From “Angels” to “Vice Smugglers”: the Criminalization of Sea Rescue NGOs in Italy

Non-governmental organizations (NGOs) have played a crucial role in conducting Search and Rescue (SAR) operations off the Libyan coast, assisting almost 120,000 migrants between 2014 and 2019. Their activities, however, have been increasingly criticized. The accusation that NGOs facilitate irregular migration has escalated into investigations by Italian and Maltese courts and various policy initiatives restricting non-governmental ships and their access to European ports. Although all NGOs investigated to date have been acquitted, the combination of criminal investigations and policy restrictions that has taken place in Italy since 2017 has severely hindered non-governmental SAR operations. Given the humanitarian repercussions of reducing NGOs’ presence at sea, the merits and shortcomings of the arguments underlying the criminalization of non-governmental maritime rescue warrant in-depth research. To that end, this article fulfils two interrelated tasks. First, it provides a genealogy of the accusation against NGOs and the ensuing combination of legal criminalization, policy restrictions, and social stigmatization in restraining their activities. Second, it uses quantitative data to show that empirically verifiable accusations like the claim that NGOs serve as a pull factor of migration, thereby causing more people to day at sea, are not supported by available evidence. By doing so, our study sheds new light onto the criminalization of humanitarianism and its implications.


Introduction
The non-governmental organizations (NGOs) seeking to relieve the humanitarian implications of large-scale human mobility have come under close scrutiny. In Italy, Greece, Hungary, and the USA alike, humanitarian workers operating along migratory routes have been accused of incentivizing people to cross borders irregularly, thereby aiding and abetting illegal immigration.
Sea rescue NGOs operating in the Mediterranean Sea have been the target of especially heated criticism. Initially praised as "angels", maritime humanitarian workers have eventually been stigmatized as "sea taxis" and "vice smugglers". By 2017, the suspicion that NGOs were serving as a "pull factor" of illegal immigration or even operating in collusion with human smugglers escalated into both criminal investigations and several policy initiatives restricting non-governmental rescue missions. While all judicial proceedings to date have resulted in acquittals, persisting risks of criminal sanctions, policy restrictions on search and rescue (SAR), and stigmatization by media have increasingly reduced NGOs' ability to assist migrants in distress at sea. Examining the ongoing efforts to restrain and criminalize humanitarian operations has therefore important theoretical and policy implications.
The existing scholarship offers important insights into efforts to "police humanitarianism" (Carrera, Vosyliute, Allsopp, and Valsamis 2019), criminalize migration and solidarity to migrants (Mezzadra 2020;Atak and Simeon 2018;Fekete 2018, Heller andPezzani 2018), enforce more restrictive border policies (Cuttitta 2018b, Moreno-Lax 2017, and "straightjacket" migrant rescuers (Cusumano 2019a). No scholarship to date, however, has provided a systematic overview of the criminalization of sea rescue NGOs in Italy and its implications. Moreover, most studies are been exclusively qualitative and have not systematically examined whether the accusations formulated against humanitarian workers are confirmed by available quantitative evidence. As it enacted several policy restrictions and launched 18 investigations on NGOs between 2017 and February 2020, Italy is a crucial case for the study of the criminalization of humanitarianism. Moreover, since the Central Mediterranean corridor connecting Italy to Libya is the deadliest migratory route worldwide, Rome's attempt at restricting non-governmental SAR operations may have especially problematic implications for human security at sea. Consequently, this study will focus on the criminalization of sea rescue NGOs in Italy.
Specifically, we seek to fill the blind spots left by the existing literature by conducting three interrelated tasks. First, we provide a genealogy of the critiques deployed against NGOs. Second, we offer a systematic overview of the legal proceedings and policy measures initiated against non-governmental maritime rescuers until February 2020. Last, we assess to what extent these accusations are corroborated by quantitative data on irregular crossings and casualties along the Central Mediterranean route in the period between 2014 and 2019. To that end, we rely on official documents as well as qualitative and quantitative evidence of irregular migration across the Mediterranean and NGOs' involvement in maritime rescue.
By doing so, we provide a novel contribution to the scholarship on the criminalization of humanitarianism, border control, and migration across the Mediterranean by substantiating three claims that have often been put forward by existing studies, but were rarely empirically tested. First, we confirm scholars' argument that the claims used to stigmatize humanitarian work at borders, while often taken for granted by decision-makers, are often unsupported by available data. Second, we illustrate the importance of dedicated maritime rescue missions in enhancing human security at sea and reducing the human costs of irregular migration. Third, we add to the literature questioning the disproportionate importance attached by decision-makers to the alleged pull factors of migration, showing that activities like SAR operations may ultimately play a negligible role in facilitating irregular border crossings. Although additional in-depth research on these issues remains warranted, our work provides a starting point for future scholarship on the criminalization and broader implications of humanitarian action.
The article is divided as follows. The first section reviews maritime rescue operations along the Central Mediterranean migratory route and of the important role played by NGOs therein. The second section provides a diachronic overview of how the discursive delegitimization of maritime rescue NGOs has escalated into judicial criminalization and the formulation of policies seeking to restrain NGOs' activities. The third section and ensuing conclusion appraise the accusation that sea rescue NGOs facilitate irregular migration and have unintended humanitarian externalities, summarize the findings of the article, and outline some avenues for future research.

Maritime Rescue NGOs: an Overview
The large numbers of migrants dying at sea while trying to reach Italy has turned the Southern Mediterranean Sea into the deadliest border worldwide. Between 2014 and the end of 2018, at least 15,000 died off the coast of Libya (International Organization for Migration n.d.; United Nations Refugee Agency n.d.).
The humanitarian crisis in the Central Mediterranean deteriorated in October 2014, after the Italian government suspended its Search and Rescue (SAR) operation Mare Nostrum, replaced by the European Border and Coast Guard (still better known as Frontex) operation Triton. Due to its narrower mandate and smaller operational area, Triton was ill-equipped to address the ongoing humanitarian emergency (Cusumano 2019b). Since then, fourteen different aid organizations have attempted to fill this gap by conducting their own SAR operations. Table 1 lists all the NGOs operating at sea and the rescue vessels they used at different moments in time (only MOAS and Médecins Sans Frontieres simultaneously operated two ships in 2015 and 2016). An indirect role in SAR operations has also been played by Alarm Phone, an NGO operating a hotline for migrants in distress in the Mediterranean Sea.
Neither the non-governmental provision of maritime rescue nor its criminalization is an entirely new phenomenon. In 2005, the German NGO Cap Anamur conducted a SAR operation in the Strait of Sicily, but its ship was confiscated by Italian authorities, who charged and later acquitted the captain, first officer, and head of mission for aiding and abetting illegal immigration (Cuttitta 2018a, Basaran 2015. Non-governmental rescue missions in the Mediterranean recommenced in September 2014 with the creation of the Migrant Offshore Aid Station (MOAS), which offered an example for several other organizations to follow. The operational branches of Médecins Sans Frontières (MSF) headquartered in Barcelona and Brussels developed independent SAR capabilities using their own ships, while MSF Amsterdam operated the 77 m Aquarius in partnership with SOS Méditerranée. In addition, two German and one Spanish NGOs, Sea-Watch, Sea-Eye, and ProActiva Open Arms, deployed smaller vessels offshore Libya (Cuttitta 2018a). Between August and September 2016, three other organizations, namely the Berlin youth association Jugend Rettet, the Dutch NGO Boat Refugee Foundation, and Save the Children, also started SAR missions in the Central Mediterranean. Figure 1 shows the number of migrants rescued by each of the organizations mentioned above between 2016 and 2017, the only years when the MRCC collected publicly available data divided by NGOs.  These organizations do not only vary significantly in their ability to rescue migrants in distress, as epitomized by Fig. 1 but also display different rescue models. In 2015 and 2016, only MOAS and MSF conducted fully fledged SAR operations and disembarked migrants in a place of safety on Italian soil indicated by Rome's Maritime Rescue Coordination Centre (MRCC) in agreement with the Ministry of Interior. Due to the limited size and speed of their first ships, Sea-Watch, Sea-Eye, Jugend Rettet, Lifeboat, Proactiva, and the Boat Refugee Foundation decided to limit their activities to temporarily assisting migrants in need while awaiting the arrival of a larger vessel. Only in 2017 did all NGOs start to directly disembark migrants in Italian ports in response to requests from Rome (Cusumano 2019b).
NGOs also differ widely in their political stance. MOAS, for instance, deliberately adopted an apolitical approach to migration to Europe, enshrined in slogans such as "Save lives first. Sort out the politics later". By contrast, organizations like MSF, Sea-Watch, and Jugend Rettet deliberately sought to combine the direct provision of humanitarian relief with advocacy, whistleblowing, and naming and shaming by using their presence at sea to denounce the suffering stemming from European border policies (Cuttitta 2018a, Stierl 2018). Cultural differences between NGOs have also translated into different conceptions of humanitarian work. Most notably, organizations with a long history of operating in conflict environments like MSF developed a stricter interpretation of the principles of neutrality and independence that underlie humanitarian action (Cusumano 2019c).
Despite some organizations' stronger commitment to neutrality and independence, all NGOs cooperated effectively with the Italian MRCC, as repeatedly acknowledged by Italian Coast Guard and Navy officers (Cuttitta 2020;Cusumano 2019c). In fact, the SAR operations conducted by NGOs between 2014 and 2017 were all coordinated and for the most part initiated by the Italian MRCC, which gathered distress calls and urged ships in the vicinity to conduct rescue operations in accordance with the international law of the sea. It was only in 2018-when the Italian government stopped accepting responsibility for SAR operations off the coast of Libya and began denying NGOs entry to its ports-that cooperation between NGOs and the MRCC faltered (Cuttitta 2020). Before then, NGOs were seen by the Italian MRCC as a multiplier of European rescue capabilities, stretched thin by the end of operation Mare Nostrum (Cusumano 2019c). Indeed, as showed by Fig. 2, NGOs played a crucial role in complementing the rescue efforts of European Navy and Coast Guard missions, assisting over 110,000 people between 2014 and 2017.
As shown by the figure, however, by the end of 2017 the number of migrants rescued by NGOs had plummeted. This decrease was not solely caused by the drop in irregular departures from Libya that occurred after July 2017. From late 2016, NGOs have also experienced a number of policy restrictions and criminal investigations that severely limited their presence at sea. In the wake of this criminalization process, several of the NGOs listed in table one suspended their operations. Jugend Rettet was investigated for allegedly abetting illegal immigration in September 2017, and its ship has been impounded by Italian authorities since (Camilli 2017). MOAS and Save the Children decided spontaneously to suspend operations in the Mediterranean. Sea-Watch, Proactiva, and the newly established charity Mission Lifeline also saw their ships temporarily confiscated by Italian and Maltese authorities, while MSF decided to reduce its involvement in SAR. The only MSF ship left in 2018, the Aquariusoperated in partnership with SOS Méditerranée-was stripped of its flag and forced to suspend operations (Del Valle 2019).
However, non-governmental rescue activities did not cease completely. ProActiva, Mission Lifeline, and Sea-Watch all had their ships eventually released and restarted SAR operations. MSF and SOS Méditerranée acquired a new ship, the Norwegian-flagged Ocean Viking, which reached the Southern Mediterranean in August 2019. In the same period, two other small charities-Mediterranea Saving Humans and Salvamento Marítimo Humanitario-also started SAR missions. Notwithstanding these new developments, far fewer NGOs are now present at sea than in 2016 and 2017 (Cuttitta 2020;Del Valle 2019). In September 2019, the formation of a new cabinet slightly softened but did not substantially change Italy's approach to non-governmental maritime rescue. A combination of discursive delegitimization, judicial action, and Italian government initiatives has severely eroded NGOs' capabilities. The next section examines each of these processes in detail.

The Delegitimization, Criminalization, and Restriction of NGO Activities
As epitomized by the 2005 case against Cap Anamur, the criminalization of maritime rescue operations is not entirely new. Since the early 2000s, Italian and Tunisian fishermen have been repeatedly investigated when conducting rescue operations (Basaran 2015). More recently, the accusation that maritime rescue operations incentivize irregular immigration has been forcefully raised against public SAR operations conducted by European military and law enforcement agencies as well. Most notably, operation Mare Nostrum was criticized as "an unintended pull factor, encouraging more migrants to attempt the dangerous sea crossing and thereby leading to more tragic and unnecessary deaths" (House of Lords 2016).
In 2015 and 2016, Italian media and society largely perceived NGOs as supporting public rescue efforts. Media coverage of SAR focused on the humanitarian duty to assist migrants in distress and rarely distinguished between public and private SAR providers. When they did, media portrayed NGOs in a very positive light, defining them as "heroes" or even "angels of the sea" (Barretta, Milazzo, Pascali & Chichi 2017). NGOs, however, soon became victims of their own success. As non-governmental assets became the largest provider of SAR, the same suspicions that had been levelled against Mare Nostrum turned into off-the-shelf accusations to be used against non-governmental rescue assets. Accusations leaked from European authorities had a crucial role in triggering a shift in Italian public opinion, already concerned by the large number of irregular migrants reaching Italian coasts. In December 2016, the Financial Times leaked excerpts from a Frontex confidential report raising concerns about the interaction between private rescuers and smugglers. These quotes stated that NGOs had indirectly communicated with human smugglers by, for instance, using light signals visible from the Libyan coast at night and suggested that migrants had been given "clear indications before departure on the precise direction to be followed in order to reach the NGOs' boats" (Robinson 2016). The Financial Times later released a correction, reporting that Frontex had only raised concerns but was in no way accusing humanitarian workers of collusion with smugglers (Financial Times 2016). However, in February 2017, then Frontex Director Fabrice Leggeri indirectly reiterated some of these accusations in an interview for the German newspaper Die Welt, stating that the presence of European ships off the Libyan coast would make it more likely for smugglers to force migrants onto unseaworthy vessels, and even openly called for "more police investigations" (Bewarder 2017). Another indirect accusation came from a confidential report signed by the Head of the EU naval mission EUNAVFOR MED Operation Sophia, Admiral Credendino, leaked to the public through Wikileaks. The report argued that "smugglers were "relying on an increasing number of NGO rescue vessels operating close to, and sometimes within, Libyan territorial waters" (EEAS 2016: 3). By early 2017, these accusations had trickled into the European political debate. In response to a parliamentary question by Marine Le Pen, EU Migration and Home Affairs Commissioner Dimitris Avramopoulos stated in March 2017 that "nothing [in the Frontex report] could be interpreted as allegation of collaboration between the smugglers and the NGOs", but acknowledged that SAR missions "close to, or within, the 12-mile territorial waters of Libya" could have "unintended consequences" (European Parliament 2017).
In October 2016, the private security contractors operating aboard Save the Children's ship started denouncing alleged contacts between human smugglers and humanitarian workers. The guards, hired by the Italian security company IMI Security Services, sent reports to the Italian judiciary and intelligence services and reached out to Italian opposition leaders. These accusations initiated the investigation against the German charity Jugend Rettet, which was apprehended by Trapani's public prosecutors in September 2017. In April 2017, another prosecutor-Catania's attorney general Carmelo Zuccaro-publicly claimed he had proof of direct contacts between NGOs and human smugglers (La Stampa 2017). When asked to corroborate his statements, Zuccaro acknowledged that he had no admissible court evidence, downplaying his own previous statements as "working hypotheses" (Scavo 2017). Nevertheless, these accusations were immediately appropriated by Italian opposition leaders, who called for a crackdown on sea rescue NGOs. Most notably, Luigi Di Maio, soon to become leader of the Five Stars Movement, popularized the expression "taxis of the sea", arguing that NGOs were not rescuing people in distress but facilitating smuggling, and suggesting that "someone" was behind their actions (Huffington Post 2018). These accusations found widespread currency in the media discourse. Between 2017 and October 2019, for instance, the right-wing outlet Il Giornale referred to NGOs as "taxis" a total of 30 times.
In response to these mounting suspicions, the Italian Senate Defence Committee initiated a parliamentary inquiry. While acknowledging that no evidence of rescuers' misbehaviour had yet been found, senators called for NGOs to be regulated in order to preserve Italy's control over its borders (Senato della Repubblica Italiana 2017). Following the Parliament's directions, the Italian cabinet led by Paolo Gentiloni-which had simultaneously negotiated an agreement urging Libyan militias to curb departures in exchange for financial aid-issued a code of conduct on maritime rescue on 1 August 2017 (Camilli 2017). Organizations refusing to sign were threatened with having the authorization to disembark migrants in Italian ports denied (Cuttitta 2020;Cusumano 2019a). Sea rescue NGOs were asked to refrain from entering Libyan territorial waters, always keep their geolocalization tracking devices on, not use light signals that could guide migrants to their ships, demonstrate that their personnel and vessels are properly trained and equipped, and constantly communicate with both flag states and the Italian Maritime Rescue Cooperation Centre. Moreover, rescuers were asked to provide information on suspect smugglers to Italian law enforcement authorities, collect makeshift boats and engines, and accept the presence of Italian policemen aboard their vessels (Ministero dell'Interno 2017). Although these measures were criticized as redundant, insinuatory, or incompatible with the humanitarian principles of neutrality, impartiality, and independence from political authorities, most NGOs signed the code (Del Valle 2019; Cusumano 2019a). Both signatory and non-signatory organizations, however, would soon see their activities severely limited.
The March 2018 national elections marked a clear victory for the parties that had demanded a stricter approach to irregular migration-the abovementioned League and the Five Stars Movement, which formed a coalition government led by the independent Giuseppe Conte (Conte I). Shortly after his appointment as Interior Minister in June 2018, the League's leader Matteo Salvini declared Italian ports "closed" to all foreign-flagged vessels that had rescued migrants off the coast of Libya (Carrera and Cortinovis 2019; Del Valle 2019). Despite the absence of formal laws to enact this warning, the government's new stance immediately had a strong impact on NGOs' activities. While in previous years Rome had allowed NGOs to dock quickly after a rescue operation, the new interior minister used his institutional role in identifying an appropriate place of safety to veto or at least delay the disembarkation of migrants on Italian territory. During the Conte I government (June 2018-August 2019), Italy's government engaged in several standoffs with NGO ships, using the refusal to authorize disembarkation to showcase its tough approach to irregular migration and obtain greater support from the EU in relocating those rescued at sea (Cuttitta 2020; Carrera and Cortinovis 2019).
As shown by Fig. 3, NGO ships and the people on board were forced to wait for an average of 9 days after requesting the authorization to land in Italy (or, much less frequently, Malta) before being allowed to dock. In two striking cases (concerning the Sea-Watch 3 vessel in June 2019, and the Open Arms in August 2019), the time at sea between rescue and disembarkation reached 20 days. In June 2019, Sea-Watch's captain Carola Rackete's decision to enter the port of Lampedusa without authorization escalated tension between the Italian government and NGOs, prompting Salvini and the right-wing press to refer to NGOs as "pirates" (Micalessin 2019).
Disembarkation delays subsided but did not entirely stop with the formation of a new coalition cabinet that no longer includes Salvini's League (Conte II) in September 2019. Since then, waiting time before disembarkation decreased from 9 to 4.2 days. Italy's new cabinet has, however, maintained the policy of not authorizing disembarkation immediately upon the reception of a "place of safety" request.
In early 2019, the Conte I cabinet complemented its power to veto disembarkation on the grounds of public safety concerns with some more specific legal provisions. In March, interior minister Salvini adopted a ministerial directive instructing border enforcement authorities to deny entry to any vessels carrying out SAR "improperly, violating the law of the sea and, therefore, jeopardizing public policy and national security" (Ministero dell'Interno 2019a). Three other directives followed, each addressing specific cases where NGOs had carried out SAR operations and had requested a place of safety in Italy . While these documents did not explicitly ban rescue ships' entry into Italian territorial waters, they paved the way for legal proceedings by suggesting that "entry should be deemed as non-innocent passage" (Ministero dell'Interno 2019d), arguing that NGOs were "exploiting international law" (Ministero dell'Interno 2019b), or blaming rescuers with "indirectly cooperating with smugglers, de facto encouraging sea crossings [and] objectively aiding and abetting entry on national soil" (Ministero dell'Interno 2019c).
Finally, in June 2019, the Italian government issued decree-law 14 June 2019, n. 53 (converted into Law n. 77 on 8 August 2019). This provided the interior minister with the power to restrict or prohibit access to Italian territorial waters to any private vessel for reasons of national security or public order. Shipmasters disobeying these provisions could be levied fines between 150,000 and 1 million EUR and have their ships confiscated (Cuttitta 2020; Carrera and Cortinovis 2019). Between June and August 2019, NGO ships were banned from entering Italian territorial waters in five separate cases. In four of these cases (concerning the Sea-Watch 3 in June; and the Open Arms, the Eleonore, and the Alan Kurdi in August), the ships were authorized to disembark after waiting for an average of 14 days at sea. In the remaining case (concerning the Ocean Viking), the ship disembarked 356 migrants in Malta after 13 days. Although the new Italian cabinet (Conte II) pledged to eventually repeal these provisions, which have never been applied, the decree remains in force as of August 2020.
Besides being the target of policy restrictions, NGOs were also increasingly subjected to judicial proceedings. Most accusations consist of aiding and abetting illegal immigration, a crime introduced into Italian legislation by the legislative decree 286/1998 and tightened by From "Angels" to "Vice Smugglers": the Criminalization of Sea Rescue... law 189/2002 and legislative decree 241/2004. However, other charges have sometimes been brought, including "violence against warship" and "illegal dumping of dangerous and infected waste". As Table 2 shows, 18 separate investigations have been launched in Italy or Malta against NGOs operating in the Central Mediterranean route between 2017 and February 2020. Although the independence of the judiciary from the executive branch is firmly embedded in the Italian political system, the wider climate of suspicion against NGOs is likely to have affected public prosecutors' decision to initiate investigations. These indictments increased in frequency over the years: four were launched in 2017, five in 2018, and nine in 2019. These proceedings resonate with the government stance towards NGOs. When the centre-left Gentiloni cabinet was in office, 0.3 investigations were initiated every month. The number of investigations against NGOs grew to 0.7 per month during the Conte I government and dropped again to 0.3 during the Conte II government. In most cases (11 out of 16), NGO ships were impounded while the investigation was ongoing. Vessel confiscations lasted an average of 6.6 months, grounding most NGO ships in the summer months, the busiest time of the year for SAR operations. By late 2019, five NGO ships had been simultaneously under seizure (see Fig. 4). The timeframe of the figure illustrates the role played by the wider political climate in the escalation and ensuing decrease of legal proceedings against NGOs.
However, it is worth acknowledging that not all investigations were motivated by the conviction that NGOs were involved in wrongdoings or the attempt to stop their activities. When rescue ships were denied the authorization to dock, some public prosecutors who considered it illegal to leave migrants at sea for too long arguably enacted a strategy of seizing ships in order to facilitate the disembarkation of those on board. This was the case for Sea-Watch 3 in May 2019 and Open Arms in August 2019, both impounded and released soon thereafter. Moreover, some prosecutors and civil society organizations challenged the closing of Italian ports on legal grounds, construing Salvini's decision to deny the disembarkation of migrants and keep ships on hold at sea as an abuse of power and a form of arbitrary detention.
As of February 2020, almost none of the investigations initiated against NGOs found sufficient incriminating evidence to start a trial. In fact, five public prosecutors closed their investigations without bringing any formal charges. Humanitarian workers have consistently been seen as operating under a state of necessity dictated by the duty to protect human life and rescue those in distress at sea, two obligations enshrined by international law that prevail over the domestic prohibition of abetting illegal immigration (Carrera and Cortinovis 2019). In the only investigation that will result in a trial-that against Jugend Rettet-humanitarian workers are suspected of having rescued migrants that were not in a situation of distress. Consequently, the crime of aiding and abetting illegal immigration could not be immediately waived on the grounds of a state of necessity. Public prosecutors, however, immediately dismissed the accusation that Jugend Rettet's personnel were in direct collusion with human smugglers and acknowledged that their motivations were "essentially humanitarian" (Camilli 2017). A trial was also initiated in Malta against Mission Lifeline's captain Claus-Peter Reisch, who was first found guilty of incorrect ship registration and later acquitted in the appeal trial (Brincat 2020).
Since the start of the Conte II government, the criminalization of non-governmental SAR has eased, but not completely subsided. Indeed, a persisting tendency to question, delegitimize, restrain, and possibly criminalize NGOs is visible both at the European and the Italian level. At the domestic level, as already mentioned, the number of investigations against NGOs has decreased sharply, and new interior minister Lamorgese largely refrained from using the power to deny rescue boats access to Italian territorial waters. The ensuing reduction in disembarkation delays has increased the time that NGOs have been able to spend conducting SAR operations off the Libyan coast (from around 25% of days during the Conte I government, to 56% of days in the first 5 months of the Conte II government). Some relevant figures within the cabinet, however, remain highly critical of NGOs' work. Foreign Minister Di Maio-who first referred to NGOs as "sea taxis"-is a case in point. At the European level, in September 2019, the Ministers of Interior of five EU countries (France, Germany, Italy, Malta, and Finland) adopted a declaration for a "predictable temporary solidarity mechanism", committing to accepting the relocation of part of the migrants rescued at sea by Italy or Malta. On the one hand, the declaration contributed to reducing disembarkation delays by replacing the lengthy ad hoc negotiations with other European partners previously initiated by Rome before authorizing ships' entry into Italian ports. On the other hand, however, the document indirectly reiterated criticism and suspicions against NGOs. Most notably, the declaration mentioned "avoiding the creation of new pull factors" as one of the official objectives of EU member states. Moreover, it also incorporates an almost literal translation of the 2017 Italian Code of Conduct, reiterating the restrictions and implicit accusations contained therein (Ministero dell'Interno 2017, Carrera and Cortinovis 2019).
In conclusion, while efforts to prosecute and obstruct NGOs decreased between late 2019 and early 2020, the broader discursive stigmatization of non-governmental migrant rescue did not completely subside. While all court decisions so far have dismissed all charges of collusion with smugglers as well as of aiding and abetting illegal immigration, NGOs have continued to be accused of serving as a pull factor of migration by media, politicians, and official documents. The article will now turn to assess these accusations.

The Gap Between Discourse and Evidence: an Appraisal
As illustrated in the previous section, no investigation to date has proven NGOs' direct involvement in unlawful activities. According to both Frontex and Italian law enforcement authorities, however, NGOs may still indirectly facilitate irregular migration and thereby contribute to increasing casualties at sea by serving as a pull factor of irregular migration. This seemingly plausible suspicion has served as the cornerstone of the discourses stigmatizing non-governmental maritime rescue and played a key role in enabling the policy restrictions and criminal investigations mentioned in the previous section. In this section, we show that these claims are largely unfounded. In order to do so, we rely on datasets focusing on irregular migrant departures from Libya and the outcome of their journey (i.e. arrival in Europe, death at sea, interception, and push-back to Libya), obtained by combining official figures from the International Organization for Migration (IOM), the UN Refugee Agency (UNHCR), and the Italian Coast Guard. Our datasets are either monthly (for the period January 2014-December 2018) or daily (between 1 January 2019 and 15 January 2020).
One of the main accusations is that NGOs' SAR activity did not contribute to reducing casualties and may have even increased deaths at sea. This alleged failure to reduce fatalities among migrants is mainly because rescue missions have allegedly encouraged smugglers to rely on increasingly unseaworthy boats (Deiana, Maheshri and Mastrobuoni 2019). By operating too close to Libyan coasts, NGO ships have been held responsible for this deterioration in the quality of migrant boats (Frontex 2017, Robinson 2016. While there is some evidence that migrants departing from Libya have used increasingly unseaworthy vessels, this deterioration occurred very early in the period of high departures (2014)(2015)(2016)(2017), probably as early as mid-2015. Data from the Italian Coast Guard shows that irregular crossing aboard dinghies-usually more rickety than the wooden or metal boats used by migrants in previous years-were already 67% of the total in 2015, when NGOs only conducted 13% of the total number of SAR operations. The share of dinghies over total migrant boats increased to 82% in 2016, on par with an increase in SAR by NGOs (25%), but then dropped again to 67% of the total in 2017, when NGOs' rescue operations peaked to 37%.
Rather than being solely encouraged by the presence of NGO ships close to Libyan shores, the use of unseaworthy rubber dinghies is likely to primarily be an unintended consequence of states' efforts to combat human smugglers. Specifically, European warships conducting the operation EUNAVFOR Med "Sophia" sought to disrupt the smuggling network by destroying at least 545 boats between 2015 and 2018. Arguably, this activity has incentivized the resort to dinghies, thereby indirectly increasing the deadliness of the crossing (Heller andPezzani 2018, House of Lords 2016). As stated in a confidential report to the European Council by the Head of EUNAVFOR Med "Wooden boats… can be re-used if recovered by smugglers. However, following operation SOPHIA entering into Phase 2A (High Seas), smugglers can no longer recover smuggling vessels on the High seas, effectively rendering them [wooden boats] a less economic option… and thereby hampering it" (EEAS 2015: 6).
Partly due to the increasing use of unseaworthy dinghies, the risk of crossing from Libya to Europe remained stable between 2014 and 2017, with dead and missing hovering at around 2% of total irregular departures despite the increasing presence of NGO ships in the area. The risk did not increase even after the first sudden drop in migrant departures after mid-July 2017. As monthly migrant departures plummeted from an average of close to 17,000 per month in the 12 months prior to the drop to just 4000 per month in the 12 months after that, the risk of crossing remained stable at around 2.1% (see Fig. 5).
The figures provided ostensibly vindicate the argument that non-governmental SAR in the Central Mediterranean fails to reduce the risk of death. The period between June 2018 and August 2019, however, shows that the presence of non-governmental SAR assets does appear to play an important role in decreasing the deadliness of the crossings. When most NGOs had suspended or reduced their operations due to the abovementioned criminalization process, the risk of crossing more than tripled relative to previous years, peaking at 6.1%. As Fig. 5 shows, this risk rose swiftly in the second half of 2018, after Salvini declared Italy's ports "closed", and further increased in the first 9 months of 2019. Finally, as Salvini's restrictive policies came to an end, the risk dropped abruptly to the lowest levels ever (1.0%), even if the last phase we examined (September-February 2020) corresponds to the autumn and winter months and therefore to rougher weather conditions. However, the small sample available warrants further verification of these findings. Overall, existing evidence shows that the presence of NGOs at sea, while unable to eliminate the risk of crossings, plays a key role in preventing casualties at sea from escalating irrespective of weather conditions. Even if NGOs reduced the relative risk of death attached to irregular crossings, they may have still indirectly contributed to increasing the total number of casualties by encouraging more migrants to make the journey to Europe. Hence, a comprehensive analysis of the impact of NGOs' activities on casualties at sea inevitably requires an appraisal of the accusation most frequently raised against non-governmental SAR operations: that of serving as a pull factor of irregular migration. In order to test this accusation, we extend the dataset on daily irregular crossings from Libya used in a previous paper (Cusumano and Villa 2019) until February 2020. We estimate daily departures from Libya by combining multiple datasets from UNHCR and IOM, and then proceed to directly track every SAR mission carried out by NGOs. Looking at daily departures and NGOs' SAR activity in 2019 and 2020 can be considered a "quasi experiment", because, during this time span, there were no other assets carrying out SAR operations off Libya apart from Tripoli's Coast Guard, which returned migrants' to Libyan coasts. Consequently, if NGO operations close to the Libyan coast serve as a "pull factor", their presence or absence off the Libyan coast should affect the number of irregular departures. We thus developed a model that tries to explain migrant departures by looking at the presence or absence of NGOs off Libya's coast, weather conditions (temperature and wind) in Libya's capital Tripoli, political conditions in Libya, the change in government in Italy, and the day of   (n.d.), and own data the week when the crossing occurred. In order to account for uncertainty over the actual day of departure from Libya (migrants sometimes depart several days before being rescued), we use a 3-day moving average for all our variables. Table 3 shows the results.
We find that the only factors with a strongly significant statistical effect on the number of irregular departures from Libya are weather conditions and the level of political stability in the country, which we calculate by using daily oil production as a proxy. By contrast, the presence of NGO vessels off Libya's coast had no significant effect on driving irregular departures. In fact, not only is the recovered effect of NGOs' presence insignificant, but the sign of the effect is itself negative. To be sure, more research remains needed on this subject. However, our evidence suggests that even if the alleged pull effect of maritime rescue operations enjoys widespread currency and has served a key role in delegitimizing NGOs, this argument is not supported by existing data.

Conclusions
The suspicion that NGOs facilitate irregular migration or even directly colluded with human smugglers has shifted European decision-makers and public opinion against non-governmental SAR operations. These allegations have caused sea rescue NGOs to face policy restrictions, judicial criminalization, and broader social delegitimization.
This article has sought to contribute to the scholarship on the criminalization of humanitarianism by conducting two interrelated tasks. First, we have thoroughly examined the origins of existing critiques to non-governmental SAR and their escalation into judicial indictments and policy restrictions. Second, we have appraised the merits of the arguments formulated against NGOs by using available evidence. As all court investigations to date have disproven the existence of a direct collusion between NGOs and human smugglers, we have chosen to examine the two main indirect, empirically verifiable accusations that have enabled this delegitimization process: the allegation that NGOs have not reduced (and may have even increased) casualties off the coast of Libya and the suspicion that non-governmental sea rescue has served as a pull factor of irregular migration. While additional research is still needed, the quantitative evidence we collected questions both arguments. Non-governmental SAR missions appear to have played an important (although not decisive) role in reducing the deadliness of sea crossings without significantly contributing to incentivizing irregular migration.
Our findings have important policy implications. Most notably, evidence showing that SAR operations do not incentivize irregular departures as often assumed may prompt European Note: *** represents significance at the .001 level From "Angels" to "Vice Smugglers": the Criminalization of Sea Rescue... governments to reconsider their approach to non-governmental SAR operations in the Mediterranean. NGOs' presence at sea has increased as a result of European law enforcement assets' disengagement from the Southern Mediterranean. As military and law enforcement assets are even more unlikely than NGOs to incentivize irregular crossings, decision-makers should consider gradually restoring operations combining SAR and border control tasks. Such missions could both rescue lives at sea and prevent undetected arrivals, thereby reconciling humanitarian imperatives and national security concerns.
Although this article has advanced the existing debate on the delegitimization and criminalization of humanitarianism and solidarity, the subject still warrants follow-up research from lawyers and social scientists alike. Most notably, legal scholars can more systematically examine the decisions to prosecute and acquit humanitarian workers briefly presented above, using Italian case law as a window into the tension between the obligations arising from international human rights law as well as the law of the sea and domestic legal provisions against aiding and abetting illegal immigration. Social scientists, on the other hand, should both continue to examine quantitative evidence on irregular departures to corroborate or disprove our results, and to systematically examine the discourses underlying the criminalization of maritime rescue in Italian, Maltese, and other countries' media.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.