Human trafficking is a widespread, fast-growing crime with victims in every country in the world. Understanding of the problem is hampered by the difficulty of collecting statistically valid data. In this pilot study we employ a different data collection method. Rather than relying on news accounts, police reports, returning victims, stakeholders and other often-used methods, we instead survey migrant sending households —entities that are the source of many trafficking victims. As migration decisions are often household decisions in rural Vietnam, we examine whether such household surveys can provide statistically valid and reliable information on both migrants and trafficking victims. We survey 496 households yielding information on 2,394 individuals in the Trà Vinh province in the Mekong Delta region of Vietnam. A sampling frame was constructed using the 2009 Vietnamese census data for the region, and villages were selected randomly from this frame. Households were randomly selected from within villages and given a short pre-survey. We oversampled households with migrant members (406 households) for the full survey and randomly sampled another 90 non-migrant households for control purposes. The survey solicited information on all household members, including questions on migration status, income, education, occupation and remittances in addition to standard demographic information. The survey also used anchoring vignettes to ensure that there was a common understanding of the concept of trafficking in contrast to migration or human smuggling. Our survey identified 38 persons reported as trafficked. This number yields a trafficking rate of 9.6 trafficking victims per 1,000 people, more than
three times the trafficking rate of 3 persons per 1,000 reported for Asia by the International Labour Organization (ILO).
We use a rare events logistic model to examine the predictors of trafficking vulnerability. The survey found no difference in the probability of being trafficked by sex nor income. The most important predictor was household and migrant education. Average household or village education was negatively correlated with the likelihood of being trafficked. More educated households perhaps have superior migration opportunities that reduces members’ susceptibility to trafficking. Conversely, at the individual migrant level, higher education is associated with a higher likelihood of being trafficked. This result may be an artifact of households selecting their most educated members to migrate.
Our results suggest that a household survey method is a valuable tool in examining the extent and nature of human trafficking. A fuller survey from multiple provinces is needed to verify this conclusion.
