Statelessness Statistics and IROSS: The UN Statistical Commission Grapples with Definitions

Bronwen Manby (LSE). This blogpost was originally published on the Critical Statelessness Studies Blog.

In this blog, Bronwen Manby, Visiting Senior Fellow at the LSE’s Firoz Lalji Institute for Africa and GLOBALCIT Regional Coordinator for Africa, explores the history of statistical data on statelessness and the ongoing process to develop International Recommendations on Statelessness Statistics (IROSS). What definitions and data should IROSS use to decide who is stateless? And when might an emphasis on statistical data not be the best use of resources?

How many stateless people are there in the world? In 2014, as UNHCR’s #Ibelong campaign to end statelessness was launched, I published a blog commenting on the unscientific nature of statelessness statistics reported by UNHCR for Africa. Rashly, I “confidently predicted” that the result of this campaign would be to “increase statelessness by many millions of people” – though only because the statistics on stateless persons were so inadequate that one of the main impacts of greater attention to the issue would be that uncounted populations would come into view.

It turns out that I was wrong: the challenge of arriving at a figure for stateless persons is sufficiently difficult that only in one country in Africa – Côte d’Ivoire – have the reported numbers increased as I predicted. Côte d’Ivoire is the only African state (and one of very few in the world) where a serious effort has been made to develop a methodological approach that would capture the total population that is “stateless or at risk of statelessness”. A national household survey, conducted in 2019 by the National Institute of Statistics with the support of UNHCR, was (ironically) commissioned in part because the government was concerned that the figure of 700,000 stateless persons previously reported as an estimate was too high – the result of the survey was an overall estimate of 1.6 million people stateless or at risk; 955,000 stateless people were reported in the 2020 Global Refugee Trends report.

The first time UNHCR reported country-level statistical data on statelessness was in the 2004 global trends report, when it indicated that for the 30 countries where data had been collected, the total number of stateless persons was estimated to be 1.5 million (excluding Palestinians). In a 2011 report UNHCR used the figure of “up to 12 million” stateless people in the world. When the #Ibelong campaign was launched three years later, the figure used was “at least 10 million”.  Since 2019, the total figure has been replaced with the more general reference to “millions” of people globally, in recognition of the problems with the statistics. In the 94 countries where data on stateless populations are now collected, there are said to be 4.2 million stateless people. As argued by Brad Blitz on this blog last year (and acknowledged by UNHCR), this more specific number is itself based on very doubtful methodologies, while clearly omitting large numbers of people even in countries for which data is reported (for example, India).

In recognition of the poor quality of existing numbers, Action 10 of the Global Action Plan accompanying the #Ibelong campaign was to “improve quantitative and qualitative data on stateless populations”. A number of UNHCR documents have discussed problems with, and efforts to improve quantitative data, starting from a 2011 report on measuring stateless populations; most recently a 2019 discussion of UNHCR statistical reporting on statelessness. There are currently two parallel efforts to develop better methodologies: an initiative led by the Expert Group on Refugee, Internally Displaced Persons and Statelessness Statistics (EGRISS) of the UN Statistical Commission to develop International Recommendations on Statelessness Statistics (IROSS), and a separate but complementary Inter-agency Group on Statelessness Estimation (IGSE).  A progress report on IROSS will be discussed at the March 2022 annual meeting of the Statistical Commission, and the ambition is for a final version of the IROSS to be adopted by the Commission in 2023.

The IROSS process highlights once again the critical question of definitions in the field of statelessness studies. It proposes three population groups falling within the statistical framework: stateless persons, persons of undetermined nationality, and stateless-related persons. The first two categories are based on the 2006 ExCom Conclusion on Identification, Prevention and Reduction of Statelessness and Protection of Stateless Persons, which revived UNHCR’s mandate on statelessness and called on the agency “to continue to work with interested Governments to engage in or to renew efforts to identify stateless populations and populations with undetermined nationality residing in their territory.”

The idea of “stateless-related persons” draws on the concept of “refugee-related persons” in the International Recommendations on Refugee Statistics adopted in 2018: it includes those who “previously experienced statelessness, are descended from one or more stateless parents, or are currently living in a household with at least one stateless person”. If a “stateless person” can be clearly identified, it is a useful addition to the research toolbox, expanding the possibilities of research into the broader impacts of statelessness.

The central question in collecting statistics about people fitting abstract concepts, however, is the definition of the abstract concept that they may fit. A stateless person is “a person who is not considered as a national by any state under the operation of its law” (Art 1(1), 1954 Convention relating to the Status of Stateless Persons). Since the adoption of its Guidelines on the Definition of Stateless Person in 2012 (now included in the Handbook on Protection of Stateless Persons), UNHCR has warned against a distinction between “de jure” and “de facto” statelessness: if a person is not recognised as a matter of fact by any state as its national (for example because of lack of supporting documents) the person is stateless even if in legal principle he or she may be a citizen of one or more states.  Statelessness is therefore a “mixed question of fact and law” (para.23 of the Handbook); and to find out if a person is stateless requires a combination of factual and legal analysis. The design of a methodology to find the number of stateless persons is thus a complex task, necessarily context specific; one which must also take into account that it may take several years of failed applications for documents for a person to find out that they are not, as it turns out, “considered as a national” by any state. Moreover, although statistical definitions have different purposes from legal definitions, any design should also take into account that data practices have political effects.

The IROSS progress report does not take this definitional discussion on board: the current proposal is to identify stateless persons “either through self-declaration, or through recognition by competent government authorities, on an individual or group basis.” That is, it requires people who believe themselves to be nationals of their country of birth and residence to declare themselves stateless in order for their (arbitrary) non-recognition as nationals to be recorded in statistics; or, alternatively, it depends on official government recognition of stateless status, including potentially a collective imposition of the status on a group who may (like many of the Rohingya) reject the label. It does not consider collection of statistics about, for example, rejected applications for documents recognising nationality. The representatives of national statistics offices debating IROSS apparently see the Côte d’Ivoire example as a warning: among the challenges noted are that “collecting statistics on statelessness may lead to large estimates of those potentially stateless or of undetermined nationality” (para.33(f) of the progress report).

The category of “undetermined nationality” has no definition in international law. The IROSS report follows a working definition adopted by UNHCR: “A person may be assessed as being with undetermined nationality following a review that verifies the following: they lack proof of possession of any nationality; and either A) have links to more than one State (on the basis of birth, descent, marriage or habitual residence); or B) are perceived and treated by authorities in the State of residence as possessing such links to other States.”  This definition is close to an idea of “at risk of statelessness” used in other contexts to refer to potentially vulnerable groups; a concept which UNHCR’s quick guides on researching statelessness, published in 2021, do not recommend for quantitative research because of the lack of a clear definition. Yet the “undetermined nationality” working definition itself seems to miss the opportunity to ask a purely factual question: does the person have an official document recognising nationality of this or any other state? Instead, like the definition of a stateless person (or the concept of “at risk of statelessness”), it requires an assessment of surrounding information regarding links to different states and the perception and treatment of the person by the authorities.

The IROSS process responds to the frequent assertion by states that they can only act on statelessness when they have proof that statelessness is in fact a problem. Better statistics are therefore important. In some contexts, however, a fixation on improving quantitative data about statelessness risks diverting attention from the more urgent task of fixing nationality law and administration. And qualitative research is often more suited to find out what needs fixing. Establishing accurate statistics on statelessness may be an important investment in countries where statelessness and/or lack of documentation affect only a small minority of the population. In any country, factual information on official identity documents held (according to national context) could provide a useful baseline of information from which to plan next steps. But, where a very large number of people do not have a documented nationality, the resources needed to determine a plausible figure for the sub-group who are stateless may be more valuably directed elsewhere. In these cases, as I noted in 2014, “better to concentrate on addressing the problems than trying to get the ‘correct’ statistics”.