The history of the 20th-century refugee typically unfolds as a tale of increasing legal protection. My dissertation challenges this narrative by retelling refugees' history from the viewpoint of statistics and demography. Doing so, I argue, reveals that international law and the social sciences developed as competing attempts to identify, classify, and control refugees between the 1920s and the 1960s. Much like lawyers, demographers set out to fill the gap of knowledge surrounding refugees, albeit with methodologies that centered on quantification. Pointing towards the constitutive nature of statistics, demographers drew on cross-national data sets and outlined refugees as a comprehensive, international category. As opposed to reactive laws, their surveys also offered predictions of future displacement and were directly involved in refugee policy as it was enacted. I further argue that demographic census-taking activities not only paralleled legislation but often preceded it, effectively undergirding a legal-bureaucratic regime that aimed to register and control human mobility. As part of this regime, the now perennial refugee camp emerged as a site of institutionalized control rather than temporary care and protection. Tracing the life cycle of refugee statistics across archival sites in Britain, France, Germany, and Switzerland, my dissertation illustrates how speculative and often incomplete statistics became a legitimate basis for altering field operations and fine-tuning policies. Theories and practices of quantitative data collection, I suggest, shaped who was deemed to fall within the classificatory boundaries of "the refugee" and entailed substantial material consequences. Studying the history of refugee statistics thus reveals how quantification served as a targeted management device in the early 20th century, but it also points to the long shadow data-driven mechanisms cast upon contemporary relief efforts, including the ongoing European "refugee crisis."