Why scam compounds survive the raids

Dr Nicola Harding, Financial Crime Lab

On 27 August 2026, Cambodia's anti-scam commission told around fifty diplomats and international organisations that its crackdown had achieved the "complete eradication of large-scale online scam operations". Between July 2025 and 20 August 2026, it said, authorities had investigated 793 suspected scam locations, raided 624 sites, detained nearly 30,000 suspects of 39 nationalities, and rescued and deported 58,266 foreign nationals (Associated Press, 2026).

Six weeks later, police in Papua New Guinea told OCCRP that a suspected scam operation they had raided in September, where 39 Indonesian workers were rescued, had arrived in June "basically escaping out from Cambodia" (OCCRP, 2026a). Displacement is not deterrence. In this piece I want to ask a harder question. Criminological evidence suggests that displacement is usually partial and often outweighed by the benefits of prevention. Why, then, does the scam compound industry appear to relocate so completely, and what would it take to change that?

What the displacement literature actually says

The fear that crime prevention simply moves crime elsewhere is as old as situational prevention itself. Reppetto (1976) set out the forms that displacement can take, including movement to other places, times, methods, targets and types of offence. Crucially, however, his conclusion was not that displacement is inevitable. He argued that there are definite limits to displacement, that the costs of moving lessen the frequency of offending even when offenders do adapt, and that the displacement potential of an anti-crime strategy can be gauged in advance and, in some cases, minimised (Reppetto, 1976).

The empirical record has largely vindicated him. Guerette and Bowers (2009) reviewed 102 evaluations of situational crime prevention, containing 574 observations. Displacement was observed in 26 per cent of observations, while the opposite effect, a diffusion of benefits to areas and offences not directly targeted, was observed in 27 per cent. Where spatial displacement did occur, it tended to be smaller than the treatment effect, so the intervention was still beneficial overall (Guerette and Bowers, 2009).

The scam compound industry therefore presents a puzzle. If displacement is usually partial, the evidence from Southeast Asia looks unusually complete.

Three forms of displacement in one crackdown

The evidence of adaptation is now substantial, and it spans several of Reppetto's categories at once.

The first is territorial. The UN Office on Drugs and Crime's 2026 threat assessment for Southeast Asia found that law enforcement pressure had displaced but not dismantled operations, with networks relocating from Myanmar and Laos to Timor-Leste, Pacific Island states and parts of Africa. The report describes syndicates "jurisdiction shopping" for places with weak governance, and identifies corruption as the primary enabler of technology-driven organised crime in the region (The Business Standard / Reuters, 2026). The Papua New Guinea case is one concrete example.

The second is tactical, and it happened within Cambodia during the crackdown itself. Amnesty International identified 86 compounds and found evidence of state intervention at only 24. Survivors described compound managers who knew about police interventions in advance, moved workers between sites, and in one account loaded people onto a bus and drove them into the mountains (Amnesty International, 2026).

The third is a displacement of targets, both of the people recruited and of the people defrauded. UNODC has documented recruitment advertisements aimed at German, Polish, Dutch, Spanish, Italian, French, Swedish, Norwegian and English speakers, and people from at least 80 countries and territories have been identified inside compounds in the region (OCCRP, 2026b).

Why the limits on displacement are not biting

Reppetto's insight was that displacement depends on cost. It is limited when the thing an intervention removes is difficult to replace. Read through that lens, the scam compound crackdown has largely targeted the most replaceable element of the operation.

A compound is a building, a network connection and a supply of devices. These are cheap to reconstitute. The Papua New Guinea operation was reportedly run from a villa (OCCRP, 2026a). The scarce resources are elsewhere: protection from the state, which the Amnesty testimony suggests could include advance warning of raids; laundering capacity, which converts victim losses into usable assets; recruitment pipelines, which supply labour; and the trafficked workers themselves. UNODC's emphasis on corruption as the primary enabler, and on networks sharing financial and operational infrastructure across multiple criminal markets, points in the same direction (The Business Standard / Reuters, 2026; OCCRP, 2026b).

The implication is practical rather than pessimistic. Displacement is not a law of nature that makes enforcement futile. It is a measure of how easily offenders can replace what has been taken from them. A crackdown that removes buildings will displace readily. A crackdown that removes protection, laundering routes and recruitment channels should, on Reppetto's own logic, displace far less, because those elements are harder and more costly to rebuild in a new jurisdiction.

This also changes how success should be measured. Counting raided sites, detained suspects and deported foreign nationals measures what the state did. It does not tell us whether the costs of operating have risen, whether losses have fallen, or whether the operation has simply moved. A credible assessment of any crackdown would track the industry's footprint across borders, not only within them.

The victims who do not fit

The second problem with the eradication narrative concerns the people it counts. Cambodia reported that it had rescued and deported 58,266 foreign nationals (Associated Press, 2026). Amnesty interviewed 73 survivors and assessed that all met the internationally recognised definition of trafficking victims, yet none had been formally recognised as a victim of human trafficking. Some were left sleeping on the streets or held in overcrowded immigration centres, and some described being extorted or threatened by police (Amnesty International, 2026).

Christie's (1986) concept of the ideal victim helps explain why. He described the ideal victim as the person who, when hit by crime, most readily receives the complete and legitimate status of being a victim. That status is easiest to claim for someone who is weak, engaged in a respectable activity, blameless for where they were, and harmed by a big and bad offender who is a stranger (Christie, 1986). A person trafficked into a scam compound fails almost every test. They often travelled voluntarily in response to a job advertisement. They are working-age adults. Above all, they were made to commit fraud against other victims. In public narratives, and frequently in official processes, they are rendered as offenders who happened to be rescued rather than as victims who were forced to offend.

International law already recognises this problem. Article 26 of the Council of Europe Convention on Action against Trafficking in Human Beings requires parties to provide for the possibility of not imposing penalties on victims for their involvement in unlawful activities, to the extent that they have been compelled to do so (Council of Europe, 2005). The non-punishment principle exists precisely because forced criminality produces victims who look like offenders. A crackdown that processes such people primarily through detention and deportation risks discarding the most valuable intelligence source the system has: the people who know how the compounds recruit, who protects them, and where they moved next.

As with money mule activity, the people at the visible edge of the operation are simultaneously exploited and instrumental, and systems built to sort people into victims or offenders struggle to see them clearly. Within liminal spaces, as my colleagues and I have argued, individuals can be transformed into victims or potential co-offenders (Harding et al., 2025). The scam compound is that transformation operating at industrial scale.

What would make displacement harder

The Financial Action Task Force now estimates global scam losses at more than $500 billion and is conducting a focused review of end-to-end illicit financial flows from scam compounds ahead of a flagship fraud event in 2027 (AML Intelligence, 2026). That focus on financial flows is the right one, because money is one of the elements of the operation that cannot simply be rebuilt in a new villa.

A displacement-aware strategy would follow the same logic across the other scarce resources. It would treat protection and corruption as enforcement targets in their own right. It would disrupt recruitment at source, including on the platforms where job advertisements are placed. It would identify and protect trafficked workers as witnesses rather than processing them as immigration cases. And it would judge success by whether the industry's operating costs rise across the whole region, rather than by whether one country can declare its own territory clear.

Cambodia may well have closed many compounds, and that matters to the people who were held in them. But the question for prevention is not whether a building has been emptied. It is whether the next one has become harder to fill.

References

AML Intelligence (2026) 'LATEST: FATF will hold "flagship event" in 2027 to tackle fraud', 7 October. Available at: https://www.amlintelligence.com/2026/10/latest-fatf-will-hold-flagship-event-in-2027-to-tackle-fraud/

Amnesty International (2026) Cambodia: Evidence suggests scamming compounds bypassed despite high-profile crackdown. News release accompanying Falling Through the Cracks: Cambodia's 'Crackdown' on Scamming Compounds, 8 June. Available at: https://www.amnesty.org/en/latest/news/2026/06/cambodia-evidence-suggests-scamming-compounds-bypassed-despite-high-profile-crackdown/

Associated Press (2026) 'Cambodia claims it has wiped out online scam compounds', 28 August. Available at: https://apnews.com/article/cambodia-online-scam-compounds-cybercrime-ae52ba73f813548d8a0cd9f81cd4469a

Christie, N. (1986) 'The ideal victim', in Fattah, E. A. (ed.) From Crime Policy to Victim Policy: Reorienting the Justice System. London: Macmillan, pp. 17-30. Available at: https://doi.org/10.1007/978-1-349-08305-3_2

Council of Europe (2005) Council of Europe Convention on Action against Trafficking in Human Beings, CETS No. 197. Warsaw, 16 May. Available at: https://rm.coe.int/168008371d

Guerette, R. T. and Bowers, K. J. (2009) 'Assessing the extent of crime displacement and diffusion of benefits: A review of situational crime prevention evaluations', Criminology, 47(4), pp. 1331-1368. Available at: https://doi.org/10.1111/j.1745-9125.2009.00177.x

Harding, N., Cooper, E., Sales, T., McDonald, A. and Kingston, S. (2025) 'The liminality of fraud: Reimagining fraud theory to inform financial crime prevention', British Journal of Criminology, 65(3), pp. 618-638. Available at: https://doi.org/10.1093/bjc/azae069

OCCRP (2026a) 'Police say scam center busted in PNG had moved from Cambodia', 8 October. Available at: https://www.occrp.org/en/news/police-say-scam-center-busted-in-png-had-moved-from-cambodia

OCCRP (2026b) 'UN report details Southeast Asia's interconnected criminal economy', 21 July. Available at: https://www.occrp.org/en/news/un-report-details-southeast-asias-interconnected-criminal-economy

Reppetto, T. A. (1976) 'Crime prevention and the displacement phenomenon', Crime & Delinquency, 22(2), pp. 166-177. Available at: https://doi.org/10.1177/001112877602200204

The Business Standard / Reuters (2026) 'Crime gangs snare more than $88b in scams in Asia-Pacific: UN', 21 July. Available at: https://www.tbsnews.net/worldbiz/asia/crime-gangs-snare-more-88b-scams-asia-pacific-un-1493826