key: cord-0024272-vyqiargj
authors: Pocock, Nicola S; Stöckl, Heidi; Tadee, Reena; Rongrongmuang, Wansiri; Tharawan, Kanokwan; B Adamson, Fiona; Zimmerman, Cathy
title: Victims or suspects? Identifying and assisting potentially trafficked fishermen: A qualitative study with stakeholders and first responders in Thailand
date: 2021-11-16
journal: J Migr Health
DOI: 10.1016/j.jmh.2021.100074
sha: 800324751bba8661ab1d6ccf9a3675a90657ab29
doc_id: 24272
cord_uid: vyqiargj

Prompted by reports of ‘sea slavery’ in the fishing industry and threats of sanctions, Thailand has faced pressure to eradicate human trafficking the fishing sector. Although the Thai government has responded with anti-trafficking policies, there remains little understanding about their implementation. Specifically, little is known about how government agencies, NGOs or industry perceive “trafficking”, and no research examines how trafficked fishermen are identified and assisted. This study aimed to: 1) explore how stakeholders described trafficking in the fishing sector and their perceptions of trafficking indicators; and 2) identify challenges encountered by frontline responders to identify and assist trafficked fishermen. We conducted interviews with 33 key informants, which were analysed thematically. Findings indicate that authorities and industry representatives believed migrant brokers caused employers to “inadvertently” traffic men. Trafficking was perceived to take place primarily outside of Thai waters, beyond the government's jurisdiction. Most stakeholders considered violence and being confined as key indicators of trafficking. Officials expressed confusion about whether debt bondage and document confiscation “counted” as indicators. Ambiguity and confusion about trafficking indicators in screening forms, combined with perceived “deservingness” of official victim status, underpinned frontline responders’ decisions about who was a victim of trafficking (VoT). Practical and structural constraints included interpreter shortages, and expanded civil servant remits without commensurate staff increases, which hindered officials’ responses to trafficking. This study addresses a critical knowledge gap on the implementation of anti-trafficking policies and offers findings to assist policymakers to address the challenges faced by frontline responders to improve victim identification and assistance.

Thailand has faced pressure to eradicate human trafficking in the fishing sector. An EU ban on seafood imports from the world's 4th largest seafood exporter has loomed large, prompted by media exposés on the extent of slavery, violence and abuse in the commercial fishing sector ( EJF, 2013 ; Hodal and Kelly, 2014 ; McDowell et al., 2015 ) . Following a downgrade to Tier 3 (worst offender status) in the US State Department's Trafficking in Persons report in 2014, the Thai government implemented wide-ranging reforms and policies. These included labour inspections, vessel monitoring, improved victim identification procedures and mi-ficked enough "), which is criminalized. In contrast, "banal exploitation " involving deceptive recruitment, debt bondage and exploitative working arrangements of the larger migrant workforce is normalized ( Yea, 2015 ( Yea, , p. 1092 . Migrants often undertake precarious jobs where unequal bargaining relations with employers are common. Bonded by debt, employers have no need to physically confine workers. Even if migrants leave economically exploitative jobs without wages, they are not guaranteed compensation, because men who experience banal exploitation alone are not deemed "trafficked enough " to warrant state intervention ( Yea, 2015 ( Yea, , p. 1096 .

Dominated by short term guest worker programs that discourage permanent settlement, migration regimes in Asia are governed by the common ethical belief that it is legitimate for the state to discriminate against non-citizens in favour of citizens. Employers are innocent and migrants are suspect in disputes ( Ong, 2009 ) . Migrant workers are ascribed lesser value than native workers, with correspondingly fewer rights and caring regimes, while facing greater disciplinary and security regimes ( Ong and Collier, 2008 ) that directly affect their wellbeing (e.g. no healthcare entitlements). In Thailand, migrants face restrictions on freedom of movement and routinely report arbitrary arrest and extortion by authorities ( Derks, 2013 ) . Migrant workers' political rights are also constrained, where they do not have the right to vote or to self-organize in trade unions ( Derks, 2013 ) .

Policies including labour inspections, vessel monitoring, improved victim identification procedures and temporary migrant worker registration centers arguably do not aim to improve the substantive freedoms of fishermen. Men remain immobilized at sea and on land: at sea via the captain's disciplinary techniques of violence; and on land where indirect forms of coercion, including debt bondage and direct ones, such as arrest and extortion by the police, combine to ensure that migrant fishermen's agency is severely constrained ( Derks, 2010 ) . The fishing sector has fewer rights protections than land-based work regarding working time, salary payment and time off ( Thanachaisethavut, 2011 ) .

In the Mekong context of sex trafficking, both traffickers and the trafficked person may not necessarily see the relationship as exploitative, due to historically grounded patron-client relationships whereby debt bondage is not seen as a moral problem ( Molland, 2012 ) . Similarly, slaveholders in India do not see bonded labour as problematic, but as a mutually beneficial relationship. Labourers receive "care " (the provision of food, shelter, minimal financial recompense) in exchange for their continued labour and respect ( Choi-Fitzpatrick, 2017 ) . However, workers are now voting with their feet and slaveholders yearn for the past when workers were considered grateful, hardworking and honest. This nostalgia and lamenting of workers' new job choices hints at slaveholders' sense of loss and decline. In the Mekong, deceptive recruitment often takes place within a village's network of social relationships by "accidental traffickers ". Debt bondage may sometimes be a conscious choice by the migrant to reduce the risks associated with migration ( O'Connell Davidson, 2013 ) . While traffickers and brokers are often depicted as organized crime networks by media and authorities, they are more often fluid, informal, familial networks, with some migration in the Greater Mekong Subregion beginning as voluntary but ending up as trafficking ( Molland, 2012 ) .

The United Nations Protocol to Prevent, Suppress and Punish Trafficking in Persons (2003) defines human trafficking to involve:

• Recruitment, transportation, transfer, harbouring or receipt of persons. • Using the threat or use of force, coercion, abduction, fraud, deception, abuse of power, vulnerability, or giving payments or benefits to a person in control of the victim.

• For the purpose of exploitation, including for the purpose of prostitution of others, sexual exploitation, forced labour, slavery or similar practices and the removal of organs ( UN, 2003 ) .

The term "human trafficking " sometimes appears interchangeably with the term forced labour. The International Labour Organization's (ILO) Forced Labour definition is the most widely operationalized measure in victim screening tools ( Stringer et al., 2016 ) . Its indicators have informed the Thai Ministry of Labour's inspection form (Supplementary  Table 2) . Similarly, Thailand's Anti-Trafficking in Persons Act (2008) is closely based on the definition offered in the UN protocol. This Thai legislation informs the trafficking indicators used in the national Preliminary Victim Identification Process (PVIP) form (Supplementary Table  1 ). The definition of "forced labour or service " in Thai legislation included the following elements:

• Compelling the person to work or provide services (Act). • Putting the person in fear of injury to life, body, liberty, reputation or property, by means of intimidation, by use of force or any other means causing person to be unable to resist (Means) ( RTG, 2008 , sec. 4) .

Migrants who arrive legally are considered as more legitimate members of the national community in immigration assimilation research ( Wright et al., 2016 ) . Among Thai citizens in states bordering Myanmar, few believed that unregistered migrants (18%) compared to registered migrants (52%) should have the same labour protections as Thais ( Sunpuwan and Niyomsilpa, 2014 , p. 25) . Thais more frequently believed that registered migrants should receive the same standard health treatments compared to unregistered migrants (75% and 40% respectively) ( Sunpuwan and Niyomsilpa, 2014 , p. 30) . Documentation affects perceived deservingness for migrants' labour and health protections in Thailand.

Despite inspections of over 10,000 vessels conducted by authorities, no trafficking cases were identified in the year preceding Thailand's downgrade to Tier 3 ( USDOS, 2014 ). And despite the repatriation of 1917 men suspected to be trafficked from Indonesian islands between 2014 and 16, just 53 were classified as VoTs by the authorities ( RTG, 2017a , p. 66) . Low victim identification rates worldwide have been attributed to: the covert nature of trafficking; low disclosure by victims due to high levels of trauma and distrust of law enforcement; poor awareness among law enforcement about trafficking indicators; and limited awareness among potential VoTs about their legal rights ( Gallagher and Holmes, 2008 ) . Law enforcement officers are rarely aware of key trafficking indicators, with trafficking mainly perceived as sex trafficking ( Barrick et al., 2014 ) . VoTs who initially consented to exploitative jobs may be considered at fault by law enforcement ( Jones, 2012 ) .

While guidelines for victim identification processes are available for first responders across law enforcement, health and social welfare ( Andrees, 2008 ; IACP, 2007 ; IOM, 2009 ) , these guidelines predominantly draw on limited empirical studies from the United States, Europe and Australia. These studies are often normative, outlining best practices, and do not always examine how victim identification and assistance are being conducted. Among frontline government responders implementing policies, exercising discretionary power can help them to cope with on-the ground challenges, to resist authority, or to achieve performance gains ( Erasmus and Gilson, 2008 ) .

This study aimed to: (1) explore how stakeholders described trafficking in the fishing sector and their perceptions of trafficking indicators, and; (2) identify challenges frontline responders faced in identifying and assisting trafficked fishermen. 

Semi-structured interviews were conducted between August to October 2014 with 33 key informants ( Table 1 ). Most ( n = 24) were frontline responders to potentially trafficked fishermen and nine were policy stakeholders. Purposive and snowball sampling were used to recruit participants, based on an initial sample frame of known service providers. Government agency participants were a mix of senior and frontline officials involved in various aspects of rescues, victim screening and labour inspections forming the Multi-Disciplinary Team (MDT) primarily responsible for victim screening but also joint labour inspections at the time of data collection. NGO participants provided health and welfare services to migrant and Thai fishermen. For two larger NGOs, multiple participants ( n = 2,3) were interviewed where the participants' roles differed drastically (e.g. legal services or frontline response). Industry participants were industry representatives who had previously employed fishermen. A provincial port research site and major fishing hub was chosen as the primary location to interview frontline responders, followed by Bangkok, where senior officials and NGO management were usually based. The port research site is not named to preserve the anonymity of participants.

Twenty interviews were conducted in Thai with one of two research assistant interpreters, who were trained on topic guide content and interview technique. One interview was conducted with assistance from a Burmese interpreter working at the organization, with the remaining 12 interviews conducted in English ( Table 1 ) . Interviews lasted 1.5 h on average. Questions focused on participant's experiences of directly assisting fishermen, or on colleagues' experiences of doing so in their organization (for senior officials who guided policy). Participants were also asked their opinions about labour inspections, migrant registration drives and victim screening processes, and the challenges they faced conducting any of these tasks.

Written informed consent was obtained from each participant. Except for one interview, interviews were digitally recorded, transcribed verbatim to English. Ethics approval was obtained from the Institute of Population and Social Research Institutional Review Board, Mahidol University (reference 2014-1-1-22), Thailand and the London School of Hygiene and Tropical Medicine Observational Research Ethics Committee (reference 8368), UK.

Qualitative data were analysed using thematic analysis. Transcripts were read and re-read to gain familiarity with the data and generate initial codes. A priori themes identified from the topic guides and conceptual framework (deductive approach), e.g. labour inspections and political context of trafficking response, shaped earlier versions of a coding framework. New codes were also identified from the data (inductive approach). Together, emergent codes and a priori codes were collated un-der overarching themes. Themes were continually reviewed for internal consistency and distinctions between themes until refined themes were developed, taking care to include negative cases and less prominent themes ( Braun and Clarke, 2006 ) . Themes were also explored by organization type ( Table 1 ) as a possible source of heterogeneity in perceptions and responses, with findings written up accordingly. Qualitative data were coded and analyzed by hand, in NVivo 11 and OneNote, by one researcher (NP). Following thematic analysis of interview data, policy documents including the Anti-Trafficking in Persons Act (2008), Labour Protection Act (1998), National Action Plans and Royal Thai Government progress reports on Anti-Human Trafficking (2012-17), fisheries legislation and victim screening forms used by officials were analyzed to assess how legal and policy definitions of trafficking intersected with participant's perceptions of trafficking. Final themes were reviewed and discussed among study authors in early drafts of the paper, which were written up by one researcher in a final draft (NP).

Findings are grouped by the main themes of: perceptions of the trafficking problem; perceptions of trafficking indicators, and; structural policy constraints. Table 2 summarizes the themes and sub-themes:

Notions of "inadvertent trafficking " implied that employers were deceived by brokers and fishermen, which led to employers accidentally trafficking fishermen. Labour shortages in fishing were intrinsic to "inadvertent trafficking ", which emerged regularly among employers and officials. Boat captains paid fishermen one-month's salary in advance to convince crew to join their boat, in order to compete with land-based sectors which did not pay advances where migrants preferred to work. Problems arose when fishermen ran away soon after receiving advances, leaving employers out-of-pocket, when they had paid the advance to fishermen directly or believed they had done so via brokers. One official suggested that the labour policy was deliberately designed to make it difficult to switch from fishing to land-based work:

"If we let them change [jobs] all of the fishing labourers will go to work on land. We have different colour cards for [land based work and fishing]. We do this because most of the boat business employers will give advance money to the workers. But… some workers work for only one trip, 10-15 days, and simply quit or disappear. The employers are at loss here." (Labour official, 20) Because of the advance payment scheme, a system of exploitation had "grown organically" in response to labour shortages according to an NGO participant. Boat captains were unapologetic about holding men on boats, forcibly, or by default because it was impossible to escape at sea after paying the advance. One participant explained how smaller operators paid advances to recruit fishermen quickly, so they could set sail sooner and earn needed profits to alleviate business debts, compared to wealthier operators who could afford to keep vessels grounded for longer (Ind., 21). A captain with less ability to find fish would be under pressure to work crew harder and demand backbreaking work (INGO, 2). A police officer explained how his friends, who were boat owners, might commit "inadvertent trafficking " because they believed that locking up workers was simply a way of ensuring they would recoup their advance payments:

"Many of my friends are owners and they would tell me that they will follow these workers and lock them on the boat so that they can pay for what they took [advance payment]. So I told them they couldn't do that, it would be a human trafficking case. Your assets can be seized…The owners did not know this. " (Law enforcement officer, 29)

Brokers were intrinsically tied to inadvertent trafficking. Captains had to use brokers due to language barriers, but they may be unaware that brokers enforced debt bondage. One industry participant suggested This industry participant's view that the trafficking law only covered sex work is common worldwide. Across participants, migrant brokers especially were blamed for trafficking. Thai brokers were infrequently mentioned or considered less of a problem. Usually, employers elicited sympathy:

"The agents who are the same nationality as the workers are the problem, there are Thai agents too but only a few of them. Those foreign agents trick the workers and also Thai business owners." (Law enforcement officer, 29)

Brokers were a convenient and fast way to recruit crew, compared to the Memorandum of Understanding (MOU) recruitment process. Recruitment via the MOU was expensive, bureaucratic and timeconsuming and employers did not think it worthwhile for sectors like fishing where worker retention was low. Consequently, most used brokers, which brought with it the possibility of "inadvertent trafficking ". Migrants might also be "inadvertently trafficked " when they did not pay the brokers' fees upfront before the journey to Thailand. When they became indebted to brokers, migrants were more likely to be sold or persuaded to work in the fishing sector.

Most participants felt that trafficking took place on long-haul boats outside of Thailand, often in cases where fishermen "changed their minds " about being in the sector. While officials who conducted rescues cited cases in Thailand, officials conducting labour inspections said it was rare to find trafficking cases within Thai waters. The short-haul boats that officials inspected docked frequently, offering more opportunities for men to escape (Off., 16). Contrary to the perception that trafficking was widespread in Thailand, labour shortages meant that captains had to treat crew extremely well to retain them:

"The workers here are being pampered like sons of the owners. They take them to restaurants to treat them, afraid that they would run away."

(Law enforcement officer, 22) This participant went on to describe how men who changed their minds on long-haul boats could "turn into " trafficking cases when they had been at sea for years. Fishermen might claim that they'd experienced trafficking because they found the work too hard, did not like it and wanted to come home:

"Sometimes people agree to go and work on the boat voluntarily, but after a year or two they change their minds and want to come back. But the boat is not due back and there are no return vessels… So they file a complaint that they were forced to work and haven't been home for 4,5 years. That is the problem today."

(Law enforcement officer, 22) Fishermen changing their minds on long-haul boats was a perceived problem among some officials and law enforcement ( n = 4) and industry ( n = 3).

Labour inspections were conducted by the Multi Disciplinary Team (MDT) and were either pre-announced or random and unannounced when conducted in Thai waters. Boats departing and returning to Thai waters were meant to be inspected by Immigration and the Marine department after employers notified these departments in advance.

Victim screening took place at two stages, a preliminary victim identification form (PVIP) (Supplementary Table 1 ) used by police and an in-depth case interview at the government shelter. Men who were classified as VoTs were sent to government shelters. Men who were not classified as VoTs and who did not have documents were considered "illegal" migrants, and were sent to Immigration Detention Centers (IDCs) for deportation. The perception that migrants would either be victims of trafficking or suspects for immigration offences is a commonly held binary categorization reflected in state responses to trafficking worldwide.

Men who were not VoTs, but who had documents and were working in Thailand legally, were released (NGO, 12), while Thai men were assisted to return home by the Ministry of Social Development and Human Security (MSDHS). Men who were classified as VoTs but who did not want to proceed with legal cases or go to the shelter could "choose" to be deported according to some NGOs (NGO, 12).

Verifying facts and being "sure " about the veracity of a trafficking case was an important theme throughout the assistance process. NGOs described having to verify that potential trafficking cases were genuine when they worked with authorities, who would demand further information before agreeing to conduct a rescue. NGOs and police may be blamed for acting against employers if it were not a "true" trafficking case (NGO, 17), which corresponds with employers' perceptions about fabricated cases. Some NGOs conducted rescues alone because of authorities' collusion with employers:

"Even though this is already hard for me, I cannot raid with the po-lice… Sometimes when they arrested, police raid and send them back to the boat. So how can I call the police?" (NGO service provider, 7)

Because of the risk associated with transporting undocumented migrants, NGOs who intervened alone also needed to be sure it was a genuine trafficking case beforehand.

Physical abuse and confinement were considered indicators of genuine trafficking cases according to this official. For law enforcement and officials, it was important to ascertain the facts when serious allegations of trafficking were made, so that employers were not wrongly accused (LE, 22) . Following up with employers the next day was preferable for one police officer, despite the possibility that employers could influence workers' responses during that time. Social workers had specialist interviewing skills which enabled them to discern facts from potential VoTs:

"We must send in a social worker. The cases that come in are not always true. We must confirm them." (Welfare official, 23)

One police officer described how police needed social skills to get the facts from victims alongside social workers, as well as build trust (LE, 29) . According to some law enforcement officers, challenges encountered during victim screening included VoT lying, which made it difficult for police to reconcile conflicting decisions about VoT status with social workers (LE, 18) . While the police often had the final say and legal authority to bring victims to government shelters, MSDHS social workers and NGOs played a key role gathering more facts from the interview to come to a mutual decision. But evidence of trafficking was not usually visible, posing a challenge for the investigation and legal case (Off., 33). When VoT had physical wounds, photos and doctor's reports were used in court as evidence of trafficking; such forensic medical evidence was important to claim compensation. The shelter case interview features a long section on physical and mental health conditions assessed by a doctor.

During victim screening, police sometimes disagreed with NGOs and social workers about VoT status, often because local police did not perceive debt bondage as an indicator of trafficking, or were less familiar with trafficking law. For officials, physical violence was a strong indicator, but cases where pay was withheld needed to be carefully assessed as to whether they counted as trafficking:

"If they do not pay the worker then we have to analyze if it is a human trafficking case or not. But if they physically hurt the worker then it is 100% a human trafficking case. According to Thai law [withholding pay] is just almost there, but not yet. But for foreigners and NGO this is already human trafficking. " (Labour official, 20)

The relatively new anti-trafficking law (eight years at time of interview) did not include debt bondage, leading to the perception among police and even courts that debt bondage alone did not indicate trafficking (NGO, 4). While PVIP indicators do not explicitly include debt bondage, the use of "force by other means (body or mind) which causes person to not resist " could be interpreted to include financial coercion. Officials would have to consult the PVIP form addendum for an explicit definition of debt bondage:

"The act of binding oneself in a contract which creates excessive and unreasonable debts, that is not possible to release, may be another category of exploitation in trafficking in persons. " (Point 6, PVIP form addendum)

Temporary identity documents were not necessarily protective against trafficking according to NGO participants. For some, OSSCs were a visible, short-term policy designed to show the international community that Thailand was addressing trafficking. Principle flaws included brokers not being required to register, and insufficient enforcement of migrant's rights after they had obtained documents (NGO, 11). Even migrants with documents could be taken advantage of by agents and employers because of corruption. An NGO participant discussed how fishermen felt that they didn't get benefits from registering. It was cheaper to make under-the-table payments to authorities than it was to formally register. One industry participant suggested that corruption by authorities was another reason why the OSSCs were not open year-round. During periods of closure, undocumented migrants would have to continue making under-the-table payments to authorities (Ind., 8) . This participant felt that year-round opening of the registration centers would address recruitment issues. In contrast to NGO participants, law enforcement and officials felt that OSSCs and registration of undocumented workers was the primary way to "solve " the trafficking problem: "When [migrants register] they have an identity, a card, a number. You know where they are, you can track them. There will be no human trafficking. They will also have access to their rights… We want them to come in legally so that they can be safe…" (Health official, 27).

Withholding of documents was not always perceived as a trafficking indicator. Wording of regulations was sometimes ambiguous, and frontline officials used discretion to make decisions. For example, fisheries law (at the time of interview) did not specify a minimum wage, written contracts or maximum working hours. Boats with fewer than 20 employees did not have to document pay, and the law excluded long-haul boats fishing outside of Thailand for one year or more. Most inspectors found it difficult to ascertain what acceptable employment practices involved. Researching the going salary rate locally in cases of wage disputes was a strategy to ascertain what was fair according to one official (Off., 16). In particular, inspectors found it difficult to ascertain whether document confiscation was an unfair labour practice, or indicator for trafficking. Some captains claimed to keep fishermen's documents to ensure they didn't get wet or spoiled at sea. "Legally they are wrong, but we give them exceptions… I'm not telling the employers to confiscate the workers' cards, I'm just telling them to hold on to them… I think that confiscating the cards of the workers partially qualifies as a human trafficking crime. Not 100% but bordering." (Labour official, 20) This official interpreted the relevant laws flexibly allowing exceptions for some captains, but was unsure whether document confiscation constituted trafficking according to the law. His-confusion was warranted; document confiscation was not explicitly included in the Act, PVIP form or even mentioned in the case shelter interview. Yet it was included in the labour inspectors form (Q's 9,31,32, Supplementary Table 2). Establishing whether documents were voluntarily deposited was crucial to determine whether a case qualified as trafficking. One NGO participant suggested that labour shortages in fishing prompted authorities to not check documents thoroughly:

"I understand that in the fishing industry, [authorities] can't check everything… Thai people do not want to do these jobs. We still need [migrants] . We have to sometimes… turn a blind eye. There are no labours." (NGO service provider, 15).

Language barriers plagued the entire assistance process, compounded by lack of interpreters during inspections and screening. During initial phone contact, fishermen who couldn't read Thai were unable to specify boat names for authorities to locate them. Instead, men could only describe the color of the boat, its number or the type of boat they were on (NGO, 13, 17) . Language barriers were also a challenge during inspections:

"For example, there are 20 workers in one ship. [Interpreters] only ask one worker for interview and they don't allow other workers to answer or confirm the facts. When there is an issue, they ask the entire ship to come back to shore. This is a waste of time and opportunity." (Industry representative, 21) This industry participant expressed resentment for being called back to shore, based on a flawed inspection process which relied on one fish-erman's testimony. He went on to describe how the presence of many officials may intimidate workers and cause them to mistakenly answer questions, even when an interpreter was present (Ind., 21). Conversely, one official suggested that interpreters put migrant workers at ease when there were many officials on-board (Off., 16). Elsewhere, it was suggested that interpreter shortages might discourage officials from inspecting migrant labour heavy industries like fishing. For NGO participants, inspections were unlikely to find trafficking as migrant workers were often told how to respond to officials by their employers. Interpreters colluding with employers was also a problem:

"…the workers are told what to say and they're forced to speak … just going in and doing some random inspection, often using the translators of the company, they're often brokers or very influential bad people. There's the management relying on these kinds of wicked people to be their translators, it's a big issue." (INGO stakeholder, 2) One social worker described the difficulties of conducting victim screening when potential VoTs spoke minority languages and no interpreters were available (Off., 24). Even when interpreters were available, officials could not rule out miscommunication. Some officials mentioned how social workers building trust with interpreters was essential, to avoid miscommunication when interpreters were not trained in specialist interview skills as social workers were. Being unsure about what was being interpreted was a source of unease for officials. NGOs were frequently contacted to provide interpreter services during victim screening, but they were usually unpaid. Budget constraints meant that one NGO participant had to turn down interpreter requests from police, who assumed the NGO had funding for such requests.

While government employed interpreters were available from the DLPW, some officials indicated a preference for NGO interpreters (Off., 23). The government shelter described using registered NGO interpreters, who were international students with MSDHS training, paid up to 1200 baht/day. But there were not enough registered interpreters and the shelter often had to wait.

Several procedural flaws with inspections meant that these were not necessarily effective at detecting human trafficking. Both officials and NGOs discussed how pre-announced inspections, or those conducted at port, were unlikely to find forced labour, as they could not observe actual working conditions (NGO, 12) . For boats departing Thai waters, immigration officers cited very low numbers of fishing boat inspections (20/month) which were usually for mother ships used in transshipment. Another problem was that fishermen with fake documents required for fishermen departing Thai waters were smuggled onto long-haul boats after inspections (Off., 20) .

A further procedural flaw was that inspectors could be sued by employers for giving them a bad inspection report, or when they disagreed with the inspector's order to pay the employee (Off., 16). The chance of being sued may act as a disincentive to inspect employers who were likely to be incompliant (INGO, 1). Additional practical challenges included labour inspectors not being used to being on boats and experiencing seasickness (Off., 16), as well as coupling inspection boats with fishing vessels at sea due to waves (Off., 28).

Registration systems lacked coherence and integration, which sometimes meant men could re-register under different names. There was no unified government database which tracked men individually. This problem, combined with separate databases for land-based work and fishing and a registration backlog, meant migrant workers could register twice at OSSCs with no penalty (Ind., 26). Migrant fishermen could therefore register and runaway with advances multiple times, which was a perceived problem among employers. Language barriers added to the confusion and meant that migrants could easily re-register, according to another industry participant:

"When I make a list of my workers, I have to write the way they look instead of their names such as the bald man, the small guy, etc.… These [Burmese migrants] always change their names when they make a new [registration] card." (Industry representative, 21) Workloads, new remits and resource constraints New demands to inspect boats on top of existing workloads put pressure on officials, alongside budget constraints and personnel constraints linked to gender. Most officials cited structural constraints, including increased workloads now that their remits had expanded to include fishing inspections. One law enforcement officer discussed how news media, trade embargoes and the relatively new trafficking law meant that inspections were now more stringent than before:

"We have to increase both the quality and quantity of our work, noticeably different from 2 to 3 years ago… Today instead of checking only the crew members we have to do more during the inspection… now we have to take pictures, go down to the engine room and much more. That is the problem resulted from the news."

(Law enforcement officer, 32) Labour inspectors were considered reactive by some NGO participants, only investigating when complaints were received:

"… the government official, they're working inside the airconditioned room, sitting in the chair. That's all. If they go, they just see, and take photos. They will not talk direct with the fishermen, they never have." (NGO service provider, 17).

An official discussed the difficulty of juggling existing responsibilities with inspections, as well as gender barriers to conducting inspections when most of the agency's staff were female:

"Sometimes the women [officers] are not flexible. And it is rather dangerous to inspect these boats. There are no restrooms either. If it is a male officer then it is easier, but for ladies, it is difficult." (Labour official, 20).

Female staff were sometimes unwilling, and perceived as unsuitable to conduct boat inspections. Women's presence on boats was considered bad luck in Thailand and captains were often reluctant to have women on board (INGO, 1). Budget constraints and rising fuel costs also limited the number of inspections that could be carried out.

Policies and procedural flaws restricted officials' capacity to address trafficking. Only government officials could conduct inspections (Off., 20). With just a handful of officials to some 80 contract staff trying to fulfil existing responsibilities in one agency, inadequate numbers of officials was a limitation:

"An inspection requires officers… We can't employ anyone to go. The law states that only government officials [who have] served for 4 years or more can go on boats. This is a new law. If you just came in for a year then you can't go to the inspection." (Labour official, 20) .

Similar constraints were faced at the MSDHS, where social workers needed a special permit to conduct victim screening (Off., 23). In addition, some officials discussed how civil servant distribution across provinces was decided centrally with uniform allocations. They implied that higher numbers of migrant workers, and perception that their province was targeted for trafficking activities, meant that more officials should be allocated to help. With a large quota of inspections to complete each month, labour inspectors might be inclined to inspect nearly compliant businesses. Fishing boats were likely to have many violations and it would be better not to check them as "everything is illegal" (INGO, 1), which would increase the inspectors' workload as they had to follow up and ensure compliance within two weeks (INGO, 1).

The civil service rotation policy was a further impediment to identifying TIP cases discussed by NGOs and officials. Frequent rotations in specialist trafficking units lead to gaps in knowledge when incoming officials or police officers did not know much about trafficking: (Law enforcement officer, 31) Some participants implied that officials would pay large sums of money to rotate to the central trafficking unit, because police could exert greater influence in the national unit and perhaps demand more bribes. Another policy constraint was no time limit on investigations by the special investigation department who took on many fishing TIP cases, unlike the central police anti trafficking division which had a sixmonth investigation limit. The DSI might take a long time to collect evidence, leading to legal delays, which would prolong men's stay in shelters (NGO, 4) .

Findings from this study show that victim identification is difficult when recent legislation pits practices including precarious working conditions, confinement and broker procurement of workers as trafficking, where these practices are not locally understood to be exploitative or harmful. Resource constraints and structural flaws also hinder victim identification rates.

In defining the problem, brokers were integral to employers "inadvertently and unknowingly " trafficking fishermen. Because migrants are not "fully forced " in some cases, employers see fishermen as job-seeking migrants rather than victims of trafficking when they engage in practices amounting to debt bondage defined in legislation. Locating the problem "outside of Thailand " allowed industry and authorities to avoid culpability. Employers' narratives conflict with accounts of NGOs and rights activists about who is at fault for fishing trafficking, which has emerged amidst a perfect storm of labour shortages, declining fish stocks and increased fuel costs. NGO reports suggest the impossibility of making profits without slave labour ( Robertson, 2011 ) . Smaller operators have gone out of business, while only businesses large enough to be compliant with new regulations have survived ( Nasueroh, 2015 ) . Our findings support those of Choi-Fitzpatrick (2017) where employer and traffickers yearn for a past when workers had fewer job choices and had to accept exploitative working arrangements ( Choi-Fitzpatrick, 2017 ).

Participant's narratives of trafficking may have been informed by the legal definition. Brokers are blamed and debt bondage and document confiscation are not problems when not explicitly defined in law or in screening forms. The mismatch between the law and indicators on screening forms may cause confusion among officials. Since this study was carried out, a plethora of legal and policy updates has ensued ( Stride, 2016 ). An expanded definition of "forced labour " now includes document confiscation and debt bondage in the revised Anti-Trafficking Act ( RTG, 2017b ). A new eight-page screening form used by officials identified over 800 individuals during 2016, of whom 524 (including 305 labour-trafficked persons) accepted assistance from the MSDHS ( RTG, 2017a , p. 74) . It is unclear how many of them were fishermen. However, since 2016, victim identification rates declined before peaking in 2019. Of 230 victims identified in 2020, only 13 were trafficked fishermen, and between January to March 2021, just 6 trafficked fishermen were being supported in MSDHS shelters ( RTG, 2021 ) . The Thai government has made significant progress in strengthening worker protections in legislation, such as extending the same protections that land based workers have to fishermen ( Ng, 2020 ) , and ratifying the ILO's Work in Fishing Convention ( MFA Thailand, 2019 ). Reforms have improved working conditions, particularly reducing violence at sea and ensuring a minimum wage standard in the sector (2020 Marschke). However, worker abuses have persisted, including forced labour ( Ng, 2020 ) . A 2019 survey found that 14% of fishermen were in forced labour, mainly linked to living and working in degrading conditions with limited freedom to leave ( ILO, 2020 ), compared to 17% forced labour among fishermen in 2013 ( ILO and ARCM, 2014 ). Re-ports suggest that instead of violence, fishermen are now being coerced by employers withholding their wages ( Ng, 2020 ) .

Our findings indicated flaws in the labour inspection process that contributed to few VoTs being identified in the fishing sector, corroborating reports elsewhere ( EJF, 2015 ; Stride and Murphy, 2016 ) . Since this study was conducted, at least 30 "Port In Port Out " (PIPO) centers have been established, where MDTs conduct inspections of longhaul vessels departing and re-entering Thai waters. In some cases, men were not screened individually and were screened with captains present ( EJF, 2015 ; Greenpeace Southeast Asia, 2016 ) . Inspections have been inadequate to identify labour exploitation, when inspections consisted of cursory checks of crew lists against migrant workers' documents ( Stride and Murphy, 2016 ) . Since inspections began, there have been reports of migrant fishermen being pre-selected to give rote responses to inspectors, being paid THB 1000 (USD 32) per inspection by employers ( ILO, 2020 ) . Inspectors are somewhat incentivized against conducting thorough document checks because of labour shortages, which in May 2020 were estimated at 50,000 ( Marschke et al., 2021 ) . Moreover, finding trafficked men or labour dispute cases would increase their workload as they would need to conduct victim screening and filing reports.

Findings raise questions about whether possessing documents prevented migrant fishermen from being trafficked. Officials in our study assumed that registration drives and documents were protective, which is contradicted in other research. For example, mobile unit registration of fishermen in 2015 regularized victims of exploitation and abuse, with some officials assuming that fishermen with formal identification documents did not need to be screened for indicators of forced labour ( EJF, 2015 ) . Similar perceptions of officials have been noted during inspections ( Stride and Murphy, 2016 ) . Documents themselves are not necessarily protective when laws are not enforced due to corruption and bribery. For example, among Lao migrants migrating via legal channels to Thailand, documents were not protective against exploitation or debt bondage ( Sosamphanh et al., 2008 ) . The Thailand-Myanmar government MOU recruitment scheme ostensibly removes unscrupulous brokers, but NGOs have suggested that recruitment via the MOU has led to "legal human trafficking " in cases where government-endorsed agencies have exploited workers ( Lun, 2016 ) . Findings suggest that interventions and advocacy are necessary around enforcing the rule of law and rights associated with documents.

Addressing resource constraints may improve victim identification rates. As labour inspectors' remits have expanded, increases in numbers of inspectors is not commensurate across Asian countries ( Lee et al., 2011 ) . Time constraints prevented inspectors from speaking with crew ( Stride and Murphy, 2016 ) . The DLPW have taken steps to increase numbers of staff with the authority to conduct inspections. With coastal provinces under greater pressure to identify VoTs compared to inland provinces, civil service staffing allocations should be based on provincial need not uniform numbers for each province. Civil servants in specialist trafficking units should not rotate frequently to avoid loss of institutional memory in handling TIP cases.

Language barriers were a major concern during inspections and during victim screening, where untrained local migrants often step in to provide interpreter services. Dependent on legal changes, foreigners could be hired in the civil service and be trained as interpreters for TIP cases. Policies enacted since this study was conducted indicate progress: by end 2016, 254 interpreters were registered with the government ( RTG, 2017a , p. 90); DLPW were hiring two migrant language coordinators for each PIPO inspection center (ibid); multilingual hotlines and mobile applications are now available; and complaints forms are now produced in Khmer, Laos and Burmese languages ( ILO, 2017 ) . However, findings also raise questions about why languages are not learned by Thai stakeholders, e.g. one operator could not speak adequate Burmese to learn crew's names, despite hiring Burmese workers for 20 years. The operator's attitude is perhaps reflective of prejudice against doing so in the native population ( Traitongyoo, 2008 ) . Prior survey research with Thais conveys an expectation to speak Thai: 74% and 84% be-lieved that unregistered and registered migrants respectively should use Thai to communicate with them ( Sunpuwan and Niyomsilpa, 2014 , p. 27) . Language interventions with both crew and captains could enhance communication about working conditions and positively affect fishermen's wellbeing, including for when they seek medical treatment for injuries and illness incurred at work ( Pocock et al., 2018 ) .

This study has some limitations. Trafficked fishermen were not interviewed about their experiences of being identified and assisted, which may limit conclusions drawn. However, our study builds on research conducted with Cambodian fishermen ( Ry, 2014 ) , and offers compelling findings about how frontline authorities respond to trafficking. Within the government category of participants, our sample was heterogeneous by Ministry or Department, which may raise concerns about representativeness of our findings. However, we interviewed all stakeholders involved in the Multi-Disciplinary Team (MDT), who conducted victim screening and who were beginning to conduct inspections at the time of interview. Theoretical saturation was achieved with our final sample. A final limitation was the rapidly changing policy landscape since data were collected in 2014 for the relevance of our findings. However, a complaint to the ILO by trade unions suggests that forced labour, use of brokers and debt bondage continued in the Thai fishing industry ( ILO, 2017 ). Furthermore, fishermen in a recent survey were asked whether pay, safety and working conditions had gotten better, worse or were unchanged since the first major reforms in 2015 ( ILO, 2020 ). While salaries and housing conditions had improved, aggregate improvements were outweighed by issues that remained unchanged including salary deductions, lack of work contracts, long working hours and limited changes in safety at work ( ILO, 2020 ) .

This study provides evidence of an understudied policy implementation process, victim identification and assistance of trafficked fishermen, in a middle-income country setting. The unique data presented in this study furthers our understanding of how stakeholders' perceptions of the "problem " affects whether and how trafficked fishermen are assisted. Our findings should remain useful for policymakers seeking to understand why victim identification rates remain low. Findings can also help policymakers to understand the challenges faced by frontline responders in trying to assist trafficked men.

This study addresses a significant knowledge gap about how trafficked fishermen are identified and assisted. Much is at stake in victim identification and assistance procedures, for both governments and potential VoTs. Threatened with trade sanctions, the Thai government have reason to improve victim identification rates, but this is challenging when local perceptions of what constitutes trafficking does not follow legal definitions. Where VoT status provides access to tangible services that are state financed, states have an incentive to exceptionalize trafficking to the most heinous cases marked by violence and physical confinement. But for potentially trafficked fishermen, not being identified or assisted can mean the difference between life or death. With fishermen's lives at stake, researchers and activists should continue to monitor the implementation of policies ostensibly aimed at reducing trafficking or improving identification rates, to ensure that fishermen get the support and assistance they need.

Forced labour and human trafficking: a handbook for labour inspectors / Beate Andrees

When farmworkers and advocates see trafficking but law enforcement does not: challenges in identifying labor trafficking in North Carolina

Using thematic analysis in psychology

What Slaveholders think : How Contemporary Perpetrators Rationalize What They do

Human rights and (Im)mobility: migrants and the state in Thailand. Sojourn

Migrant labour and the politics of immobilisation: cambodian fishermen in Thailand

Thailand's seafood slaves: human trafficking, slavery and murder in kantang's fishing industry

Sold to the Sea: human trafficking in thailand's fishing industry

How to start thinking about investigating power in the organizational settings of policy implementation

Developing an effective criminal justice response to human trafficking: lessons from the front line

Turn the tide: human rights abuses and illegal fishing in Thailand's overseas fishing industry

Trafficked into slavery on Thai trawlers to catch food for prawns. the Guardian

The crime of human trafficking: a law enforcement guide to identification and investigation

Endline research findings on fishers and seafood workers in Thailand. ILO ship to shore

Sixth supplementary report: report of the committee set up to examine the representation alleging non-observance by Thailand of the forced labour convention, 1930 (No. 29), made under article 24 of the ILO constitution by the international trade union con

Employment practices and working conditions in Thailand's fishing sector, ILO tripartite action to protect the rights of migrant workers within and from the greater mekong subregion (GMS TRIANGLE Project). International labour organization (ILO) & Asian research center for migration

Caring for trafficked persons: guidance for health providers

Human trafficking victim identification: should consent matter?

Research on occupational safety and health for migrant workers in five Asia and the pacific countries

After state counsellor's visit, overhaul of thai migrant worker scheme expected

COVID-19, instability and migrant fish workers in Asia

AP Investigation: slaves may have caught the fish you bought

Thailand ratifies ILO's work in fishing convention no

Thai Fishermen Reeling From Regulations

Thailand's seafood slavery: why the abuse of fishermen just won't go away

Troubling freedom: migration, debt, and modern slavery

A bio-cartography: maids, neo-slavery, and NGOs. Migration and Mobilities: Citizenship. Borders and Gender

Global Assemblages: Technology, Politics, and Ethics as Anthropological Problems

Because if we talk about health issues first, it is easier to talk about human trafficking "; findings from a mixed methods study on health needs and service provision among migrant and trafficked fishermen in the Mekong

Trafficking of fishermen in Thailand

Royal Thai government's progress report on anti-human trafficking effort s

Thailand's country report on anti-human trafficking response

Amendment to the prevention and suppression of human trafficking act B.E. 2560 (3rd version). Royal Thai Government (RTG)

Anti-trafficking in persons act

Deceived Into the Fishing industries: Case Studies of Trafficked Cambodian workers

Whether you go illegally or legally in the end it's the same, you're cheated' -a study of formal and informal recruitment practices of Lao workers migrating to Thailand

Current and potential impacts of legal reforms on businesses and workers in Thailand's fishing industry, focus on labour issues in the fishing industry

Assessing government and business responses to the thai seafood crisis

New Zealand's turbulent waters: the use of forced labour in the fishing industry

Survey of thai public opinion toward myanmar refugees and migrant workers: an overview

Informal workers and legal protection in Thailand

The Management of Irregular Migration in Thailand: Thainess

United nations protocol to prevent, suppress and punish trafficking in persons, especially women and Children, general as

Trafficking in Persons Report

Public attitudes toward immigration policy across the legal/illegal divide: the role of categorical and attribute-based decision-making

Trafficked enough? Missing bodies, migrant labour exploitation, and the classification of trafficking victims in singapore

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.jmh.2021.100074 .