Despite efforts in data collection (including asking surveyed individuals about their family networks), reliable data on the entire population subjected to modern slavery is not available.

Survey data cannot capture an adequate number of cases for the estimation of certain populations – namely, children in modern slavery and individuals subjected to forced commercial sexual exploitation. It is challenging to interview a representative sample of these populations due to the ethical issues that arise when interviewing children, as well as the sensitivities and stigma that may be associated with forced commercial sexual exploitation and modern slavery of children.

For these reasons, the survey data are deemed to be sufficiently reliable for measuring adult forced labour exploitation and forced marriage, but not sufficiently reliable for measuring forced commercial sexual exploitation and forced labour exploitation of children. Therefore, odds ratio is used to extrapolate the prevalence of these hidden populations from the CTDC data to complement the survey data.

To estimate the odds of being subjected to forced commercial sexual exploitation relative to being subjected to forced labour exploitation (model 1), we apply a logit model regressing the type of exploitation (forced commercial sexual exploitation versus forced labour exploitation) (as the dependent or Y variable) against sex, majority status, and an interaction term of sex and majority status (as the independent or X variables). See the methodology report (forthcoming) for more information.

To estimate the odds of being victims of forced labour exploitation as a child relative to that of an adult (model 2), we apply a logit model regressing the proportion of child victims of forced labour exploitation versus the proportion of adult victims of forced labour exploitation (as the dependent or Y variable) against sex (as the independent or X variable). We do the same for sexual exploitation, as a child rather than as an adult.

The odds ratios are then applied to the corresponding estimates of forced labour exploitation of adults derived from the national surveys to estimate forced labour exploitation of children, forced commercial sexual exploitation of adults, and commercial sexual exploitation of children.

An example

In model 1 (below), we use a logit model to calculate the odds ratio of falling victim to forced commercial sexual exploitation relative to being subjected to forced labour exploitation in the region of exploitation r.

Here is the calculation process for model 1.

Source: Methodology report, forthcoming.

The odds ratios are reported in the last column of the table. They are greater than 1 for children and smaller than 1 for adults. This means that children, boys or girls, are more likely to fall victim to forced commercial sexual exploitation than forced labour exploitation. By contrast, adults, particularly men, are more likely to fall victim to forced labour exploitation than forced commercial sexual exploitation.

Here is a summary of the odds ratio computed for the Global Estimates.

Notes: For simplicity, adult female is classified as woman; children female is girl; adult male is man; children male is boy. Source: Methodology report, forthcoming.

To estimate the forced commercial sexual exploitation of adults, let’s use the odds ratio of woman and man as examples. In Table 3 (below), you can see how we compute the estimates. We multiply the estimates of adult forced labour exploitation (obtained from surveys) to the odds ratios of adult forced commercial sexual exploitation relative to adult forced labour exploitation (obtained from the logit model and the CTDC dataset).

Source: Methodology report, forthcoming.

Let’s use the odds ratio of 0.0492 and 0.7685 for males and females as examples.

For men, the odds of being a victim of forced commercial sexual exploitation are 0.05 times lower than being a victim of forced labour exploitation. This means that for every 100 men who were victims of forced labour exploitation, it is likely that 5 other men were victims of forced commercial sexual exploitation.

For women, the odds of being a victim of forced commercial sexual exploitation are 0.77 times lower than being a victim of forced labour exploitation. This means that for every 100 women who were victims of forced labour exploitation, it is likely that 77 other women were victims of forced commercial sexual exploitation.

There are 27.6 million people in situations of forced labour on any given day from 2017 to 2021. This absolute number translates to 3.5 people in forced labour for every thousand people in the world.

Table 4 (below) is a summary table with the Global Estimates that are computed using different data sources and methods. Forced labour exploitation of adults is computed based on national surveys. Forced commercial sexual exploitation of adults and forced labour of children (including forced labour exploitation of children and commercial sexual exploitation of children) are estimated using the odds ratio based on the CTDC dataset.

Notes: (a) Number is expressed in thousands. (b) “‰” denotes cases per 1,000 population. (c) “TOTAL FORCED LABOUR” comprises forced labour exploitation and state-imposed forced labour. (d) “TOTAL MODERN SLAVERY” comprises privately-imposed forced labour, state-imposed forced labour, and forced marriage. (e) “Income grouping” refers to the income group of the country where forced labour occurs. Source: Table 1 of the main report.

While the Global Estimates did not explicitly estimate the prevalence of trafficking in persons, the report underlines those individuals subject to forced labour and forced commercial sexual exploitation, as well as migrant workers, are vulnerable to unscrupulous recruitment and abuse by traffickers.

The methodology underlying the Global Estimates is the product of an ongoing process of refinement made possible by an improved understanding of forced labour and forced marriage and of how best to measure them. Reliable information and statistics on forced labour, forced marriage, and human trafficking are critical to promoting awareness and understanding of the problem, and to informing policy responses. The methodological refinements in measurement, combined with a more balanced geographical coverage in data, mean that the Global Estimates are the most complete and comprehensive to date.

We invite you to read the Global Estimates of Modern Slavery: Forced Labour and Forced Marriage in greater detail. If you’re interested in the estimation methodology, please refer to the methodology report (forthcoming).