Policy Simulations Could Help Combat Sex Trafficking

August 27, 2012

CREDIT: Mr. Theklan (CC).

Human trafficking is the fastest growing criminal industry in the world. It is also the third largest, exceeded only by the drug trade and the illegal arms industry. According to the State Department's 2010 Trafficking in Persons Report, 12.3 million adults and children are in forced labor, bonded labor, and forced prostitution around the world.

Since the fall of the Iron Curtain, former Eastern bloc countries have become a major source for trafficked women and children. Estimates of the extent of trafficking in Eastern Europe and Central Asia vary widely, from 200,000 people total to 175,000 persons annually. The UNODC reports that Eastern European trafficking victims are primarily young women who are exploited in the commercial sex industry.

Despite a growing body of research on trafficking, many agencies with a role in combating it do not systematically collect and analyze data that would allow them to judge whether their programs are successful. Thus assessment of the long-term impacts of counter-trafficking is at an early stage, due to program design characteristics that impede evaluation, and a lack of quality data.

My hope is to address this gap in knowledge on the effectiveness of efforts to reduce sex trafficking. What is needed is a quantitative assessment of the impacts of different interventions and policy responses.

The Connection between Trafficking and HIV

Among the most significant health consequences of sex trafficking is the risk of HIV infection. Multiple factors account for the increased vulnerability of trafficked persons to infection. They are often forced to endure multiple sex partners and have no ability to insist on condom use or to refuse high-risk sex acts. Their young and immature bodies are extremely fragile and even more vulnerable to injury during forced sex. Injuries and abrasions further increase the victims' chances of becoming infected if exposed to the virus.

Another factor that puts these victims at risk of HIV infection is drug use. Victims may be injected with drugs to increase their compliance, or they may choose to inject drugs as a coping mechanism. Victims may also receive surgical treatments such as forced or voluntary pregnancy terminations in unsanitary conditions, by unqualified practitioners, using unscreened blood supplies.

Empirical investigation to date into HIV prevalence among survivors of sex trafficking is limited to South Asia. A study of sex-trafficked women and girls rescued from brothels in Mumbai, India found that 22.9 percent were HIV positive. Another study of girls and women who were sex-trafficked from Nepal to India and then repatriated found that 38 percent were HIV positive, with the figure rising above 60 percent among those who had been trafficked before age 15. The high rates of HIV prevalence among survivors of sex trafficking in South Asia support concerns that sex trafficking may be a significant factor in the expansion of the HIV epidemic in other countries as well.

Not only does sex trafficking increase the risk of exposure to HIV and AIDS, the HIV and AIDS crisis itself drives greater victimization. HIV and AIDS increases the number of children trafficked because there is an increased demand for sex with young girls, since they are perceived to be HIV negative. Young girls are more vulnerable to HIV and AIDS, both biologically and because of their lack of power to negotiate the use of condoms. In many of the countries where AIDS takes its heaviest toll, children who have lost both their parents to AIDS are particularly vulnerable to traffickers. The HIV and AIDS pandemic can be seen as both a cause and a consequence of trafficking.

A Microsimulation Model to Test Different Scenarios

Microsimulation modeling could be used to assess the impact of efforts to combat sex trafficking and related HIV infections and thus help build better knowledge-based policies. It is an evaluation method that has been used for similar interventions in other areas of health and welfare. This work would enable policy makers and practitioners to focus limited anti-trafficking resources more effectively.

Microsimulations are computer models for analyzing activities such as highway traffic flowing through an intersection, financial transactions, or pathogens spreading through a population. They are often used to evaluate the effects of proposed interventions before they are implemented in the real world. For example, a traffic microsimulation could be used to evaluate the effectiveness of lengthening a turning lane at an intersection, and thus help decide whether lengthening the lane would be cost effective.

Microsimulation models look at the interaction of individual "units" such as people, families, or households. Each unit is treated as an autonomous entity, and the interaction of the units is allowed to vary depending on randomized parameters. These parameters are intended to represent individual preferences and tendencies. For example, in a public health model individuals could vary in their resistance to a virus, as well as in personal habits that contribute to the spread of the virus.

The model I propose would simulate a population of households longitudinally and track whether the primary respondent (the person with the most recent birthday) or someone in their close family has ever experienced a situation that would be classified as sex trafficking. A 2006 International Organization for Migration human trafficking survey is taken as the starting population and provides individual, household, and regional characteristics. The household population is dynamic, with a continual process of family members aging, dying, and being born. Primary respondents who die are replaced by the person in the household with the next most recent birthday. As the household ages, individual, household, and regional variables are updated.

Each virtual year, the microsimulation model would expose each household to an annual risk of whether the interviewed person reported a victim of sex trafficking among close family members. The risk is conditional on a set of individual, household, and regional variables (independent variables). These variables raise or lower the likelihood of trafficking. The individual variables include age, gender, level of education, and marital status of the respondent. Examples of household-level variables include the number of children below the age of 17, a dummy for households living in rural areas, and a dummy for households living in the district around the capital city. The regional variables include migration prevalence, awareness of trafficking, the rural fraction of the population, density of physicians, infant mortality, and crime rate. A dummy variable is one that takes the value 0 or 1 to indicate the absence or presence of some categorical effect that may be expected to shift the outcome.

A variation on the analysis performed by Toman Mahmoud and Christoph Trebesch will inform the model. Data to support the microsimulation come from two unique cross-sectional household surveys commissioned by the International Organization for Migration. Fieldwork for the surveys was conducted in Belarus, Bulgaria, Moldova, Romania, and Ukraine. The relatively large sample and the survey design make it the most comprehensive micro-level survey of human trafficking available worldwide.

This microsimulation model, while specific to Eastern Europe, could be applied to the United States and elsewhere when comparable data for those regions become available.

Project Outputs and Policy Interventions

The output of the microsimulation will be the prevalence of families that suffer from sex trafficking, as well as the prevalence of families that suffer from sex trafficking–related HIV infections according to year.

Some of the policy interventions that could be modeled include raising awareness of human trafficking, raising access to public information and news, raising education among female respondents, improving the local economy, and reducing gender-based discrimination. The effect of these interventions would be estimated using Population Attributable Fractions (PAF). Conceptually, the PAF is the fraction by which sex trafficking would be reduced under a more favorable alternative. For example, we can estimate the effect of raising awareness of human trafficking by assigning awareness to households that reported no awareness in a "what-if" scenario.

By running the microsimulation in both the absence and presence of various interventions and comparing the difference in sex trafficking prevalence, we can estimate the effect of each intervention on reducing the prevalence of sex trafficking and related HIV infections.

While HIV infection is clearly not the only health consequence of sex trafficking, it is among the most significant potential health consequences and is the one for which the best data are available. Hence, it makes sense from both a modeling and response perspective to focus on HIV when assessing the health consequences of sex trafficking.

There are other consequences of sex trafficking for which the microsimulation model can act as a substrate on which to perform further analyses. For example, from a health perspective, victims face serious physical internal injuries such as head injuries and broken bones, STDs, tuberculosis, permanent damage to reproductive systems, and post-traumatic stress disorder. Sex trafficking also has widespread negative non-health consequences for individuals and societies. For example, sex trafficking helps to promote societal breakdown by removing women and girls from their families and communities. If and when victims are able to return to their communities, they often find themselves doubly victimized by social stigmatization, discrimination, and rejection.

Sex trafficking also fuels organized crime groups that usually participate in many other illegal activities, including drug and weapons trafficking and money laundering. In addition, sex trafficking negatively impacts local and national labor markets, due to the loss of human resources. The effects include depressed wages, fewer individuals left to care for elderly persons, and an undereducated generation. Sex trafficking also erodes government authority, encourages widespread corruption, and threatens the security of vulnerable populations. As data on these health and non-health consequences of sex trafficking become increasingly available, the microsimulation model could be adapted to include them.

By identifying the most effective interventions, microsimulation modeling would help guide counter-trafficking policies and programs and contribute to the ultimate goal of eliminating human trafficking. This would enable policy makers and practitioners to focus limited anti-TIP resources more effectively.

Stephanie L. Bailey is a project lead in the Center for Health Policy at the Stanford University School of Medicine and an adjunct professor at Dominican University of California.