South Africa’s trafficking Act punishes harshly on paper but protects victims and convicts traffickers poorly
Introduction
South Africa had no single trafficking statute until 2015. Before the Prevention and Combating of Trafficking in Persons Act 7 of 2013 began operating on 9 August 2015, prosecutors worked with scattered common-law and statutory offences, and the Act was meant to replace that patchwork with something modelled on the 2000 UN Trafficking Protocol (Mollema, 2024). It was also meant to do more than punish. It sets up a system for identifying victims, housing them, regularising the stay of foreign nationals and paying compensation (Republic of South Africa, 2013; Iroanya, 2025). Eleven years on, the picture is not flattering. The United States State Department’s 2026 report, to take the most recent snapshot available, records 28 prosecutions and no convictions in its latest period, down from 57 prosecutions and 27 convictions in the one before (United States Department of State, 2026).
This paper argues that the Act is a reasonably good piece of criminal law and a much weaker piece of victim protection, and that the gap between the two explains a good deal of the poor enforcement record. The penalties are heavy and the definitions are wide, which most of the commentary accepts (Kruger, 2024; Mollema, 2024). What fails is the machinery around the offence: identification, certification, shelter, immigration relief and data. Some of that machinery was never switched on at all (Ramsamooj, 2023; Iroanya, 2025). The sections below set out what the Act does, look at the prosecution record, examine victim identification and protection, deal with the provisions that remain inactive, and then ask whether more enforcement is really the answer.
What the Act promises
On paper, the offence is wide. Section 4 criminalises delivering, recruiting, transporting, harbouring or receiving a person through means that include threats, force, deception, abduction and abuse of vulnerability, for the purpose of exploitation. Consent is no defence once any of those means is shown, and the definition of abuse of vulnerability reaches undocumented presence, pregnancy, addiction and economic circumstances (Republic of South Africa, 2013). The maximum penalty is life imprisonment, a fine of up to R100 million, or both, with debt bondage and using a victim’s services carrying up to 15 years (Republic of South Africa, 2013). Ramsamooj (2023) contrasts this with Canada, where sentences are often lenient, and concludes that South African sentencing is the stricter of the two.
Kruger (2024) is the most positive of the commentators. Reviewing court decisions between 2009 and 2023, she finds that the Act meets international minimum standards, covers all forms of trafficking, and has produced binding and clarifying decisions, including support for the one legally binding definition of trafficking. Mollema (2024) points to the heaviest sentence reported so far, 19 life terms handed down in S v Seleso in 2019 for trafficking an orphaned child. Chipofya (2024) is more cautious. In a comparison with Malawi and Uganda, he suggests that some duty-to-act provisions in the South African Act use vague terms, or ignore whether the person bound by the duty had the capacity to comply, which may weaken enforcement and sit awkwardly with the principle of legality. That matters because the Act creates offences of failing to report, punishable by up to five years (Republic of South Africa, 2013).
So the commentary does not agree on how good the statute is. Kruger (2024) leans towards the view that the law is sound and the courts are making it work. Ramsamooj (2023) says some provisions remain ineffective, while Frankel, as summarised by Emser (2024), is reported to see deficiencies in the Act that undermine its implementation structure. Iroanya (2025) takes a middle line, arguing that the Act supplies sufficient legal instruments and that protection is weak anyway. I think the middle line is closest to the evidence, for reasons the next sections give.
Few prosecutions and fewer convictions
The enforcement numbers swing so widely that they are hard to read. According to the State Department, South Africa started 123 prosecutions in one period, 57 in the next and 28 in the one after that, while convictions went from 12 to 27 and then to none (United States Department of State, 2025; 2026). Iroanya (2025) cites the same source for 2022, when 29 cases were investigated, prosecution began in 15 and 74 suspects were arrested. These are a foreign government’s compilation, and the pages do not state the reporting periods, so I treat them as indicative. Swings of that size on a small base probably say as much about reporting and case clustering as about real change. They also tell us that South Africa does not publish a consolidated annual record of its own, since Ramsamooj (2023) works from individual cases and general estimates and offers no national totals.
Why do so few cases end in conviction? The sources point to several things at once. Prosecutions depend heavily on victim testimony, and Iroanya (2025) reports that victims rarely testify because of intimidation and fear tied to traditional oaths. Botha and Warria (2021) found that social workers in shelters faced court delays, complicated trauma, security problems and mistrust, all of which affect whether a survivor stays engaged with a case. Motseki (2022) interviewed 40 officials from the Directorate for Priority Crime Investigation, the police, Home Affairs and Social Development in Gauteng, and the findings pointed to corruption, weak commitment, limited awareness, no clear strategy for investigating and prosecuting, and a lack of capacity, resources and training. The sample is small and regional, and the journal is modest, so it is best read as indicative.
Complicity is the hardest part to measure. The 2026 report says no officials were reported as investigated, prosecuted or convicted for complicity, and that observers described corruption as widespread in Home Affairs and the police (United States Department of State, 2026). I would treat that with some care, because it rests on what observers told the report’s authors and not on tested evidence. Even so, it is consistent with Motseki (2022) and with Frankel’s view, as Emser (2024) summarises it, that the criminal justice system seems reluctant to punish offenders in line with their crimes. Heavy maximum penalties do little if the cases never get that far.
When convictions do happen, the sentences are severe. The 2025 report records sentences from 10 years to life, and the 2026 report mentions, from media coverage, seven Chinese nationals sentenced to at least 20 years each (United States Department of State, 2025; 2026). Prosecutors also sometimes add organised crime charges, which raises the penalty again. The difficulty is not leniency at the sentencing stage. The 2026 report says police and prosecutors struggled with cases involving many victims and lacked training and resources, which is a capacity problem that no penalty clause can fix (United States Department of State, 2026; Motseki, 2022).
Victim identification and protection
The Act builds protection on identification. Anyone who knows or reasonably suspects that an adult they meet in their work is a trafficking victim must report to the police at once, the police must refer the person to an accredited organisation and notify the provincial Department of Social Development within 24 hours, and the department must decide without delay whether to issue a letter of recognition (Republic of South Africa, 2013). A refusal can be appealed to the MEC within 14 days. It is a sensible design. Everything downstream, including shelter and services, depends on someone spotting the victim and the paperwork moving quickly, which is why Iroanya (2025) treats identification and referral as the point where protection either starts or does not.
In practice it does not move quickly. Iroanya (2025) describes a multi-agency system that is not working efficiently, with police, social workers and others applying standard operating procedures inconsistently across agencies and provinces, officers who lack training in identifying, referring and certifying victims, and Home Affairs staff who confuse victims with illegal migrants. The State Department says the same thing in blunter terms. It reports inconsistent screening, delays in government certification that held up access to emergency shelter, and officers who sometimes kept victims overnight at police stations (United States Department of State, 2025; 2026). Ramsamooj (2023) adds that police sometimes left victims with NGOs without telling the Department of Social Development, so the NGOs could not claim the per-night funding.
Shelter capacity has grown but remains thin. The 2025 report counted 20 accredited NGO shelters, and the 2026 report counts 29, six run by government and 23 by NGOs, though only four multipurpose shelters accept men, no shelter takes foreign child victims, and private donors reportedly cover more than half of shelter budgets (United States Department of State, 2025; 2026). Sambo and Sibanda (2025) interviewed 14 service providers at five organisations and found that a multi-disciplinary approach helped women survivors rebuild their lives, but that it ran into trouble when some stakeholders did not deliver on their mandates. That is a modest finding from a small qualitative study. Still, it lines up with Botha and Warria (2021) and with Iroanya (2025), and it suggests that the weak point is not the NGOs at the end of the chain but the state agencies feeding into it.
Foreign and undocumented victims seem to fall through most easily. The 2025 report says officials sometimes required victims to take part in an investigation before they could get services, which the report says breaches the law, and that undocumented victims were barred from working even while cooperating with the authorities (United States Department of State, 2025). Foreign victims also feared deportation if they reported. Put differently, the system sometimes asks for cooperation from the people it gives the least reason to cooperate, and it does that at the very moment prosecutors most need their evidence (Iroanya, 2025).
Provisions that were never switched on
The most telling feature of the Act is that parts of it have never operated. Chapter 3 gives foreign victims who may help an investigation a recovery and reflection visa of up to three months, extendable once by another three, and shields them from deportation while a decision is pending. The SAFLII consolidation, updated to October 2024, still shows sections 15 and 16 and section 31(2)(b)(ii) as awaiting proclamation (Republic of South Africa, 2013). Ramsamooj (2023) says they have stayed ineffective since the Act began, and Iroanya (2025) says the regulations behind them have not been promulgated. The 2026 report puts it more sharply, saying the Minister of Home Affairs approved implementing regulations but Parliament has not adopted them (United States Department of State, 2026).
There is an oddity here. Section 19 tells the police to inform a foreign suspected victim of the right to apply for the very visa that section 15 would create, so officers are required to describe a protection that cannot yet be granted (Republic of South Africa, 2013). Mollema (2024) notes in the same vein that the regulations on relief from deportation had not been issued. It is hard to see how a foreign national, often undocumented, would come forward to a police officer under those conditions, and that is exactly the witness the prosecution needs. This is an inference, and none of the sources tests it directly. Even so, it fits Iroanya’s (2025) argument that weak immigration protection and the confusion of victims with illegal migrants undercut both protection and prosecution.
Other delays follow the same pattern. The Act required a national policy framework to be tabled within a year of commencement, and Ramsamooj (2023) reports that it was adopted on 25 April 2019, nearly four years after the Act began. She also argues that compensation under section 29, which lets a court order a convicted trafficker to pay a victim for injury and lost income, should be actively enforced (Ramsamooj, 2023). The State Department notes that judges can order restitution (United States Department of State, 2026), but none of the sources reports how often they do. The provisions that need only a court to act appear to be used rarely, and the provisions that need ministers and Parliament to act have not been used at all.
Is more enforcement the answer?
There is a reasonable objection to my argument, and Hüncke (2023) supplies it. Her chapter suggests that the trafficking concept in South Africa has been used to support a restrictive migration policy and the criminalisation of commercial sex, and that separating cross-border migration and sex work from trafficking can strengthen calls for tighter border control. On that view, pressing for more arrests and more prosecutions is not obviously a gain for the people the Act is meant to protect, and Iroanya (2025) makes a related complaint about confusing trafficking victims with irregular migrants. I take the point, although the evidence behind it is not visible from the chapter’s abstract.
A second worry is about balance between forms of exploitation. Emser (2024) reports that research on South African trafficking has leaned towards sex trafficking at the expense of other forms, and labour trafficking, which affects men disproportionately, gets less attention. The enforcement figures are less lopsided than that. The 2025 report records 15 convictions for sex trafficking and 11 for labour trafficking (United States Department of State, 2025). The 2026 report, however, says inspections are insufficient, that the government reported no efforts to regulate labour recruiters, and that the ban on worker-paid recruitment fees is enforced weakly, especially in mining and the informal sector (United States Department of State, 2026). Labour trafficking is prosecuted, then, but the preventive tools that would find it earlier look underused.
The other sources offer a partial answer. Iroanya (2025) points to the way Home Affairs officials treat victims as illegal migrants as a real harm of the legalistic approach, and the State Department reports that victims have been penalised for unlawful acts committed because they were trafficked (United States Department of State, 2026). Those findings suggest the risk Hüncke describes is real. They also point away from heavier penalties and towards the unglamorous work of training, screening and regulations that put protection ahead of enforcement. So the reform that fits the evidence best is to finish the Act, not to toughen it.
Conclusion
The Act punishes harshly and defines trafficking widely, and the commentary is largely content with that (Kruger, 2024; Ramsamooj, 2023). The weaknesses sit elsewhere. Prosecutions and convictions are few and erratic, with the State Department’s latest report recording no convictions at all (United States Department of State, 2026). Victim identification depends on agencies that apply procedures inconsistently (Iroanya, 2025), and certification delays have held back emergency shelter (United States Department of State, 2025).
The clearest failure is that the immigration protections for foreign victims have never been brought into force, more than a decade after the Act was passed (Republic of South Africa, 2013; Ramsamooj, 2023). Proclaiming sections 15 and 16 and adopting the regulations would cost very little compared with a new statute, and they would address the group of victims the system currently handles worst (Iroanya, 2025; United States Department of State, 2026). After that, three modest steps seem to follow from the evidence: binding and consistent identification procedures with real training, an annual public record of investigations, prosecutions, convictions and certified victims, and routine use of the compensation provision. Without them the next report will probably describe the same distance between what the Act says and what the state does.
Reference list
Botha, R. and Warria, A. (2021) Challenges of social workers providing social services to adult victims of human trafficking in select shelters in South Africa. Social Work/Maatskaplike Werk, 57(1). doi:https://doi.org/10.15270/57-1-906
Chipofya, M.V. (2024) Duty to act provisions and omissions offences under the anti-human trafficking statutes of Malawi, Uganda and South Africa. Journal of Comparative Law in Africa, 11(2), pp. 80-108. doi:https://doi.org/10.47348/jcla/v11/i2a3
Emser, M. (2024) Human trafficking: misery and myopia in South Africa. South African Journal of Science, 120(3/4), article 17393. doi:https://doi.org/10.17159/sajs.2024/17393
Hüncke, A.S. (2023) ‘Negotiating around “human trafficking” in South Africa and the border area’, in Migration and making an income in the context of ‘human trafficking’. Wiesbaden: Springer VS, pp. 157-191. doi:https://doi.org/10.1007/978-3-658-41670-6_5
Iroanya, R.O. (2025) ‘The challenge of legalistic approach to protecting the rights of trafficking victims in South Africa’, in Kaime, T. and Dinokopila, B. (eds) Intractable problems of human rights. Pretoria: Pretoria University Law Press, pp. 182-206. doi:https://doi.org/10.29053/978-1-0672373-3-2_9
Kruger, B. (2024) Enhancing the criminal justice response to human trafficking in South Africa: legislation and case law in the spotlight. South African Journal of Criminal Justice, 37(1), pp. 48-83. doi:https://doi.org/10.47348/sacj/v37/i1a3
Mollema, N. (2024) ‘Combating human trafficking in South Africa’, in Rodrigues, A.M. and Guia, M.J. (eds) New forms of human trafficking. Cham: Springer, pp. 53-74. doi:https://doi.org/10.1007/978-3-031-39732-5_4
Motseki, M.M. (2022) Justice for victims of human trafficking in Gauteng Province, South Africa. Technium Social Sciences Journal, 28, pp. 360-369. doi:https://doi.org/10.47577/tssj.v28i1.5098
Ramsamooj, J. (2023) A comparative study on human trafficking as a crime in South Africa. Southern African Public Law, 38(2), article 12459. doi:https://doi.org/10.25159/2522-6800/12459
Sambo, J. and Sibanda, S. (2025) Multi-disciplinary initiatives to rendering services to women survivors of human trafficking in South Africa. Journal of Social Service Research, 52(1), pp. 4-14. doi:https://doi.org/10.1080/01488376.2025.2468949
Legislation and case law
Republic of South Africa (2013) Prevention and Combating of Trafficking in Persons Act 7 of 2013. Available at: https://www.saflii.org/za/legis/consol_act/pacotipa2013472/index.html (Accessed: 18 October 2026).
United States Department of State (2025) 2025 Trafficking in Persons Report: South Africa. Available at: https://www.state.gov/reports/2025-trafficking-in-persons-report/south-africa/ (Accessed: 18 October 2026).
United States Department of State (2026) 2026 Trafficking in Persons Report: South Africa. Available at: https://www.state.gov/reports/2026-trafficking-in-persons-report/south-africa/ (Accessed: 18 October 2026).