The Slave Market by Gustave Boulanger

Over the past three decades a radical, worldwide social and political revolution has been slowly building steam. This revolution rejects the exploitation of human beings for private profit and seeks to eliminate these practices that have built our world and that continue to power global economic growth. The leaders of this revolution, including those involved in the Global Slavery Index, deserve to be applauded for their courage in waging what often feels like an impossible battle. But for everyone working in this field, good intentions are not enough. Dialogue, criticism and constructive debate are also needed to move this revolution forward.

The Global Slavery Index (GSI) rose without a trace in 2013: the flagship of an organization established the previous year with the explicit mandate of ending modern slavery in the lifetime of its founder-funder. A revised version was produced in 2014 and the latest Index was launched in London on June 1st, 2016 at an event headlined by Australian actor Russell Crowe. The rationale for the Index was clear from the outset. As Bill Gates advised fellow billionaire philanthropist Andrew Forrest, in order to secure traction on this issue he had to find some way of quantifying the problem that Forrest had committed his reputation – and an undisclosed slice of his personal fortune - to eradicating. In Gates’ words: “if you can’t measure it, it doesn’t exist”. Just as importantly, if you can’t measure it, you cannot reliably demonstrate impact of any interventions. Through rigorous measurement, Gates has been able to show, very convincingly, that the work of his Foundation is making a real difference to the global disease toll.

But measuring modern slavery and quantifying disease are worlds apart. Determining whether someone has malaria or HIV; the extent to which a community is affected; and even the vulnerability of a given population to contracting that disease is relatively straightforward and State, private and academic institutions exist whose sole function is to do just that. They all use the same, replicable method and criteria and they all come up with results that can generally be relied upon. (While establishing vulnerability to such diseases can be more difficult, the protocols are now well established). Extrapolation to estimate disease prevalence works well in the field of public health because our understanding of disease, our definitions and our diagnostic tools are sound and universally accepted, having emerged from a long history of inquiry, analysis and refinement. We also have the benefit of a willing and to some extent, captive test population. Of course these factors can come together in very different contexts. The Index’s authors note that in the mid-nineteenth century the United States was able to accurately measure the size of its slave population. But this is because everyone involved knew exactly what was being counted: slavery back then was a legal fact as well as a physical reality; records were kept and the subjects of measurement were clearly not going anywhere.

‘Modern slavery’ is something quite different - a made-up term that seeks to encompass under its expansive umbrella a raft of exploitative practices and myriad victims: from the Nigerian schoolgirls abducted by Boko Haram to the abused maids of diplomatic households in London and Washington; from orphanage tourism in Cambodia to the recruitment of child soldiers by the so-called Islamic State.  The authors of the Index have recalibrated their conception of ‘modern slavery’ from year to year, which makes comparison between their own reports a challenging exercise. To take the disease analogy further, there is no epidemiology of what is referred to as modern slavery: no scientific or rational basis for studying patterns, causes and effects. We don't yet have universally accepted diagnostic criteria or credible tools of measurement - which means that universal, reliable calculation of the size of the problem, while an important goal to strive for, is not yet possible.

These difficulties are well known to those working in the field, which makes the appearance of an indicator attached to the new UN Sustainable Development Goals “number of victims of human trafficking per 100,000 population, by sex, age group and form of exploitation” curious and, at the very least, highly premature.  After being called to task for its persistent and amateurish attempts to measure global rates of “human trafficking”, the US Government eventually gave up trying.

In recent years, the International Labour Organization (ILO) has felt pressure to affirm the scale of a problem that is central to its raison d’etre. The ILO has a considerable head start: forced labour is complex but, unlike modern slavery, it is defined in international law and generally well understood. Even so, the hurdles of its measurement are immense and despite its opportunistic and legally indefensible adoption of the new language of ‘modern slavery’, the ILO has tried to be open about the difficulties of measuring the number of those in forced labour; the fragility of the resulting data; and the highly provisional nature of any conclusions based on that information.

The Index, by contrast, sets itself three Herculean tasks: (i) calculating the vulnerability of individuals within each country to enslavement; (ii) measuring the total number, or prevalence, of slaves in every country; and (iii) assessing the overall quality of government responses to modern slavery. Seeking to work out how – and how well - this has been done is itself a daunting prospect. The methodology used by the Index is complex and in parts opaque and incomplete, requiring considerable persistence to unravel and analyze. But that task, which I can only begin here, is an important one. The purpose of the Index is not just to raise awareness about human exploitation - it is also intended to provide an evidence base for “[advocating and building] sound policies that will eradicate modern slavery” as well as a raft of initiatives loosely linked to Walk Free including a number of funds that are expected to disburse millions of dollars to support anti-slavery efforts. Governments and the many civil society bodies operating in this field will certainly use the Index to advance their own agendas and interests. For these reasons the Index deserves rigorous scrutiny.

This article briefly explores each of the three elements of the Index (vulnerability measurement; prevalence measurement; and response measurement), making some preliminary findings about the quality of the methodology and its application under each heading. While recognising that an analysis of this brevity can only scratch the surface (for example I forego, in this piece, any discussion of the value and validity of measuring a global, internationalized phenomenon at the national level), I return to a theme addressed in a previous critique: the lack of critical engagement with the Index among those who have the capacity - and I would argue, the responsibility - to interrogate this work carefully, openly and honestly.

Assessing country-level vulnerability to modern slavery

The vulnerability measure assigned to each country is important because it is a major element in assessing the prevalence of slavery. The measure uses several dozen variables divided into four categories to work out the risk of enslavement in a particular country: (i) civil and political protections; (ii) economic, health and other social rights; (iii) personal security; and (iv) refugees and conflict. A country’s final “vulnerability score” is calculated by averaging the sum of the variables for each category and then averaging the four resulting scores. For example, the country with the lowest vulnerability rating, Denmark, scores highly across all categories and most variables. The country with the highest vulnerability rating, the Central African Republic, receives low scores across the board.

The categories themselves generally make sense. Individuals and communities caught up in conflict will inevitably be more vulnerable than others to exploitation. Poverty and personal insecurity (measured in the Index by proxies such as being in debt and inability to raise emergency funds) also seems to be relevant as a vulnerability factor because we know that fewer resources inevitably compromise life choices and lead to risky strategies such as irregular and debt-financed migration. But some factors chosen to assess vulnerability are less easy to reconcile and no explanation is offered. Where is the evidence to support the assertion that “less cell phone subscriptions per 100 people in each country is considered to be … a vulnerability factor to slavery” or that greater access to weapons correlates with (let alone causes) increased vulnerability to slavery? Except as three of many available proxies for poverty, what do the rate of tuberculosis, undernourishment and “access to an improved water source” tell us about vulnerability to slavery? Sexual orientation and disability rights are now staples of the international human rights agenda but we are a long way from establishing a credible link between denial of these rights and increased risk of modern slavery. To turn this analysis on its head, a government that adopts an arms-reduction programme, or that actively works to reduce its tuberculosis rate; or that works to increase access to cell-phones through subscription can be said to be reducing people’s vulnerability to slavery - even though none of these measures have been concretely linked to the national rate of forced labour, forced marriage or other forms of exploitation associated with modern slavery.

There are compelling indications that the availability of data points influenced and in some cases determined the selection of vulnerability factors and not just in relation to the examples cited above of cell phone subscriptions, weapons availability, health and access to water. For example, confidence in the judiciary – rather than the capacity of the State to protect citizens from violence (including violent crime, perhaps the most well-established and defensible measure of vulnerability to forms of exploitation associated with modern slavery) - is cited as a vulnerability factor, presumably because the former is supported by a contemporary data source and the latter is not.

Other oddities in the final vulnerability assessment appear to affirm fundamental weaknesses in the both the GSI methodology and in its application. For example, Brunei makes it on to the list of the top ten most vulnerable countries to modern slavery, above Iraq and Pakistan. This is baffling. A relatively poor score on the category of civil and political rights is probably justified for Brunei. But the assertion of the Index that tiny, oil-rich, low-crime Brunei ties with Papua New Guinea as worst in the world for personal security is nonsensical. It also comes as a surprise to learn that this country, which has not experienced internal conflict for at least a generation and neither produces nor accepts refugees, is considered by the GSI to be in poorer shape than Lebanon, Turkey and Yemen in the “Conflict and Refugees” category. Another highly questionable assertion is that, of the 167 countries surveyed, only Iceland, Denmark and New Zealand rank higher than Singapore as the countries where one is most safe from modern slavery. Certainly it is true that Singaporeans (much like their citizen-counterparts in Brunei) are very well insulated from the risk of exploitation. But the same cannot be said of Singapore’s 870,000 non-residents, many of whom are denied the right to unionize as well as (in some sectors) the protection of a mandated minimum wage and standardized working hours. Exploitation of migrant workers in Singapore, especially in the construction and domestic service industries is a widely acknowledged and well-documented reality.

The second part of this article explores the Index’s measurements of country-level prevalence and government responses to ‘modern slavery’.