removal to the failure to secure travel documents in 948 of the 3,053 cases, including in cases of statelessness in which citizenship could not be established (DHS-OIG 2019, 7). To date, DHS/OIS has not responded to CMS’s broad data request. However, CMS has secured refugee and immigration data from other sources. To complement its ACS estimates, CMS secured data from DOS’s WRAPS database, which tracks refugees admitted for resettlement (RPC 2019), and contains records on country of birth, admission date, religion, ethnicity, age, education, and native language. As above, CMS estimates individuals whose countries of birth, ethnic backgrounds, and years of arrival to the United States, or language proficiencies match its profiles of US residents who are potentially stateless or potentially at risk of statelessness.65 These estimates are in Table 3 below. To facilitate comparison with ACS data, CMS analyzed WRAPS data from calendar year 2002, the first year of its availability, through 2017. Since WRAPS data does not indicate whether admitted refugees eventually naturalized, CMS used ACS data from 2017 to calculate estimated naturalization rates by country of birth (US Census Bureau 2018b). To ensure that these counts do not include persons who have naturalized, CMS multiplied WRAPS counts of refugees fitting each missing profile by the 2017 average naturalization rate of all US residents born in those refugees’ birth countries.66 Table 3: WRAPS Estimates of Profiles of Those who are Potentially Stateless or Potentially at Risk of Statelessness Profile Description Ethnic Russians, Belarussians, Ukrainians from Estonia Ethnic Russians, Belarussians, Poles from Latvia South Caucasus Armenians from Azerbaijan Meskhetian Turks Russia Uzbekistan Roma, Ashkali, and Balkan Egyptians Count 34 38 143 2,442 37 65 WRAPS data is only available in aggregated form. It is not possible to filter the data by multiple categories (e.g., language and ancestry). Thus, CMS researchers used aggregated counts by ethnicity, when available, to produce estimates. When ancestry records were not available, CMS used aggregated counts by language. It is also not possible to filter WRAPS data based on year of birth (RPC 2019) 66 Data limitations did not permit the calculation of a naturalization rate for Mauritania or Estonia. For Mauritania, CMS researchers applied a global average naturalization rate for individuals born in Western Africa. For Estonia, CMS researchers applied a global average naturalization rate for individuals born in “Other Europe,” a category including many small European countries. 61