UNHCR, Statistical Reporting, and the Representation of Stateless People 81 of stateless persons to the definition in the 1954 Convention and focused on the obligations of States that are party to the Convention – again, the plight of de facto stateless persons was pushed to the margins. One consequence of UNHCR’s advice in the handbook was that the authors of UNHCR-commissioned mapping studies did not try to estimate the number of that country’s (i.e., the selected country included in the mapping studies) stateless population.52 One further dilemma, which arguably applies to all international agencies whose mandate relies on the implementation of international law, is the overt bias toward states – a tendency bolstered by the technocratic logic that further informed the work of UNHCR. Like other UN agencies that must present “results” to donors, over the past fifteen years UNHCR has been building up its statistical capacity, including most recently by establishing a joint data center with the World Bank.53 In the 2005 Statistical Yearbook (published in 2007), for the first time the Agency set out its rationale for providing better statistical coverage: “In an effort to ensure evidencebased resource allocation and policy formulation, a variety of information sources are utilized to quantify and profile UNHCR’s population of concern.”54 This ambition was reflected in its increased coverage of groups that had been underreported. For example, while UNHCR’s coverage of stateless people included just thirty states in 2004, by 2015, UNHCR was publishing data on seventy-nine states.55 Most important, UNHCR also started to explain its methodological processes for estimating people who fell under its mandate and then, like other agencies, sought to make its data more operational.56 It offered more comprehensive definitions of the indicators used, the various categories of concern, and its main sources of data. In its notes to the published statistics, UNHCR provides some important small print. In the 2011 Global Appeal, UNHCR recorded that “the data are generally provided by governments, based on their own definitions and methods of data collection.”57 This statement was repeated to UNHCR’s donors as recently as February 2020.58 A close reading of the footnotes in UNHCR’s principal publications records the methodological difficulties involved in data estimation and the challenges facing UNHCR and member states. For example, its 2020 report on its financial performance lists the different approaches it has taken to produce data, which include relying on national estimates, excluding data, and adjusting based on censuses and Kerwin et al., “Statelessness in the United States.” See Strategic Advisory Council, Strategy for the Joint Data Center on Forced Displacement 2021–2023: Zero Draft (August 27, 2020), www.jointdatacenter.org/wp-content/uploads/2020/08/ JDC-strategy_zero-draft_August-2020.pdf. 54 UNHCR, UNHCR Statistical Yearbook 2005 (Geneva: UNHCR, 2007), p. 19. 55 See UNHCR, 2004 UNHCR Statistical Yearbook (Geneva: UNHCR, 2005); UNHCR, UNHCR Statistical Yearbook 2015 (Geneva: UNCHR, 2017). 56 See UNHCR, Guidance Document on Measuring Stateless Populations (May 2011), www .refworld.org/docid/4f6887672.html. 57 UNHCR, Global Appeal 2010–11. 58 UNHCR Executive Committee, Update on Budgets and Funding (2019, 2020–2021), p. 30. 52 53 https://doi.org/10.1017/9781108914994 Published online by Cambridge University Press