FATF terms) employed, the assets seized, the business sectors involved and the countries in which ML was conducted. The results confirm that criminals tend to employ unsophisticated typologies, as well as prefer Italy or jurisdictions that are close (geographically and culturally) to Italy for laundering their illicit proceeds. Tangible assets (first real estate and registered vehicles) are more frequent than financial assets. Finally, differences exist between the laundering by mafia-related ML offenders and non-mafia ones. The study provides empirical ground to progress in the knowledge of how ML offenders behave, and supports the idea that criminals, when laundering their proceeds, do not act as legitimate entrepreneurs, but may be driven by other constraints and drivers.

Keywords

Cash Italy mafia money laundering organised crime

Introduction

Contrary to other crime types, money laundering (ML hereinafter) ‘is notable for the diversity of its forms, participants, and settings’ (Levi and Reuter, 2006: 312), and can be carried out in different modi operandi, ranging from basic to highly sophisticated ones (Arnone and Borlini, 2010). To counter it, international bodies and national governments have put in place since the early 1990s an articulated regulatory framework, with a number of recommendations – first of all those issued by the Financial Action Task Force (FATF hereafter) – and obligations imposed on both the public and private sector.

Despite the width of these interventions, and increasing media attention (fostered by leaks such as Panama Papers), the history of ML ‘has been more supported by righteousness than by empirical facts’ (Van Duyne et al., 2018: 10). The existing knowledge is still disproportionately based on journalistic exposés and sensationalistic claims by institutional bodies while empirical research is lagging behind, resulting in the persistence of ‘folk theories’ about how illicit proceeds are laundered (Halliday, 2018).

This study contributes to addressing this knowledge gap by analysing the ML activities of a sample of 2818 Italian offenders extracted from the LexisNexis’ WorldCompliance database (LN WoCo hereinafter). By means of quantitative content analysis (QCA), we coded information on offences, laundering methods, assets and countries involved in the ML schemes from the textual data associated with each profile, providing insights and empirical evidence about the behaviour of Italian ML offenders.

The study is structured as follows. In the second section, we review the empirical studies which analysed patterns and behaviours associated with ML activities, and we formulate the research questions. In the third section, we describe the data and the methodology used in the analysis. In the fourth section, we present the results of the analysis which are further discussed in the fifth section, together with both research and policy implications, and the study's limitations.

Literature review

Literature on ML may be broadly classified into three main branches (Kruisbergen et al., 2015). The first one focuses on estimating the size of the phenomenon. The second branch of literature focuses on assessing to what extent anti-money laundering (AML) policies are effective and cost-effective, a central question in ML research (see, e.g., Barone et al., 2022; Ferwerda and Reuter, 2019; Halliday et al., 2020; Harvey et al., 2021). The last branch focuses on investigating how criminals launder and integrate their illicit proceeds into the legal economy. Most of the contributions to this branch come from economists (Van Duyne et al., 2018). Following Gary Becker's 1968 seminal work ‘Crime and punishment: an economic approach’ (Becker, 1968), several economists applied neoclassical economic principles to model criminals' decision-making (see for a review Ferwerda, 2009). In particular, a number of works focused on money launderers’ behaviour (Ferwerda, 2009; Masciandaro, 1999; McCarthy et al., 2015; Unger, 2007).

The monopoly of economists in the ML/AML literature assumes that criminals, similarly to rational economic actors, aim at maximising their economic returns when laundering their proceeds. However, criminologists – as well as numerous investigations worldwide – suggest that criminals may follow different drivers and constraints, for example, among others, the willingness to stay geographically close to their proceeds (Kruisbergen et al., 2015), avoid business sectors or assets with entry and exit barriers (Riccardi, 2014), exploit existing social ties (Van de Bunt et al., 2014), limit the involvement of third parties to minimise principal–agency costs (Levi and Soudijn, 2020; Reuter, 1983) and minimise the risk of being detected and/or having their assets seized (Riccardi and Levi, 2018). For this reason, classical economic models have been criticised for being too abstract and failing to properly describe criminal behaviour (Cornish and Clarke, 1985). As also advocated by economists themselves, these models should be complemented with insights from other social sciences to make them more realistic (Unger et al., 2006).

In this literature review (see Table 1), we focused on the empirical studies which analysed the behaviour of individuals arrested or convicted due to ML or whose assets have been seized because purchased with proceeds of crime. These studies are supplemented by those few which, by employing suspicious transaction reports (STRs) or suspicious activity reports (SARs), provided insights into the patterns, destination, use and modi operandi employed by criminals when laundering money. Eventually, empirical studies of the behaviour of ML offenders are few. They cover several jurisdictions and types of offenders, but their low number is symptomatic of the limited empirical knowledge on money launderers to date. Such scarcity is simply striking if we consider the number of recommendations, controls, sanctions and orders that have been issued globally in the AML domain over the last 30 years.

Table 1.

Previous empirical studies that investigated the behaviour of money launderers, grouped by country.

Country	Empirical studies on ML behaviour	Theme
Bulgaria	Petrunov (2011)	Money laundering methods
Canada	Schneider (2004)	Money laundering methods
	Beare and Schneider (2007)	Money laundering methods
	Malm and Bichler (2013)	Professional money laundering
Finland	Petrell and Houtsonen (2016)	Money laundering methods
Germany	Suendorf (2001)	Money laundering methods
Italy	Riccardi (2014)	Investments by mafia groups
	Cassetta et al. (2014)	Illicit financial flows
	Dugato et al. (2015)	Investments by mafia groups
	Gara and De Franceschis (2015)	Illicit financial flows
Spain	Steinko (2012)	Money laundering methods
Spain	Palomo et al. (2015)	Money laundering methods
Sweden	Skinnari et al. (2007)	Money laundering methods
The Netherlands	van Duyne and Soudijn (2009)	Money laundering methods
	Kruisbergen et al. (2015)	Investments by OC groups
	Soudijn (2014)	Professional money laundering
	Soudijn (2018)	Money laundering methods
	Custers et al. (2018)	Money laundering methods
	Custers et al. (2020)	Money laundering methods
	Kruisbergen et al. (2019)	Money laundering methods
	Kramer et al. (2023)	Professional money laundering
The United Kingdom	Matanky-Becker and Cockbain (2021)	Money laundering methods
The United States	Cummings and Stepnowsky (2010)	Professional money laundering

Source: Authors’ elaboration.

Previous empirical studies have focused on several areas of interest in this research domain, such as the type of ML methods (see, e.g., Irwin et al., 2011; Matanky-Becker and Cockbain, 2021), the investment strategy of organised crime groups (see, e.g., Dugato et al., 2015; Kruisbergen et al., 2015; Riccardi, 2014), the extent of the involvement of professionals and financial facilitators (see, e.g., Cummings and Stepnowsky, 2010; Kramer et al., 2023; Malm and Bichler, 2013), and transnational illicit financial flows (see, e.g., Cassetta et al., 2014; Gara and De Franceschis, 2015). Key trends and patterns emerging from the existing literature are discussed below.

The proximity of money laundering

Transfers of monetary values is almost costless in an increasingly globalised and digitalised economy. According to the standard economic approach, that would hold for crime money too: criminals would choose far distant countries – offshore jurisdictions, remote islands – that would offer financial and corporate opacity, while hampering the asset recovery by judicial authorities.

On the contrary, most of the empirical works mentioned above reveal that this happens only in exceptional cases. Proximity – in both geographic and cultural terms – is the most frequent evidence when it comes to ML across territories (Riccardi, 2022). In their analysis of around 1200 individual assets of organised crime offenders identified by Dutch authorities, Kruisbergen et al. (2015) showed that the majority were located either in the country of origin or in the country of residence of the criminals. Petrell and Houtsonen (2016) found that most of the assets held by Finnish motorcycle gangs were located in Finland, with the exception of a few assets in neighbouring countries such as Russia, Estonia and Sweden. Similarly, Steinko (2012), who analysed 367 cases of ML judged between 1995 and 2011 in Spain, concluded that, out of the total, only 23 (6.2%) of the cases had an international dimension. Similar results also stemmed from studies on Italian mafias showing that criminal assets were almost exclusively located in Italy and, more specifically, in those regions and provinces where the presence of mafia groups was the highest (Dugato et al., 2015; Riccardi, 2014).

Although these results may mirror the difficulties faced by law enforcement in recovering assets in distant places, proximity can also be observed when analysing STRs and SARs. As reported by Riccardi (2022), only a minor share of STRs in Italy, the Netherlands and Peru concern foreign countries; when it happens, most reports involve bordering jurisdictions (e.g., Germany, Luxembourg and Belgium for the Netherlands; Switzerland for Italy; other Latin American countries for Peru) or to countries with strong communities active in the country (e.g., China for Italy; Turkey for the Netherlands).

The importance of proximity may be determined by various reasons: the need by criminals to keep control over the territory where the money is generated or integrated, to minimise costs and the involvement of third parties (e.g., professionals, international tax advisers) who could increase principal–agency costs, and to have close locations in which to physically move dirty cash (Riccardi, 2022). In summary, all the other factors being equal, geographic and cultural distance between two countries are strong deterrents for money launderers (Ferwerda et al., 2020).

Unsophisticated money laundering

Despite the claims by policymakers and the media, a large body of literature suggests that ML modi operandi may not be particularly sophisticated (Kruisbergen et al., 2015; Riccardi, 2014). For example, in his analysis of ML trends in the last 20 years of Dutch police reports, Soudijn (2018) highlighted that, despite technological innovations, traditional ML methods keep returning over the years. This means, first of all, heavy reliance on cash-based schemes, trade-based ML and real estate investments. While ML through cryptocurrency is highlighted as on rise, the literature and the available judiciary evidence seem to stress that it still represents the minority of the cases.

Regarding assets, literature has pointed out the relevant role of real estate (Dugato et al., 2015; Unger and Ferwerda, 2011), vehicles, boats and helicopters (Calderoni et al., 2013; Savona and Riccardi, 2015) and high-value goods, such as precious metals and jewels (Petrunov, 2011; Steinko, 2012). Non-tangible and financial assets seem to be less widespread. But this result may be again due, on the one side, to the difficulty for asset recovery offices to trace and seize electronic funds; and, on the other, to the type of offenders covered by the analyses – mostly members of traditional organised crime groups.

Business sectors

The study of ML through legitimate firms and business sectors is mainly related to the study of the infiltration of organised crime groups into the legal economy, which has given rise in recent years to a number of works (for a review, see Savona et al., 2016). Following the standard economic approach, several authors suggest that criminals launder money in profitable business sectors (Masciandaro et al., 2007; Unger and Rawlings, 2008). However, the correlation between criminal infiltration and the business sector profitability has not been substantiated by empirical evidence yet (Riccardi, 2014).

On the contrary, various evidence demonstrates that criminals tend to prefer traditional sectors, which are not capital intensive, have low entry and exit barriers and are characterised by relatively simple legal forms (e.g., limited liability companies or cooperatives). In this framework, cash-intensive businesses, such as restaurants, hotels, construction and retail companies are historically preferred as they facilitate the comingling of illicit proceeds with firms’ turnover and facilitate the setting up of trade-based ML schemes (Riccardi and Levi, 2018).

Obviously, the choice of the business sector may also depend on other drivers and criminal purposes, such as the role played by the firm in the predicate offence itself. For example, sectors that are usually used for value-added tax (VAT) frauds (e.g., IT services) may also be exploited for laundering the resulting proceeds; similarly, import/export and transportation companies may be useful to cover for the international trade of illicit goods by organised crime groups.

Money laundering methods across different offenders: Mafia versus non-mafia actors

Studies of how ML methods change depending on the offender are very few. Only Irwin et al. (2011), through an analysis of 300 SARs from various countries, attempted to break down ML schemes according to the predicate offence. They demonstrated that fraud-related ML shows more complex patterns than the laundering of drug proceeds, among others. Other studies, including those reviewed in the previous section, focused on single predicate crimes or did not distinguish among them. However, some hypotheses can be formulated.

Relevant differences can be presumed to exist based on the nature of the source crime, and on the characteristics of the offender, namely his/her background and his/her social network. White-collar criminals (such as tax fraudsters) or corrupt politicians, if compared to other offenders (e.g., drug or human traffickers) usually rely on greater financial and accounting knowledge, and greater ‘social capital’, with widespread links with professionals, third-party advisers, and corporate service providers, which are often employed to carry out the predicate crime itself, but which are frequently involved in the ML scheme, too. This makes it more likely they will employ more complex and transational ML schemes (see e.g. van der Does de Willebois et al., 2011 on the transnational laundering of grand corruption). On the contrary, members of organised crime groups, including mafia-type ones, may rely on poorer financial skills, and may want to minimise the involvement of third-party professionals and brokers so as to avoid tipping off (Savona et al., 2016; Transcrime, 2018). This can force these offenders to employ more basic and ‘local’ ML schemes which are characterised by lower levels of transnationality and complexity, and lower reliance on banking channels.

Research problem and research questions

The empirical literature on the behaviour of ML offenders highlighted relevant patterns that do not fit the standard economic approach and conflict with the dominant narrative. While being the dominant theoretical framework for modelling ML activities, economic theory has been criticised by scholars for not acknowledging that criminals may follow other drivers when laundering their illicit proceeds, rather than always trying to maximise their economic returns. This study attempts to contribute to this branch of empirical studies by analysing the ML behaviour of a group of Italian offenders reported in the LN WoCo database. In particular, it addresses four main research questions:

Does proximity (in geographic or cultural terms) matter for offenders when choosing where to launder illicit proceeds?

Do offenders prefer basic ML typologies or sophisticated ones?

Are traditional and cash-intensive business sectors preferred for laundering illicit proceeds?

Are there differences between the ML typologies employed by mafia actors and non-mafia actors?

Based on the previous literature review, four hypotheses can be formulated:

H1. Criminals mainly launder illicit proceeds in the country where they committed the predicate offence and/or in neighbouring countries;

H2. Criminals mainly employ basic ML typologies and invest in tangible assets;

H3. Criminals mainly invest in traditional and cash-intensive business sectors;

H4. ML schemes by mafia members show more basic patterns (in terms of destination, typologies, and assets) compared to other offenders.

Enhancing our understanding of how criminals launder their illicit proceeds is essential to advance criminological knowledge on the economic dimension of offenders. Second, it also serves to better inform and support those scholars involved in assessing the effectiveness of the AML regime, showing if AML policies are identifying risks correctly.

Data and methodology

Data and methodological approach

To test these hypotheses, we analyse the ML behaviour of a group of Italian offenders as derived from a QCA of reports by official government sources and media news, previously collected and processed by LexisNexis (LN) and then made available through its WorldCompliance (WoCo) database (described in detail below). While the use of LN WoCO represents an innovation with respect to the previous literature, QCA follows the same approach successfully employed in previous studies in the criminological domain (Benson and Gottschalk, 2015; Calderoni et al., 2016; Koslicki, 2021) and more specifically in the study of ML (Matanky-Becker and Cockbain, 2021).

In particular, our empirical data consist of criminal cases associated with 2818 Italian individuals arrested and/or sentenced due to ML charges and listed in the LN WoCo database. This is a daily updated database that provides information about more than 2.5 million entities (both individuals and companies) that are linked to 60 crime or threat categories (e.g., ML, drug trafficking, terrorist financing, corruption, environmental crimes).¹ It is widely used by AML-obliged entities in customer due diligence activities (e.g., for screening clients against subsistence of previous enforcement measures or checking whether they may be classified as politically exposed persons – PEPs), but it has never been employed in ML research. The employment of this repository is a good alternative – and the only option – when accessing prosecution or sentencing data is not possible (e.g., due to personal data protection constraints or sensitivity issues).

For the present study, the database has been filtered following three criteria:

Offence category: ML (if the individual was associated with further offences beyond ML, this was noted and considered in the analysis)²;

Country: Italy. See below for a discussion on this choice.;

Database segment (data source): ‘Adverse media’ and ‘Enforcement’. This filter refers to the sources of the information provided in the database:

Enforcement: it includes individuals or companies who have been associated with illicit activities by LexisNexis according to information provided by state government authorities and enforcement agencies (e.g., law enforcement agencies, financial intelligence units, securities and exchange commissions, central banks, other supervisors).

Adverse media: it includes individuals or companies who have been associated with illicit activities by LexisNexis according to information provided by other public or news sources worldwide (e.g., international, national and local newspapers, broadcasts, press releases).

‘Country’ refers to the jurisdiction which is associated by LN to the individual. It usually refers to the country of birth or nationality of the individual. We selected here all the individuals in the dataset classified as ‘Italy’.³ Italy is an important country which, due to the great amount of mafia-type organised crime and of tax crime, represents a significant source of criminal proceeds available for ML (CSF, 2019; FATF, 2016). We decided to focus only on Italian offenders (and not, e.g., on foreign offenders active in Italy) for two reasons, one substantial, and the other methodological. First, as stressed in the previous section and by previous studies and theories (e.g., social embeddedness), ML offenders may prefer countries that are neighbouring or close to the place of origin, or in which the community of compatriots is abundant and can play a role in the ML scheme. This is particularly true for members of Italian mafias, as demonstrated by wide literature on mafia movements (see Calderoni et al., 2016 for a review). Second, based on the available data, place of birth is an information element that has a much higher level of certainty and coverage if compared to the information on offenders’ movements abroad. The latter is often missing or in the best scenario is only partial. A comprehensive and validated list of foreign offenders active in Italy would have been impossible to generate based on the available sources.

All the individuals responding to the aforementioned criteria were collected, and no sampling was carried out. The above-mentioned filtering resulted in 2983 profiles of individuals and legal persons that have been further skimmed due to:

Removal of aliases: WoCo also includes known aliases of offenders (associable with a unique ID identifier). 147 aliases have been removed to avoid double-counting;

Removal of other entities: 18 legal persons have been removed to consider only individuals in the analysis.

The final sample consisted of 2818 ML offenders. About 81% (2284) are included in the Adverse media database segment while 19% (534) are in the Enforcement database segment. This unbalance does not strictly pose issues in terms of reliability of evidence used for the analysis. Although most cases have not received a final judgement by a court, judicial decisions may take a long time – especially considering the three sets of proceedings in the Italian criminal system – and often exclusions of evidence are only due to formal errors (Leukfeldt and Kleemans, 2021). The issue is even more relevant in the case of ML where judicial authorities face several challenges to prosecute ML and trace (and seize) criminal proceeds, therefore justifying the inclusion of information stemming from news for research purposes.

An individual is included by LexisNexis in the database if reported by two independent sources. The information gathered from open sources (e.g., law enforcements’ press release, sanction releases, newspaper articles, judicial documents), counterchecked by LexisNexis, are summarised in a textual field associated with each profile (‘Remarks’), providing a comprehensive overview of the criminal activities the individual was involved in. The WoCo database also contains the links of the sources which the information was extracted from. Whenever possible (e.g., links not protected by a paywall), we also read the full-text sources and employed them to integrate the ‘Remarks’ fields if needed.

The textual data of the 2818 profiles were analysed through a QCA. This methodology allows pieces of text to be classified and treated as variables in statistical analyses (Kort-Butler, 2016). Data for each offender in the sample were classified in categories related to five variables (see Table 2 below for details): (i) offences; (ii) laundering methods; (iii) assets; (iv) business sectors; (v) countries involved in the ML scheme. As for offences, we identified and classified all the crimes – including predicate ones (i.e., crimes that have generated the illicit proceeds then laundered) – the offenders have been charged with. In several cases, it was not possible to clearly distinguish between predicate offences and other crimes in which the ML offender was involved in. The laundering methods were instead classified relying on FATF typologies and previous literature, while business sectors were classified according to NACE economic classification.⁴ For each category, several dichotomous but non-mutually exclusive variables have been considered, as one reference may provide information on more than one element (e.g., an ML offender may employ several laundering methods and invest in different business sectors and countries). In the case of the category related to the countries involved in the ML schemes, a categorical variable has been computed. Details on categories and related variables are presented in Table 2.

Table 2.

Overview of the coding framework employed in the analysis.

Variable	Operationalisation and categories
Offence	The offences the individual was charged with in addition to ML.
ML methods	Cash: presence of cash at any stage of the laundering process (e.g., cash withdrawal, cash handover, cash smuggling); False documents: use of false documents at any stage of the laundering process (e.g., to open bank accounts); False invoicing: issuing of false invoices schemes to move illicit proceeds between companies, both nationally and abroad; Figureheads: disguise of ownership of companies/assets by employing third individuals/figureheads/strawmen; Financial investments: purchase of government or companies’ bonds, company shares, shares of investment funds, insurance policies and other financial instruments; Foreign bank accounts: use of bank accounts opened in a foreign country; High-value goods: purchase of high-value goods (e.g., artworks, jewels, luxury watches); Misuse of companies: employment of a company at any stage of the laundering process (e.g., acquisition/financing a company, justifying illicit proceeds as company turnover, etc); Real estate investments: purchase of real estate (e.g., commercial buildings, dwellings, lands); Structuring: breakdown of a large amount of illicit proceeds in smaller amounts to hamper tracing; Value transfers: use of wire transfers and money transfer services to move illicit proceeds; Virtual currencies: conversion illicit proceeds in virtual currencies at any stage of the laundering process.
Assets seized during the criminal investigations	Cash and bank accounts: cash, bank accounts and other current assets; Companies: company shares and company assets; Financial assets: bonds, shares, insurance policies, other financial instruments; High-value goods: artworks, jewels, luxury watches; Real estate: commercial buildings, dwellings, lands; Registered assets: cars, trucks, motorbikes, boats, helicopters planes
Business sectors	Business sectors in which the ML offender invested or in which his companies, employed in the ML scheme, were active.
Countries involved in the ML schemes	The countries mentioned in the textual field of each individual were classified into three groups: Countries involved in the commission of the predicate/other offences; Countries involved, at any stage, in the ML scheme; Countries mentioned for other reasons (e.g., place of execution of an arrest). In particular, for group (b), which was ultimately used in the analysis, a country was deemed to be involved in a ML scheme if it was: (i) the jurisdiction of the authority that arrested/prosecuted/sentenced an individual due to ML; (ii) a location where assets belonging to the individual were seized or confiscated; (iii) explicitly mentioned in the summary or linked in news reports as being part of the ML scheme, such as, for example, being known as the place where either bank accounts were opened to deposit dirty money or shell firms incorporated.

Source: Authors’ elaboration.

Table 3.

Individuals laundering in Italy versus abroad (N = 2818).

	Number	% on relevant total
A. Individuals in the dataset (total)	2818	100
Individuals for which there is no information on the country where ML is carried out	1102	39
Individuals for which there is information on the country where ML is carried out	1716	61
B. Individuals for which there is information on the country where ML is carried out	1716	100
Individuals laundering only in Italy	1032	60
Individuals laundering both in Italy and abroad	417	24
Individuals laundering only abroad	267	16
C. Countries mentioned in the sample as the location of ML activity	3092	100
Mentions of Italy	1399	45
Mentions of foreign countries	1693	55
No. of foreign countries mentioned	75

Source: Authors’ elaboration of LN WoCo.

Once classified, frequency and correlation analyses were carried out as reported and discussed in the next section. We decided to limit the analysis to descriptive statistics which – given the low amount of available empirical evidence on ML, as discussed – already provides useful insights and a contribution to the current knowledge of the phenomenon.

Representativeness of the sample and limitations

The present study suffers from some limitations. The first limitation refers to case selection. The sample employed in the present study only includes ML cases involving Italian offenders. It is debatable to what degree Italy is a generalisable case, especially given the significant role of local mafia groups which is not always applicable to other countries. However, as noted by previous scholars, ‘[how] laundering is carried out depends on local circumstances and changes from crime to crime, from criminal group to criminal group, and from country to country’ (Levi and Soudijn, 2020: 10). For its own nature, ML is context-specific and has a relational nature (Riccardi, 2022).

Second, news media articles (i.e., ‘Adverse media’) are the main data source of the sample (81%). However, newspapers rarely write about ML and, when they do, focus on describing the predicate offences rather than the ML process itself (Unger et al., 2006). Also, minor ML cases, or those related to non-mafia criminals (e.g., tax evaders), may be overlooked by media news. This may raise the issue of how representative this sample is if compared to the universe of individuals prosecuted in Italy for ML. According to the aggregate statistics provided by the Italian Min. Interior (reported by the National Statistical Office (ISTAT)), 10,818 ML offences have been ported to the police since 2013, while the number of individuals included in WoCo since 2013 is 2532. While the difference seems sensible, it should also be kept in mind that the Min. Interior records, in contrast to this WoCo excerpt, also include (a) non-Italian individuals and (b) fencing crimes (because of the broader parameters of the ISTAT crime classification). The regional distribution of the two samples is overall satisfying (Pearson's R = .32) considering that in some regions (e.g., Tuscany) ML prosecution statistics include numerous Chinese individuals, while we focus on Italians only, and that administrative data refer to the location of prosecutors' offices, while we look at ML offenders’ birthplace.

Third, it was not always possible to trace the timespan of the single ML schemes. For this reason, we decided not to conduct any time series analysis, based on the assumption that no relevant change over time could be observed in the ML strategies of Italian offenders. In general, the analysed timespan covers the 1995–2020 period, but 96.3% of the cases refer to the 2010–2020 period.

Lastly, inter-coder issues do not affect the results of the present study as only one of the two authors has been the coder of the cases. The employment of a clear codebook, agreed between the two researchers, limited the subjectivity of the coders' choices Kort-Butler, 2016: 5). The adoption of a hand-coded content analysis was motivated by the complexity of the unstructured textual data under analysis but potentially exposes to potential errors of classification.

Results and discussion

Characteristics of the individuals involved in the ML schemes

The final sample employed in the analysis consisted of 2818 individuals. 90% (2525) are male while only 10% (293) are female. The result is in line with previous research showing that females generally commit less white-collar crimes compared to men (Benson and Simpson, 2018). Also, it is in line with the percentage of females reported to the Italian authority for ML (average of 15% between 2008 and 2018) and much higher than that of females prosecuted for mafia association (about 5%).

About 83.9% of the individuals in the sample (2365) have been charged with organised crime offences. About 56.7% (1597) of the individuals in the sample are associated with Italian mafia organisations, namely Cosa Nostra (368), ‘Ndrangheta (680), Camorra (471) and Sacra Corona Unita (78). The specific ML behaviour of these actors – and the differences with non-mafia fellows – will be discussed in detail below.

While for all the individuals in the sample, the Italian nationality was proven, we do not have information on the exact municipality or province of birth for the 41% (1143) of individuals. For the remaining 1675, 87% of them (1465) were born in only five Italian regions (out of 20): Calabria, Campania, Sicily, Apulia and Lazio – the first four being the historical areas of origin and influence of Italian mafias, respectively ‘Ndrangheta, Camorra, Cosa Nostra and Sacra Corona Unita. Regarding the Italian provinces of birthplace, the first three per number of individuals – Vibo Valentia (284), Napoli (255) and Reggio Calabria (187) – are among the top provinces in Italy per risk of ML as already highlighted by previous literature (Riccardi et al., 2019b).

Offences

Almost 34% of the individuals committed multiple offences. Overall, Figure 1 shows the relevant role of white-collar crimes, such as fraud (45%), tax crimes (29%), fictitious registration of assets (20%) and corruption (10%). More than half of the individuals in the sample (52%) were involved in extortion – which is not surprising given the high percentage of individuals involved in mafia-type OC (see above). Also, usury is particularly relevant (30%). In addition to being a profitable criminal activity, it also allows criminals to launder their illicit proceeds by lending them to individuals in need (Barone and Masciandaro, 2019).

Geographical and cultural proximity in ML activity

Information of the countries involved in the ML process is available for 61% of the individuals in the sample (1716 offenders), while for the remaining 39% information is available only on the countries where the predicate offence was committed. Focusing only on the offenders for which information on the ML country is available, most of them (60%, 1032 individuals) laundered their proceeds only in Italy, while those laundering both in Italy and abroad correspond to a further 24%. This means that 85% of the ML offenders engaged, at least partially, in domestic ML schemes. This result itself can be read as a confirmation of the proximity hypothesis (Table 3).

Foreign countries involved provide further confirmation. 1693 references of foreign countries involved in the ML process have been identified, for a total of 75 unique countries. The top 10 countries per number of references are all European (Figure 2). Among them, three (Austria, San Marino and Switzerland) border with Italy (a fourth, Slovenia, is among the top 25) while most of the remaining are geographically close and can be reached within a few hours of driving or via ferry boat. In general, the average distance of the typical foreign country employed by Italians in ML is 2158 km (taken between the population-weighted centres, and weighted by the number of ML references), but, considering the first 10 countries in terms of references, it drops to only 508 km. The limited distance may be explained also by the need to have trusted places in which, for example, to move physically dirty cash (Riccardi, 2022). Also, the correlation is significant between the number of references and the geographical distance, the contiguity and with the ‘cultural proximity’, here operationalised in terms of common language. Also, most of these countries have the same currency – the euro – which eventually facilitates the laundering of cash proceeds (Table 4).

Table 4.

Correlation of references for the foreign countries involved in ML schemes with measures of proximity and presence in blacklists/greylists.

	Pearson's correlation
ML references	1
Geographical proximity	0.60***
Contiguity	0.48***
Common currency	0.39***
Common official language	0.52***
Blacklisted	−0.08
Greylisted	−0.1

Note: ML references = number of mentions of country being involved in ML schemes; Geographical proximity = reciprocal of the physical distance (in km) between the population-weighted centres of Italy and the third country (source: CEPII). Contiguity = dummy signalling if countries share the same border (source CEPII); Common currency = dummy indicating if the third country shares the same currency of Italy, that is, euro; Common official language = variable measuring the percentage of the population in the third country which speaks Italian; Blacklisted = dummy indicating if the third country appears at least once since 2000 in the FATF blacklist; Greylisted = dummy indicating if the third country appears at least once since 2000 in the FATF greylist. *** Coefficient significantly different from zero at the 99.9% confidence level.

Source: Authors’ elaboration of various sources.

Only two offshore jurisdictions, Curacao and Panama, appear if we consider top 25 countries per number of references, while ‘onshore’ countries in the European area seem to be more relevant (e.g., Malta, Switzerland, San Marino, Luxembourg). Furthermore, only four countries appear in AML lists (updated as of the end of 2021): North Korea in the FATF blacklist (High-Risk Jurisdictions Subject to a Call for Action) and Albania, Malta⁵ and Panama in the FATF greylist (Jurisdictions under Increased Monitoring).

Results confirm that, as already highlighted by previous works, Italian offenders prefer countries which, ceteris paribus, can guarantee a relatively high level of secrecy but are also close – geographically and culturally – to Italy. This choice may minimise the need to involve third parties (e.g., professionals, international tax advisers or law firms) who could help setting up international ML schemes, but also constitute a vulnerability for the criminals in terms of tipping off or fraud. Moreover, countries in Europe attract less attention from AML authorities and obliged entities compared to blacklisted/greylisted countries and exotic offshore jurisdictions which are subject to enhanced due diligence.

Finally, results show that destinations or countries involved in ML schemes are highly correlated with those in which predicate offences are committed. 60% of the offenders laundered their illicit proceeds in the same country where the associate offences were committed. This result demonstrates the relevance of proximity not only with respect to offenders’ origin but also with their ‘location of activity’.

Modi operandi and assets

Results show that the ML schemes employed by Italian criminals are often basic, involve few modi operandi, and have limited geographical scope. For the 1711 individuals for which information on modi operandi is available (61% of the total), 46% employed only one or two methods (or ‘typologies’, in FATF terms), while only 20% employed five or more methods. As shown in Figure 3, 70% of these individuals misused companies, followed by value transfers (54%) and the employment of figureheads to disguise the ownership of assets and companies (34%). Also, cash-based methods are widely employed (25%), while few references for financial investments and virtual currencies can be found (respectively 2% and 0.3% of the sample).

The extent of the use of companies is a further confirmation of the key role played by firms today in organised and financial crime schemes (Savona and Riccardi, 2018). According to the latest Europol SOCTA, 80% of organised crime groups in Europe make use of legitimate businesses in conducting their criminal activity (Europol, 2021). In this analysis, it is difficult to always discern whether the firms were employed for committing the predicate offences (e.g., tax crimes), for laundering the money themselves, or for both; but ML schemes appear to be strictly related to the use of legal persons.

In terms of assets confiscated during the criminal investigations, we can see a relevant ‘asset diversification’. For more than half (54%) of the offenders for which information on assets is available, three or four different asset types were confiscated, while for 6% of them, five or more asset types were confiscated. Regarding the typologies (Figure 4), tangible assets, such as real estate (78%), cash and bank accounts (61%) and companies (57%) are the most relevant. Despite being frequently associated with criminal habits high-value goods are limited (16%).

Business sectors

Although it was not always easy to distinguish among businesses and business sectors employed by offenders to commit predicate offences, as opposed to those in which offenders laundered/integrated their illicit proceeds, certain economic sectors appear more frequently than others (Figure 5). Those with the highest number of references are R – entertainment (33%); I – hotel and bars (25%); G – wholesale and retail (20%); Q – health and social work (20%) and F – construction (13%).

This list of sectors is not surprising. They are traditional sectors, characterised by low entry and exit barriers, low capital-intensiveness and companies with relatively simple legal forms – for example, limited liability companies or cooperatives (Fabrizi et al., 2017; Ravenda et al., 2015; Riccardi et al., 2019a). In addition, they are usually classified as cash-intensive sectors, in the sense that (Riccardi and Levi, 2018): (i) they mainly manage cash payments, allowing criminals to easily comingle illicit proceeds with businesses’ legitimate revenues and deposit them in bank accounts on a daily basis; (ii) they are mainly characterised by current assets, allowing criminals to rapidly sell them (contrary to non-current assets) in case of criminal investigations to avoid potential confiscations and seizures.

Mafia versus non-mafia laundering

Finally, results show that the ML patterns change based on the type of actor involved. As mentioned, 59% of the individuals in the sample are associated with Italian mafias. Some differences with the non-mafia sample can be highlighted. The first regards the propensity to use foreign countries to launder money. Only 30.9% of mafia actors laundered their illicit proceeds abroad, contrary to 48.3% of non-mafia actors (t = −12.67, p = 0.01).⁶ The second difference refers to ML methods. While false invoicing and figureheads are relevant for both types of actors (22% vs. 29% and 33% vs. 35%, respectively), in the case of mafia actors the overall misuse of companies is far more relevant (84% vs. 57%). Non-mafia actors seem to exploit more widely the banking system, through ML schemes involving financial institutions, as demonstrated by high number of references for value transfers (66%), structuring (37%) and the use of foreign bank accounts (30%) (Figure 6).

The importance of false invoicing in the current mafias’ business has been demonstrated by a large number of investigations. False invoices are a ‘multi-purpose’ financial crime which does not only allow to launder money, but also to create slush funds (useful e.g., for corruptive purposes), reduce taxable income, produce VAT credits and also provide criminal services to entrepreneurs in difficulty that, in the medium term, may become a target of mafia acquisition (Riccardi, 2022; Saenz and Lewer, 2022). While figureheads, as a way to conceal beneficial ownership, are usually preferred by mafias if compared to other more sophisticated strategies such as the employment of shell companies established offshore or complex corporate schemes. Mafia organisations tend to keep control ‘in-house’ to maximise the tenure of the association. In this sense, non-mafia ML offenders may rely more heavily on the banking sector (and on foreign bank accounts) because, on the one side, they may be less sensitive to the involvement of third-party professionals and intermediaries and, on the other side, because they can count on a smaller number of supporters and affiliates who could act as figureheads.

When it comes to business sectors, mafia actors, with respect to non-mafia ones, make significantly higher employment of firms active in gaming and betting (40% vs. 4%), bars, restaurants, and hotels (30% vs. 11%) and Construction (16% vs. 6%) (Figure 7). The (legal) gaming and betting industry (especially online gaming) now represents a key sector for both ‘Ndrangheta and Cosa Nostra, as shown by numerous investigations and studies. For example, research project MORE by Transcrime reported, between 2016 and 2018, seven police investigations that involved gaming companies directly or indirectly owned by mafia organisations. In most of these cases, Malta was involved as location of the firm's registered seat (Savona and Riccardi, 2018).

Conclusions and future research directions

The analysis of the behaviours and modi operandi of Italian ML offenders confirmed the four hypotheses presented in the second section. First, Italian ML offenders mainly launder their illicit proceeds in Italy, in the country where predicate offence was committed, or in countries that are geographically or culturally close to Italy (or which have the same currency, the euro).

Second, ML typologies mainly entail traditional methods and tangible assets. Misuse of companies is abundant, as well as related methods such as the employment of figureheads and false invoicing schemes (used in trade-based ML) while cases with financial instruments and virtual currencies are few. Third, the hypothesis on the use of traditional business sectors – cash-intensive, labour-intensive and with low entry/exit barriers – is also confirmed.

Last, although some ML methods are common between mafia and non-mafia ML offenders, such as the widespread use of companies and cash, relevant differences can be identified. First, there is relatively higher employment of false invoicing and figureheads by mafia-related individuals, and higher reliance of the banking system by non-mafia offenders, in the form of higher use of value transfers, structuring schemes and the employment of foreign bank accounts. Also, non-mafia offenders show a stronger tendency to employ foreign countries.

Overall, results generally demonstrate that the level of sophistication of ML schemes is lower than what is portrayed by media or often in the political debate, and that ML offenders do not behave necessarily as legitimate entrepreneurs, but may instead follow other drivers. However, results could mirror the difficulty of law enforcement in investigating and tracing complex and transnational ML schemes. An Italian authority may much more easily trace (and seize) a real estate placed somewhere in Milan or Rome than detect a trade-based ‘round tripping’ ML scheme involving, for example, a firm located in the United Arab Emirates and a legal arrangement in Cyprus. Similarly, a law enforcement agency can more easily freeze 10 million euros held in cash or in a national bank account rather than in several wallets of virtual currencies. In other words, the limited sophistication of the ML behaviour that emerges from the analysis could potentially reflect the challenges that Financial intelligence units (FIUs), law enforcement and asset recovery agencies might encounter during transnational investigations. This is a limitation that should be considered in future studies – although controlling for the effectiveness of financial investigations is almost impossible.

Despite these limitations, this study contributes to the scarce branch of empirical studies of ML behaviour. The employment of the LN WoCo appears as a promising avenue for research in this field, especially when and whether data on prosecuted or convicted individuals are not largely available. This is the first study that analyses this unique database and, more generally, one of the first studies that also employs news media information to empirically investigate ML. However, the analysis should be updated and enriched by employing other data sources, possibly official ones such as judicial and police records, or the insights stemming from STRs/SARs and from intelligence services. And it should be extended to other types of criminals beyond the Italian realm, and by going more in-depth into the understanding of how ML strategies change based on the type of actor, predicate offence, regulatory framework and geographical landscape.

Footnotes

Acknowledgements

The authors would like to acknowledge the participants of the Third AML Conference of the Central Bank of Bahamas for the inputs provided, and in particular Peter Reuter.

Author contributions

This article is the result of the joint efforts of both authors M.N. and M.R. who jointly designed the study, the research concept and developed the analysis. In particular, M.N. wrote the introduction, carried out the literature review, carried out the coding and performed the analysis of modi operandi and sectors. M.R. performed the analysis of geographical scope, and representativeness of the sample and wrote discussion and conclusions.

Declaration of conflicting interests

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

Funding

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

ORCID iD

Michele Riccardi

Notes

References

Arnone

Borlini

(2010) International anti-money laundering programs: Empirical assessment and issues in criminal regulation. Journal of Money Laundering Control 13(3): 226–271.

Barone

Masciandaro

(2019) Cryptocurrency or usury? Crime and alternative money laundering techniques. European Journal of Law and Economics 47(2): 233–254.

Barone

Masciandaro

Schneider

(2022) Corruption and money laundering: You scratch my back, i’ll scratch yours. Metroeconomica 73(1): 318–342.

Beare

Schneider

(2007) Money Laundering in Canada: Chasing Dirty and Dangerous Dollars. Toronto: Toronto University Press.

Becker

(1968) Crime and punishment: An economic approach. Journal of Political Economy 76(2): 169–217.

Benson

Gottschalk

(2015) Gender and white-collar crime in Norway: An empirical study of media reports. International Journal of Law, Crime and Justice 43(4): 535–552.

Benson

Simpson

(2018) White-Collar Crime: An Opportunity Perspective. 3rd ed. New York: Routledge.

Calderoni

Aziani

Favarin

(2013) Poland. The Factbook on the Illicit Trade in Tobacco Products 4. Issue 4, The Factbook on the Illicit Trade in Tobacco Products.

Calderoni

Berlusconi

Garofalo

, et al. (2016) The Italian mafias in the world: A systematic assessment of the mobility of criminal groups. European Journal of Criminology 13(4): 413–433.

10.

Cassetta

Pauselli

Rizzica

, et al. (2014) Financial Flows to Tax Havens: Determinants and Anomalies. Quaderni dell’antiriciclaggio. Roma: Unità di Informazione Finanziaria – Banca D’Italia.

11.

Cornish

Clarke

(1985) Modeling offenders’ decisions: A framework for research and policy. In: Tonry

Morris

(eds) Crime and Justice: An Annual Review of Research. Chicago: University of Chicago Press.

12.

CSF (2019) Analisi nazionale dei rischi di riciclaggio di denaro e di finanziamento del terrorismo elaborata dal Comitato di sicurezza finanziaria. Available at: https://www.dt.mef.gov.it/export/sites/sitodt/modules/documenti_it/prevenzione_reati_finanziari/prevenzione_reati_finanziari/Analisi_dei_rischi_di_riciclaggio_e_di_finanziamento_del_terrorismo_2018-IT_-_Sintesi.pdf.

13.

Cummings

Stepnowsky

(2010) My brother’s keeper: An empirical study of attorney facilitation of money-laundering through commercial transactions. University of Maryland Legal Studies Research Paper 32. DOI: https://doi.org/10.2139/ssrn.1658604.

14.

Custers

Oerlemans

J-J

Pool

(2020) Laundering the profits of ransomware: Money laundering methods for vouchers and cryptocurrencies. European Journal of Crime, Criminal Law and Criminal Justice 28(2): 121–152.

15.

Custers

Pool

Cornelisse

(2018) Banking malware and the laundering of its profits. European Journal of Criminology 16(6): 728–745.

16.

Dugato

Favarin

Giommoni

(2015) The risks and rewards of organized crime investments in real estate. British Journal of Criminology 55(5): 944–965.

17.

Europol (2021) EU Serious and Organised Crime Threat Assessment 2021. The Hague: EUROPOL. Available at: https://www.europol.europa.eu/activities-services/main-reports/european-union-serious-and-organised-crime-threat-assessment.

18.

Fabrizi

Malaspina

Parbonetti

(2017) Caratteristiche e modalità di gestione delle aziende criminali. Rivista di Studi e Ricerche sulla Criminalità Organizzata 3(1): 47–66.

19.

FATF (2016) Anti-money Laundering and Counter-terrorist Financing Measures – Italy. Fourth Round Mutual Evaluation Report. Paris: FATF. Available at: https://www.fatf-gafi.org/en/publications/mutualevaluations/documents/mer-italy-2016.html.

20.

Ferwerda

(2009) The economics of crime and money laundering: Does anti-money laundering policy reduce crime? Review of Law & Economics 5(2). DOI: https://doi.org/10.2202/1555-5879.1421.

21.

Ferwerda

Reuter

(2019) Learning from money laundering national risk assessments: The case of Italy and Switzerland. European Journal on Criminal Policy and Research 25(1): 5–20.

22.

Ferwerda

van Saase

Unger

, et al. (2020) Estimating money laundering flows with a gravity model-based simulation. Scientific Reports 10(1). DOI: https://doi.org/10.1038/s41598-020-75653-x.

23.

Gara

De Franceschis

(2015) I paradisi fiscali: Caratteristiche operative, evidenze empiriche e anomalie finanziarie. Quaderni dell’antiriciclaggio. Roma: Unità di Informazione Finanziaria – Banca D’Italia.

24.

Halliday

(2018) Plausible folk theories: Throwing veils of plausibility over zones of ignorance in global governance: Plausible folk theories. The British Journal of Sociology 69(4): 936–961.

25.

Halliday

Levi

Reuter

(2020) Why do transnational legal orders persist? The curious case of money-laundering controls. In: Shaffer

Aaronson

(eds) Transnational Legal Ordering of Criminal Justice. 1st ed. Cambridge: Cambridge University Press, pp. 51–83. DOI: https://doi.org/10.1017/9781108873994.002.

26.

Harvey

Turner

Ward

, et al. (2021) Transnational criminal law: A case study of the international anti-money laundering framework as applied in Nigeria, a case of smoke and mirrors. Global integrity anti-corruption evidence project: Hiding the beneficial owner and the proceeds of corruption. Global Integrity. Available at: https://ace.globalintegrity.org/wp-content/uploads/2021/03/Jackie_Harvey-Smoke_Mirrors-1.pdf.

27.

Irwin

ASM

Raymond Choo

Liu

(2011) An analysis of money laundering and terrorism financing typologies. Journal of Money Laundering Control 15(1): 85–111.

28.

Kort-Butler

(2016) Content Analysis in the Study of Crime, Media, and Popular Culture. Oxford Research Encyclopedia. Criminology and Criminal Justice. Oxford: Oxford University Press. DOI: https://doi.org/10.1093/acrefore/9780190264079.013.23.

29.

Koslicki

(2021) Recruiting warriors or guardians? A content analysis of police recruitment videos. Policing and Society 31(6): 702–720.

30.

Kramer

J-A

Blokland

AAJ

Kleemans

, et al. (2023) Money laundering as a service: Investigating business-like behavior in money laundering networks in the Netherlands. Trends in Organized Crime. DOI: https://doi.org/10.1007/s12117-022-09475-w.

31.

Kruisbergen

Kleemans

Kouwenberg

(2015) Profitability, power, or proximity? Organized crime offenders investing their money in legal economy. European Journal on Criminal Policy and Research 21(2): 237–256.

32.

Kruisbergen

Leukfeldt

Kleemans

, et al. (2019) Money talks: Money laundering choices of organized crime offenders in a digital age. Journal of Crime and Justice 42(5): 569–581.

33.

Leukfeldt

Kleemans

(2021) Breaking the walls of silence: Analyzing criminal investigations to improve our understanding of cybercrime. In: Lavorgna

Holt

(eds) Researching Cybercrimes. Cham: Springer International Publishing, pp. 127–144. DOI: https://doi.org/10.1007/978-3-030-74837-1_7.

34.

Levi

Reuter

(2006) Money laundering. Crime and Justice: A Review of Research 34(1): 289–375.

35.

Levi

Soudijn

(2020) Understanding the laundering of organized crime money. In: Tonry

Reuter

(eds) Organizing Crime: Mafias, Markets, and Networks. Crime and Justice. Chicago: University of Chicago Press, pp. 579–631. Available at: https://www.journals.uchicago.edu/doi/10.1086/708047 (accessed 31 August 2020).

36.

Malm

Bichler

(2013) Using friends for money: The positional importance of money-launderers in organized crime. Trends in Organized Crime 16(4): 365–381.

37.

Masciandaro

(1999) Money laundering: The economics of regulation. European Journal of Law and Economics 7(3): 225–240.

38.

Masciandaro

Takáts

Unger

(2007) Black Finance: The Economics of Money Laundering. Cheltenham, UK; Northampton, MA: Edward Elgar.

39.

Matanky-Becker

Cockbain

(2021) Behind the criminal economy: Using UK tax fraud investigations to understand money laundering myths and models. Crime, Law and Social Change 77: 405–429. DOI: https://doi.org/10.1007/s10611-021-09997-4.

40.

McCarthy

van Santen

Fiedler

(2015) Modeling the money launderer: Microtheoretical arguments on anti-money laundering policy. International Review of Law and Economics 43: 148–155.

41.

Palomo

Marquez

Ruiz

(2015) Organized crime investments in Spain. In: Savona

Riccardi

(eds) From Illegal Markets to Legitimate Businesses: The Portfolio of Organised Crime in Europe. Trento: Transcrime – Università degli Studi di Trento, pp. 202–209.

42.

Petrell

Houtsonen

(2016) From bikers to entrepreneurs: The infiltration of motorcycle gangs into legitimate business in Finland. In: Organised Crime in European Businesses. Abingdon: Routledge, pp. 159–174.

43.

Petrunov

(2011) Managing money acquired from human trafficking: Case study of sex trafficking from Bulgaria to Western Europe. Trends in Organized Crime 14(2–3): 165–183.

44.

Ravenda

Argilés-Bosch

Valencia-Silva

(2015) Detection model of legally registered mafia firms in Italy. European Management Review 12(1): 23–29.

45.

Reuter

(1983) Disorganized Crime: The Economics of the Visible Hand. Cambridge, Mass: MIT Press.

46.

Riccardi

(2014) When criminals invest in businesses: Are we looking in the right direction? An exploratory analysis of companies controlled by Mafias. In: Caneppele

Calderoni

(eds) Organized Crime, Corruption and Crime Prevention. Cham: Springer International Publishing, pp. 197–206. DOI: https://doi.org/10.1007/978-3-319-01839-3_23.

47.

Riccardi

(2022) Money Laundering Blacklists. London: Routledge. DOI: https://doi.org/10.4324/9781003212867.

48.

Riccardi

Levi

(2018) Cash, crime and anti-money laundering. In: King

Walker

Gurulé

(eds) The Palgrave Handbook of Criminal and Terrorism Financing Law. Cham: Springer International Publishing, pp. 135–163. DOI: https://doi.org/10.1007/978-3-319-64498-1_7.

49.

Riccardi

Maggioni

Ferluga

(2019a) I soldi dei ‘bravi ragazzi’: controllo e occultamento nelle decisioni di investimento delle mafie italiane. Rassegna Economica In corso di pubblicazione.

50.

Riccardi

Milani

Camerini

(2019b) Assessing money laundering risk across regions. An application in Italy. European Journal on Criminal Policy and Research 25(1): 21–43.

51.

Saenz

Lewer

(2022) Estimates of trade based money laundering within the European Union. Applied Economics 55(51): 5991–6003. DOI: https://doi.org/10.1080/00036846.2022.2141444.

52.

Savona

Riccardi

(eds) (2015) From Illegal Markets to Legitimate Businesses: The Portfolio of Organized Crime in Europe. Final Report of Project OCP – Organized Crime Portfolio, Milano: Transcrime - Università Cattolica del Sacro Cuore.

53.

Savona

Riccardi

(eds) (2018) Mapping the Risk of Serious and Organised Crime Infiltration in European Businesses Final Report of the MORE Project. Milano: Transcrime – Università Cattolica del Sacro Cuore.

54.

Savona

Riccardi

Berlusconi

(2016) Organised Crime in European Businesses. Abingdon: Routledge.

55.

Schneider

(2004) Money laundering in Canada: A quantitative analysis of Royal Canadian Mounted Police cases. Journal of Financial Crime 11(3): 282–291.

56.

Skinnari

Versterhav

Korsell

(2007) Vart tog alla pengarna vägen? En studie av narkotikabrottslighetens ekonomihantering. Bra°-rapport. Stockholm: Brottsförebyggande ra°det.

57.

Soudijn

(2014) Using strangers for money: A discussion on money-launderers in organized crime. Trends in Organized Crime 17(3): 199–217.

58.

Soudijn

(2018) Using police reports to monitor money laundering developments. Continuity and change in 12 years of Dutch money laundering crime pattern analyses. European Journal on Criminal Policy and Research 25(1): 83–97.

59.

Steinko

(2012) Financial channels of money laundering in Spain. British Journal of Criminology 52(5): 908–931.

60.

Suendorf

(2001) Geldwasche: Eine kriminologische Untersuchung. Available at: https://www.ojp.gov/ncjrs/virtual-library/abstracts/geldwasche-eine-kriminologische-untersuchung.

61.

Transcrime (2018) Mapping the Risk of Serious and Organised Crime Infiltration in Europe – Final Report of the MORE Project. Milano: Università Cattolica del Sacro Cuore.

62.

Unger

(2007) The Scale and Impacts of Money Laundering. Cheltenham, UK; Northampton, MA: Edward Elgar.

63.

Unger

Ferwerda

(2011) Money Laundering in the Real Estate Sector: Suspicious Properties. Edward Elgar. DOI: https://doi.org/10.4337/9781781000915.

64.

Unger

Ferwerda

Siegel

, et al. (2006) The amounts and the effects of money laundering. Report for Dutch Ministry of Finance.

65.

Unger

Rawlings

(2008) Competing for criminal money. Global Business and Economics Review 10(3): 331.

66.

van der Does de Willebois

Halter

Harrison

, et al. (2011) The Puppet Masters: How the Corrupt Use Legal Structures to Hide Stolen Assets and What to Do About It. The World Bank. http://elibrary.worldbank.org/doi/book/10.1596/978-0-8213-8894-5.

67.

Van de Bunt

Siegel

Zaitch

(2014) The social embeddedness of organized crime. In: Paoli

(ed.) The Oxford Handbook of Organized Crime. Oxford: Oxford University Press.

68.

Van Duyne

Harvey

Gelemerova

(2018) The Critical Handbook of Money Laundering: Policy, Analysis and Myths. London: Palgrave Macmillan UK. DOI: https://doi.org/10.1057/978-1-137-52398-3.

69.

Van Duyne

Soudijn

MRJ

(2009) Hot money, hot stones and hot air: Crime-money threat, real estate and real concern. Journal of Money Laundering Control 12(2): 173–188.