The Sorry State of Sex Trafficking Research
Last month, Northern Ireland saw its first prosecution for paying for sexual services (alongside violent crimes) brought before the courts. Having adopted the ‘Nordic Model’ of criminalizing the purchase of sex in June 2015, Northern Ireland seems to have inspired SNP politicians to advocate for a similar approach in Scotland. The debate about sex work legislation also rages on in the rest of the United Kingdom, where marginal progress is being made in some areas.
Research into the impact of different forms of sex work legislation and sex trafficking is a fascinating but highly-fraught area. The prevalence of sex trafficking and its relationship to different legal regimes pertaining to sex work is one of the chief battlegrounds for those who seek reform, and reliable evidence on this emotive issue is hard to come by.
A new paper published online last week by sociologist Ronald Weitzer—who wrote this excellent article on sex work policy and sex trafficking in 2011—gives new insight into the sorry state of the academic literature on these topics. First on his list of grievances is the unreliability of the data that is often used:
Using information on 161 countries from the United Nations Office on Drugs and Crime (UNODC), Cho et al. (2013) and Jakobsson and Kotsadam (2013) attempted to determine whether national prostitution laws were related to the prevalence of human trafficking. Yet UNODC had warned against using its figures either for one nation or cross-nationally, because “the report does not provide information regarding actual numbers of victims” (UNODC 2006, pp. 37, 44–45). UNODC’s caution was based on the unstandardized definitions of trafficking across countries (with some conflating trafficking, smuggling, and irregular migration); the widespread lack of transparency in data collection and reporting; and the reliance on different sources across the 161 countries (the media, research institutes, government agencies, NGOs, IOs). For some countries, only one of these sources was available. The authors concede that “the underlying data may be of bad quality” and are “limited and unsatisfactory in many ways” (Jakobsson and Kotsadam 2013, p. 93) and that it is “difficult, perhaps impossible, to find hard evidence” of a relationship between trafficking and any other phenomenon (Cho et al. 2013, p. 70). Yet they nevertheless treat the UNODC report as a data source and draw profound conclusions about the relationship between trafficking and national prostitution laws, concluding that human trafficking is more prevalent in countries with legal prostitution than countries where prostitution is criminalized.
Flawed data is just the tip of the iceberg. In the case of the two studies cited above, the authors’ approach to study design also leaves a lot to be desired:
A cross-sectional design (at a single time point) is used to measure something that should be examined longitudinally: the amount of trafficking before and after legalization. The latter approach would require reliable baseline figures to compare to reliable recent figures—neither of which exist.
The authors use aggregate national trafficking estimates (which combine labor, sex, and other kinds of trafficking) in their attempt to assess whether legal prostitution makes a difference. This means that there is a gross mismatch between the trafficking figures and prostitution law: In assessing whether prostitution law is related to the incidence of trafficking, figures on sex trafficking alone should be used, not the totals for all types if trafficking.
It is quite possible that nations where some type of prostitution is legal may have superior mechanisms for detecting sex trafficking, a variable missing in both studies.
A subsequent study by Cho (2016) using a different data source contains another howler. It “uses information on countries’ level of protection for [overall] human trafficking victims, which is then correlated with whether prostitution is allowed in a country.” The justification for this sleight of hand?
Without citing any source, Cho claims that “prostitution is closely linked to human trafficking because sex trafficking for the purpose of prostitution is the most common form of human trafficking and constitutes the largest fraction of trafficking victims” (Cho 2016, pp. 325–326). This claim is contradicted by the International Labour Organization (ILO) and the U.S. State Department. According to the ILO, “Forced commercial sexual exploitation represents 11% of all cases” of forced labor worldwide (ILO 2005, p. 12), and the State Department declares that “the majority of human trafficking in the world takes the form of forced labor” (USDS 2010, pp. 8–9).
All of these problems are endemic to the field of sex trafficking research. Weitzer’s paper would be an interesting academic exercise in the perils of low-quality data and poor research methodology, except for the fact that policies being enacted on the basis of the former two studies are hurting marginalized women:
...these two studies were embraced by politicians and policy makers in several countries, and used to justify new criminalization laws.
Weitzer also offers some thoughts on how the current state of research and the public debate based upon it can persist without significant challenge:
It is easy to make blanket and cavalier assertions regarding both trafficking and prostitution when (1) solid data are lacking, (2) the media simply recapitulate “official” claims without questioning or fact-checking them, (3) experts who challenge official assertions are ignored or denounced, and (4) the participants in sexual commerce are highly stigmatized and marginalized. This unfortunate pattern can be seen in both prohibitionist nations (e.g., Sweden) and in nations facing resistance to their current [comparatively] liberal policies (e.g., Germany, the Netherlands).
I’m sure readers of this blog approach every sensationalist media headline with a critical eye. However, when it comes to scaremongering stories about sex trafficking epidemics and accompanying calls for the ‘Nordic Model’, extra caution should be exercised.